WO2020059741A1 - Planning apparatus, method, and program - Google Patents

Planning apparatus, method, and program Download PDF

Info

Publication number
WO2020059741A1
WO2020059741A1 PCT/JP2019/036524 JP2019036524W WO2020059741A1 WO 2020059741 A1 WO2020059741 A1 WO 2020059741A1 JP 2019036524 W JP2019036524 W JP 2019036524W WO 2020059741 A1 WO2020059741 A1 WO 2020059741A1
Authority
WO
WIPO (PCT)
Prior art keywords
period
model
unit
operation plan
power generation
Prior art date
Application number
PCT/JP2019/036524
Other languages
French (fr)
Japanese (ja)
Inventor
豪秀 奈木野
Original Assignee
旭化成株式会社
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 旭化成株式会社 filed Critical 旭化成株式会社
Publication of WO2020059741A1 publication Critical patent/WO2020059741A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Energy or water supply
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications

Definitions

  • the present invention relates to a planning device, a method, and a program.
  • an electrolysis apparatus that generates hydrogen by electrolyzing water has been known.
  • a power supply unit a power generation device that generates power using renewable energy, or a power system in which an electricity rate fluctuates according to weather or power supply cost is known.
  • a first aspect of the present invention provides a planning device.
  • the planning device uses a power generation forecasting model that predicts a change in the amount of renewable energy generated by the power generating device during the target period based on the value of the first factor that is available before the target period.
  • a power generation amount prediction unit for predicting a transition of the power generation amount may be provided.
  • the planning device prepares an operation plan of the electrolysis device that satisfies the usage plan of the product of the electrolysis device in the first period in the future, based on the predicted change in the amount of generated renewable energy and the electricity rate of the power system.
  • An operation plan generation unit that generates the operation plan may be provided.
  • the power generation amount prediction model calculates a transition of the renewable energy power generation amount in the target period based on the value of the first factor including at least one of the renewable energy power generation amount of the power generation device and the weather information before the target period. You can predict.
  • the planning device may include a first model updating unit that updates the power generation prediction model by learning based on the value of the first factor in the past period and the actual change in the amount of renewable energy generated in the past period.
  • the operation plan generation unit during the first period, in order to satisfy the use plan of the product of the electrolysis device, the operation plan to operate the electrolysis device using the electricity from the power generation device in preference to the electricity from the power system. May be generated.
  • the use plan of the product of the electrolyzer is at least one of a plan to maintain the storage amount of the product of the electrolyzer in the first period within the reference range and a plan to satisfy the demand for the product of the electrolyzer in the first period. May be included.
  • the operation plan generation model includes a value of a second factor including at least one of operation data of the electrolysis apparatus, a demand amount of a product of the electrolysis apparatus, and a storage amount of a product of the electrolysis apparatus before the target period, An operation plan of the electrolysis device in the target period may be generated based on the prediction result of the change in the amount of generated renewable energy in the target period and the electricity rate.
  • the planning device calculates the value of the second factor in the past period, the change in the amount of renewable energy generated in the past period or the predicted result of the change in the amount of renewable energy generated in the past period, the change in the electricity rate after the past period, and the change in the past period.
  • a second model updating unit that updates the operation plan generation model by learning based on the target operation plan of the electrolysis apparatus may be provided.
  • the planning device uses a power rate prediction model that predicts a change in the power rate of the power system in the target period based on the value of the third factor available before the target period, and calculates a future change in the power rate.
  • An electricity rate prediction unit for predicting the electricity rate may be provided.
  • the operation plan generation unit generates a use plan of the product of the electrolytic device in the first period in the future based on the predicted future change in the amount of renewable energy generated and the predicted future change in the electricity rate.
  • An operation plan for the electrolyzer to be filled may be generated.
  • the electricity rate prediction model calculates at least one of the electricity rate, the power demand, the power supply, the renewable energy generation, the predicted value of the renewable energy generation, and the weather information in the power system before the target period. Based on the value of the third factor included, the transition of the electricity rate of the power system in the target period may be predicted.
  • a method includes using a power generation forecasting model that predicts a change in the amount of renewable energy generated by a power generator during a target period based on a value of a first factor that is available before the target period.
  • a step of predicting a change in the amount may be provided.
  • the method generates an operation plan of an electrolyzer that satisfies an electrolyzer product usage plan for a first period in the future, based on a predicted future change in renewable energy generation and an electricity rate of a power system.
  • the method may include a step of:
  • FIG. 1 shows a system 10 according to the present embodiment.
  • 4 shows a detailed configuration example of a planning device 40 according to the present embodiment.
  • 5 shows an example of an operation flow of the planning device 40 according to the present embodiment.
  • 14 illustrates an example of a computer 1900 in which aspects of the embodiments can be wholly or partially embodied.
  • Electrolysis device 20 is connected to power generation device 30, planning device 40, and power system 50.
  • the electrolytic device 20 may be a device that generates a product using electric energy.
  • the electrolysis device 20 is, for example, a hydrogen generation device that generates hydrogen as a product by electrolysis.
  • the electrolysis device 20 operates according to the operation plan generated by the planning device 40.
  • the electrolysis device 20 operates by being supplied with power from the power generation device 30 and the power system 50.
  • the planning device 40 generates an operation plan for the electrolysis device 20.
  • the planning device 40 may be a computer such as a personal computer, a tablet computer, a smartphone, a workstation, a server computer, or a general-purpose computer, or may be a computer system to which a plurality of computers are connected.
  • the planning device 40 may generate the operation plan by a process in a CPU, a GPU (Graphics Processing Unit), and / or a TPU (Tensor Processing Unit) of the computer.
  • the planning device 40 may perform various processes on a cloud provided by a server computer.
  • the operation plan generated by the planning device 40 may be a table or data describing a state in which the electrolysis device 20 should be operated in the first period.
  • the operation plan includes, for example, a time period in which the electrolysis device 20 is operated (or not operated), a time period in which the electrolysis device 20 is operated, and a time period in which the electrolysis device 20 is operated with the power from the power generation device 30 and the operation from the power system 50. It may be a table or data or the like that defines a time zone for powering, a time zone for selling the power generated by the power generation device 30 to the power system 50, and / or an operation rate of the electrolysis device 20 for each time zone.
  • the planning device 40 includes an acquisition unit 100, a storage unit 110, a model generation unit 120, a learning processing unit 130, a prediction unit 140, an operation plan generation unit 150, and a control unit 160.
  • the storage unit 110 is connected to the model generation unit 120, the learning processing unit 130, the prediction unit 140, and the operation plan generation unit 150, and stores information acquired by the acquisition unit 100.
  • the storage unit 110 may store data processed by the planning device 40.
  • the storage unit 110 may store intermediate data, calculation results, parameters, and the like calculated (or used) in the process of generating the operation plan by the planning device 40. Further, the storage unit 110 may supply the stored data to a request source in response to a request from each unit in the planning device 40.
  • the storage unit 110 supplies the stored data to the model generation unit 120 in response to a request from the model generation unit 120, for example.
  • the ⁇ model generation unit 120 is connected to the learning processing unit 130, and generates a learning model that the planning device 40 learns.
  • the model generation unit 120 generates a learning model according to the first factor, the second factor, and the third factor stored in the storage unit 110.
  • the model generation unit 120 may generate one or a plurality of learning models.
  • the model generation unit 120 supplies the generated learning model to the learning processing unit 130.
  • the learning processing unit 130 is connected to the prediction unit 140 and learns the generated learning model based on the learning data acquired by the acquisition unit 100.
  • the learning processing unit 130 may execute the reinforcement learning to update the learning model.
  • the learning processing unit 130 may update one or a plurality of learning models.
  • the learning processing unit 130 supplies the updated learning model to the prediction unit 140.
  • the operation plan generation unit 150 is connected to the control unit 160, and generates an operation plan of the electrolysis device 20 in the first period in the future.
  • the operation plan generation unit 150 generates, for example, an operation plan that minimizes or reduces the product manufacturing cost while satisfying the product use plan of the electrolytic device 20 in the first period in the future.
  • the operation plan generation unit 150 supplies the generated operation plan to the control unit 160.
  • the use plan of the product of the electrolyzer 20 includes a plan to maintain the storage amount of the product of the electrolyzer 20 within the reference range and a plan to satisfy the demand or supply of the product of the electrolyzer 20. May be included.
  • the demand amount or supply amount of the product may be, for example, the total amount of the product to be supplied by the electrolysis device 20 in the first period or the amount per predetermined time.
  • the storage amount may be the amount of the product stored inside or outside the electrolysis device 20.
  • Such a reference range of the storage amount, the demand amount of the product, or the supply amount is determined based on data input from the outside to the planning device 40, past data, or past data in the planning device 40 in the future first period. May be expected.
  • the control unit 160 controls the operation of the electrolysis device 20 using the operation plan of the electrolysis device 20 in the first period.
  • the control unit 160 may operate the electrolysis device 20 by selectively using the electric power from the power generation device 30 and the electric power from the power system 50 according to a time zone according to an operation plan.
  • the control unit 160 may operate each of the plurality of electrolyzers 20.
  • Control unit 160 may instruct stop and start of operation of electrolysis apparatus 20 when the operation of electrolysis apparatus 20, the storage amount of the product, and the like fall outside the expected ranges.
  • the control unit 160 may control the power generated by the power generation device 30 to be sold to the power system 50 using the operation plan of the electrolysis device 20 in the first period.
  • the planning device 40 of the present embodiment described above it is possible to generate the operation plan of the electrolysis device 20 according to the fluctuation of the renewable energy power generation amount of the power generation device 30 and the fluctuation of the electricity rate of the power system 50,
  • the product can be produced at a lower production cost than the amount predetermined by the usage plan.
  • a more specific configuration example of such a planning device 40 will be described below.
  • the planning device 40 includes a first model generation unit 200, a first model updating unit 210, and a power generation amount prediction unit 220, and predicts a renewable energy power generation amount of the power generation device 30 in the future.
  • the planning device 40 includes a second model generation unit 230 and a second model update unit 240, and the operation plan generation unit 150 generates a future operation plan of the electrolytic device 20.
  • the planning device 40 includes a third model generation unit 250, a third model update unit 260, and an electricity rate prediction unit 270, and predicts a future electricity rate of the power system 50.
  • the second factor may include information on the operation of the electrolysis device 20 or the usage plan of the product of the electrolysis device 20.
  • the second factor is at least one of the operation data of the electrolysis device 20, the demand of the product of the electrolysis device 20, and the storage amount of the product of the electrolysis device 20, which are available before the target period of the operation plan to be generated. Including one.
  • the second factor may include an operation plan of the electrolysis device 20 generated by the planning device 40 in the past.
  • the second factor may include virtual data calculated from a physical model of the electrolysis device 20.
  • the second factor may include information obtained by the obtaining unit 100 from the electrolysis device 20.
  • the operation data may be a product generation amount per unit time in the electrolysis device 20 and / or a product generation amount per unit power.
  • the demand may be the amount of product that needs to be supplied by the electrolyzer 20.
  • the storage amount may be an amount stored in a tank or the like that stores a product of the electrolysis device 20.
  • the third factor may include information that affects fluctuations in the electricity rate of the power system 50.
  • the third factor is the electricity price, the power demand, the power supply, the renewable energy power generation, the predicted value of the renewable energy power generation, and the weather information in the power system 50 before the target period for predicting the electricity price.
  • At least one of The electricity rate in the power system 50 is the actual electricity rate for the power supplied from the power system 50 to the electrolysis device 20 at the place where the electrolysis device 20 is installed, and / or the power generation device 30 sells the power to the power system 50. It may be the power sale price at the time.
  • the weather information may be weather information of a region where the electrolysis device 20 is installed.
  • the information of the first factor, the second factor, and the third factor may be time-series information at approximately constant time intervals.
  • the information of the first factor, the second factor, and the third factor may be added or updated over time, respectively.
  • the acquisition unit 100 may acquire and update each piece of information at predetermined intervals.
  • the acquisition unit 100 may acquire the information at substantially the same or different periods according to the information to be acquired, and may add or update each.
  • the information of the first factor, the second factor, and the third factor may include information supplied from an external device or the like.
  • the first model generation unit 200 is connected to the first model update unit 210.
  • the first model generation unit 200 generates a power generation prediction model that predicts a change in the amount of renewable energy generated by the power generation device 30 in the target period based on a value of a first factor available before the target period. .
  • the first model generation unit 200 may generate a power generation amount prediction model by a process called pre-learning or off-line learning using information past the target period.
  • the first model generation unit 200 generates a power generation prediction model using, for example, regression analysis, Bayesian inference, a neural network, a Gaussian mixture model, a hidden Markov model, or the like.
  • the first model generation unit 200 supplies the generated power generation prediction model to the first model update unit 210 as a first model.
  • the first model updating unit 210 is connected to the power generation amount prediction unit 220.
  • the first model updating unit 210 updates the power generation prediction model by learning based on the value of the first factor in the past period and the actual change in the amount of renewable energy generated in the past period.
  • the first model updating unit 210 has a first model learning unit 215, and updates the power generation amount prediction model according to the learning result of the first model learning unit 215.
  • the first model update unit 210 may update the power generation amount prediction model learned by the first model learning unit 215 as a new power generation amount prediction model every predetermined first update period.
  • the first model updating unit 210 may respond to various conditions such as that the first model learning unit 215 has learned a predetermined number of times, or that an error difference due to learning is below a predetermined threshold.
  • the power generation amount prediction model may be updated.
  • the power generation amount prediction unit 220 is connected to the operation plan generation unit 150.
  • the power generation amount prediction unit 220 predicts a future transition of renewable energy power generation of the power generation device 30 using a power generation amount prediction model.
  • the power generation amount prediction unit 220 predicts, for example, for each predetermined period, the amount of renewable energy generated by the power generation device 30 in the future for the predetermined period.
  • the power generation amount prediction unit 220 predicts the power generation amount using the power generation amount prediction model and the information of the first factor.
  • the power generation amount prediction unit 220 applies, for example, the information of the first factor in the period immediately before the period in which the renewable energy power generation is to be predicted to the power generation prediction model to calculate the renewable energy power generation Predict.
  • the power generation amount prediction unit 220 supplies the prediction result to the operation plan generation unit 150.
  • the second model generation unit 230 is connected to the second model update unit 240.
  • the second model generation unit 230 generates the operation plan generation model based on the value of the second factor before the target period, the transition of the amount of renewable energy generated by the power generation device 30, and the transition of the electricity rate of the power system 50. Generate.
  • the operation plan generation model by learning, calculates the operation plan in the target period, the value of the second factor available before the target period, and the prediction result of the transition of the renewable energy power generation amount of the power generator 30 in the target period. , A model generated based on the power rate of the power system 50.
  • the second model generation unit 230 may use the value of the first factor as the past transition of renewable energy power generation, and may use the value of the third factor as the past transition of electricity rates.
  • the second model generation unit 230 may generate an operation plan generation model by a learning process called pre-learning or offline learning using information past the target period.
  • the second model generating unit 230 generates an operation plan by executing reinforcement learning using an arbitrary machine learning model such as a regression analysis, Bayesian inference, a neural network, a Gaussian mixture model, and a hidden Markov model as an identification model. Generate a model.
  • the second model generation unit 230 supplies the generated operation plan generation model as a second model to the second model update unit 240.
  • the second model updating unit 240 is connected to the operation plan generating unit 150.
  • the second model updating unit 240 calculates the value of the second factor in the past period and the prediction result of the transition of the renewable energy generation amount of the power generation device 30 or the transition of the renewable energy generation amount of the power generation device 30 after the past period,
  • the operation plan generation model is updated by learning based on the transition of the electricity rate of the power system 50 after the past period and the operation plan of the electrolytic device 20 to be targeted after the past period.
  • the second model updating unit 240 has a second model learning unit 245, and updates the operation plan generation model according to the learning result of the second model learning unit 245.
  • the second model updating unit 240 may update the operation plan generation model learned by the second model learning unit 245 as a new operation plan generation model every predetermined second update period.
  • the second model updating unit 240 may update the operation plan generation model in response to the second model learning unit 245 learning a predetermined number of times.
  • the second model learning unit 245 may learn the operation plan generation model by a process called adaptive learning or online learning.
  • the second model learning unit 245 executes an operation plan generation by executing reinforcement learning using an arbitrary machine learning model such as a regression analysis, Bayesian inference, a neural network, a Gaussian mixture model, and a hidden Markov model as an identification model. Learn the model. By performing such machine learning, the second model learning unit 245 can predict a value corresponding to the second factor with an accuracy corresponding to a model to be applied, using the second factor as an input. .
  • the second model learning unit 245 further learns using information that is temporally later than the information of the second factor used by the second model generation unit 230 to generate the operation plan generation model.
  • the second model learning unit 245 includes information on the first factor updated by the change in the actual amount of renewable energy generated by the power generation device 30, information on the second factor updated by the actual operation of the electrolysis device 20,
  • the operation plan generation model is learned using the information on the third factor updated according to the actual transition of the electricity rate.
  • the prediction result of the power generation amount prediction unit 220 may be used instead of the actual renewable energy power generation amount transition.
  • the prediction result of the electricity rate prediction unit 270 may be used instead of the transition of the actual electricity rate. That is, the second model learning unit 245 calculates the value of the second factor in the past period and the transition result of the renewable energy generation amount and the electricity rate or the prediction result of the transition of the renewable energy generation amount and the electricity rate in the past period and thereafter. Based on, the operation plan generation model is learned.
  • the second model learning unit 245 may execute the learning of the operation plan generation model in accordance with the update of the information of the second factor.
  • the second model learning unit 245 executes learning one or more times during the second update period of the second model update unit 240.
  • the second model updating unit 240 supplies the updated operation plan generation model to the operation plan generation unit 150.
  • the third model generation unit 250 is connected to the third model update unit 260.
  • the third model generation unit 250 generates an electricity rate prediction model that predicts a change in the electricity rate of the power system 50 in the target period based on a value of a third factor available before the target period.
  • the third model generation unit 250 may generate an electricity price prediction model by using a process called pre-learning or off-line learning using information past the target period.
  • the third model generation unit 250 generates an electricity bill prediction model using, for example, regression analysis, Bayesian inference, a neural network, a Gaussian mixture model, a hidden Markov model, or the like.
  • the third model generation unit 250 supplies the generated electricity price prediction model as a third model to the third model update unit 260.
  • the third model learning unit 265 may learn the electricity bill prediction model by a process called adaptive learning or online learning.
  • the third model learning unit 265 executes, for example, an electric charge prediction by executing reinforcement learning using an arbitrary machine learning model such as a regression analysis, Bayesian inference, a neural network, a Gaussian mixture model, and a hidden Markov model as an identification model. Learn the model.
  • the third model learning unit 265 can input the third factor and predict the electricity rate according to the third factor with accuracy according to the model to be applied. Become. Further, if a model having LSTM (Long short-term memory), RNN (Recurrent Neural Network), and other storages is used as the third model, for example, the electricity rate can be predicted from the time series of the third factor. Can also.
  • LSTM Long short-term memory
  • RNN Recurrent Neural Network
  • the third model learning unit 265 learns further using information that is temporally later than the information of the third factor used by the third model generation unit 250 to generate the electricity rate prediction model.
  • the third model learning unit 265 learns the electricity rate prediction model using the information on the third factor updated based on the actual transition of the electricity rate of the power system 50.
  • the third model learning unit 265 may execute the learning of the electricity rate prediction model according to the update of the information of the third factor.
  • the third model learning unit 265 learns the electricity rate prediction model based on the value of the third factor in the past period and the actual transition of the electricity rate after the past period.
  • the third model learning unit 265 performs learning one or more times during the third update period of the third model update unit 260.
  • the third model updating unit 260 supplies the updated electricity bill prediction model to the electricity bill prediction unit 270.
  • the electricity rate prediction unit 270 is connected to the operation plan generation unit 150.
  • the electricity rate prediction unit 270 predicts a future transition of the electricity rate of the power system 50 using the updated electricity rate prediction model.
  • the electricity rate prediction unit 270 for example, predicts the electricity rate in the future for the predetermined period for each predetermined period.
  • the electricity rate prediction unit 270 predicts the electricity rate using the electricity rate prediction model and the information of the third factor.
  • the electricity rate prediction unit 270 predicts the electricity rate by applying, for example, the information of the third factor in the period immediately before the period in which the electricity rate is to be predicted to the electricity rate prediction model.
  • the electricity rate prediction unit 270 supplies the prediction result to the operation plan generation unit 150.
  • the operation plan generation unit 150 determines the electrolysis device 20 for the first period in the future based on the predicted change in the amount of renewable energy generated by the power generation device 30 and the predicted electricity rate of the power system 50 in the future.
  • the operation plan of the electrolysis device 20 that satisfies the use plan of the product of the above is generated.
  • the operation plan generation unit 150 may generate an operation plan of the electrolysis device 20 in the future first period by using the operation plan generation model.
  • the operation plan generation unit 150 uses the electrolysis device 20 by giving priority to the power from the power generation device 30 over the power from the power system 50 so as to satisfy the usage plan of the product of the electrolysis device 20.
  • An operation plan to be operated may be generated.
  • the operation plan generation unit 150 generates an operation plan of the electrolysis apparatus 20 with a period such as several days or ten and several days, one or several weeks as a first period.
  • the operation plan generation unit 150 generates an operation plan for N days, for example.
  • the control unit 160 operates the electrolysis device 20 according to the operation plan generated by the operation plan generation unit 150.
  • the control unit 160 may control the power generation device 30, the amount of power to be supplied from the power generation device 30 to the electrolysis device 20, the time period to be supplied, the amount of power to sell the power generated by the power generation device 30, and / or the power sale.
  • the time period of the operation may be controlled.
  • the planning device 40 generates the operation plan of the electrolysis device 20 while predicting the amount of renewable energy generated by the power generation device 30 and the electricity rate of the power system 50 by learning. The operation of the planning device 40 will be described below.
  • FIG. 3 shows an example of an operation flow of the planning device 40 according to the present embodiment.
  • the planning device 40 may execute the operation flow illustrated in FIG. 3 to operate the electrolysis device 20.
  • the acquisition unit 100 acquires information on the amount of renewable energy generated by the power generation device 30, the electricity rate of the power system 50, and the first factor, the second factor, and the third factor that are past trends of the electrolysis device 20 ( S310).
  • the acquisition unit 100 acquires, for example, information on a first factor, a second factor, and a third factor from time t0 to time t1.
  • the period from time t0 to time t1 is a second period before the first period.
  • the acquiring unit 100 causes the storage unit 110 to store the acquired information of the first factor, the second factor, and the third factor. Further, the acquisition unit 100 may directly supply the information of the first factor, the second factor, and the third factor to the model generation unit 120.
  • the model generation unit 120 generates a learning model (S320).
  • the model generation unit 120 generates a learning model based on the values of the first factor, the second factor, and the third factor in the second period.
  • the first model generation unit 200 generates a power generation amount prediction model using a value of a first factor including at least one of the amount of renewable energy generated by the power generation device 30 and weather information in the second period.
  • the third model generation unit 250 includes a third model that includes at least one of the electricity rate, the power demand, the power supply, the renewable energy generation, the predicted value of the renewable energy generation, and the weather information in the second period. Using the values of the factors, an electricity price prediction model is generated.
  • the second model generation unit 230 generates an operation plan generation model based on the values of the first factor, the second factor, and the third factor. For example, the second model generation unit 230 outputs the renewable energy power generation amount of the power generation device 30, the electricity rate of the power system 50, the operation data of the electrolysis device 20, the storage amount of the product of the electrolysis device 20, and the second period.
  • An operation plan generation model is generated using at least one of the virtual data of the operation plan of the electrolysis device 20.
  • the second model generation unit 230 sets virtual data based on the physical model of the electrolysis device 20 as prediction data to be targeted, and compares the prediction data with actual data obtained by operation of the electrolysis device 20 in the past.
  • an operation plan generation model may be generated.
  • the second model generation unit 230 executes the reinforcement learning to generate the operation plan generation model so that the difference between the target predicted data and the past actual data is 0 or less than a predetermined value. .
  • the second model generation unit 230 sets the period of M days in the second period as a virtual prediction period, for example.
  • the M days may be, for example, a period of several days, ten and several days, one or several weeks. It is desirable that M days coincide with the first period (N days).
  • the second model generation unit 230 calculates the prediction result of the operation operation in the prediction period based on the values of the first factor, the second factor, and the third factor in the period before the prediction period in the second period, The reinforcement learning is performed so that an error between the actual data and the virtual data is minimized.
  • the second model generation unit 230 satisfies the usage plan of the product of the electrolysis device 20 (first condition), and uses the electricity from the power generation device 30 in preference to the electricity from the power system 50 to use the electrolysis device.
  • the reinforcement learning may be performed so as to reduce the operation cost while satisfying the condition of operating 20 (second condition).
  • the first condition that the use plan of the product of the electrolysis device 20 is satisfied is, for example, a range in which the storage amount of the product in the electrolysis device 20 varies from 0 to the maximum storage amount, or The condition may be such that a product equal to or more than the total demand or supply amount of the product is generated in the electrolytic device 20.
  • the second condition is that the total amount of renewable energy generated by the power generation device 30 predicted in a predetermined period (for example, the second period) is used for the operation of the electrolysis device 20 or the power generation device 30
  • the condition may be such that more electric power is supplied for the operation of the electrolysis device 20.
  • the generation of the learning model by the model generation unit 120 may be executed before the planning device 40 acquires the actual data of the electrolysis device 20 with the operation of the electrolysis device 20.
  • the first model learning unit 215 adaptively learns the power generation prediction model based on the value of the first factor.
  • the first model learning unit 215 may adaptively learn the power generation prediction model using at least one of the renewable energy power generation amount of the power generation device 30 and weather information in the third period.
  • the first model learning unit 215 performs reinforcement learning so that the result of predicting the amount of renewable energy generated in the third period using the generated amount prediction model matches the obtained amount of renewable energy generated in the third period. Good.
  • the first model learning unit 215 sets the period of M days in the third period as a virtual prediction period, for example.
  • the M days may be, for example, a period of several days, ten and several days, one or several weeks. It is desirable that M days coincide with the first period (N days).
  • the first model learning unit 215 determines that the difference between the prediction result of the prediction period based on the value of the first factor in the period before the prediction period in the third period and the actual data of the prediction period is 0 or a predetermined value. Reinforcement learning is performed so as to be less than the set value.
  • the second model learning unit 245 may apply and learn the operation plan generation model based on the values of the first factor, the second factor, and the third factor. For example, in the third period, the second model learning unit 245 calculates the transition of the renewable energy power generation amount of the power generation device 30 or the prediction result of the transition, the transition of the electricity rate of the power system 50 or the prediction result of the transition, the electrolysis device.
  • the operation plan generation model may be learned using at least one of the operation data of the operation plan 20, the storage amount or demand amount of the product of the electrolysis device 20, and the actual data of the operation plan.
  • the second model learning unit 245 determines that the difference between the result of predicting the operation of the electrolysis device 20 in the third period using the operation plan generation model and the acquired actual data in the third period is 0 or less than a predetermined value.
  • the reinforcement learning may be performed so that
  • the second model learning unit 245 sets the period of M days in the third period as a virtual prediction period, for example.
  • the M days may be, for example, a period of several days, ten and several days, one or several weeks. It is desirable that M days coincide with the first period (N days).
  • the second model learning unit 245 calculates the prediction result of the operation operation in the prediction period based on the values of the first factor, the second factor, and the third factor in the period before the prediction period in the third period, and The reinforcement learning is performed such that the difference between the actual data and the actual data becomes 0 or less than a predetermined value.
  • the second model learning unit 245 may similarly use the first condition, the second condition, and the like used by the second model generation unit 230 to generate the operation plan generation model. That is, the second model learning unit 245 may perform the reinforcement learning of the operation plan generation model so as to reduce the production cost of the product while satisfying the two conditions.
  • the third model learning unit 265 sets the period of M days in the third period as a virtual prediction period, for example.
  • the M days may be, for example, a period of several days, ten and several days, one or several weeks. It is desirable that M days coincide with the first period (N days).
  • the third model learning unit 265 determines that the difference between the prediction result of the prediction period based on the value of the third factor in the period before the prediction period in the third period and the actual data of the prediction period is 0 or a predetermined value. Reinforcement learning is performed so as to be less than the set value.
  • the learning processing unit 130 updates the learning model (S340).
  • the learning processing unit 130 may update the learning model every predetermined time. For example, the learning processing unit 130 performs the first update of the learning model after continuing the adaptive learning for an initial update period necessary for the update after the start of the adaptive learning, and thereafter updates the learning model at regular intervals. repeat.
  • the initial update period is N days or more, which is the period of the operation plan to be generated.
  • the fixed period for repeating the update may be several hours, ten and several hours, one day, several tens of hours, or several days.
  • the prediction unit 140 predicts the amount of renewable energy generated by the power generation device 30 and the electricity rate of the power system 50 using the updated learning model (S350).
  • the power generation amount prediction unit 220 predicts the transition of the renewable energy power generation amount of the power generation device 30 in the first period using the updated power generation amount prediction model and the value of the first factor.
  • the power generation amount prediction unit 220 applies the value of the first factor for N days acquired by the acquisition unit 100 during the initial update period to the power generation amount prediction model, and reproduces N days after the initial update period. Forecast the transition of available energy generation.
  • the electricity rate prediction unit 270 predicts the transition of the electricity rate in the first period using the updated electricity rate prediction model and the value of the third factor. As an example, the electricity rate prediction unit 270 applies the value of the third factor for N days acquired by the acquisition unit 100 during the initial update period to the electricity rate prediction model, and generates electricity for N days after the initial update period. Predict changes in rates.
  • the operation plan generation unit 150 may similarly use the first condition and the second condition used by the second model generation unit 230 to generate the operation plan generation model. That is, the operation plan generation unit 150 may generate an operation plan that minimizes the manufacturing cost while satisfying the two conditions.
  • the operation plan generation unit 150 may generate an operation plan including a period during which the electrolysis device 20 is operated and a period during which the electrolysis device 20 is not operated in the first period. Further, the operation plan generation unit 150 may generate an operation plan indicating a period during which the electrolysis apparatus 20 is operated, together with an operation rate. It is desirable that the operation plan generation unit 150 generates an operation plan in which the operation rate changes in a time series. The operation plan generation unit 150 generates, for example, an operation plan for each fixed time. The operation plan generation unit 150 may generate an operation plan every tens of minutes, every hour, or every several hours.
  • the operation plan generator 150 may generate an operation plan for each of the plurality of electrolyzers 20.
  • the operation plan generator 150 may generate substantially the same operation plan.
  • the operation plan generation unit 150 controls the control unit 160 to control different types of electrolyzers 20, electrolyzers 20 purchased at different times, electrolyzers 20 of different manufacturers, or a plurality of electrolyzers 20 including a combination thereof. In this case, different operation plans may be generated for each of the electrolyzers 20.
  • the second model generation unit 230 generates one operation plan generation model corresponding to the plurality of electrolyzers 20, and the second model update unit 240 generates the operation plan generation model learned by the second model learning unit 245. May be updated.
  • the operation plan generation model may be a model that generates an operation plan for cooperatively operating the plurality of electrolyzers 20.
  • the operation start timing and the operation period of each of the plurality of electrolyzers 20 may be used.
  • the process returns to S330, and the learning processing unit 130 adaptively learns the learning model.
  • the acquiring unit 100 sequentially acquires the information of the first factor and the third factor in the first period and the information of the second factor that changes due to the operation of the electrolysis device 20 in the first period, and stores the information in the storage unit 110.
  • the planning device 40 includes the information of the first period in the past information, and sets the target period to be predicted to be a period later than the first period (for example, a fourth period).
  • the planning device 40 repeats the adaptive learning of the model using the information of the first period, updates the model as the certain period elapses, generates an operation plan of the electrolysis device 20 for the fourth period, and generates The electrolysis device 20 is operated according to the operation plan thus set.
  • the planning device 40 continues the electrolysis device 20 while updating the learning model by repeatedly generating the operation plan of the electrolysis device 20 for the target period and operating the target period. Can operate.
  • the planning device 40 is operated in chronological order in the second period, the third period, the first period, and the fourth period.
  • the second period, the third period, the first period, and the fourth period may be temporally continuous in this order. It is preferable that at least the first period and the fourth period are continuous periods.
  • the planning device 40 predicts the amount of renewable energy generated by the power generation device 30 and the electricity rate of the power system 50 by learning, and generates a quantity of products according to the usage plan in the electrolysis device 20 at low cost. Create a work plan that can be generated.
  • Programmable circuits include logical AND, logical OR, logical XOR, logical NAND, logical NOR, and other logical operations, memory elements such as flip-flops, registers, field programmable gate arrays (FPGAs), programmable logic arrays (PLAs), etc. And the like, and may include reconfigurable hardware circuits.
  • Computer readable media may include any tangible device capable of storing instructions for execution by a suitable device, such that computer readable media having instructions stored thereon is specified in a flowchart or block diagram.
  • Product comprising instructions that can be executed to create a means for performing the specified operation.
  • Examples of the computer readable medium may include an electronic storage medium, a magnetic storage medium, an optical storage medium, an electromagnetic storage medium, a semiconductor storage medium, and the like.
  • the computer readable instructions may be provided to a processor or programmable circuit of a general purpose computer, special purpose computer, or other programmable data processing device, either locally or over a wide area network (WAN) such as a local area network (LAN), the Internet, or the like. ) May be executed to create means for performing the operations specified in the flowcharts or block diagrams.
  • WAN wide area network
  • LAN local area network
  • processors include computer processors, processing units, microprocessors, digital signal processors, controllers, microcontrollers, and the like.
  • FIG. 4 illustrates an example of a computer 1900 in which aspects of the present invention may be wholly or partially implemented.
  • the programs installed on the computer 1900 can cause the computer 1900 to function as one or more sections of the operation or the device associated with the device according to the embodiment of the present invention, or the operation or the one or more of the one or more devices. Sections may be executed and / or computer 1900 may execute a process or steps of a process according to an embodiment of the present invention.
  • Such programs may be executed by CPU 2000 to cause computer 1900 to perform certain operations associated with some or all of the blocks in the flowcharts and block diagrams described herein.
  • the input / output controller 2084 connects the host controller 2082 to the communication interface 2030, the hard disk drive 2040, and the DVD drive 2060, which are relatively high-speed input / output devices.
  • the communication interface 2030 communicates with another device via a network by wire or wirelessly.
  • the communication interface functions as hardware for performing communication.
  • the hard disk drive 2040 stores programs and data used by the CPU 2000 in the computer 1900.
  • the DVD drive 2060 reads a program or data from the DVD 2095 and provides it to the hard disk drive 2040 via the RAM 2020.
  • the ROM 2010, the flash memory drive 2050, and the relatively low-speed input / output device of the input / output chip 2070 are connected to the input / output controller 2084.
  • the ROM 2010 stores a boot program executed by the computer 1900 at the time of startup, and / or a program depending on hardware of the computer 1900.
  • the flash memory drive 2050 reads a program or data from the flash memory 2090 and provides it to the hard disk drive 2040 via the RAM 2020.
  • the input / output chip 2070 connects the flash memory drive 2050 to the input / output controller 2084 and inputs / outputs various input / output devices via, for example, a parallel port, a serial port, a keyboard port, a mouse port, and the like. Connect to controller 2084.
  • the program provided to the hard disk drive 2040 via the RAM 2020 is stored in a recording medium such as a flash memory 2090, a DVD 2095, or an IC card and provided by a user.
  • the program is read from the recording medium, installed on the hard disk drive 2040 in the computer 1900 via the RAM 2020, and executed by the CPU 2000.
  • the information processing described in these programs is read by the computer 1900 and provides for cooperation between the software and the various types of hardware resources described above.
  • An apparatus or method may be configured for implementing manipulation or processing of information according to the use of computer 1900.
  • the communication interface 2030 may transfer the transmission / reception data to / from the storage device by the DMA (Direct Memory Access) method, and instead, the CPU 2000 may use the transfer source storage device or the communication interface 2030.
  • the data may be read from the communication interface 2030 or the data may be written to the communication interface 2030 or the storage device of the transfer destination to transfer the transmission and reception data.
  • DMA Direct Memory Access
  • the CPU 2000 transfers all or a necessary portion from a file or a database stored in an external storage device such as a hard disk drive 2040, a DVD drive 2060 (DVD 2095), or a flash memory drive 2050 (flash memory 2090) to a DMA.
  • the data is read into the RAM 2020 by transfer or the like, and various processes are performed on the data on the RAM 2020.
  • the CPU 2000 writes the processed data back to the external storage device by DMA transfer or the like.
  • the RAM 2020 can be regarded as temporarily holding the contents of the external storage device. Therefore, in this embodiment, the RAM 2020 and the external storage device are collectively referred to as a memory, a storage unit, or a storage device.
  • Various information such as various programs, data, tables, and databases in the present embodiment are stored on such a storage device and are subjected to information processing.
  • the CPU 2000 can also hold a part of the RAM 2020 in a cache memory and perform reading and writing on the cache memory. Even in such a form, the cache memory plays a part of the function of the RAM 2020. Therefore, in the present embodiment, the cache memory is also included in the RAM 2020, the memory, and / or the storage device unless otherwise indicated. I do.
  • the CPU 2000 performs various calculations, information processing, condition determination, information search / replacement, and the like described in the present embodiment on the data read from the RAM 2020, as specified by the instruction sequence of the program. And write it back to the RAM 2020.
  • the CPU 2000 determines whether the various variables described in the present embodiment satisfy conditions such as larger, smaller, greater than, less than, equal to, and the like as compared with other variables or constants. Then, if the condition is satisfied (or not satisfied), a branch is made to a different instruction sequence or a subroutine is called.
  • the CPU 2000 can search for information stored in a file or a database in the storage device. For example, in the case where a plurality of entries in which the attribute value of the second attribute is associated with the attribute value of the first attribute are stored in the storage device, the CPU 2000 determines whether the plurality of entries stored in the storage device By searching for an entry in which the attribute value of the first attribute matches the specified condition and reading the attribute value of the second attribute stored in the entry, the entry is associated with the first attribute satisfying the predetermined condition. The attribute value of the obtained second attribute can be obtained.
  • elements other than the listed elements may be used. For example, if "X performs Y using A, B, and C", X may perform Y using D in addition to A, B, and C.
  • System 20 Electrolysis device 30 Power generation device 40 Planning device 50 Power system 100 Acquisition unit 110 Storage unit 120 Model generation unit 130 Learning processing unit 140 Prediction unit 150 Operation plan generation unit 160 Control unit 200 First model generation unit 210 First model update Unit 215 first model learning unit 220 power generation prediction unit 230 second model generation unit 240 second model update unit 245 second model learning unit 250 third model generation unit 260 third model update unit 265 third model learning unit 270 electricity Charge prediction unit 1900 Computer 2000 CPU 2010 ROM 2020 RAM 2030 Communication interface 2040 Hard disk drive 2050 Flash memory drive 2060 DVD drive 2070 Input / output chip 2075 Graphic controller 2080 Display device 2082 Host controller 2084 Input / output controller 2090 Flash memory 2095 DVD

Landscapes

  • Business, Economics & Management (AREA)
  • Health & Medical Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Economics (AREA)
  • Public Health (AREA)
  • Water Supply & Treatment (AREA)
  • General Health & Medical Sciences (AREA)
  • Human Resources & Organizations (AREA)
  • Marketing (AREA)
  • Primary Health Care (AREA)
  • Strategic Management (AREA)
  • Tourism & Hospitality (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Supply And Distribution Of Alternating Current (AREA)

Abstract

It is necessary to reduce costs for the operating power of an electrolysis apparatus. Provided is a planning apparatus comprising: a power generation amount prediction unit that predicts a trend for the future amount of renewable energy generation by using a power generation prediction model that predicts the trend in the amount of renewable energy generation of a power generation apparatus during a target period on the basis of the value of a first factor available before the target period; and an operation plan generation unit that generates an operation plan for the electrolysis apparatus that satisfies a use plan for a product of the electrolysis apparatus in a future first period, on the basis of power system electricity rates and the predicted trend of the amount of future renewable energy generation.

Description

計画装置、方法、およびプログラムPlanning device, method, and program
　本発明は、計画装置、方法、およびプログラムに関する。 The present invention relates to a planning device, a method, and a program.
　従来、水を電気分解することにより水素を生成する電解装置が知られている。また、電力供給手段として、再生可能エネルギーによる発電を行う発電装置、または天気や電力の供給コストに応じて電気料金が変動する電力系統が知られている。 電解 Conventionally, an electrolysis apparatus that generates hydrogen by electrolyzing water has been known. In addition, as a power supply unit, a power generation device that generates power using renewable energy, or a power system in which an electricity rate fluctuates according to weather or power supply cost is known.
解決しようとする課題Issues to be solved
　このような発電装置および電力系統から電解装置に電力供給を行う場合、稼働電力のコストを低減し、水素の製造コストを低減させることが望ましい。しかしながら、電解装置により供給すべき水素の量が定められている場合等がある。この場合、水素の供給期間に、発電装置の発電量が十分でないと、高い電気料金の時間帯で電力系統からの電力により電解装置を稼働させなければならなくなる。 行 う When power is supplied from such a power generation device and a power system to an electrolysis device, it is desirable to reduce the cost of operating power and reduce the cost of producing hydrogen. However, there are cases where the amount of hydrogen to be supplied by the electrolytic device is determined. In this case, if the power generation amount of the power generation device is not sufficient during the hydrogen supply period, the electrolysis device must be operated by the power from the power system in a time period when the electricity rate is high.
一般的開示General disclosure
　上記課題を解決するために、本発明の第１の態様においては、計画装置を提供する。計画装置は、対象期間における発電装置の再生可能エネルギー発電量の推移を対象期間よりも前に入手可能な第１因子の値に基づいて予測する発電量予測モデルを用いて、将来の再生可能エネルギー発電量の推移を予測する発電量予測部を備えてよい。計画装置は、予測された将来の再生可能エネルギー発電量の推移と電力系統の電気料金とに基づいて、将来の第１期間における、電解装置の生成物の使用計画を満たす電解装置の稼働計画を生成する稼働計画生成部を備えてよい。 た め In order to solve the above problems, a first aspect of the present invention provides a planning device. The planning device uses a power generation forecasting model that predicts a change in the amount of renewable energy generated by the power generating device during the target period based on the value of the first factor that is available before the target period. A power generation amount prediction unit for predicting a transition of the power generation amount may be provided. The planning device prepares an operation plan of the electrolysis device that satisfies the usage plan of the product of the electrolysis device in the first period in the future, based on the predicted change in the amount of generated renewable energy and the electricity rate of the power system. An operation plan generation unit that generates the operation plan may be provided.
　発電量予測モデルは、対象期間よりも前の、発電装置の再生可能エネルギー発電量および天気情報の少なくとも１つを含む第１因子の値に基づいて、対象期間における再生可能エネルギー発電量の推移を予測してよい。 The power generation amount prediction model calculates a transition of the renewable energy power generation amount in the target period based on the value of the first factor including at least one of the renewable energy power generation amount of the power generation device and the weather information before the target period. You can predict.
　計画装置は、過去期間における第１因子の値と過去期間以降の再生可能エネルギー発電量の現実の推移とに基づいて、発電量予測モデルを学習により更新する第１モデル更新部を備えてよい。 The planning device may include a first model updating unit that updates the power generation prediction model by learning based on the value of the first factor in the past period and the actual change in the amount of renewable energy generated in the past period.
　稼働計画生成部は、第１期間中において、電解装置の生成物の使用計画を満たすように、発電装置からの電気を電力系統からの電気より優先して用いて電解装置を稼働させる稼働計画を生成してよい。 The operation plan generation unit, during the first period, in order to satisfy the use plan of the product of the electrolysis device, the operation plan to operate the electrolysis device using the electricity from the power generation device in preference to the electricity from the power system. May be generated.
　電解装置の生成物の使用計画は、第１期間における電解装置の生成物の貯蔵量を基準範囲内に維持する計画、および第１期間における電解装置の生成物の需要量を満たす計画の少なくとも１つを含んでよい。 The use plan of the product of the electrolyzer is at least one of a plan to maintain the storage amount of the product of the electrolyzer in the first period within the reference range and a plan to satisfy the demand for the product of the electrolyzer in the first period. May be included.
　稼働計画生成部は、対象期間における稼働計画を、対象期間よりも前に入手可能な第２因子の値と、対象期間における再生可能エネルギー発電量の推移の予測結果と、電気料金とに基づいて生成する稼働計画生成モデルを用いて、将来の第１期間における電解装置の稼働計画を生成してよい。 The operation plan generation unit calculates the operation plan in the target period based on the value of the second factor available before the target period, the predicted result of the change in the amount of renewable energy generated in the target period, and the electricity rate. The operation plan of the electrolysis apparatus in the first period in the future may be generated using the generated operation plan generation model.
　稼働計画生成モデルは、対象期間よりも前の、電解装置の稼働データ、電解装置の生成物の需要量、および電解装置の生成物の貯蔵量の少なくとも１つを含む第２因子の値と、対象期間における再生可能エネルギー発電量の推移の予測結果と、電気料金とに基づいて、対象期間における電解装置の稼働計画を生成してよい。 The operation plan generation model includes a value of a second factor including at least one of operation data of the electrolysis apparatus, a demand amount of a product of the electrolysis apparatus, and a storage amount of a product of the electrolysis apparatus before the target period, An operation plan of the electrolysis device in the target period may be generated based on the prediction result of the change in the amount of generated renewable energy in the target period and the electricity rate.
　計画装置は、過去期間における第２因子の値と、過去期間以降における再生可能エネルギー発電量の推移または再生可能エネルギー発電量の推移の予測結果と、過去期間以降における電気料金の推移と、過去期間以降において目標とすべき電解装置の稼働計画とに基づいて、稼働計画生成モデルを学習により更新する第２モデル更新部を備えてよい。 The planning device calculates the value of the second factor in the past period, the change in the amount of renewable energy generated in the past period or the predicted result of the change in the amount of renewable energy generated in the past period, the change in the electricity rate after the past period, and the change in the past period. A second model updating unit that updates the operation plan generation model by learning based on the target operation plan of the electrolysis apparatus may be provided.
　計画装置は、対象期間における電力系統の電気料金の推移を、対象期間よりも前に入手可能な第３因子の値に基づいて予測する電気料金予測モデルを用いて、将来の電気料金の推移を予測する電気料金予測部を備えてよい。稼働計画生成部は、予測された将来の再生可能エネルギー発電量の推移と、予測された将来の電気料金の推移とに基づいて、将来の第１期間における、電解装置の生成物の使用計画を満たす電解装置の稼働計画を生成してよい。 The planning device uses a power rate prediction model that predicts a change in the power rate of the power system in the target period based on the value of the third factor available before the target period, and calculates a future change in the power rate. An electricity rate prediction unit for predicting the electricity rate may be provided. The operation plan generation unit generates a use plan of the product of the electrolytic device in the first period in the future based on the predicted future change in the amount of renewable energy generated and the predicted future change in the electricity rate. An operation plan for the electrolyzer to be filled may be generated.
　電気料金予測モデルは、対象期間よりも前の、電力系統における電気料金、電力需要量、電力供給量、再生可能エネルギー発電量、再生可能エネルギー発電量の予測値、および天気情報の少なくとも１つを含む第３因子の値に基づいて、対象期間における電力系統の電気料金の推移を予測してよい。 The electricity rate prediction model calculates at least one of the electricity rate, the power demand, the power supply, the renewable energy generation, the predicted value of the renewable energy generation, and the weather information in the power system before the target period. Based on the value of the third factor included, the transition of the electricity rate of the power system in the target period may be predicted.
　本発明の第２の態様においては、方法を提供する。方法は、対象期間における発電装置の再生可能エネルギー発電量の推移を対象期間よりも前に入手可能な第１因子の値に基づいて予測する発電量予測モデルを用いて、将来の再生可能エネルギー発電量の推移を予測する段階を備えてよい。方法は、予測された将来の再生可能エネルギー発電量の推移と電力系統の電気料金とに基づいて、将来の第１期間における、電解装置の生成物の使用計画を満たす電解装置の稼働計画を生成する段階を備えてよい。 に お い て In a second aspect of the present invention, a method is provided. The method includes using a power generation forecasting model that predicts a change in the amount of renewable energy generated by a power generator during a target period based on a value of a first factor that is available before the target period. A step of predicting a change in the amount may be provided. The method generates an operation plan of an electrolyzer that satisfies an electrolyzer product usage plan for a first period in the future, based on a predicted future change in renewable energy generation and an electricity rate of a power system. The method may include a step of:
　本発明の第３の態様においては、コンピュータに、本発明の第１の態様の計画装置として機能させるプログラムを提供する。 In the third aspect of the present invention, there is provided a program for causing a computer to function as the planning device of the first aspect of the present invention.
　なお、上記の発明の概要は、本発明の必要な特徴の全てを列挙したものではない。また、これらの特徴群のサブコンビネーションもまた、発明となりうる。 The summary of the invention described above does not enumerate all necessary features of the invention. Further, a sub-combination of these feature groups can also be an invention.
本実施形態に係るシステム１０を示す。1 shows a system 10 according to the present embodiment. 本実施形態に係る計画装置４０の詳細な構成例を示す。4 shows a detailed configuration example of a planning device 40 according to the present embodiment. 本実施形態に係る計画装置４０の動作フローの一例を示す。5 shows an example of an operation flow of the planning device 40 according to the present embodiment. 本実施形態の複数の態様が全体的または部分的に具現化されうるコンピュータ１９００の例を示す。14 illustrates an example of a computer 1900 in which aspects of the embodiments can be wholly or partially embodied.
　以下、発明の実施の形態を通じて本発明を説明するが、以下の実施形態は請求の範囲にかかる発明を限定するものではない。また、実施形態の中で説明されている特徴の組み合わせの全てが発明の解決手段に必須であるとは限らない。 Hereinafter, the present invention will be described through embodiments of the invention, but the following embodiments do not limit the invention according to the claims. In addition, not all combinations of the features described in the embodiments are necessarily essential to the solution of the invention.
　図１は、本実施形態に係るシステム１０を示す。システム１０は、電気料金および再生可能エネルギーの発電量に応じた電解装置２０の稼働計画を生成し、稼働計画に従って電解装置２０を稼働させる。システム１０は、電解装置２０と、発電装置３０と、計画装置４０とを備える。 FIG. 1 shows a system 10 according to the present embodiment. The system 10 generates an operation plan of the electrolysis device 20 according to the electricity rate and the amount of generated renewable energy, and operates the electrolysis device 20 according to the operation plan. The system 10 includes an electrolysis device 20, a power generation device 30, and a planning device 40.
　電解装置２０は、発電装置３０と計画装置４０と電力系統５０とに接続される。電解装置２０は、電気エネルギーを用いて生成物を生成する装置でよい。電解装置２０は、一例として、生成物として水素を、電気分解によって生成する水素生成装置である。電解装置２０は、計画装置４０が生成した稼働計画に従って稼働する。電解装置２０は、発電装置３０および電力系統５０から電力供給されることによって稼働する。 Electrolysis device 20 is connected to power generation device 30, planning device 40, and power system 50. The electrolytic device 20 may be a device that generates a product using electric energy. The electrolysis device 20 is, for example, a hydrogen generation device that generates hydrogen as a product by electrolysis. The electrolysis device 20 operates according to the operation plan generated by the planning device 40. The electrolysis device 20 operates by being supplied with power from the power generation device 30 and the power system 50.
　ここで、電力系統５０は、一例として、原子力発電、火力発電、または風力等の再生可能エネルギーによる発電等を行う１または複数の発電所から、送電網を介して多数の需要家に電力を供給するシステムである。電力系統５０は、発電量及び需要量に応じて、例えば、所定時間毎、１日毎、または１ヶ月毎等に電気料金（例えば、売電料金および買電料金）が変動しうるものである。 Here, as an example, the power system 50 supplies power to a large number of consumers via a power transmission network from one or a plurality of power plants that generate power using renewable energy such as nuclear power, thermal power, or wind power. System. The electric power system 50 can change an electric power rate (for example, a power sale rate and a power purchase rate) according to a power generation amount and a demand amount, for example, every predetermined time, every day, or every month.
　発電装置３０は、電解装置２０と計画装置４０と電力系統５０とに接続され、風力、太陽光、熱、地熱、水力、および／またはバイオマス等のような再生可能エネルギーを用いて発電した電力を電解装置２０に供給するローカルな電源であってよい。発電装置３０は、電解装置２０に電力系統５０を介さずに直接、および／または電力系統５０の送電網を介して、電解装置２０に電力供給してよい。発電装置３０は、発電した電力を電力系統５０に売電してよい。 The power generation device 30 is connected to the electrolysis device 20, the planning device 40, and the power system 50, and generates electric power generated using renewable energy such as wind, solar, heat, geothermal, hydro, and / or biomass. It may be a local power supply to be supplied to the electrolysis device 20. The power generation device 30 may supply power to the electrolysis device 20 directly to the electrolysis device 20 without passing through the power system 50 and / or via the power grid of the power system 50. The power generation device 30 may sell the generated power to the power system 50.
　計画装置４０は、電解装置２０の稼働計画を生成する。計画装置４０は、パーソナルコンピュータ、タブレット型コンピュータ、スマートフォン、ワークステーション、サーバコンピュータ、または汎用コンピュータ等のコンピュータであってよく、複数のコンピュータが接続されたコンピュータシステムであってもよい。計画装置４０は、コンピュータのＣＰＵ、ＧＰＵ（Ｇｒａｐｈｉｃｓ　Ｐｒｏｃｅｓｓｉｎｇ　Ｕｎｉｔ）、および／またはＴＰＵ（Ｔｅｎｓｏｒ　Ｐｒｏｃｅｓｓｉｎｇ　Ｕｎｉｔ）における処理によって稼働計画を生成してよい。また、計画装置４０は、サーバコンピュータにより提供されるクラウド上で各種の処理を行うものであってよい。 The planning device 40 generates an operation plan for the electrolysis device 20. The planning device 40 may be a computer such as a personal computer, a tablet computer, a smartphone, a workstation, a server computer, or a general-purpose computer, or may be a computer system to which a plurality of computers are connected. The planning device 40 may generate the operation plan by a process in a CPU, a GPU (Graphics Processing Unit), and / or a TPU (Tensor Processing Unit) of the computer. The planning device 40 may perform various processes on a cloud provided by a server computer.
　計画装置４０が生成する稼働計画は、第１期間における、電解装置２０の稼働すべき状態を記述したテーブルまたはデータ等であってよい。稼働計画は、例えば、電解装置２０を稼働させる（または稼働させない）時間帯、電解装置２０の稼働時間帯のうち、発電装置３０からの電力で稼働させる時間帯と電力系統５０からの電力で稼働させる時間帯、発電装置３０の発電した電力を電力系統５０に売電する時間帯、および／または、電解装置２０の時間帯毎の稼働率を定めたテーブルまたはデータ等でよい。計画装置４０は、取得部１００と、記憶部１１０と、モデル生成部１２０と、学習処理部１３０と、予測部１４０と、稼働計画生成部１５０と、制御部１６０とを備える。 The operation plan generated by the planning device 40 may be a table or data describing a state in which the electrolysis device 20 should be operated in the first period. The operation plan includes, for example, a time period in which the electrolysis device 20 is operated (or not operated), a time period in which the electrolysis device 20 is operated, and a time period in which the electrolysis device 20 is operated with the power from the power generation device 30 and the operation from the power system 50. It may be a table or data or the like that defines a time zone for powering, a time zone for selling the power generated by the power generation device 30 to the power system 50, and / or an operation rate of the electrolysis device 20 for each time zone. The planning device 40 includes an acquisition unit 100, a storage unit 110, a model generation unit 120, a learning processing unit 130, a prediction unit 140, an operation plan generation unit 150, and a control unit 160.
　取得部１００は、電解装置２０と発電装置３０と記憶部１１０とに接続され、学習に用いるパラメータおよび学習データ等を取得してよい。取得部１００は、ネットワーク等に接続され、当該ネットワークを介してデータを取得してもよい。取得部１００は、取得すべきデータの少なくとも一部が外部のデータベース等に記憶されている場合、当該データベース等にアクセスして、取得してよい。取得部１００は、取得したデータを、記憶部１１０に供給してよい。 The acquisition unit 100 may be connected to the electrolysis device 20, the power generation device 30, and the storage unit 110, and may acquire parameters used for learning, learning data, and the like. The acquisition unit 100 may be connected to a network or the like, and acquire data via the network. When at least a part of the data to be acquired is stored in an external database or the like, the acquiring unit 100 may access the database or the like and acquire the data. The acquisition unit 100 may supply the acquired data to the storage unit 110.
　記憶部１１０は、モデル生成部１２０と学習処理部１３０と予測部１４０と稼働計画生成部１５０とに接続され、取得部１００が取得した情報を記憶する。記憶部１１０は、当該計画装置４０が処理するデータを記憶してよい。記憶部１１０は、計画装置４０が稼働計画を生成する過程で算出する（または利用する）中間データ、算出結果、およびパラメータ等をそれぞれ記憶してもよい。また、記憶部１１０は、計画装置４０内の各部の要求に応じて、記憶したデータを要求元に供給してよい。記憶部１１０は、一例として、モデル生成部１２０の要求に応じて、記憶したデータを当該モデル生成部１２０に供給する。 The storage unit 110 is connected to the model generation unit 120, the learning processing unit 130, the prediction unit 140, and the operation plan generation unit 150, and stores information acquired by the acquisition unit 100. The storage unit 110 may store data processed by the planning device 40. The storage unit 110 may store intermediate data, calculation results, parameters, and the like calculated (or used) in the process of generating the operation plan by the planning device 40. Further, the storage unit 110 may supply the stored data to a request source in response to a request from each unit in the planning device 40. The storage unit 110 supplies the stored data to the model generation unit 120 in response to a request from the model generation unit 120, for example.
　モデル生成部１２０は、学習処理部１３０に接続され、計画装置４０が学習する学習モデルを生成する。モデル生成部１２０は、記憶部１１０に記憶された第１因子、第２因子、および第３因子に応じて、学習モデルを生成する。モデル生成部１２０は、１または複数の学習モデルを生成してよい。モデル生成部１２０は、生成した学習モデルを学習処理部１３０に供給する。 The 生成 model generation unit 120 is connected to the learning processing unit 130, and generates a learning model that the planning device 40 learns. The model generation unit 120 generates a learning model according to the first factor, the second factor, and the third factor stored in the storage unit 110. The model generation unit 120 may generate one or a plurality of learning models. The model generation unit 120 supplies the generated learning model to the learning processing unit 130.
　学習処理部１３０は、予測部１４０に接続され、取得部１００で取得された学習データに基づいて、生成した学習モデルを学習する。学習処理部１３０は、強化学習を実行して、学習モデルを更新してよい。学習処理部１３０は、１または複数の学習モデルを更新してよい。学習処理部１３０は、更新した学習モデルを予測部１４０に供給する。 The learning processing unit 130 is connected to the prediction unit 140 and learns the generated learning model based on the learning data acquired by the acquisition unit 100. The learning processing unit 130 may execute the reinforcement learning to update the learning model. The learning processing unit 130 may update one or a plurality of learning models. The learning processing unit 130 supplies the updated learning model to the prediction unit 140.
　予測部１４０は、稼働計画生成部１５０に接続される。予測部１４０は、学習モデルに基づいて、計画装置４０が稼働計画を生成する際に用いられる将来の各種データの推移を予測する。予測部１４０は、例えば、発電装置３０における将来の再生可能エネルギー発電量の推移、および／または電力系統５０の電気料金の推移を予測する。予測部１４０は、予測したデータを稼働計画生成部１５０に供給する。 The prediction unit 140 is connected to the operation plan generation unit 150. The prediction unit 140 predicts, based on the learning model, changes in various future data used when the planning device 40 generates the operation plan. The prediction unit 140 predicts, for example, a future change in the amount of renewable energy generated in the power generation device 30 and / or a change in the electricity rate of the power system 50. The prediction unit 140 supplies the predicted data to the operation plan generation unit 150.
　稼働計画生成部１５０は、制御部１６０に接続され、将来の第１期間における電解装置２０の稼働計画を生成する。稼働計画生成部１５０は、例えば、将来の第１期間において、電解装置２０の生成物の使用計画を満たしつつ、生成物の製造コストを最小化または低減させる稼働計画を生成する。稼働計画生成部１５０は、生成した稼働計画を制御部１６０に供給する。 The operation plan generation unit 150 is connected to the control unit 160, and generates an operation plan of the electrolysis device 20 in the first period in the future. The operation plan generation unit 150 generates, for example, an operation plan that minimizes or reduces the product manufacturing cost while satisfying the product use plan of the electrolytic device 20 in the first period in the future. The operation plan generation unit 150 supplies the generated operation plan to the control unit 160.
　ここで、電解装置２０の生成物の使用計画は、電解装置２０の生成物の貯蔵量を基準範囲内に維持する計画、および電解装置２０の生成物の需要量または供給量を満たす計画のうちの少なくとも１つを含んでよい。生成物の需要量または供給量は、例えば、第１期間における電解装置２０が供給すべき生成物の累計の量または所定時間毎の量であってよい。貯蔵量は、電解装置２０の内部または外部に貯蔵される生成物の量であってよい。このような貯蔵量の基準範囲、生成物の需要量、または供給量は、計画装置４０に外部から入力されるデータ、過去のデータ、または、計画装置４０において過去のデータから将来の第１期間について予測されるものであってよい。 Here, the use plan of the product of the electrolyzer 20 includes a plan to maintain the storage amount of the product of the electrolyzer 20 within the reference range and a plan to satisfy the demand or supply of the product of the electrolyzer 20. May be included. The demand amount or supply amount of the product may be, for example, the total amount of the product to be supplied by the electrolysis device 20 in the first period or the amount per predetermined time. The storage amount may be the amount of the product stored inside or outside the electrolysis device 20. Such a reference range of the storage amount, the demand amount of the product, or the supply amount is determined based on data input from the outside to the planning device 40, past data, or past data in the planning device 40 in the future first period. May be expected.
　制御部１６０は、第１期間における電解装置２０の稼働計画を用いて、当該電解装置２０を稼働させる制御を行う。例えば、制御部１６０は、発電装置３０からの電力と電力系統５０からの電力とを、稼働計画により時間帯に応じて選択的に用いて、電解装置２０を稼働させてよい。制御部１６０は、複数の電解装置２０をそれぞれ稼働させてよい。また、制御部１６０は、電解装置２０の動作および生成物の貯蔵量等が想定とは異なる範囲となった場合に、電解装置２０の稼働の停止および開始を指示してもよい。制御部１６０は、第１期間における電解装置２０の稼働計画を用いて、発電装置３０の発電した電力を電力系統５０に売電するように制御してよい。 The control unit 160 controls the operation of the electrolysis device 20 using the operation plan of the electrolysis device 20 in the first period. For example, the control unit 160 may operate the electrolysis device 20 by selectively using the electric power from the power generation device 30 and the electric power from the power system 50 according to a time zone according to an operation plan. The control unit 160 may operate each of the plurality of electrolyzers 20. Control unit 160 may instruct stop and start of operation of electrolysis apparatus 20 when the operation of electrolysis apparatus 20, the storage amount of the product, and the like fall outside the expected ranges. The control unit 160 may control the power generated by the power generation device 30 to be sold to the power system 50 using the operation plan of the electrolysis device 20 in the first period.
　以上の本実施形態の計画装置４０によれば、発電装置３０の再生可能エネルギー発電量の変動と電力系統５０の電気料金の変動とに応じた電解装置２０の稼働計画を生成することができ、使用計画により予め定められた量以上の生成物を、より低い製造コストで生成することができる。このような計画装置４０のより具体的な構成例について、次に説明する。 According to the planning device 40 of the present embodiment described above, it is possible to generate the operation plan of the electrolysis device 20 according to the fluctuation of the renewable energy power generation amount of the power generation device 30 and the fluctuation of the electricity rate of the power system 50, The product can be produced at a lower production cost than the amount predetermined by the usage plan. A more specific configuration example of such a planning device 40 will be described below.
　図２は、本実施形態に係る計画装置４０の構成例を示す。図２の計画装置４０において、図１に示された本実施形態に係る計画装置４０の動作と略同一のものには同一の符号を付け、説明を省略する。図２は、図１における計画装置４０のモデル生成部１２０、学習処理部１３０、および予測部１４０をより詳細に示す。 FIG. 2 shows a configuration example of the planning device 40 according to the present embodiment. In the planning device 40 of FIG. 2, those substantially the same as those of the operation of the planning device 40 according to the present embodiment shown in FIG. 1 are denoted by the same reference numerals, and description thereof will be omitted. FIG. 2 shows the model generation unit 120, the learning processing unit 130, and the prediction unit 140 of the planning device 40 in FIG. 1 in more detail.
　計画装置４０は、第１モデル生成部２００、第１モデル更新部２１０、および発電量予測部２２０を備え、将来の発電装置３０の再生可能エネルギー発電量を予測する。計画装置４０は、第２モデル生成部２３０、および第２モデル更新部２４０を備え、稼働計画生成部１５０において将来の電解装置２０の稼働計画を生成する。計画装置４０は、第３モデル生成部２５０、第３モデル更新部２６０、および電気料金予測部２７０を備え、将来の電力系統５０の電気料金を予測する。 The planning device 40 includes a first model generation unit 200, a first model updating unit 210, and a power generation amount prediction unit 220, and predicts a renewable energy power generation amount of the power generation device 30 in the future. The planning device 40 includes a second model generation unit 230 and a second model update unit 240, and the operation plan generation unit 150 generates a future operation plan of the electrolytic device 20. The planning device 40 includes a third model generation unit 250, a third model update unit 260, and an electricity rate prediction unit 270, and predicts a future electricity rate of the power system 50.
　記憶部１１０は、取得部１００が取得した第１因子、第２因子、および第３因子を記憶する。第１因子（発電量予測因子）は、発電装置３０の再生可能エネルギー発電量に影響を及ぼす情報を含んでよい。第１因子は、例えば、発電装置３０の再生可能エネルギー発電量の推移を予測する対象期間よりも前の、発電装置３０の再生可能エネルギー発電量および天気情報の少なくとも１つを含む。第１因子は、発電装置３０から取得部１００が取得した情報を含んでよい。例えば、天気情報は、風速、風向き、晴れ、雨、温度、波の高さ、および日照時間等の少なくとも１つを含んでよい。天気情報は、発電装置３０が設置された地域の天気情報でよい。発電装置３０の再生可能エネルギー発電量は、過去の所定期間における、略一定時間毎の発電量、発電量の累計、平均値、上限値、および下限値の少なくとも１つを含んでよい。 The storage unit 110 stores the first factor, the second factor, and the third factor acquired by the acquisition unit 100. The first factor (power generation amount prediction factor) may include information that affects the amount of renewable energy generated by the power generation device 30. The first factor includes, for example, at least one of the renewable energy power generation amount of the power generation device 30 and the weather information before the target period in which the transition of the renewable energy power generation amount of the power generation device 30 is predicted. The first factor may include information obtained by the obtaining unit 100 from the power generation device 30. For example, the weather information may include at least one of wind speed, wind direction, sunny, rain, temperature, wave height, sunshine duration, and the like. The weather information may be weather information of an area where the power generation device 30 is installed. The renewable energy power generation amount of the power generation device 30 may include at least one of a power generation amount, a cumulative total of the power generation amount, an average value, an upper limit value, and a lower limit value at approximately constant time intervals in a past predetermined period.
　第２因子（稼働予測因子）は、電解装置２０の稼働または電解装置２０の生成物の使用計画に関する情報を含んでよい。第２因子は、生成する稼働計画の対象期間よりも前に入手可能な、電解装置２０の稼働データ、電解装置２０の生成物の需要量、および電解装置２０の生成物の貯蔵量の少なくとも１つを含む。また、第２因子は、当該計画装置４０が過去に生成した電解装置２０の稼働計画を含んでもよい。また、第２因子は、電解装置２０の物理モデルから算出される仮想データを含んでよい。また、第２因子は、電解装置２０から取得部１００が取得した情報を含んでよい。例えば、稼働データは、電解装置２０における単位時間当たりの生成物の生成量、および／または単位電力当たりの生成物の生成量であってよい。需要量は、電解装置２０により供給する必要のある生成物の量であってよい。貯蔵量は、電解装置２０の生成物を貯蔵するタンク等に貯蔵された量であってよい。 The second factor (operation prediction factor) may include information on the operation of the electrolysis device 20 or the usage plan of the product of the electrolysis device 20. The second factor is at least one of the operation data of the electrolysis device 20, the demand of the product of the electrolysis device 20, and the storage amount of the product of the electrolysis device 20, which are available before the target period of the operation plan to be generated. Including one. Further, the second factor may include an operation plan of the electrolysis device 20 generated by the planning device 40 in the past. Further, the second factor may include virtual data calculated from a physical model of the electrolysis device 20. Further, the second factor may include information obtained by the obtaining unit 100 from the electrolysis device 20. For example, the operation data may be a product generation amount per unit time in the electrolysis device 20 and / or a product generation amount per unit power. The demand may be the amount of product that needs to be supplied by the electrolyzer 20. The storage amount may be an amount stored in a tank or the like that stores a product of the electrolysis device 20.
　第３因子（電気料金予測因子）は、電力系統５０の電気料金の変動に影響を及ぼす情報を含んでよい。第３因子は、電気料金を予測する対象期間よりも前の、電力系統５０における電気料金、電力需要量、電力供給量、再生可能エネルギー発電量、再生可能エネルギー発電量の予測値、および天気情報の少なくとも１つを含む。電力系統５０における電気料金は、電解装置２０が設置された場所において、電力系統５０から電解装置２０に供給される電力に対する実際の電気料金、および／または発電装置３０が電力系統５０に電力を売る際の売電料金であってよい。天気情報は、電解装置２０が設置された地域の天気情報でよい。電力系統５０における電力需要量、電力供給量、再生可能エネルギー発電量、および再生可能エネルギー発電量の予測値は、電解装置２０が設置された地域に電力系統５０の送電網を介して電力を供給する発電所等の情報でよい。当該発電所等は、電解装置２０に接続されたローカルな電源である発電装置３０とは異なってよい。 The third factor (electricity rate prediction factor) may include information that affects fluctuations in the electricity rate of the power system 50. The third factor is the electricity price, the power demand, the power supply, the renewable energy power generation, the predicted value of the renewable energy power generation, and the weather information in the power system 50 before the target period for predicting the electricity price. At least one of The electricity rate in the power system 50 is the actual electricity rate for the power supplied from the power system 50 to the electrolysis device 20 at the place where the electrolysis device 20 is installed, and / or the power generation device 30 sells the power to the power system 50. It may be the power sale price at the time. The weather information may be weather information of a region where the electrolysis device 20 is installed. The power demand, the power supply, the renewable energy power generation, and the predicted value of the renewable energy power generation in the power system 50 are used to supply power to the area where the electrolysis device 20 is installed via the power grid of the power system 50. Information on the power station to be used may be used. The power plant or the like may be different from the power generator 30 which is a local power supply connected to the electrolyzer 20.
　第１因子、第２因子、および第３因子の情報は、略一定時間毎の時系列の情報でよい。第１因子、第２因子、および第３因子の情報は、時間の経過と共にそれぞれ追加または更新されてよい。例えば、取得部１００は、予め定められた期間毎に、それぞれの情報を取得して更新してよい。また、取得部１００は、取得すべき情報に応じて、略同一または異なる期間毎に取得して、それぞれ追加または更新してよい。第１因子、第２因子、および第３因子の情報は、外部の装置等から供給される情報を含んでよい。 情報 The information of the first factor, the second factor, and the third factor may be time-series information at approximately constant time intervals. The information of the first factor, the second factor, and the third factor may be added or updated over time, respectively. For example, the acquisition unit 100 may acquire and update each piece of information at predetermined intervals. In addition, the acquisition unit 100 may acquire the information at substantially the same or different periods according to the information to be acquired, and may add or update each. The information of the first factor, the second factor, and the third factor may include information supplied from an external device or the like.
　第１モデル生成部２００は、第１モデル更新部２１０に接続される。第１モデル生成部２００は、対象期間における発電装置３０の再生可能エネルギー発電量の推移を、対象期間よりも前に入手可能な第１因子の値に基づいて予測する発電量予測モデルを生成する。第１モデル生成部２００は、対象期間よりも過去の情報を用いて、事前学習またはオフライン学習等と呼ばれる処理により、発電量予測モデルを生成してよい。第１モデル生成部２００は、例えば、回帰分析、ベイズ推論、ニューラルネットワーク、ガウシアン混合モデル、および隠れマルコフモデル等を用いて、発電量予測モデルを生成する。第１モデル生成部２００は、生成した発電量予測モデルを第１モデルとして第１モデル更新部２１０に供給する。 The first model generation unit 200 is connected to the first model update unit 210. The first model generation unit 200 generates a power generation prediction model that predicts a change in the amount of renewable energy generated by the power generation device 30 in the target period based on a value of a first factor available before the target period. . The first model generation unit 200 may generate a power generation amount prediction model by a process called pre-learning or off-line learning using information past the target period. The first model generation unit 200 generates a power generation prediction model using, for example, regression analysis, Bayesian inference, a neural network, a Gaussian mixture model, a hidden Markov model, or the like. The first model generation unit 200 supplies the generated power generation prediction model to the first model update unit 210 as a first model.
　第１モデル更新部２１０は、発電量予測部２２０に接続される。第１モデル更新部２１０は、過去期間における第１因子の値と過去期間以降の再生可能エネルギー発電量の現実の推移とに基づいて、発電量予測モデルを学習により更新する。第１モデル更新部２１０は、第１モデル学習部２１５を有し、第１モデル学習部２１５の学習結果に応じて、発電量予測モデルを更新する。第１モデル更新部２１０は、例えば、予め定められた第１更新期間毎に、第１モデル学習部２１５が学習した発電量予測モデルを、新たな発電量予測モデルとして更新してよい。これに代えて、第１モデル更新部２１０は、第１モデル学習部２１５が予め定められた回数だけ学習したこと、または学習による誤差差分があらかじめ定められた閾値を下回る等の諸条件に応じて、発電量予測モデルを更新してもよい。 The first model updating unit 210 is connected to the power generation amount prediction unit 220. The first model updating unit 210 updates the power generation prediction model by learning based on the value of the first factor in the past period and the actual change in the amount of renewable energy generated in the past period. The first model updating unit 210 has a first model learning unit 215, and updates the power generation amount prediction model according to the learning result of the first model learning unit 215. For example, the first model update unit 210 may update the power generation amount prediction model learned by the first model learning unit 215 as a new power generation amount prediction model every predetermined first update period. Instead, the first model updating unit 210 may respond to various conditions such as that the first model learning unit 215 has learned a predetermined number of times, or that an error difference due to learning is below a predetermined threshold. Alternatively, the power generation amount prediction model may be updated.
　第１モデル学習部２１５は、適応学習またはオンライン学習等と呼ばれる処理により、発電量予測モデルを学習してよい。第１モデル学習部２１５は、例えば、回帰分析、ベイズ推論、ニューラルネットワーク、ガウシアン混合モデル、および隠れマルコフモデル等の任意の機械学習モデルを識別モデルとして、強化学習を実行することによって、発電量予測モデルを学習する。このような機械学習を行うことにより、第１モデル学習部２１５は、第１因子を入力として、第１因子に応じた再生可能エネルギー発電量を、適用するモデルに応じた精度で予測することができるようになる。また、第１モデルとして、例えば、ＬＳＴＭ（Ｌｏｎｇ　ｓｈｏｒｔ－ｔｅｒｍ　ｍｅｍｏｒｙ）、ＲＮＮ（Ｒｅｃｕｒｒｅｎｔ　Ｎｅｕｒａｌ　Ｎｅｔｗｏｒｋ）、およびその他の記憶を有するモデルを使用すれば、第１因子の時系列から再生可能エネルギー発電量を予測することもできる。 The first model learning unit 215 may learn the power generation prediction model by a process called adaptive learning or online learning. The first model learning unit 215 performs power generation prediction by executing reinforcement learning using an arbitrary machine learning model such as a regression analysis, Bayesian inference, a neural network, a Gaussian mixture model, or a hidden Markov model as an identification model. Learn the model. By performing such machine learning, the first model learning unit 215 can predict a renewable energy power generation amount corresponding to the first factor with an accuracy corresponding to a model to be applied, using the first factor as an input. become able to. Further, if a model having LSTM (Long short-term memory), RNN (Recurring Neural Network), and other memories is used as the first model, the amount of renewable energy generated from the time series of the first factor can be calculated. You can also predict.
　第１モデル学習部２１５は、第１モデル生成部２００が発電量予測モデルの生成に用いた第１因子の情報よりも時間的に後の情報を更に用いて学習することが望ましい。第１モデル学習部２１５は、実際の再生可能エネルギー発電量の推移によって更新された第１因子の情報を用いて、発電量予測モデルを学習する。第１モデル学習部２１５は、第１因子の情報が更新されたことに応じて、発電量予測モデルの学習を実行してよい。一例として、第１モデル学習部２１５は、過去期間における第１因子の値と過去期間以降の再生可能エネルギー発電量の現実の推移とに基づいて、発電量予測モデルを学習する。第１モデル学習部２１５は、第１モデル更新部２１０の第１更新期間の間に、１または複数回の学習を実行してよい。第１モデル更新部２１０は、更新した発電量予測モデルを発電量予測部２２０に供給する。 It is desirable that the first model learning unit 215 further learns using information that is temporally later than the information of the first factor used by the first model generation unit 200 to generate the power generation prediction model. The first model learning unit 215 learns the power generation prediction model using the information of the first factor updated based on the actual transition of the renewable energy power generation. The first model learning unit 215 may execute the learning of the power generation amount prediction model according to the update of the information of the first factor. As an example, the first model learning unit 215 learns a power generation prediction model based on the value of the first factor in the past period and the actual change in the amount of renewable energy generated in the past period. The first model learning unit 215 may execute learning one or more times during the first update period of the first model update unit 210. The first model updating unit 210 supplies the updated power generation amount prediction model to the power generation amount prediction unit 220.
　発電量予測部２２０は、稼働計画生成部１５０に接続される。発電量予測部２２０は、発電量予測モデルを用いて、発電装置３０の将来の再生可能エネルギー発電量の推移を予測する。発電量予測部２２０は、例えば、予め定められた期間毎に、将来における当該予め定められた期間の発電装置３０の再生可能エネルギー発電量を予測する。発電量予測部２２０は、発電量予測モデルと第１因子の情報を用いて、発電量を予測する。発電量予測部２２０は、例えば、再生可能エネルギー発電量を予測すべき期間の直前までの期間における第１因子の情報を、発電量予測モデルに適用して発電装置３０の再生可能エネルギー発電量を予測する。発電量予測部２２０は、予測結果を稼働計画生成部１５０に供給する。 The power generation amount prediction unit 220 is connected to the operation plan generation unit 150. The power generation amount prediction unit 220 predicts a future transition of renewable energy power generation of the power generation device 30 using a power generation amount prediction model. The power generation amount prediction unit 220 predicts, for example, for each predetermined period, the amount of renewable energy generated by the power generation device 30 in the future for the predetermined period. The power generation amount prediction unit 220 predicts the power generation amount using the power generation amount prediction model and the information of the first factor. The power generation amount prediction unit 220 applies, for example, the information of the first factor in the period immediately before the period in which the renewable energy power generation is to be predicted to the power generation prediction model to calculate the renewable energy power generation Predict. The power generation amount prediction unit 220 supplies the prediction result to the operation plan generation unit 150.
　第２モデル生成部２３０は、第２モデル更新部２４０に接続される。第２モデル生成部２３０は、対象期間よりも前の第２因子の値、発電装置３０の再生可能エネルギー発電量の推移、および電力系統５０の電気料金の推移に基づいて、稼働計画生成モデルを生成する。稼働計画生成モデルは、学習により、対象期間における稼働計画を、対象期間よりも前に入手可能な第２因子の値と、対象期間における発電装置３０の再生可能エネルギー発電量の推移の予測結果と、電力系統５０の電気料金とに基づいて生成するモデルであってよい。なお、第２モデル生成部２３０は、過去の再生可能エネルギー発電量の推移として、第１因子の値を用いてよく、過去の電気料金の推移として、第３因子の値を用いてよい。第２モデル生成部２３０は、対象期間よりも過去の情報を用いて、事前学習またはオフライン学習等と呼ばれる学習処理により、稼働計画生成モデルを生成してよい。 The second model generation unit 230 is connected to the second model update unit 240. The second model generation unit 230 generates the operation plan generation model based on the value of the second factor before the target period, the transition of the amount of renewable energy generated by the power generation device 30, and the transition of the electricity rate of the power system 50. Generate. The operation plan generation model, by learning, calculates the operation plan in the target period, the value of the second factor available before the target period, and the prediction result of the transition of the renewable energy power generation amount of the power generator 30 in the target period. , A model generated based on the power rate of the power system 50. Note that the second model generation unit 230 may use the value of the first factor as the past transition of renewable energy power generation, and may use the value of the third factor as the past transition of electricity rates. The second model generation unit 230 may generate an operation plan generation model by a learning process called pre-learning or offline learning using information past the target period.
　第２モデル生成部２３０は、例えば、回帰分析、ベイズ推論、ニューラルネットワーク、ガウシアン混合モデル、および隠れマルコフモデル等の任意の機械学習モデルを識別モデルとして、強化学習を実行することによって、稼働計画生成モデルを生成する。第２モデル生成部２３０は、生成した稼働計画生成モデルを第２モデルとして第２モデル更新部２４０に供給する。 The second model generating unit 230 generates an operation plan by executing reinforcement learning using an arbitrary machine learning model such as a regression analysis, Bayesian inference, a neural network, a Gaussian mixture model, and a hidden Markov model as an identification model. Generate a model. The second model generation unit 230 supplies the generated operation plan generation model as a second model to the second model update unit 240.
　第２モデル更新部２４０は、稼働計画生成部１５０に接続される。第２モデル更新部２４０は、過去期間における第２因子の値と、過去期間以降における発電装置３０の再生可能エネルギー発電量の推移または発電装置３０の再生可能エネルギー発電量の推移の予測結果と、過去期間以降における電力系統５０の電気料金の推移と、過去期間以降において目標とすべき電解装置２０の稼働計画とに基づいて、稼働計画生成モデルを学習により更新する。第２モデル更新部２４０は、第２モデル学習部２４５を有し、第２モデル学習部２４５の学習結果に応じて、稼働計画生成モデルを更新する。第２モデル更新部２４０は、例えば、予め定められた第２更新期間毎に、第２モデル学習部２４５が学習した稼働計画生成モデルを、新たな稼働計画生成モデルとして更新してよい。これに代えて、第２モデル更新部２４０は、第２モデル学習部２４５が予め定められた回数だけ学習したことに応じて、稼働計画生成モデルを更新してもよい。 The second model updating unit 240 is connected to the operation plan generating unit 150. The second model updating unit 240 calculates the value of the second factor in the past period and the prediction result of the transition of the renewable energy generation amount of the power generation device 30 or the transition of the renewable energy generation amount of the power generation device 30 after the past period, The operation plan generation model is updated by learning based on the transition of the electricity rate of the power system 50 after the past period and the operation plan of the electrolytic device 20 to be targeted after the past period. The second model updating unit 240 has a second model learning unit 245, and updates the operation plan generation model according to the learning result of the second model learning unit 245. For example, the second model updating unit 240 may update the operation plan generation model learned by the second model learning unit 245 as a new operation plan generation model every predetermined second update period. Alternatively, the second model updating unit 240 may update the operation plan generation model in response to the second model learning unit 245 learning a predetermined number of times.
　第２モデル学習部２４５は、適応学習またはオンライン学習等と呼ばれる処理により、稼働計画生成モデルを学習してよい。第２モデル学習部２４５は、例えば、回帰分析、ベイズ推論、ニューラルネットワーク、ガウシアン混合モデル、および隠れマルコフモデル等の任意の機械学習モデルを識別モデルとして、強化学習を実行することによって、稼働計画生成モデルを学習する。このような機械学習を行うことにより、第２モデル学習部２４５は、第２因子を入力として、第２因子に応じた値を、適用するモデルに応じた精度で予測することができるようになる。また、第２モデルとして、例えば、ＬＳＴＭ（Ｌｏｎｇ　ｓｈｏｒｔ－ｔｅｒｍ　ｍｅｍｏｒｙ）、ＲＮＮ（Ｒｅｃｕｒｒｅｎｔ　Ｎｅｕｒａｌ　Ｎｅｔｗｏｒｋ）、およびその他の記憶を有するモデルを使用すれば、第２因子の時系列から電解装置２０の稼働状態を予測することもできる。 The second model learning unit 245 may learn the operation plan generation model by a process called adaptive learning or online learning. The second model learning unit 245 executes an operation plan generation by executing reinforcement learning using an arbitrary machine learning model such as a regression analysis, Bayesian inference, a neural network, a Gaussian mixture model, and a hidden Markov model as an identification model. Learn the model. By performing such machine learning, the second model learning unit 245 can predict a value corresponding to the second factor with an accuracy corresponding to a model to be applied, using the second factor as an input. . Also, if a model having LSTM (Long short-term memory), RNN (Recent Neural Network), or other storage is used as the second model, for example, the operating state of the electrolysis apparatus 20 is determined from the time series of the second factor. Can also be predicted.
　第２モデル学習部２４５は、第２モデル生成部２３０が稼働計画生成モデルの生成に用いた第２因子の情報よりも時間的に後の情報を更に用いて学習することが望ましい。第２モデル学習部２４５は、発電装置３０の実際の再生可能エネルギー発電量の推移によって更新された第１因子の情報と、実際の電解装置２０の稼働によって更新された第２因子の情報と、実際の電気料金の推移によって更新された第３因子の情報とを用いて、稼働計画生成モデルを学習する。なお、再生可能エネルギー発電量の推移については、例えば、実際の再生可能エネルギー発電量の推移に代えて、発電量予測部２２０の予測結果を用いてもよい。また、電気料金の推移については、例えば、実際の電気料金の推移に代えて、電気料金予測部２７０の予測結果を用いてもよい。即ち、第２モデル学習部２４５は、過去期間における第２因子の値と、過去期間以降における再生可能エネルギー発電量および電気料金の推移、または再生可能エネルギー発電量および電気料金の推移の予測結果とに基づいて、稼働計画生成モデルを学習する。 ２ It is desirable that the second model learning unit 245 further learns using information that is temporally later than the information of the second factor used by the second model generation unit 230 to generate the operation plan generation model. The second model learning unit 245 includes information on the first factor updated by the change in the actual amount of renewable energy generated by the power generation device 30, information on the second factor updated by the actual operation of the electrolysis device 20, The operation plan generation model is learned using the information on the third factor updated according to the actual transition of the electricity rate. For the transition of the renewable energy power generation amount, for example, the prediction result of the power generation amount prediction unit 220 may be used instead of the actual renewable energy power generation amount transition. Further, regarding the transition of the electricity rate, for example, the prediction result of the electricity rate prediction unit 270 may be used instead of the transition of the actual electricity rate. That is, the second model learning unit 245 calculates the value of the second factor in the past period and the transition result of the renewable energy generation amount and the electricity rate or the prediction result of the transition of the renewable energy generation amount and the electricity rate in the past period and thereafter. Based on, the operation plan generation model is learned.
　第２モデル学習部２４５は、第２因子の情報が更新されたことに応じて、稼働計画生成モデルの学習を実行してよい。第２モデル学習部２４５は、第２モデル更新部２４０の第２更新期間の間に、１または複数回の学習を実行する。第２モデル更新部２４０は、更新した稼働計画生成モデルを稼働計画生成部１５０に供給する。 The second model learning unit 245 may execute the learning of the operation plan generation model in accordance with the update of the information of the second factor. The second model learning unit 245 executes learning one or more times during the second update period of the second model update unit 240. The second model updating unit 240 supplies the updated operation plan generation model to the operation plan generation unit 150.
　第３モデル生成部２５０は、第３モデル更新部２６０に接続される。第３モデル生成部２５０は、対象期間における電力系統５０の電気料金の推移を、対象期間よりも前に入手可能な第３因子の値に基づいて予測する電気料金予測モデルを生成する。第３モデル生成部２５０は、対象期間よりも過去の情報を用いて、事前学習またはオフライン学習等と呼ばれる処理により、電気料金予測モデルを生成してよい。第３モデル生成部２５０は、例えば、回帰分析、ベイズ推論、ニューラルネットワーク、ガウシアン混合モデル、および隠れマルコフモデル等を用いて、電気料金予測モデルを生成する。第３モデル生成部２５０は、生成した電気料金予測モデルを第３モデルとして第３モデル更新部２６０に供給する。 The third model generation unit 250 is connected to the third model update unit 260. The third model generation unit 250 generates an electricity rate prediction model that predicts a change in the electricity rate of the power system 50 in the target period based on a value of a third factor available before the target period. The third model generation unit 250 may generate an electricity price prediction model by using a process called pre-learning or off-line learning using information past the target period. The third model generation unit 250 generates an electricity bill prediction model using, for example, regression analysis, Bayesian inference, a neural network, a Gaussian mixture model, a hidden Markov model, or the like. The third model generation unit 250 supplies the generated electricity price prediction model as a third model to the third model update unit 260.
　第３モデル更新部２６０は、電気料金予測部２７０に接続され、電気料金予測モデルを学習により更新する。第３モデル更新部２６０は、第３モデル学習部２６５を有し、第３モデル学習部２６５の学習結果に応じて、電気料金予測モデルを更新する。第３モデル更新部２６０は、例えば、予め定められた第３更新期間毎に、第３モデル学習部２６５が学習した電気料金予測モデルを、新たな電気料金予測モデルとして更新してよい。これに代えて、第３モデル更新部２６０は、第３モデル学習部２６５が予め定められた回数だけ学習したことに応じて、電気料金予測モデルを更新してもよい。 The third model updating unit 260 is connected to the electricity rate prediction unit 270 and updates the electricity rate prediction model by learning. The third model updating section 260 has a third model learning section 265, and updates the electricity rate prediction model according to the learning result of the third model learning section 265. For example, the third model updating unit 260 may update the electricity rate prediction model learned by the third model learning unit 265 as a new electricity rate prediction model every predetermined third update period. Alternatively, the third model updating unit 260 may update the electricity rate prediction model in response to the third model learning unit 265 learning a predetermined number of times.
　第３モデル学習部２６５は、適応学習またはオンライン学習等と呼ばれる処理により、電気料金予測モデルを学習してよい。第３モデル学習部２６５は、例えば、回帰分析、ベイズ推論、ニューラルネットワーク、ガウシアン混合モデル、および隠れマルコフモデル等の任意の機械学習モデルを識別モデルとして、強化学習を実行することによって、電気料金予測モデルを学習する。このような機械学習を行うことにより、第３モデル学習部２６５は、第３因子を入力として、第３因子に応じた電気料金を、適用するモデルに応じた精度で予測することができるようになる。また、第３モデルとして、例えば、ＬＳＴＭ（Ｌｏｎｇ　ｓｈｏｒｔ－ｔｅｒｍ　ｍｅｍｏｒｙ）、ＲＮＮ（Ｒｅｃｕｒｒｅｎｔ　Ｎｅｕｒａｌ　Ｎｅｔｗｏｒｋ）、およびその他の記憶を有するモデルを使用すれば、第３因子の時系列から電気料金を予測することもできる。 The third model learning unit 265 may learn the electricity bill prediction model by a process called adaptive learning or online learning. The third model learning unit 265 executes, for example, an electric charge prediction by executing reinforcement learning using an arbitrary machine learning model such as a regression analysis, Bayesian inference, a neural network, a Gaussian mixture model, and a hidden Markov model as an identification model. Learn the model. By performing such machine learning, the third model learning unit 265 can input the third factor and predict the electricity rate according to the third factor with accuracy according to the model to be applied. Become. Further, if a model having LSTM (Long short-term memory), RNN (Recurrent Neural Network), and other storages is used as the third model, for example, the electricity rate can be predicted from the time series of the third factor. Can also.
　第３モデル学習部２６５は、第３モデル生成部２５０が電気料金予測モデルの生成に用いた第３因子の情報よりも時間的に後の情報を更に用いて学習することが望ましい。第３モデル学習部２６５は、電力系統５０の実際の電気料金の推移によって更新された第３因子の情報を用いて、電気料金予測モデルを学習する。第３モデル学習部２６５は、第３因子の情報が更新されたことに応じて、電気料金予測モデルの学習を実行してよい。一例として、第３モデル学習部２６５は、過去期間における第３因子の値と過去期間以降の電気料金の現実の推移とに基づいて、電気料金予測モデルを学習する。第３モデル学習部２６５は、第３モデル更新部２６０の第３更新期間の間に、１または複数回の学習を実行する。第３モデル更新部２６０は、更新した電気料金予測モデルを電気料金予測部２７０に供給する。 It is desirable that the third model learning unit 265 learns further using information that is temporally later than the information of the third factor used by the third model generation unit 250 to generate the electricity rate prediction model. The third model learning unit 265 learns the electricity rate prediction model using the information on the third factor updated based on the actual transition of the electricity rate of the power system 50. The third model learning unit 265 may execute the learning of the electricity rate prediction model according to the update of the information of the third factor. As an example, the third model learning unit 265 learns the electricity rate prediction model based on the value of the third factor in the past period and the actual transition of the electricity rate after the past period. The third model learning unit 265 performs learning one or more times during the third update period of the third model update unit 260. The third model updating unit 260 supplies the updated electricity bill prediction model to the electricity bill prediction unit 270.
　電気料金予測部２７０は、稼働計画生成部１５０に接続される。電気料金予測部２７０は、更新された電気料金予測モデルを用いて、電力系統５０の将来の電気料金の推移を予測する。電気料金予測部２７０は、例えば、予め定められた期間毎に、将来における当該予め定められた期間の電気料金を予測する。電気料金予測部２７０は、電気料金予測モデルと第３因子の情報を用いて、電気料金を予測する。電気料金予測部２７０は、例えば、電気料金を予測すべき期間の直前までの期間における第３因子の情報を、電気料金予測モデルに適用して電気料金を予測する。電気料金予測部２７０は、予測結果を稼働計画生成部１５０に供給する。 The electricity rate prediction unit 270 is connected to the operation plan generation unit 150. The electricity rate prediction unit 270 predicts a future transition of the electricity rate of the power system 50 using the updated electricity rate prediction model. The electricity rate prediction unit 270, for example, predicts the electricity rate in the future for the predetermined period for each predetermined period. The electricity rate prediction unit 270 predicts the electricity rate using the electricity rate prediction model and the information of the third factor. The electricity rate prediction unit 270 predicts the electricity rate by applying, for example, the information of the third factor in the period immediately before the period in which the electricity rate is to be predicted to the electricity rate prediction model. The electricity rate prediction unit 270 supplies the prediction result to the operation plan generation unit 150.
　稼働計画生成部１５０は、予測された将来の発電装置３０の再生可能エネルギー発電量の推移と予測された将来の電力系統５０の電気料金とに基づいて、将来の第１期間における、電解装置２０の生成物の使用計画を満たす電解装置２０の稼働計画を生成する。稼働計画生成部１５０は、稼働計画生成モデルを用いて、将来の第１期間における電解装置２０の稼働計画を生成してよい。稼働計画生成部１５０は、第１期間中において、電解装置２０の生成物の使用計画を満たすように、発電装置３０からの電気を電力系統５０からの電気より優先して用いて電解装置２０を稼働させる稼働計画を生成してよい。稼働計画生成部１５０は、例えば、数日または十数日、１または数週間といった期間を第１期間として、電解装置２０の稼働計画を生成する。稼働計画生成部１５０は、一例として、Ｎ日分の稼働計画を生成する。 The operation plan generation unit 150 determines the electrolysis device 20 for the first period in the future based on the predicted change in the amount of renewable energy generated by the power generation device 30 and the predicted electricity rate of the power system 50 in the future. The operation plan of the electrolysis device 20 that satisfies the use plan of the product of the above is generated. The operation plan generation unit 150 may generate an operation plan of the electrolysis device 20 in the future first period by using the operation plan generation model. During the first period, the operation plan generation unit 150 uses the electrolysis device 20 by giving priority to the power from the power generation device 30 over the power from the power system 50 so as to satisfy the usage plan of the product of the electrolysis device 20. An operation plan to be operated may be generated. The operation plan generation unit 150 generates an operation plan of the electrolysis apparatus 20 with a period such as several days or ten and several days, one or several weeks as a first period. The operation plan generation unit 150 generates an operation plan for N days, for example.
　制御部１６０は、稼働計画生成部１５０が生成した稼働計画に応じて、電解装置２０を稼働させる。制御部１６０は、発電装置３０を制御してよく、発電装置３０から電解装置２０に供給する電力量、供給する時間帯、発電装置３０の発電電力を売電する電力量、および／または売電する時間帯等を制御してよい。 The control unit 160 operates the electrolysis device 20 according to the operation plan generated by the operation plan generation unit 150. The control unit 160 may control the power generation device 30, the amount of power to be supplied from the power generation device 30 to the electrolysis device 20, the time period to be supplied, the amount of power to sell the power generated by the power generation device 30, and / or the power sale. The time period of the operation may be controlled.
　以上の本実施形態に係る計画装置４０は、発電装置３０の再生可能エネルギー発電量と電力系統５０の電気料金とを学習によって予測しつつ、電解装置２０の稼働計画を生成する。このような計画装置４０の動作について、次に説明する。 The planning device 40 according to the above-described embodiment generates the operation plan of the electrolysis device 20 while predicting the amount of renewable energy generated by the power generation device 30 and the electricity rate of the power system 50 by learning. The operation of the planning device 40 will be described below.
　図３は、本実施形態に係る計画装置４０の動作フローの一例を示す。計画装置４０は、図３に示す動作フローを実行し、電解装置２０を稼働させてよい。 FIG. 3 shows an example of an operation flow of the planning device 40 according to the present embodiment. The planning device 40 may execute the operation flow illustrated in FIG. 3 to operate the electrolysis device 20.
　取得部１００は、発電装置３０の再生可能エネルギー発電量、電力系統５０の電気料金、および電解装置２０の過去のトレンドとなる第１因子、第２因子、および第３因子の情報を取得する（Ｓ３１０）。取得部１００は、例えば、時刻ｔ０から時刻ｔ１における、第１因子、第２因子、および第３因子の情報を取得する。ここで、時刻ｔ０から時刻ｔ１の間の期間は、第１期間よりも前の第２期間とする。取得部１００は、取得した第１因子、第２因子、および第３因子の情報を記憶部１１０に記憶させる。また、取得部１００は、第１因子、第２因子、および第３因子の情報をモデル生成部１２０に直接供給してもよい。 The acquisition unit 100 acquires information on the amount of renewable energy generated by the power generation device 30, the electricity rate of the power system 50, and the first factor, the second factor, and the third factor that are past trends of the electrolysis device 20 ( S310). The acquisition unit 100 acquires, for example, information on a first factor, a second factor, and a third factor from time t0 to time t1. Here, the period from time t0 to time t1 is a second period before the first period. The acquiring unit 100 causes the storage unit 110 to store the acquired information of the first factor, the second factor, and the third factor. Further, the acquisition unit 100 may directly supply the information of the first factor, the second factor, and the third factor to the model generation unit 120.
　次に、モデル生成部１２０は、学習モデルを生成する（Ｓ３２０）。モデル生成部１２０は、第２期間の第１因子、第２因子、および第３因子の値に基づき、学習モデルを生成する。例えば、第１モデル生成部２００は、第２期間における、発電装置３０の再生可能エネルギー発電量および天気情報の少なくとも１つを含む第１因子の値を用いて、発電量予測モデルを生成する。第３モデル生成部２５０は、第２期間における、電気料金、電力需要量、電力供給量、再生可能エネルギー発電量、再生可能エネルギー発電量の予測値、および天気情報の少なくとも１つを含む第３因子の値を用いて、電気料金予測モデルを生成する。 Next, the model generation unit 120 generates a learning model (S320). The model generation unit 120 generates a learning model based on the values of the first factor, the second factor, and the third factor in the second period. For example, the first model generation unit 200 generates a power generation amount prediction model using a value of a first factor including at least one of the amount of renewable energy generated by the power generation device 30 and weather information in the second period. The third model generation unit 250 includes a third model that includes at least one of the electricity rate, the power demand, the power supply, the renewable energy generation, the predicted value of the renewable energy generation, and the weather information in the second period. Using the values of the factors, an electricity price prediction model is generated.
　また、第２モデル生成部２３０は、第１因子、第２因子、および第３因子の値に基づき、稼働計画生成モデルを生成する。例えば、第２モデル生成部２３０は、第２期間における、発電装置３０の再生可能エネルギー発電量、電力系統５０の電気料金、電解装置２０の稼働データ、電解装置２０の生成物の貯蔵量、および電解装置２０の稼働計画の仮想データの少なくとも１つを用いて、稼働計画生成モデルを生成する。 {Circle around (2)} The second model generation unit 230 generates an operation plan generation model based on the values of the first factor, the second factor, and the third factor. For example, the second model generation unit 230 outputs the renewable energy power generation amount of the power generation device 30, the electricity rate of the power system 50, the operation data of the electrolysis device 20, the storage amount of the product of the electrolysis device 20, and the second period. An operation plan generation model is generated using at least one of the virtual data of the operation plan of the electrolysis device 20.
　また、第２モデル生成部２３０は、電解装置２０の物理モデルに基づく仮想データを目標とすべき予測データとし、当該予測データおよび過去の電解装置２０の稼働によって取得された実データとを比較することにより、稼働計画生成モデルを生成してよい。例えば、第２モデル生成部２３０は、目標とすべき予測データおよび過去の実データの差分が０または予め定められた値未満となるように、強化学習を実行して稼働計画生成モデルを生成する。 In addition, the second model generation unit 230 sets virtual data based on the physical model of the electrolysis device 20 as prediction data to be targeted, and compares the prediction data with actual data obtained by operation of the electrolysis device 20 in the past. Thus, an operation plan generation model may be generated. For example, the second model generation unit 230 executes the reinforcement learning to generate the operation plan generation model so that the difference between the target predicted data and the past actual data is 0 or less than a predetermined value. .
　第２モデル生成部２３０は、一例として、第２期間におけるＭ日間の期間を仮想的な予測期間とする。なお、Ｍ日間は、例えば、数日または十数日、１または数週間といった期間でよい。Ｍ日間は、第１期間（Ｎ日間）と一致することが望ましい。そして、第２モデル生成部２３０は、第２期間における予測期間よりも前の期間の第１因子、第２因子、および第３因子の値に基づく予測期間の稼働動作の予測結果と、予測期間の実データまたは仮想データとの誤差が、最小となるように強化学習する。 The second model generation unit 230 sets the period of M days in the second period as a virtual prediction period, for example. Note that the M days may be, for example, a period of several days, ten and several days, one or several weeks. It is desirable that M days coincide with the first period (N days). Then, the second model generation unit 230 calculates the prediction result of the operation operation in the prediction period based on the values of the first factor, the second factor, and the third factor in the period before the prediction period in the second period, The reinforcement learning is performed so that an error between the actual data and the virtual data is minimized.
　この場合、第２モデル生成部２３０は、電解装置２０の生成物の使用計画を満たし（第１条件）、かつ発電装置３０からの電気を電力系統５０からの電気より優先して用いて電解装置２０を稼働させる（第２条件）という条件を満たしながら、稼働コストを低減させるように強化学習してよい。ここで、電解装置２０の生成物の使用計画を満たすという第１条件は、例えば、電解装置２０の生成物の貯蔵量の変動範囲を０から最大貯蔵量の範囲とする、または電解装置２０の生成物の需要量または供給量の累計以上の生成物を電解装置２０で生成する等の条件であってよい。また、第２条件は、所定期間（例えば第２期間）における予測される発電装置３０の累計の再生可能エネルギー発電量を全て電解装置２０の稼働に使用する、または、発電装置３０が電力系統５０よりも多くの電力を電解装置２０の稼働のために供給する等の条件であってよい。 In this case, the second model generation unit 230 satisfies the usage plan of the product of the electrolysis device 20 (first condition), and uses the electricity from the power generation device 30 in preference to the electricity from the power system 50 to use the electrolysis device. The reinforcement learning may be performed so as to reduce the operation cost while satisfying the condition of operating 20 (second condition). Here, the first condition that the use plan of the product of the electrolysis device 20 is satisfied is, for example, a range in which the storage amount of the product in the electrolysis device 20 varies from 0 to the maximum storage amount, or The condition may be such that a product equal to or more than the total demand or supply amount of the product is generated in the electrolytic device 20. In addition, the second condition is that the total amount of renewable energy generated by the power generation device 30 predicted in a predetermined period (for example, the second period) is used for the operation of the electrolysis device 20 or the power generation device 30 The condition may be such that more electric power is supplied for the operation of the electrolysis device 20.
　なお、このようなモデル生成部１２０による学習モデルの生成は、電解装置２０の稼働に伴って計画装置４０が当該電解装置２０の実データを取得する前に、実行されてよい。 The generation of the learning model by the model generation unit 120 may be executed before the planning device 40 acquires the actual data of the electrolysis device 20 with the operation of the electrolysis device 20.
　次に、学習処理部１３０は、生成した学習モデルを適応学習する（Ｓ３３０）。ここで、取得部１００は、第１因子、第２因子、および第３因子の情報をさらに取得してよい。取得部１００は、例えば、時刻ｔ１から時刻ｔ２における、第１因子、第２因子、および第３因子の情報を取得する。なお、時刻ｔ１から時刻ｔ２の間の期間は、第１期間および第２期間の間の第３期間とする。学習処理部１３０は、取得部１００が新たに取得した第１因子、第２因子、および第３因子の情報を用いて適応学習してよい。 Next, the learning processing unit 130 adaptively learns the generated learning model (S330). Here, the acquisition unit 100 may further acquire information on the first factor, the second factor, and the third factor. The acquisition unit 100 acquires information on the first factor, the second factor, and the third factor from time t1 to time t2, for example. Note that a period from time t1 to time t2 is a third period between the first period and the second period. The learning processing unit 130 may perform adaptive learning using the information of the first factor, the second factor, and the third factor newly acquired by the acquiring unit 100.
　例えば、第１モデル学習部２１５は、第１因子の値に基づき、発電量予測モデルを適応学習する。第１モデル学習部２１５は、第３期間における、発電装置３０の再生可能エネルギー発電量および天気情報の少なくとも１つを用いて、発電量予測モデルを適応学習してよい。第１モデル学習部２１５は、発電量予測モデルを用いて第３期間における再生可能エネルギー発電量を予測した結果が、取得した第３期間の再生可能エネルギー発電量と一致するように強化学習してよい。 {For example, the first model learning unit 215 adaptively learns the power generation prediction model based on the value of the first factor. The first model learning unit 215 may adaptively learn the power generation prediction model using at least one of the renewable energy power generation amount of the power generation device 30 and weather information in the third period. The first model learning unit 215 performs reinforcement learning so that the result of predicting the amount of renewable energy generated in the third period using the generated amount prediction model matches the obtained amount of renewable energy generated in the third period. Good.
　第１モデル学習部２１５は、一例として、第３期間におけるＭ日間の期間を仮想的な予測期間とする。なお、Ｍ日間は、例えば、数日または十数日、１または数週間といった期間でよい。Ｍ日間は、第１期間（Ｎ日間）と一致することが望ましい。第１モデル学習部２１５は、一例として、第３期間における予測期間よりも前の期間の第１因子の値に基づく予測期間の予測結果と、予測期間の実データとの差分が０または予め定められた値未満となるように強化学習する。 The first model learning unit 215 sets the period of M days in the third period as a virtual prediction period, for example. Note that the M days may be, for example, a period of several days, ten and several days, one or several weeks. It is desirable that M days coincide with the first period (N days). As an example, the first model learning unit 215 determines that the difference between the prediction result of the prediction period based on the value of the first factor in the period before the prediction period in the third period and the actual data of the prediction period is 0 or a predetermined value. Reinforcement learning is performed so as to be less than the set value.
　また、第２モデル学習部２４５は、第１因子、第２因子、および第３因子の値に基づき、稼働計画生成モデルを適用学習してよい。例えば、第２モデル学習部２４５は、第３期間における、発電装置３０の再生可能エネルギー発電量の推移または当該推移の予測結果、電力系統５０の電気料金の推移または当該推移の予測結果、電解装置２０の稼働データ、電解装置２０の生成物の貯蔵量または需要量、および稼働計画の実データの少なくとも１つを用いて、稼働計画生成モデルを学習してよい。第２モデル学習部２４５は、稼働計画生成モデルを用いて第３期間における電解装置２０の稼働動作を予測した結果と、取得した第３期間の実データの差分が０または予め定められた値未満となるように、強化学習を実行してよい。 {Circle around (2)} The second model learning unit 245 may apply and learn the operation plan generation model based on the values of the first factor, the second factor, and the third factor. For example, in the third period, the second model learning unit 245 calculates the transition of the renewable energy power generation amount of the power generation device 30 or the prediction result of the transition, the transition of the electricity rate of the power system 50 or the prediction result of the transition, the electrolysis device. The operation plan generation model may be learned using at least one of the operation data of the operation plan 20, the storage amount or demand amount of the product of the electrolysis device 20, and the actual data of the operation plan. The second model learning unit 245 determines that the difference between the result of predicting the operation of the electrolysis device 20 in the third period using the operation plan generation model and the acquired actual data in the third period is 0 or less than a predetermined value. The reinforcement learning may be performed so that
　第２モデル学習部２４５は、一例として、第３期間におけるＭ日間の期間を仮想的な予測期間とする。なお、Ｍ日間は、例えば、数日または十数日、１または数週間といった期間でよい。Ｍ日間は、第１期間（Ｎ日間）と一致することが望ましい。そして、第２モデル学習部２４５は、第３期間における予測期間よりも前の期間の第１因子、第２因子、および第３因子の値に基づく予測期間の稼働動作の予測結果と、予測期間の実データとの差分が０または予め定められた値未満となるように強化学習する。 The second model learning unit 245 sets the period of M days in the third period as a virtual prediction period, for example. Note that the M days may be, for example, a period of several days, ten and several days, one or several weeks. It is desirable that M days coincide with the first period (N days). Then, the second model learning unit 245 calculates the prediction result of the operation operation in the prediction period based on the values of the first factor, the second factor, and the third factor in the period before the prediction period in the third period, and The reinforcement learning is performed such that the difference between the actual data and the actual data becomes 0 or less than a predetermined value.
　この場合、第２モデル学習部２４５は、第２モデル生成部２３０が稼働計画生成モデルの生成に用いた、第１条件、および第２条件等を同様に用いてよい。即ち、第２モデル学習部２４５は、２つの条件を満たしつつ、生成物の製造コストを低減させるように稼働計画生成モデルを強化学習してよい。 In this case, the second model learning unit 245 may similarly use the first condition, the second condition, and the like used by the second model generation unit 230 to generate the operation plan generation model. That is, the second model learning unit 245 may perform the reinforcement learning of the operation plan generation model so as to reduce the production cost of the product while satisfying the two conditions.
　また、第３モデル学習部２６５は、第３因子の値に基づき、電気料金予測モデルを適応学習する。第３モデル学習部２６５は、第３期間における、電力系統５０についての電気料金、電力需要量、電力供給量、再生可能エネルギー発電量、再生可能エネルギー発電量の予測値、および天気情報の少なくとも１つを用いて、電気料金予測モデルを適応学習してよい。第３モデル学習部２６５は、電気料金予測モデルを用いて第３期間における電気料金等を予測した結果が、取得した第３期間の電気料金と一致するように強化学習してよい。 {Circle around (3)} The third model learning unit 265 adaptively learns the electricity rate prediction model based on the value of the third factor. The third model learning unit 265 performs at least one of the electricity rate, the power demand, the power supply, the renewable energy generation, the predicted value of the renewable energy generation, and the weather information for the power system 50 in the third period. One of them may be used to adaptively learn the electricity price prediction model. The third model learning unit 265 may perform the reinforcement learning such that the result of predicting the electricity rate or the like in the third period using the electricity rate prediction model matches the acquired electricity rate in the third period.
　第３モデル学習部２６５は、一例として、第３期間におけるＭ日間の期間を仮想的な予測期間とする。なお、Ｍ日間は、例えば、数日または十数日、１または数週間といった期間でよい。Ｍ日間は、第１期間（Ｎ日間）と一致することが望ましい。第３モデル学習部２６５は、一例として、第３期間における予測期間よりも前の期間の第３因子の値に基づく予測期間の予測結果と、予測期間の実データとの差分が０または予め定められた値未満となるように強化学習する。 The third model learning unit 265 sets the period of M days in the third period as a virtual prediction period, for example. Note that the M days may be, for example, a period of several days, ten and several days, one or several weeks. It is desirable that M days coincide with the first period (N days). As an example, the third model learning unit 265 determines that the difference between the prediction result of the prediction period based on the value of the third factor in the period before the prediction period in the third period and the actual data of the prediction period is 0 or a predetermined value. Reinforcement learning is performed so as to be less than the set value.
　次に、学習処理部１３０は、学習モデルを更新する（Ｓ３４０）。学習処理部１３０は、予め定められた時間毎に学習モデルを更新してよい。例えば、学習処理部１３０は、適応学習を開始してから更新に必要な初期更新期間だけ適応学習を継続させてから、学習モデルの最初の更新を実行し、その後、一定の期間毎に更新を繰り返す。ここで、初期更新期間は、生成すべき稼働計画の期間であるＮ日以上であることが望ましい。また、更新を繰り返す一定の期間は、数時間、十数時間、１日、数十時間、または数日等でよい。 Next, the learning processing unit 130 updates the learning model (S340). The learning processing unit 130 may update the learning model every predetermined time. For example, the learning processing unit 130 performs the first update of the learning model after continuing the adaptive learning for an initial update period necessary for the update after the start of the adaptive learning, and thereafter updates the learning model at regular intervals. repeat. Here, it is desirable that the initial update period is N days or more, which is the period of the operation plan to be generated. Further, the fixed period for repeating the update may be several hours, ten and several hours, one day, several tens of hours, or several days.
　例えば、第１モデル更新部２１０は、初期更新期間後、発電量予測モデルを第１更新期間毎に更新する。また、第２モデル更新部２４０は、初期更新期間後、稼働計画生成モデルを第２更新期間毎に更新する。また、第３モデル更新部２６０は、初期更新期間後、電気料金予測モデルを第３更新期間毎に更新する。第１更新期間、第２更新期間、および第３更新期間は、異なる期間でよく、これに代えて、略同一の期間でもよい。第１更新期間、第２更新期間、および第３更新期間は、一例として１日である。 For example, after the initial update period, the first model update unit 210 updates the power generation prediction model every first update period. Further, after the initial update period, the second model update unit 240 updates the operation plan generation model every second update period. Further, after the initial update period, third model update section 260 updates the electricity rate prediction model every third update period. The first update period, the second update period, and the third update period may be different periods, and may be substantially the same periods instead. The first update period, the second update period, and the third update period are, for example, one day.
　次に、予測部１４０は、更新した学習モデルを用いて発電装置３０の再生可能エネルギー発電量および電力系統５０の電気料金を予測する（Ｓ３５０）。例えば、発電量予測部２２０は、更新された発電量予測モデルおよび第１因子の値を用いて、第１期間の発電装置３０の再生可能エネルギー発電量の推移を予測する。発電量予測部２２０は、一例として、初期更新期間に取得部１００が取得したＮ日分の第１因子の値を発電量予測モデルに適用して、初期更新期間の後のＮ日分の再生可能エネルギー発電量の推移を予測する。 Next, the prediction unit 140 predicts the amount of renewable energy generated by the power generation device 30 and the electricity rate of the power system 50 using the updated learning model (S350). For example, the power generation amount prediction unit 220 predicts the transition of the renewable energy power generation amount of the power generation device 30 in the first period using the updated power generation amount prediction model and the value of the first factor. As an example, the power generation amount prediction unit 220 applies the value of the first factor for N days acquired by the acquisition unit 100 during the initial update period to the power generation amount prediction model, and reproduces N days after the initial update period. Forecast the transition of available energy generation.
　また、電気料金予測部２７０は、更新された電気料金予測モデルおよび第３因子の値を用いて、第１期間の電気料金の推移を予測する。電気料金予測部２７０は、一例として、初期更新期間に取得部１００が取得したＮ日分の第３因子の値を電気料金予測モデルに適用して、初期更新期間の後のＮ日分の電気料金の推移を予測する。 {Circle around (2)} The electricity rate prediction unit 270 predicts the transition of the electricity rate in the first period using the updated electricity rate prediction model and the value of the third factor. As an example, the electricity rate prediction unit 270 applies the value of the third factor for N days acquired by the acquisition unit 100 during the initial update period to the electricity rate prediction model, and generates electricity for N days after the initial update period. Predict changes in rates.
　次に、稼働計画生成部１５０は、更新された学習モデルを用いて、第１期間の電解装置２０の稼働計画を生成する（Ｓ３６０）。稼働計画生成部１５０は、更新された稼働計画生成モデル、発電量予測部２２０の再生可能エネルギー発電量の予測結果、電気料金予測部２７０の電気料金の予測結果、および第２因子の値を用いて、第１期間の稼働計画を生成してよい。稼働計画生成部１５０は、一例として、初期更新期間に取得部１００が取得したＮ日分の第２因子の値と、初期更新期間の後のＮ日分の再生可能エネルギー発電量の予測結果および電気料金の推移の予測結果とを稼働計画生成モデルに適用して、初期更新期間の後のＮ日分の稼働計画を生成する。 Next, the operation plan generation unit 150 generates an operation plan of the electrolysis device 20 for the first period using the updated learning model (S360). The operation plan generation unit 150 uses the updated operation plan generation model, the prediction result of the renewable energy power generation amount of the power generation amount prediction unit 220, the prediction result of the electricity price of the electricity price prediction unit 270, and the value of the second factor. Thus, an operation plan for the first period may be generated. As an example, the operation plan generation unit 150 calculates the value of the second factor for N days acquired by the acquisition unit 100 during the initial update period, the predicted result of the renewable energy power generation amount for N days after the initial update period, and By applying the prediction result of the change in the electricity rate and the operation plan generation model, an operation plan for N days after the initial update period is generated.
　稼働計画生成部１５０は、第２モデル生成部２３０が稼働計画生成モデルの生成に用いた、第１条件および第２条件を同様に用いてよい。即ち、稼働計画生成部１５０は、２つの条件を満たしつつ、製造コストを最小化する稼働計画を生成してよい。 The operation plan generation unit 150 may similarly use the first condition and the second condition used by the second model generation unit 230 to generate the operation plan generation model. That is, the operation plan generation unit 150 may generate an operation plan that minimizes the manufacturing cost while satisfying the two conditions.
　稼働計画生成部１５０は、例えば、第１期間において、発電装置３０からの電力により電解装置２０を稼働させる期間と、電力系統５０からの電力により電解装置２０を稼働させる期間とを含む稼働計画を生成する。稼働計画生成部１５０は、一例として、稼働計画において、第１期間における発電装置３０の予測した再生可能エネルギー発電量を、当該第１期間内で使い切るように、発電装置３０からの電力により電解装置２０を稼働させる期間を設定してよい。さらに、発電装置３０からの電力が第１期間における電解装置２０の生成物の使用計画を満たすには不足している場合等に、稼働計画生成部１５０は、稼働計画において、第１期間における電解装置２０の生成物の使用計画における不足分を満たすように、電気料金が閾値より低い期間に、電力系統５０からの電力により電解装置２０を稼働させるように設定してよい。 The operation plan generation unit 150 generates, for example, an operation plan including a period in which the electrolysis device 20 is operated by electric power from the power generation device 30 and a period in which the electrolysis device 20 is operated by electric power from the power system 50 in the first period. Generate. As an example, the operation plan generating unit 150 uses the electric power from the power generation device 30 in the operation plan so that the amount of renewable energy generated by the power generation device 30 in the first period is used up in the first period. 20 may be set to operate. Further, when the power from the power generation device 30 is insufficient to satisfy the usage plan of the product of the electrolysis device 20 in the first period, for example, the operation plan generation unit 150 may perform the electrolysis in the first period in the operation plan. In order to satisfy the shortfall in the usage plan of the product of the apparatus 20, the electrolysis apparatus 20 may be set to operate with the electric power from the electric power system 50 during the period when the electricity rate is lower than the threshold.
　また、稼働計画生成部１５０は、第１期間において、電解装置２０を稼働させる期間と、稼働させない期間とを含む稼働計画を生成してよい。また、稼働計画生成部１５０は、電解装置２０を稼働させる期間を、稼働率と共に示す稼働計画を生成してよい。稼働計画生成部１５０は、時系列に稼働率が推移する稼働計画を生成することが望ましい。稼働計画生成部１５０は、例えば、一定時間ごとの稼働計画を生成する。稼働計画生成部１５０は、数十分、１時間、または数時間ごとの稼働計画を生成してよい。 The operation plan generation unit 150 may generate an operation plan including a period during which the electrolysis device 20 is operated and a period during which the electrolysis device 20 is not operated in the first period. Further, the operation plan generation unit 150 may generate an operation plan indicating a period during which the electrolysis apparatus 20 is operated, together with an operation rate. It is desirable that the operation plan generation unit 150 generates an operation plan in which the operation rate changes in a time series. The operation plan generation unit 150 generates, for example, an operation plan for each fixed time. The operation plan generation unit 150 may generate an operation plan every tens of minutes, every hour, or every several hours.
　また、稼働計画生成部１５０は、制御部１６０が複数の電解装置２０を制御する場合、当該複数の電解装置２０のそれぞれに対する稼働計画を生成してよい。稼働計画生成部１５０は、複数の電解装置２０が略同一の電解装置２０である場合は、略同一の稼働計画をそれぞれ生成してよい。また、稼働計画生成部１５０は、制御部１６０が異なる種類の電解装置２０、異なる時期に購入した電解装置２０、異なる製造メーカの電解装置２０、またはこれらの組み合わせを含む複数の電解装置２０を制御する場合、それぞれの電解装置２０に対応して、異なる稼働計画をそれぞれ生成してよい。 In addition, when the control unit 160 controls a plurality of electrolyzers 20, the operation plan generator 150 may generate an operation plan for each of the plurality of electrolyzers 20. When a plurality of electrolyzers 20 are substantially the same electrolyzer 20, the operation plan generator 150 may generate substantially the same operation plan. In addition, the operation plan generation unit 150 controls the control unit 160 to control different types of electrolyzers 20, electrolyzers 20 purchased at different times, electrolyzers 20 of different manufacturers, or a plurality of electrolyzers 20 including a combination thereof. In this case, different operation plans may be generated for each of the electrolyzers 20.
　この場合、第２モデル生成部２３０は、複数の電解装置２０の稼働台数毎または複数の電解装置２０の組み合わせ毎にそれぞれ対応する複数の稼働計画生成モデルを生成してよい。また、第２モデル学習部２４５は、複数の稼働生成モデルをそれぞれ学習してよく、第２モデル更新部２４０は、複数の稼働生成モデルをそれぞれ更新してよい。稼働計画生成部１５０は、複数の稼働計画生成モデルのうち、第１期間における複数の電解装置２０の生成物の使用計画に応じた稼働計画生成モデルを用いて、第１期間における電解装置２０の稼働計画を生成してよい。また、第２モデル生成部２３０は、複数の電解装置２０に対応する１つの稼働計画生成モデルを生成し、第２モデル更新部２４０は、第２モデル学習部２４５により学習した稼働計画生成モデルを更新してよい。この場合、稼働計画生成モデルは、複数の電解装置２０を連携稼働させるための稼働計画を生成するモデルであってよく、一例として、複数の電解装置２０のそれぞれの稼働開始のタイミングおよび稼働期間等が最適化された稼働計画を生成するモデルであってよい。 In this case, the second model generation unit 230 may generate a plurality of operation plan generation models corresponding to each of the number of operating electrolysis devices 20 or each combination of the electrolysis devices 20. Further, the second model learning unit 245 may learn each of the plurality of operation generation models, and the second model updating unit 240 may update each of the plurality of operation generation models. The operation plan generation unit 150 uses the operation plan generation model corresponding to the use plan of the products of the plurality of electrolysis devices 20 in the first period, among the plurality of operation plan generation models, to generate the electrolysis device 20 in the first period. An operation plan may be generated. Further, the second model generation unit 230 generates one operation plan generation model corresponding to the plurality of electrolyzers 20, and the second model update unit 240 generates the operation plan generation model learned by the second model learning unit 245. May be updated. In this case, the operation plan generation model may be a model that generates an operation plan for cooperatively operating the plurality of electrolyzers 20. For example, the operation start timing and the operation period of each of the plurality of electrolyzers 20 may be used. May be a model that generates an optimized operation plan.
　制御部１６０は、稼働計画生成部１５０が生成した稼働計画を用いて、電解装置２０をＮ日分稼働させる（Ｓ３７０）。これにより、電解装置２０は、第１期間において、予め定められた生成物の供給計画を満たしつつ、生成物の製造コストを低減させるように稼働することができる。また、制御部１６０は、稼働計画を用いて、発電装置３０を制御し、発電装置３０から電解装置２０への電力供給および発電装置３０から電力系統５０への電力の売電を行ってよい。 The control unit 160 operates the electrolysis apparatus 20 for N days using the operation plan generated by the operation plan generation unit 150 (S370). Thereby, the electrolysis apparatus 20 can operate so as to reduce the production cost of the product while satisfying the predetermined product supply plan in the first period. In addition, the control unit 160 may control the power generation device 30 using the operation plan, and supply power from the power generation device 30 to the electrolysis device 20 and sell power from the power generation device 30 to the power system 50.
　計画装置４０が第１期間の経過後に電解装置２０の稼働を継続させる場合（Ｓ３８０：Ｎｏ）、Ｓ３３０に戻り、学習処理部１３０は学習モデルを適応学習する。この場合、取得部１００は、当該第１期間の第１因子および第３因子の情報と、当該第１期間の電解装置２０の稼働によって推移する第２因子の情報を順次取得し、記憶部１１０に順次記憶する。即ち、計画装置４０は、第１期間の情報を過去の情報に含め、予測すべき対象期間を第１期間よりも後の期間（一例として、第４期間）とする。 When the planning device 40 continues the operation of the electrolysis device 20 after the elapse of the first period (S380: No), the process returns to S330, and the learning processing unit 130 adaptively learns the learning model. In this case, the acquiring unit 100 sequentially acquires the information of the first factor and the third factor in the first period and the information of the second factor that changes due to the operation of the electrolysis device 20 in the first period, and stores the information in the storage unit 110. Are sequentially stored. That is, the planning device 40 includes the information of the first period in the past information, and sets the target period to be predicted to be a period later than the first period (for example, a fourth period).
　そして、計画装置４０は、第１期間の情報を用いてモデルの適応学習を繰り返し、一定期間の経過に応じてモデルを更新して、第４期間の電解装置２０の稼働計画を生成し、生成した稼働計画に応じて電解装置２０を稼働させる。このように、本実施形態に係る計画装置４０は、電解装置２０の対象期間の稼働計画の生成と、当該対象期間の稼働とを繰り返すことにより、学習モデルを更新しつつ電解装置２０を継続して稼働できる。 Then, the planning device 40 repeats the adaptive learning of the model using the information of the first period, updates the model as the certain period elapses, generates an operation plan of the electrolysis device 20 for the fourth period, and generates The electrolysis device 20 is operated according to the operation plan thus set. As described above, the planning device 40 according to the present embodiment continues the electrolysis device 20 while updating the learning model by repeatedly generating the operation plan of the electrolysis device 20 for the target period and operating the target period. Can operate.
　以上の計画装置４０の動作フローにおいて、第２期間、第３期間、第１期間、および第４期間の順に、計画装置４０を時系列に動作させる例を説明した。ここで、第２期間、第３期間、第１期間、および第４期間は、この順に時間的に連続した期間でよい。少なくとも、第１期間および第４期間は、連続した期間であることが望ましい。 In the above operation flow of the planning device 40, an example has been described in which the planning device 40 is operated in chronological order in the second period, the third period, the first period, and the fourth period. Here, the second period, the third period, the first period, and the fourth period may be temporally continuous in this order. It is preferable that at least the first period and the fourth period are continuous periods.
　本実施形態に係る計画装置４０は、発電装置３０の再生可能エネルギー発電量と電力系統５０の電気料金とを学習によって予測し、電解装置２０において使用計画に沿った量の生成物を低コストで生成可能な稼働計画を作成できる。 The planning device 40 according to the present embodiment predicts the amount of renewable energy generated by the power generation device 30 and the electricity rate of the power system 50 by learning, and generates a quantity of products according to the usage plan in the electrolysis device 20 at low cost. Create a work plan that can be generated.
　なお、本実施形態に係る計画装置４０は、第３モデル生成部２５０、第３モデル更新部２６０、および電気料金予測部２７０を備えていなくてもよい。この場合、計画装置４０で用いる電力系統５０の電気料金は、例えば、外部の装置で予測した電気料金、電力系統５０の事業者から提供された電気料金、または過去の電気料金であってもよい。また、電力系統５０の電気料金は、変動せずに一定であってもよい。 Note that the planning device 40 according to the present embodiment may not include the third model generation unit 250, the third model update unit 260, and the electricity rate prediction unit 270. In this case, the electricity rate of the power system 50 used by the planning device 40 may be, for example, an electricity rate predicted by an external device, an electricity rate provided by a company of the power system 50, or a past electricity rate. . Further, the electricity rate of the power system 50 may be constant without fluctuating.
　本発明の様々な実施形態は、フローチャートおよびブロック図を参照して記載されてよく、ここにおいてブロックは、（１）操作が実行されるプロセスの段階または（２）操作を実行する役割を持つ装置のセクションを表わしてよい。特定の段階およびセクションが、専用回路、コンピュータ可読媒体上に格納されるコンピュータ可読命令と共に供給されるプログラマブル回路、および／またはコンピュータ可読媒体上に格納されるコンピュータ可読命令と共に供給されるプロセッサによって実装されてよい。専用回路は、デジタルおよび／またはアナログハードウェア回路を含んでよく、集積回路（ＩＣ）および／またはディスクリート回路を含んでよい。プログラマブル回路は、論理ＡＮＤ、論理ＯＲ、論理ＸＯＲ、論理ＮＡＮＤ、論理ＮＯＲ、および他の論理操作、フリップフロップ、レジスタ、フィールドプログラマブルゲートアレイ（ＦＰＧＡ）、プログラマブルロジックアレイ（ＰＬＡ）等のようなメモリ要素等を含む、再構成可能なハードウェア回路を含んでよい。 Various embodiments of the invention may be described with reference to flowcharts and block diagrams, wherein blocks are (1) steps in a process in which an operation is performed or (2) devices responsible for performing an operation. Section. Certain stages and sections are implemented by dedicated circuitry, programmable circuitry provided with computer readable instructions stored on computer readable media, and / or processors provided with computer readable instructions stored on computer readable media. May be. Dedicated circuits may include digital and / or analog hardware circuits, and may include integrated circuits (ICs) and / or discrete circuits. Programmable circuits include logical AND, logical OR, logical XOR, logical NAND, logical NOR, and other logical operations, memory elements such as flip-flops, registers, field programmable gate arrays (FPGAs), programmable logic arrays (PLAs), etc. And the like, and may include reconfigurable hardware circuits.
　コンピュータ可読媒体は、適切なデバイスによって実行される命令を格納可能な任意の有形なデバイスを含んでよく、その結果、そこに格納される命令を有するコンピュータ可読媒体は、フローチャートまたはブロック図で指定された操作を実行するための手段を作成すべく実行され得る命令を含む、製品を備えることになる。コンピュータ可読媒体の例としては、電子記憶媒体、磁気記憶媒体、光記憶媒体、電磁記憶媒体、半導体記憶媒体等が含まれてよい。コンピュータ可読媒体のより具体的な例としては、フロッピー（登録商標）ディスク、ディスケット、ハードディスク、ランダムアクセスメモリ（ＲＡＭ）、リードオンリメモリ（ＲＯＭ）、消去可能プログラマブルリードオンリメモリ（ＥＰＲＯＭまたはフラッシュメモリ）、電気的消去可能プログラマブルリードオンリメモリ（ＥＥＰＲＯＭ）、静的ランダムアクセスメモリ（ＳＲＡＭ）、コンパクトディスクリードオンリメモリ（ＣＤ-ＲＯＭ）、デジタル多用途ディスク（ＤＶＤ）、ブルーレイ（ＲＴＭ）ディスク、メモリスティック、集積回路カード等が含まれてよい。 Computer readable media may include any tangible device capable of storing instructions for execution by a suitable device, such that computer readable media having instructions stored thereon is specified in a flowchart or block diagram. Product comprising instructions that can be executed to create a means for performing the specified operation. Examples of the computer readable medium may include an electronic storage medium, a magnetic storage medium, an optical storage medium, an electromagnetic storage medium, a semiconductor storage medium, and the like. More specific examples of computer readable media include floppy disks, diskettes, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM or flash memory), Electrically erasable programmable read only memory (EEPROM), static random access memory (SRAM), compact disk read only memory (CD-ROM), digital versatile disk (DVD), Blu-ray (RTM) disk, memory stick, integrated A circuit card or the like may be included.
　コンピュータ可読命令は、アセンブラ命令、命令セットアーキテクチャ（ＩＳＡ）命令、マシン命令、マシン依存命令、マイクロコード、ファームウェア命令、状態設定データ、またはＳｍａｌｌｔａｌｋ、ＪＡＶＡ（登録商標）、Ｃ＋＋等のようなオブジェクト指向プログラミング言語、Ｐｙｔｈｏｎ、および「Ｃ」プログラミング言語または同様のプログラミング言語のような従来の手続型プログラミング言語を含む、１または複数のプログラミング言語の任意の組み合わせで記述されたソースコードまたはオブジェクトコードのいずれかを含んでよい。 The computer readable instructions may be assembler instructions, instruction set architecture (ISA) instructions, machine instructions, machine dependent instructions, microcode, firmware instructions, state setting data, or object oriented programming such as Smalltalk, JAVA, C ++, etc. Any source or object code written in any combination of one or more programming languages, including conventional procedural programming languages, such as the language, Python, and the "C" programming language or similar programming languages. May include.
　コンピュータ可読命令は、汎用コンピュータ、特殊目的のコンピュータ、若しくは他のプログラム可能なデータ処理装置のプロセッサまたはプログラマブル回路に対し、ローカルにまたはローカルエリアネットワーク（ＬＡＮ）、インターネット等のようなワイドエリアネットワーク（ＷＡＮ）を介して提供され、フローチャートまたはブロック図で指定された操作を実行するための手段を作成すべく、コンピュータ可読命令を実行してよい。プロセッサの例としては、コンピュータプロセッサ、処理ユニット、マイクロプロセッサ、デジタル信号プロセッサ、コントローラ、マイクロコントローラ等を含む。 The computer readable instructions may be provided to a processor or programmable circuit of a general purpose computer, special purpose computer, or other programmable data processing device, either locally or over a wide area network (WAN) such as a local area network (LAN), the Internet, or the like. ) May be executed to create means for performing the operations specified in the flowcharts or block diagrams. Examples of processors include computer processors, processing units, microprocessors, digital signal processors, controllers, microcontrollers, and the like.
　図４は、本発明の複数の態様が全体的または部分的に具現化されてよいコンピュータ１９００の例を示す。コンピュータ１９００にインストールされたプログラムは、コンピュータ１９００に、本発明の実施形態に係る装置に関連付けられる操作または当該装置の１または複数のセクションとして機能させることができ、または当該操作または当該１または複数のセクションを実行させることができ、および／またはコンピュータ１９００に、本発明の実施形態に係るプロセスまたは当該プロセスの段階を実行させることができる。そのようなプログラムは、コンピュータ１９００に、本明細書に記載のフローチャートおよびブロック図のブロックのうちのいくつかまたはすべてに関連付けられた特定の操作を実行させるべく、ＣＰＵ２０００によって実行されてよい。 FIG. 4 illustrates an example of a computer 1900 in which aspects of the present invention may be wholly or partially implemented. The programs installed on the computer 1900 can cause the computer 1900 to function as one or more sections of the operation or the device associated with the device according to the embodiment of the present invention, or the operation or the one or more of the one or more devices. Sections may be executed and / or computer 1900 may execute a process or steps of a process according to an embodiment of the present invention. Such programs may be executed by CPU 2000 to cause computer 1900 to perform certain operations associated with some or all of the blocks in the flowcharts and block diagrams described herein.
　本実施形態に係るコンピュータ１９００は、ホスト・コントローラ２０８２により相互に接続されるＣＰＵ２０００、ＲＡＭ２０２０、グラフィック・コントローラ２０７５、及び表示装置２０８０を有するＣＰＵ周辺部と、入出力コントローラ２０８４によりホスト・コントローラ２０８２に接続される通信インターフェイス２０３０、ハードディスクドライブ２０４０、及びＤＶＤドライブ２０６０を有する入出力部と、入出力コントローラ２０８４に接続されるＲＯＭ２０１０、フラッシュメモリ・ドライブ２０５０、及び入出力チップ２０７０を有するレガシー入出力部を備える。 A computer 1900 according to the present embodiment is connected to a host computer 2082 including a CPU 2000, a RAM 2020, a graphic controller 2075, and a display device 2080, which are mutually connected by a host controller 2082, and an input / output controller 2084 to the host controller 2082. Input / output unit having a communication interface 2030, a hard disk drive 2040, and a DVD drive 2060, and a legacy input / output unit having a ROM 2010, a flash memory drive 2050, and an input / output chip 2070 connected to an input / output controller 2084. .
　ホスト・コントローラ２０８２は、ＲＡＭ２０２０と、高い転送レートでＲＡＭ２０２０をアクセスするＣＰＵ２０００及びグラフィック・コントローラ２０７５とを接続する。ＣＰＵ２０００は、ＲＯＭ２０１０及びＲＡＭ２０２０に格納されたプログラムに基づいて動作し、各部の制御を行う。グラフィック・コントローラ２０７５は、ＣＰＵ２０００等がＲＡＭ２０２０内に設けたフレーム・バッファ上に生成する画像データを取得し、表示装置２０８０上に表示させる。これに代えて、グラフィック・コントローラ２０７５は、ＣＰＵ２０００等が生成する画像データを格納するフレーム・バッファを、内部に含んでもよい。 (4) The host controller 2082 connects the RAM 2020 to the CPU 2000 and the graphic controller 2075 that access the RAM 2020 at a high transfer rate. The CPU 2000 operates based on programs stored in the ROM 2010 and the RAM 2020, and controls each unit. The graphic controller 2075 acquires image data generated by the CPU 2000 or the like on a frame buffer provided in the RAM 2020, and displays the image data on the display device 2080. Alternatively, the graphic controller 2075 may include a frame buffer for storing image data generated by the CPU 2000 or the like.
　入出力コントローラ２０８４は、ホスト・コントローラ２０８２と、比較的高速な入出力装置である通信インターフェイス２０３０、ハードディスクドライブ２０４０、ＤＶＤドライブ２０６０を接続する。通信インターフェイス２０３０は、有線又は無線によりネットワークを介して他の装置と通信する。また、通信インターフェイスは、通信を行うハードウェアとして機能する。ハードディスクドライブ２０４０は、コンピュータ１９００内のＣＰＵ２０００が使用するプログラム及びデータを格納する。ＤＶＤドライブ２０６０は、ＤＶＤ２０９５からプログラム又はデータを読み取り、ＲＡＭ２０２０を介してハードディスクドライブ２０４０に提供する。 The input / output controller 2084 connects the host controller 2082 to the communication interface 2030, the hard disk drive 2040, and the DVD drive 2060, which are relatively high-speed input / output devices. The communication interface 2030 communicates with another device via a network by wire or wirelessly. The communication interface functions as hardware for performing communication. The hard disk drive 2040 stores programs and data used by the CPU 2000 in the computer 1900. The DVD drive 2060 reads a program or data from the DVD 2095 and provides it to the hard disk drive 2040 via the RAM 2020.
　また、入出力コントローラ２０８４には、ＲＯＭ２０１０と、フラッシュメモリ・ドライブ２０５０、及び入出力チップ２０７０の比較的低速な入出力装置とが接続される。ＲＯＭ２０１０は、コンピュータ１９００が起動時に実行するブート・プログラム、及び／又は、コンピュータ１９００のハードウェアに依存するプログラム等を格納する。フラッシュメモリ・ドライブ２０５０は、フラッシュメモリ２０９０からプログラム又はデータを読み取り、ＲＡＭ２０２０を介してハードディスクドライブ２０４０に提供する。入出力チップ２０７０は、フラッシュメモリ・ドライブ２０５０を入出力コントローラ２０８４へと接続するとともに、例えばパラレル・ポート、シリアル・ポート、キーボード・ポート、マウス・ポート等を介して各種の入出力装置を入出力コントローラ２０８４へと接続する。 The ROM 2010, the flash memory drive 2050, and the relatively low-speed input / output device of the input / output chip 2070 are connected to the input / output controller 2084. The ROM 2010 stores a boot program executed by the computer 1900 at the time of startup, and / or a program depending on hardware of the computer 1900. The flash memory drive 2050 reads a program or data from the flash memory 2090 and provides it to the hard disk drive 2040 via the RAM 2020. The input / output chip 2070 connects the flash memory drive 2050 to the input / output controller 2084 and inputs / outputs various input / output devices via, for example, a parallel port, a serial port, a keyboard port, a mouse port, and the like. Connect to controller 2084.
　ＲＡＭ２０２０を介してハードディスクドライブ２０４０に提供されるプログラムは、フラッシュメモリ２０９０、ＤＶＤ２０９５、又はＩＣカード等の記録媒体に格納されて利用者によって提供される。プログラムは、記録媒体から読み出され、ＲＡＭ２０２０を介してコンピュータ１９００内のハードディスクドライブ２０４０にインストールされ、ＣＰＵ２０００において実行される。これらのプログラム内に記述される情報処理は、コンピュータ１９００に読み取られ、ソフトウェアと、上記様々なタイプのハードウェア資源との間の協働をもたらす。装置または方法が、コンピュータ１９００の使用に従い情報の操作または処理を実現することによって構成されてよい。 The program provided to the hard disk drive 2040 via the RAM 2020 is stored in a recording medium such as a flash memory 2090, a DVD 2095, or an IC card and provided by a user. The program is read from the recording medium, installed on the hard disk drive 2040 in the computer 1900 via the RAM 2020, and executed by the CPU 2000. The information processing described in these programs is read by the computer 1900 and provides for cooperation between the software and the various types of hardware resources described above. An apparatus or method may be configured for implementing manipulation or processing of information according to the use of computer 1900.
　一例として、コンピュータ１９００と外部の装置等との間で通信を行う場合には、ＣＰＵ２０００は、ＲＡＭ２０２０上にロードされた通信プログラムを実行し、通信プログラムに記述された処理内容に基づいて、通信インターフェイス２０３０に対して通信処理を指示する。通信インターフェイス２０３０は、ＣＰＵ２０００の制御を受けて、ＲＡＭ２０２０、ハードディスクドライブ２０４０、フラッシュメモリ２０９０、又はＤＶＤ２０９５等の記憶装置上に設けた送信バッファ領域等に記憶された送信データを読み出してネットワークへと送信し、もしくは、ネットワークから受信した受信データを記憶装置上に設けた受信バッファ領域等へと書き込む。このように、通信インターフェイス２０３０は、ＤＭＡ（ダイレクト・メモリ・アクセス）方式により記憶装置との間で送受信データを転送してもよく、これに代えて、ＣＰＵ２０００が転送元の記憶装置又は通信インターフェイス２０３０からデータを読み出し、転送先の通信インターフェイス２０３０又は記憶装置へとデータを書き込むことにより送受信データを転送してもよい。 As an example, when performing communication between the computer 1900 and an external device or the like, the CPU 2000 executes a communication program loaded on the RAM 2020, and executes a communication interface based on the processing content described in the communication program. The communication processing is instructed to 2030. Under the control of the CPU 2000, the communication interface 2030 reads out transmission data stored in a transmission buffer area provided on a storage device such as the RAM 2020, the hard disk drive 2040, the flash memory 2090, or the DVD 2095, and transmits the data to the network. Alternatively, it writes the received data received from the network to a reception buffer area or the like provided on the storage device. As described above, the communication interface 2030 may transfer the transmission / reception data to / from the storage device by the DMA (Direct Memory Access) method, and instead, the CPU 2000 may use the transfer source storage device or the communication interface 2030. The data may be read from the communication interface 2030 or the data may be written to the communication interface 2030 or the storage device of the transfer destination to transfer the transmission and reception data.
　また、ＣＰＵ２０００は、ハードディスクドライブ２０４０、ＤＶＤドライブ２０６０（ＤＶＤ２０９５）、フラッシュメモリ・ドライブ２０５０（フラッシュメモリ２０９０）等の外部記憶装置に格納されたファイルまたはデータベース等の中から、全部または必要な部分をＤＭＡ転送等によりＲＡＭ２０２０へと読み込ませ、ＲＡＭ２０２０上のデータに対して各種の処理を行う。そして、ＣＰＵ２０００は、処理を終えたデータを、ＤＭＡ転送等により外部記憶装置へと書き戻す。このような処理において、ＲＡＭ２０２０は、外部記憶装置の内容を一時的に保持するものとみなせるから、本実施形態においてはＲＡＭ２０２０及び外部記憶装置等をメモリ、記憶部、または記憶装置等と総称する。 Further, the CPU 2000 transfers all or a necessary portion from a file or a database stored in an external storage device such as a hard disk drive 2040, a DVD drive 2060 (DVD 2095), or a flash memory drive 2050 (flash memory 2090) to a DMA. The data is read into the RAM 2020 by transfer or the like, and various processes are performed on the data on the RAM 2020. Then, the CPU 2000 writes the processed data back to the external storage device by DMA transfer or the like. In such a process, the RAM 2020 can be regarded as temporarily holding the contents of the external storage device. Therefore, in this embodiment, the RAM 2020 and the external storage device are collectively referred to as a memory, a storage unit, or a storage device.
　本実施形態における各種のプログラム、データ、テーブル、データベース等の各種の情報は、このような記憶装置上に格納されて、情報処理の対象となる。なお、ＣＰＵ２０００は、ＲＡＭ２０２０の一部をキャッシュメモリに保持し、キャッシュメモリ上で読み書きを行うこともできる。このような形態においても、キャッシュメモリはＲＡＭ２０２０の機能の一部を担うから、本実施形態においては、区別して示す場合を除き、キャッシュメモリもＲＡＭ２０２０、メモリ、及び／又は記憶装置に含まれるものとする。 各種 Various information such as various programs, data, tables, and databases in the present embodiment are stored on such a storage device and are subjected to information processing. Note that the CPU 2000 can also hold a part of the RAM 2020 in a cache memory and perform reading and writing on the cache memory. Even in such a form, the cache memory plays a part of the function of the RAM 2020. Therefore, in the present embodiment, the cache memory is also included in the RAM 2020, the memory, and / or the storage device unless otherwise indicated. I do.
　また、ＣＰＵ２０００は、ＲＡＭ２０２０から読み出したデータに対して、プログラムの命令列により指定された、本実施形態中に記載した各種の演算、情報の加工、条件判断、情報の検索・置換等を含む各種の処理を行い、ＲＡＭ２０２０へと書き戻す。例えば、ＣＰＵ２０００は、条件判断を行う場合においては、本実施形態において示した各種の変数が、他の変数または定数と比較して、大きい、小さい、以上、以下、等しい等の条件を満たすか否かを判断し、条件が成立した場合（又は不成立であった場合）に、異なる命令列へと分岐し、またはサブルーチンを呼び出す。 In addition, the CPU 2000 performs various calculations, information processing, condition determination, information search / replacement, and the like described in the present embodiment on the data read from the RAM 2020, as specified by the instruction sequence of the program. And write it back to the RAM 2020. For example, when performing the condition determination, the CPU 2000 determines whether the various variables described in the present embodiment satisfy conditions such as larger, smaller, greater than, less than, equal to, and the like as compared with other variables or constants. Then, if the condition is satisfied (or not satisfied), a branch is made to a different instruction sequence or a subroutine is called.
　また、ＣＰＵ２０００は、記憶装置内のファイルまたはデータベース等に格納された情報を検索することができる。例えば、第１属性の属性値に対し第２属性の属性値がそれぞれ対応付けられた複数のエントリが記憶装置に格納されている場合において、ＣＰＵ２０００は、記憶装置に格納されている複数のエントリの中から第１属性の属性値が指定された条件と一致するエントリを検索し、そのエントリに格納されている第２属性の属性値を読み出すことにより、所定の条件を満たす第１属性に対応付けられた第２属性の属性値を得ることができる。 The CPU 2000 can search for information stored in a file or a database in the storage device. For example, in the case where a plurality of entries in which the attribute value of the second attribute is associated with the attribute value of the first attribute are stored in the storage device, the CPU 2000 determines whether the plurality of entries stored in the storage device By searching for an entry in which the attribute value of the first attribute matches the specified condition and reading the attribute value of the second attribute stored in the entry, the entry is associated with the first attribute satisfying the predetermined condition. The attribute value of the obtained second attribute can be obtained.
　また、実施形態の説明において複数の要素が列挙された場合には、列挙された要素以外の要素を用いてもよい。例えば、「Ｘは、Ａ、Ｂ及びＣを用いてＹを実行する」と記載される場合、Ｘは、Ａ、Ｂ及びＣに加え、Ｄを用いてＹを実行してもよい。 When a plurality of elements are listed in the description of the embodiment, elements other than the listed elements may be used. For example, if "X performs Y using A, B, and C", X may perform Y using D in addition to A, B, and C.
　以上、本発明を実施の形態を用いて説明したが、本発明の技術的範囲は上記実施の形態に記載の範囲には限定されない。上記実施の形態に、多様な変更または改良を加えることが可能であることが当業者に明らかである。その様な変更または改良を加えた形態も本発明の技術的範囲に含まれ得ることが、請求の範囲の記載から明らかである。 Although the present invention has been described using the embodiments, the technical scope of the present invention is not limited to the scope described in the above embodiments. It is apparent to those skilled in the art that various changes or improvements can be made to the above embodiment. It is apparent from the description of the appended claims that embodiments with such changes or improvements can be included in the technical scope of the present invention.
　請求の範囲、明細書、および図面中において示した装置、システム、プログラム、および方法における動作、手順、ステップ、および段階等の各処理の実行順序は、特段「より前に」、「先立って」等と明示しておらず、また、前の処理の出力を後の処理で用いるのでない限り、任意の順序で実現しうることに留意すべきである。請求の範囲、明細書、および図面中の動作フローに関して、便宜上「まず、」、「次に、」等を用いて説明したとしても、この順で実施することが必須であることを意味するものではない。 The execution order of each processing such as operation, procedure, step, and step in the apparatus, system, program, and method shown in the claims, the description, and the drawings is particularly “before” or “before”. It should be noted that they can be realized in any order as long as the output of the previous process is not used in the subsequent process. Even if the operation flow in the claims, the description, and the drawings is described using “first”, “next”, etc. for convenience, it means that it is essential to implement in this order is not.
１０　システム
２０　電解装置
３０　発電装置
４０　計画装置
５０　電力系統
１００　取得部
１１０　記憶部
１２０　モデル生成部
１３０　学習処理部
１４０　予測部
１５０　稼働計画生成部
１６０　制御部
２００　第１モデル生成部
２１０　第１モデル更新部
２１５　第１モデル学習部
２２０　発電量予測部
２３０　第２モデル生成部
２４０　第２モデル更新部
２４５　第２モデル学習部
２５０　第３モデル生成部
２６０　第３モデル更新部
２６５　第３モデル学習部
２７０　電気料金予測部
１９００　コンピュータ
２０００　ＣＰＵ
２０１０　ＲＯＭ
２０２０　ＲＡＭ
２０３０　通信インターフェイス
２０４０　ハードディスクドライブ
２０５０　フラッシュメモリ・ドライブ
２０６０　ＤＶＤドライブ
２０７０　入出力チップ
２０７５　グラフィック・コントローラ
２０８０　表示装置
２０８２　ホスト・コントローラ
２０８４　入出力コントローラ
２０９０　フラッシュメモリ
２０９５　ＤＶＤ 10 System 20 Electrolysis device 30 Power generation device 40 Planning device 50 Power system 100 Acquisition unit 110 Storage unit 120 Model generation unit 130 Learning processing unit 140 Prediction unit 150 Operation plan generation unit 160 Control unit 200 First model generation unit 210 First model update Unit 215 first model learning unit 220 power generation prediction unit 230 second model generation unit 240 second model update unit 245 second model learning unit 250 third model generation unit 260 third model update unit 265 third model learning unit 270 electricity Charge prediction unit 1900 Computer 2000 CPU
2010 ROM
2020 RAM
2030 Communication interface 2040 Hard disk drive 2050 Flash memory drive 2060 DVD drive 2070 Input / output chip 2075 Graphic controller 2080 Display device 2082 Host controller 2084 Input / output controller 2090 Flash memory 2095 DVD

Claims (12)

  1. 　対象期間における発電装置の再生可能エネルギー発電量の推移を対象期間よりも前に入手可能な第１因子の値に基づいて予測する発電量予測モデルを用いて、将来の再生可能エネルギー発電量の推移を予測する発電量予測部と、
    　予測された前記将来の再生可能エネルギー発電量の推移と電力系統の電気料金とに基づいて、将来の第１期間における、電解装置の生成物の使用計画を満たす前記電解装置の稼働計画を生成する稼働計画生成部と
    　を備える計画装置。     Changes in future renewable energy power generation using a power generation prediction model that predicts the change in renewable energy power generation of the power generator during the target period based on the value of the first factor available before the target period A power generation prediction unit for predicting
    An operation plan of the electrolysis apparatus that satisfies the use plan of the product of the electrolysis apparatus in the first period in the future is generated based on the predicted transition of the future renewable energy power generation amount and the electricity rate of the power system. A planning device comprising: an operation plan generation unit.
  2. 　前記発電量予測モデルは、対象期間よりも前の、前記発電装置の再生可能エネルギー発電量および天気情報の少なくとも１つを含む前記第１因子の値に基づいて、対象期間における前記再生可能エネルギー発電量の推移を予測する
    　請求項１に記載の計画装置。     The power generation amount prediction model is based on a value of the first factor including at least one of a renewable energy generation amount of the power generation device and weather information before a target period. The planning device according to claim 1, which predicts a change in the amount.
  3. 　過去期間における前記第１因子の値と前記過去期間以降の前記再生可能エネルギー発電量の現実の推移とに基づいて、前記発電量予測モデルを学習により更新する第１モデル更新部を更に備える
    　請求項２に記載の計画装置。     A first model updating unit that updates the power generation amount prediction model by learning based on a value of the first factor in a past period and an actual transition of the renewable energy power generation amount after the past period. 3. The planning device according to 2.
  4. 　前記稼働計画生成部は、前記第１期間中において、前記電解装置の生成物の使用計画を満たすように、前記発電装置からの電気を前記電力系統からの電気より優先して用いて前記電解装置を稼働させる稼働計画を生成する
    　請求項１から３のいずれか一項に記載の計画装置。     The operation plan generation unit, during the first period, so as to satisfy the usage plan of the product of the electrolysis device, the electrolysis device by using the electricity from the power generation device with priority over the electricity from the power system A plan apparatus according to any one of claims 1 to 3, wherein the plan apparatus generates an operation plan for operating the system.
  5. 　前記電解装置の生成物の使用計画は、前記第１期間における前記電解装置の前記生成物の貯蔵量を基準範囲内に維持する計画、および前記第１期間における前記電解装置の前記生成物の需要量を満たす計画の少なくとも１つを含む
    　請求項１から４のいずれか一項に記載の計画装置。     The use plan of the product of the electrolyzer is a plan for maintaining the storage amount of the product of the electrolyzer in the first period within a reference range, and the demand for the product of the electrolyzer in the first period. The planning device according to any one of claims 1 to 4, comprising at least one of plans satisfying the amount.
  6. 　前記稼働計画生成部は、対象期間における稼働計画を、対象期間よりも前に入手可能な第２因子の値と、対象期間における前記再生可能エネルギー発電量の推移の予測結果と、前記電気料金とに基づいて生成する稼働計画生成モデルを用いて、将来の前記第１期間における前記電解装置の稼働計画を生成する
    　請求項１から５のいずれか一項に記載の計画装置。     The operation plan generation unit, the operation plan in the target period, the value of the second factor available before the target period, the predicted result of the change in the amount of renewable energy power generation in the target period, and the electricity rate The planning device according to any one of claims 1 to 5, wherein an operation plan of the electrolysis device in the future first period is generated using an operation plan generation model generated based on the operation plan.
  7. 　前記稼働計画生成モデルは、対象期間よりも前の、前記電解装置の稼働データ、前記電解装置の生成物の需要量、および前記電解装置の生成物の貯蔵量の少なくとも１つを含む前記第２因子の値と、対象期間における前記再生可能エネルギー発電量の推移の予測結果と、前記電気料金とに基づいて、対象期間における前記電解装置の稼働計画を生成する
    　請求項６に記載の計画装置。     The operation plan generation model includes at least one of operation data of the electrolysis device, a demand of a product of the electrolysis device, and a storage amount of a product of the electrolysis device before a target period. The planning device according to claim 6, wherein an operation plan of the electrolytic device in a target period is generated based on a value of a factor, a prediction result of a change in the amount of generated renewable energy in a target period, and the electricity rate.
  8. 　過去期間における前記第２因子の値と、前記過去期間以降における前記再生可能エネルギー発電量の推移または前記再生可能エネルギー発電量の推移の予測結果と、前記過去期間以降における前記電気料金の推移と、前記過去期間以降において目標とすべき前記電解装置の稼働計画とに基づいて、前記稼働計画生成モデルを学習により更新する第２モデル更新部を更に備える
    　請求項６または７に記載の計画装置。     The value of the second factor in the past period, the transition of the renewable energy generation amount or the prediction result of the transition of the renewable energy generation amount after the past period, and the transition of the electricity rate after the past period, The planning device according to claim 6, further comprising: a second model updating unit that updates the operation plan generation model by learning based on an operation plan of the electrolysis apparatus to be targeted in the past period or later.
  9. 　対象期間における前記電力系統の電気料金の推移を、対象期間よりも前に入手可能な第３因子の値に基づいて予測する電気料金予測モデルを用いて、将来の電気料金の推移を予測する電気料金予測部を更に備え、
    　前記稼働計画生成部は、予測された前記将来の再生可能エネルギー発電量の推移と、予測された前記将来の電気料金の推移とに基づいて、前記将来の第１期間における、前記電解装置の生成物の使用計画を満たす前記電解装置の稼働計画を生成する
    　請求項１から８のいずれか一項に記載の計画装置。     An electricity tariff prediction model that predicts a transition of the electricity tariff of the power system in the target period based on a value of the third factor available before the target period, and predicts a future transition of the electricity tariff. It further includes a charge forecasting section,
    The operation plan generation unit is configured to generate the electrolysis device in the first future period based on the predicted transition of the future renewable energy power generation amount and the predicted transition of the future electricity rate. The planning device according to any one of claims 1 to 8, wherein an operation plan of the electrolysis device that satisfies a use plan of the object is generated.
  10. 　前記電気料金予測モデルは、対象期間よりも前の、前記電力系統における電気料金、電力需要量、電力供給量、再生可能エネルギー発電量、再生可能エネルギー発電量の予測値、および天気情報の少なくとも１つを含む前記第３因子の値に基づいて、対象期間における前記電力系統の電気料金の推移を予測する
    　請求項９に記載の計画装置。     The electricity rate prediction model includes at least one of an electricity rate, a power demand amount, a power supply amount, a renewable energy generation amount, a predicted value of the renewable energy generation amount, and weather information in the power system before a target period. The planning device according to claim 9, wherein a transition of an electricity rate of the power system in a target period is predicted based on a value of the third factor including the following.
  11. 　対象期間における発電装置の再生可能エネルギー発電量の推移を対象期間よりも前に入手可能な第１因子の値に基づいて予測する発電量予測モデルを用いて、将来の再生可能エネルギー発電量の推移を予測する段階と、
    　予測された前記将来の再生可能エネルギー発電量の推移と電力系統の電気料金とに基づいて、将来の第１期間における、電解装置の生成物の使用計画を満たす前記電解装置の稼働計画を生成する段階と
    　を備える方法。     Changes in future renewable energy power generation using a power generation prediction model that predicts the change in renewable energy power generation of the power generator during the target period based on the value of the first factor available before the target period Predicting the
    An operation plan of the electrolysis device that satisfies the usage plan of the product of the electrolysis device in the first period in the future is generated based on the predicted transition of the future renewable energy power generation amount and the electricity rate of the power system. A method comprising steps and.
  12. 　コンピュータに、請求項１から１０のいずれか一項に記載の計画装置として機能させるプログラム。 A program that causes a computer to function as the planning device according to any one of claims 1 to 10.
PCT/JP2019/036524 2018-09-21 2019-09-18 Planning apparatus, method, and program WO2020059741A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2018177795A JP7237500B2 (en) 2018-09-21 2018-09-21 Planning device, method and program
JP2018-177795 2018-09-21

Publications (1)

Publication Number Publication Date
WO2020059741A1 true WO2020059741A1 (en) 2020-03-26

Family

ID=69887223

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2019/036524 WO2020059741A1 (en) 2018-09-21 2019-09-18 Planning apparatus, method, and program

Country Status (2)

Country Link
JP (1) JP7237500B2 (en)
WO (1) WO2020059741A1 (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112859591A (en) * 2020-12-23 2021-05-28 华电电力科学研究院有限公司 Reinforced learning control system for operation optimization of energy system

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2023190553A1 (en) * 2022-03-31 2023-10-05 旭化成株式会社 Information processing device, electrolysis device, plan creation method, and program

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2017026287A1 (en) * 2015-08-07 2017-02-16 シャープ株式会社 Control device, energy management device, system, and control method
JP2017139850A (en) * 2016-02-02 2017-08-10 株式会社日立製作所 Power supply system

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2002180281A (en) 2000-12-19 2002-06-26 Shinko Pantec Co Ltd Hydrogen utilizing system
JP2007523580A (en) 2004-01-23 2007-08-16 スチュアート エナジー システムズ コーポレーション Energy network
US8019445B2 (en) 2004-06-15 2011-09-13 Intelligent Generation Llc Method and apparatus for optimization of distributed generation
JP2006235846A (en) 2005-02-23 2006-09-07 Honda Motor Co Ltd Hydrogen supply system, hydrogen supply control server, and hydrogen supply device
US7645931B2 (en) 2007-03-27 2010-01-12 Gm Global Technology Operations, Inc. Apparatus to reduce the cost of renewable hydrogen fuel generation by electrolysis using combined solar and grid power
JP6323890B2 (en) 2014-11-14 2018-05-16 Ｎｅｃソリューションイノベータ株式会社 Schedule management device, schedule management method, and program
JP6189448B2 (en) 2015-07-21 2017-08-30 株式会社東芝 Power supply system, control device, and power supply method
DE202018006783U1 (en) 2017-09-25 2022-12-20 Asahi Kasei Kabushiki Kaisha Hydrogen production planning device and program

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2017026287A1 (en) * 2015-08-07 2017-02-16 シャープ株式会社 Control device, energy management device, system, and control method
JP2017139850A (en) * 2016-02-02 2017-08-10 株式会社日立製作所 Power supply system

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112859591A (en) * 2020-12-23 2021-05-28 华电电力科学研究院有限公司 Reinforced learning control system for operation optimization of energy system
CN112859591B (en) * 2020-12-23 2022-10-21 华电电力科学研究院有限公司 Reinforced learning control system for operation optimization of energy system

Also Published As

Publication number Publication date
JP7237500B2 (en) 2023-03-13
JP2021193471A (en) 2021-12-23

Similar Documents

Publication Publication Date Title
JP7221376B2 (en) Apparatus, method and program
JP7061616B2 (en) Planning equipment, operation plan generation method, hydrogen production method, and program
US20210399575A1 (en) Planning apparatus, controlling apparatus, method, and rcording medium
Huang et al. Deep reinforcement learning based preventive maintenance policy for serial production lines
JP7157815B2 (en) Planning equipment, planning methods and planning programs
EP4222562B1 (en) Computer system and method providing operating instructions for thermal control of a blast furnace
Zhao et al. A two-stage online prediction method for a blast furnace gas system and its application
CA3116047C (en) A planning system and method for generating and changing a maintenance work period
WO2020059741A1 (en) Planning apparatus, method, and program
JP2022065773A (en) Control device, controller, control system, control method, and control program
Wood Trend decomposition aids short-term countrywide wind capacity factor forecasting with machine and deep learning methods
Safaraliev et al. Medium-term forecasting of power generation by hydropower plants in isolated power systems under climate change
Shi et al. Rolling horizon wind-thermal unit commitment optimization based on deep reinforcement learning
CN113723717B (en) Method, device, equipment and readable storage medium for predicting short-term load before system day
JP7487704B2 (en) EVALUATION APPARATUS, EVALUATION METHOD, EVALUATION PROGRAM, CONTROL APPARATUS, AND CONTROL PROGRAM
Al-Haija et al. Time-series model for forecasting short-term future additions of renewable energy to worldwide capacity
JP7219808B2 (en) PLANNING DEVICE, OPERATION PLAN GENERATION METHOD, HYDROGEN PRODUCTION METHOD, AND PROGRAM
JP2022138621A (en) Control device, control method, and control program
EP4283412A1 (en) Model selection apparatus, model selection method, and model selection program
Devianto et al. Hidden Markov Model as the Predictive Time Series Movement of Natural Gas Price
Tang et al. A novel self-adaptive nonlinear discrete grey model based on non-homogeneous index sequence and its application
Xu et al. Bi-level hierarchical model with deep reinforcement learning-based extended horizon scheduling for integrated electricity-heat systems
Tungom et al. SWOAM: Swarm optimized agents for energy management in grid-interactive connected buildings
Luo et al. Integrating Reinforcement Learning and LSTM for Raw Material Procurement Optimization in Salt Lake Chemical Manufacturing
CN116523111A (en) Mid-term load prediction method, system and equipment based on season ARIMA and GPR

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19863431

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19863431

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: JP