CN115545429A - Multi-scene low-voltage renewable resource load prediction method based on CPSS framework - Google Patents
Multi-scene low-voltage renewable resource load prediction method based on CPSS framework Download PDFInfo
- Publication number
- CN115545429A CN115545429A CN202211147957.XA CN202211147957A CN115545429A CN 115545429 A CN115545429 A CN 115545429A CN 202211147957 A CN202211147957 A CN 202211147957A CN 115545429 A CN115545429 A CN 115545429A
- Authority
- CN
- China
- Prior art keywords
- load
- user
- cpss
- load prediction
- scene
- Prior art date
- Legal status (The legal status 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 status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 38
- ODKSFYDXXFIFQN-BYPYZUCNSA-N L-arginine Chemical compound OC(=O)[C@@H](N)CCCN=C(N)N ODKSFYDXXFIFQN-BYPYZUCNSA-N 0.000 title claims abstract 14
- 230000005611 electricity Effects 0.000 claims abstract description 52
- 238000012546 transfer Methods 0.000 claims abstract description 27
- 230000004044 response Effects 0.000 claims abstract description 16
- 230000004927 fusion Effects 0.000 claims abstract description 14
- 230000007246 mechanism Effects 0.000 claims abstract description 9
- 230000006870 function Effects 0.000 claims description 31
- 238000010248 power generation Methods 0.000 claims description 13
- 238000004146 energy storage Methods 0.000 claims description 12
- 230000008878 coupling Effects 0.000 claims description 11
- 238000010168 coupling process Methods 0.000 claims description 11
- 238000005859 coupling reaction Methods 0.000 claims description 11
- 238000009826 distribution Methods 0.000 claims description 9
- 230000005484 gravity Effects 0.000 claims description 6
- 238000004422 calculation algorithm Methods 0.000 claims description 4
- 238000005096 rolling process Methods 0.000 claims description 4
- 230000001105 regulatory effect Effects 0.000 claims description 2
- 230000006399 behavior Effects 0.000 abstract description 15
- 230000003993 interaction Effects 0.000 abstract description 15
- 230000011273 social behavior Effects 0.000 abstract description 2
- 230000008569 process Effects 0.000 description 6
- 229910052799 carbon Inorganic materials 0.000 description 5
- 230000008901 benefit Effects 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 4
- 238000004891 communication Methods 0.000 description 4
- 238000007726 management method Methods 0.000 description 4
- 238000003062 neural network model Methods 0.000 description 4
- 238000005457 optimization Methods 0.000 description 4
- 238000012549 training Methods 0.000 description 4
- 238000000342 Monte Carlo simulation Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000009977 dual effect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 238000010220 Pearson correlation analysis Methods 0.000 description 1
- 230000004931 aggregating effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000005284 excitation Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000035699 permeability Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000008255 psychological mechanism Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
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
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06315—Needs-based resource requirements planning or analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/04—Inference or reasoning models
- G06N5/048—Fuzzy inferencing
-
- 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
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Tourism & Hospitality (AREA)
- Fuzzy Systems (AREA)
- Evolutionary Computation (AREA)
- Development Economics (AREA)
- Automation & Control Theory (AREA)
- Operations Research (AREA)
- Educational Administration (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Quality & Reliability (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Game Theory and Decision Science (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses a CPSS framework-based multi-scene low-voltage renewable resource load prediction method, which comprises the steps of constructing a micro-grid based on a CPSS framework of an cyber-physical-social fusion system, considering an energy management system with double uncertainties of renewable energy access and user social behavior influence in a multi-scene fusion mode, and establishing a controllable standardized model capable of perceiving interaction and a load prediction system from three layers of a physical layer, an information layer and a social layer. The method comprises the steps of establishing a charging load prediction model considering the factors of electricity price, current battery charge and parking time for the electric automobile, establishing a functional relation between peak-valley electricity price difference and load transfer rate based on consumer psychology for a user electricity utilization scene, establishing a demand response model based on the consumer psychology mechanism, and fully considering user behavior characteristics. And (4) by combining the models, multi-scene fusion load prediction is carried out on the low-voltage renewable energy and the load under the CPSS framework.
Description
Technical Field
The invention relates to the technical field of low-voltage renewable resource load prediction, in particular to a multi-scene low-voltage renewable resource load prediction method based on a CPSS framework.
Background
In recent years, china proposes a double-carbon target, and builds a novel power system mainly based on new energy to solve the urgent energy transformation problem. One of the key points of the novel power system under the double-carbon transformation is to fully mine and expand the information value of power and optimize the user behavior of the power, and combine the renewable energy on the load side and the electric vehicle to perform comprehensive overall micro-grid load prediction, so as to provide a reliable basis for system scheduling.
With the continuous improvement of the permeability of renewable energy sources, the user side gradually embodies the source-load dual characteristics. The number of power users who spontaneously use the photovoltaic generator set is continuously increased, on one hand, practical benefits are brought to the users, the construction pressure of newly added capacity of a power distribution network is relieved, on the other hand, new challenges are provided for traditional load total prediction from top to bottom, and multiple uncertainties are brought. In addition, the uncertainty of the user side is also amplified by the vigorous development of the electric automobile, so that the research on the user behavior is very important, and a corresponding appropriate prediction strategy should be formulated for load prediction of different scenes.
Due to the coupling of system planning and operation, a single load prediction scene is difficult to fit the operation condition of mutual influence of multiple resources in the actual operation process. The prediction error based on the traditional single load prediction scene is larger, the prediction is more difficult to implement in practical application, clear description is lacked for load prediction of different scenes and user electricity utilization response behaviors, and potential risks are brought to planning and operation of an urban power grid. Therefore, the load prediction needs to integrate the conditions of various resources and multiple scenes and access the power grid to perform prediction according to a unified standard.
By searching the prior art, a power load prediction method (application number CN201811242485. X) suitable for multiple scenes is found, load prediction influence factors in an area to be predicted are obtained and ranked, historical data of the load prediction influence factors in the area to be predicted are obtained, and a load prediction model of the area to be predicted is established to predict the power load of the area to be predicted. However, under the condition that the proportion of renewable energy sources is not considered to be increased continuously, the access of renewable energy sources and electric vehicles on the load side enables the target of the traditional load prediction method to be no longer a pure electricity user. Most of the existing researches are single load prediction of each resource, and a framework for comprehensively managing each load prediction result and considering load prediction operation is not provided. The process performance is therefore to be further improved.
Disclosure of Invention
The invention aims to establish a controllable standardized model capable of perceiving interaction from three levels of a physical layer, an information layer and a social layer by constructing a micro-grid based on a CPSS (combined processing system) architecture of an cyber-physical-social fusion system and considering the double uncertainties of renewable energy access and user social behavior influence in a multi-scene fusion mode.
The invention is realized by at least one of the following technical schemes.
The multi-scene low-voltage renewable resource load prediction method based on the CPSS framework comprises the following steps:
coupling renewable energy output, electric vehicle charging load and user load, and predicting multi-scene power load spatial distribution;
establishing a fuzzy inference model of the user charging behavior;
establishing a relation model of peak-to-valley electricity price difference and load transfer rate, and establishing a demand response model based on a consumer psychology mechanism;
and (4) by combining the models, multi-scene fusion load prediction is carried out on the low-voltage renewable energy and the load under the CPSS framework.
Further, coupling is carried out by combining time-space characteristics, multi-scene power load space distribution is predicted, and a multi-scene load coupling formula is as follows:
wherein λ is L 、λ EV 、λ WT 、λ PV Respectively representing various load prediction regulation coefficients, P i，t Predicting the multi-scene space load at the time t in the corresponding area of the node,
and the predicted power of the load individual, the electric automobile individual, the wind generating set individual and the photovoltaic set individual k at the time t is respectively represented.
Furthermore, the load forecasting and adjusting coefficients of various energy individuals are set as follows according to the deviation between the forecasting data and the actual data of the previous day:
in the formula of lambda j Predicting the regulating coefficient for various loads, T is the number of rolling cycles per day,
and
the predicted load and the actual load at time t in the previous day are respectively indicated.
Further, the energy individuals comprise distributed photovoltaic, distributed fans, diesel generators, distributed energy storage, electric vehicles and user loads.
Further, for energy individuals k of different types j, generating power according to access nodes i at time t
Accumulated electric quantity
And rate of climbing
Building unified individual physical traits for a standardAnd establishing corresponding electrical constraints.
Further, the fuzzy inference model comprises three sections of fuzzy subsets and a set membership function; for the charging scene of the electric vehicle user, the electricity price, the current SOC and the parking time are used as input quantities of a fuzzy inference model of the charging behavior of the user, and the charging probability of the user is generated;
further, the membership function adopts a joint gaussian membership function:
wherein, x is input quantity including electricity price, current SOC and parking time, sigma and c are shape coefficients of a joint Gaussian membership function, and f (x, sigma, c) is user charging probability under the corresponding input quantity.
Further, for the charging scene of the electric automobile, a gravity center method of a fuzzy algorithm is adopted to solve a fuzzy inference model, and a clear value of charging probability is obtained:
wherein f is U (x) Is a fuzzy set on the continuous domain U, and C is a clear value after the fuzzy is solved.
Further, for a user electricity utilization scene, a relation model of peak-to-valley electricity price difference x and load transfer rate lambda is established based on consumer psychology:
wherein k is linear region proportionality coefficient, lambda max The maximum load transfer rate of the saturation region, a and b are power price difference nodes of a dead region, a linear region and a saturation region section, and lambda is the load transfer rate under different peak-valley power price differences x.
Further, peak-valley intervals are divided through fuzzy membership degrees of the large-scale half trapezoid and the small-scale half trapezoid, and a demand response model based on a consumer psychology mechanism is established:
P L ＝λ p P L-p +λ f P L-f +λ g P L-g
in the formula, P L-p 、P L-f 、P L-g Is the load amount of the flat period, peak period and valley period, lambda p 、λ f 、λ g Load transfer rates for flat, peak and valley periods, P L The load quantity after load transfer in each time interval is integrated, namely the electric power curve of the user is used.
Compared with the prior art, the invention has the beneficial effects that:
1) A controllable standardized model capable of sensing interaction is established from three levels of a physical layer, an information layer and a social layer, and convenience is provided for access and management of distributed energy individuals.
2) The method comprises the steps of establishing a demand response user load prediction model based on a consumer psychology mechanism for a power user, establishing a charging load prediction model considering the factors of electricity price, current SOC and parking time for an electric vehicle, establishing a functional relation between peak-valley electricity price difference and load transfer rate based on the consumer psychology for a user electricity utilization scene, establishing a demand response model based on the consumer psychology mechanism, and fully considering user behavior characteristics.
3) The load prediction of multi-scene fusion is carried out on the low-voltage renewable energy and the load, and the renewable energy utilization rate of the microgrid energy management system is enhanced.
Drawings
Fig. 1 is a micro-grid CPSS interaction architecture of a multi-scenario low-voltage renewable resource load prediction method based on a CPSS architecture according to an embodiment;
FIG. 2 is a functional block diagram of an embodiment information layer multi-scene fusion load prediction;
FIG. 3 is a flowchart illustrating load prediction of an electric vehicle according to an embodiment.
Detailed Description
The present invention will be further described with reference to the following specific examples.
Example 1
A multi-scene low-voltage renewable resource load prediction method based on a CPSS framework comprises the following steps:
establishing a uniform physical characteristic model, and generating power for energy individuals k of different types j at time t according to access nodes i
Accumulated electric quantity
And rate of climbing
Unified individual physical characteristics are constructed for the standard, and corresponding electrical constraint conditions are established;
coupling renewable energy output, electric vehicle charging load and user load, and predicting multi-scene power load spatial distribution; the multi-scenario load coupling formula is as follows:
wherein λ is L 、λ EV 、λ WT 、λ PV Respectively representing various load prediction regulation coefficients, P i，t For multi-scene space load prediction at time t in the corresponding region of the node,
and the predicted power of the load individual, the electric automobile individual, the wind generating set individual and the photovoltaic set individual k at the time t is respectively represented.
Setting load prediction adjustment coefficients of various energy individuals according to the deviation of the prediction data and the actual data of the previous day as follows:
in the formula, λ j The predicted adjustment coefficients for various types of loads j, T is the number of rolling cycles per day,
and
the predicted load and the actual load at time t in the previous day are respectively indicated.
Establishing a renewable energy output model; renewable energy mainly is photovoltaic power generation and wind power generation, and according to the existing research, the influence of weather on the power generation power of the renewable energy is large, and the correlation analysis is firstly carried out on meteorological factors influencing the output of the renewable energy by adopting a Pearson coefficient, and the Pearson coefficient is calculated by each factor and the power generation power:
in the formula, X i And Y i Indicating a certain meteorological factor and generated power,
and
the average value of a certain meteorological factor and the average value of generated power. The more the calculated Pearson correlation coefficient r is close to 1 or-1, the larger the positive correlation or the negative correlation is, and the closer r is to 0, the weaker the correlation is.
The existing common model for predicting the output of the renewable energy is a BP neural network model, the model comprises an input layer, a hidden layer and an output layer, and the model has the advantages of simplicity, high efficiency, high running time and the like. After feature optimization is carried out by combining with the Pearson correlation coefficient, factors with large correlation (generally, solar irradiance is used for photovoltaic power generation, and wind speed is used for wind power generation) are selected as feature input of an input layer of the BP neural network model, and training is carried out by combining with historical data. The setting of the training parameters of the BP neural network model and the training process are not elaborated in detail, and according to the trained model, the meteorological factors (including solar irradiance and wind speed) related to the renewable energy at a certain moment are input to obtain the generated power of the renewable energy corresponding to the moment.
Establishing a fuzzy inference model of the user charging behavior; the fuzzy inference model comprises three sections of fuzzy subsets and sets membership functions; for the charging scene of the electric vehicle user, the electricity price, the current SOC and the parking time are used as input quantities of a fuzzy inference model of the charging behavior of the user, and the charging probability of the user is generated;
the membership function adopts a combined Gaussian membership function:
wherein, x is input quantity (including electricity price, current SOC and parking time), σ and c are shape coefficients of a joint Gaussian membership function, and f (x, σ and c) is user charging membership under the corresponding input quantity and represents the probability of belonging to a certain class under the input quantity. The electricity price is taken as an example for explanation, the judgment of the user on the electricity price can be generally described in terms of "cheap, moderate and expensive", and after the real-time electricity price passes through the membership function, the judgment falls into the ranges of the three descriptions of "cheap, moderate and expensive". Similarly, the current battery capacity SOC and the parking time length can be divided into "insufficient, moderate, sufficient" and "short, medium, and long" by the membership function.
And combining the three sections of description ranges of the three input quantities to obtain a fuzzy membership function, namely an output membership function, of the charging probability of the electric automobile. For the electric automobile output membership function, solving a fuzzy inference model by adopting a gravity center method of a fuzzy algorithm to obtain a clear value of charging probability:
wherein f is U (x) Is a fuzzy set on a continuous domain U, and C is a clear value after the fuzzy is solvedI.e. the charging probability of the electric vehicle.
Therefore, after the real-time electricity price is obtained, the method is divided into 'cheap, moderate and expensive', and the user charging probability obtained by the fuzzy inference model can be obtained by combining the SOC of the user and the parking time. And according to the charging power setting of the charging station, multiplying and summing the charging probability of each user and the charging power to obtain the electric vehicle charging load prediction of the charging station.
For a user electricity utilization scene, a relation model of peak-to-valley electricity price difference x and load transfer rate lambda is established based on consumer psychology:
wherein k is linear region proportionality coefficient, λ max The maximum load transfer rate of the saturation region, a and b are power price difference nodes of a dead region, a linear region and a saturation region section, and lambda is the load transfer rate under different peak-valley power price differences x.
Dividing peak-valley intervals through semi-large semi-trapezoidal fuzzy membership degrees and semi-small semi-trapezoidal fuzzy membership degrees, and establishing a demand response model based on a consumer psychology mechanism:
P L ＝λ p P L-p +λ f P L-f +λ g P L-g
in the formula, P L-p 、P L-f 、P L-g Is the load amount of the flat period, peak period and valley period, lambda p 、λ f 、λ g Load transfer rates for flat, peak and valley periods, P L The load quantity after load transfer in each time interval is integrated, namely the electric power curve of the user is used.
By combining the models, the charging power of the electric automobile user and the power consumption power of the user can be obtained, and multi-scene fusion load prediction is carried out on the low-voltage renewable energy and the load under the CPSS framework.
The system for realizing the CPSS architecture-based multi-scenario low-voltage renewable resource load prediction method comprises the following steps:
and the data acquisition and interaction module is used for collecting physical data of the physical layer energy individual sensor to an information layer user terminal by adopting a Common Information Model (CIM), distributed energy, an advanced measurement system (AMI) and an IEC61970/61968 core standard of demand response for carrying out interaction and management on energy and data.
And the server collects and manages data collected by the user terminal and transmits the data to the microgrid operator server through a communication network.
And the prediction calculation module is used for calculating the net load output of each power generation/utilization user in the region to which the server belongs and transmitting the prediction result to the microgrid operator server. In addition, the module also can complete the functions of scene division and scene fusion, and the prediction condition of each scene is displayed on an operator server.
Example 2
A multi-scene low-voltage renewable resource load prediction method system based on CPSS framework of an cyber-physical-social fusion system is provided, a controllable standardized model capable of sensing interaction and multi-scene load prediction are established from three layers of a physical layer, an information layer and a social layer, and the method comprises the following steps:
1) Physical layer
The physical layer covers energy individuals commonly accessed to the microgrid, and the energy individuals comprise distributed photovoltaic, distributed fans, diesel generators, distributed energy storage, electric vehicles and user loads. The physical layer mainly considers a basic net rack, access positions of various energy individuals, physical models of the energy individuals, sensor equipment and communication equipment, describes the energy individuals on a physical level and reveals constraint relations between related electrical quantities and other physical quantities.
The sensors adopt IEC61970/61968 core standards including a common information model CIM, distributed energy, an advanced measurement system AMI and demand response, and collect physical data of physical layer energy individual sensors to an information layer user terminal for interaction and management of energy and data.
Establishing a unified physical characteristic model according to the resource interaction requirements among various energy individuals, and for the energy individuals k of different types j at the time tAccessing node i to generate power
Accumulated electric quantity
And rate of climbing
Unified individual physical characteristics are built for the standards and corresponding electrical constraints are established and transmitted to the information layer aggregator through the equipment terminals.
The physical model constraints for each type of energy individual are as follows:
1.1 Distributed photovoltaic and distributed wind turbine
For both the distributed photovoltaic unit and the distributed fan, the distributed photovoltaic unit and the distributed fan can be regarded as uncontrollable equipment, only the optimization of adjustable load resources is considered, and the day-ahead predicted values of the power generation power of the distributed photovoltaic unit and the power generation power of the adjustable load resources are considered in actual operation
And
and participating in optimized scheduling calculation, wherein the physical constraint is satisfied:
The distributed photovoltaic units and the distributed fans are considered to be in a local micro-grid, the climate characteristics are kept stable, the output of each unit can be regarded as consistency, and factors such as climate change and errors are ignored.
1.2 ) diesel generators
The diesel generating set is developed into a standby resource under the trend of high-proportion renewable energy sources, only participates in a peak-shaving frequency-modulation scene, is not used as a main power generation source, and is effectively supplemented and adjusted when the output of a fan and photovoltaic is insufficient, so that the safe and stable operation of a system is ensured.
1.3 ) energy storage
The energy storage is an important deterministic energy individual in the system, and has a great role of frequency modulation and peak regulation, and the energy storage characteristic is mainly described by the Charge amount (SOC) of the energy storage device at the time t:
state of charge constraints and charge-discharge power constraints:
in the formula:
the state of charge of the energy storage k accessed by the node i at the end of the t-th time period; eta ch And η dis Respectively the charge and discharge efficiency of the stored energy; e r Rated capacity for stored energy; p t ch And P t dis The discharge power of the current time interval does not exceed the maximum power limit of charge and discharge
And
and the energy storage device cannot be in a charged and discharged state simultaneously, at is the time interval (i.e., 15 minutes), SOC min And SOC max The lower and upper bounds of the amount of stored charge.
For energy storage devices, it is generally required that the end time state of charge be the same as the start time state of charge, or vary by less than a certain range:
SOC 1 ＝SOC 96
|SOC 96 -SOC 1 |≤ε ES
in the formula SOC 1 And SOC 96 A starting time state (t = 1) and an ending time state (t = 96) of a day for the amount of energy storage device charge, the difference between the two states being smaller than a given error range epsilon ES 。
1.4 ) electric vehicle
The load model of the electric automobile is similar to the energy storage, and the difference lies in that the access time and the leaving time of the electric automobile accord with the behavior habit of a user and meet certain constraint, and the SOC of the electric automobile at the leaving time should ensure reasonable travel distance. The specific user behavior scene definition is set forth at the social layer.
The electrical constraints are therefore:
in the formula: t is t l In order to end the charging timing of the EV,
the state of charge at the end of the EV charge time; alpha is alpha EV The lowest SOC coefficient at the end of charging is desired for the user.
1.5 User load)
The power users can be subdivided into three categories, namely residential users, industrial users and commercial buildings, the prediction of user loads is very critical, and the prediction has great influence on the safe operation and scientific planning of a power distribution network. In view of the fact that the power consumers have great uncertainty, it is not absolutely rational individuals, for example, residential consumers are likely to give priority to their comfort or satisfaction to perform power consumption behaviors. Since many users have shown the dual characteristics of power generation and power consumption, the prediction error of a single load prediction scenario becomes larger in load prediction on the load side, and the invention corrects the user load prediction based on multiple scenarios. The user load power accessed at the node i at the time t in the physical layer is defined as
2) Information layer
The information layer receives and processes various information of the physical layer terminal, divides various energy individuals of the physical layer into three types of sources, storage and loads for description, and has the functions of load prediction and optimized operation configuration. Under the background of high proportion of renewable energy sources, a source end and a charge end can simultaneously have superposition of double uncertainties, and stored energy is used as an optimal configuration resource.
The load forecasting function module is divided into three parts, namely renewable energy output forecasting, electric vehicle charging load forecasting and user load forecasting. The renewable energy output prediction is carried out on-line prediction based on historical data and real-time data, and the charging load of the electric vehicle and the user load need to be combined with characteristic data obtained by a social layer to carry out on-line prediction on the load.
The regional load is considered in the form of a net load, and the distributed photovoltaic and wind turbine output is considered as a negative load. The weather in continuous time in the region is stable, and the output of each distributed photovoltaic unit is consistent. After the load prediction is carried out on the three types of resources, the coupling is carried out by combining the time-space characteristics, the multi-scene power load spatial distribution is obtained, and the multi-scene-based space load prediction is completed. The multi-scene load coupling calculation formula is as follows:
wherein, w L 、w EV 、w WT 、w PV Respectively representing various load prediction regulation coefficients, P i，t Predicting the multi-scene space load at the time t in the corresponding area of the node,
and the predicted power of the load individual, the electric automobile individual, the wind generating set individual and the photovoltaic set individual k at the time t is respectively represented.
And various load prediction adjustment coefficients are set according to the deviation between the prediction data and the actual data of the previous day:
in the formula, w j T is the rolling period number of one day, T =96 is taken by taking 15min as a period,
and
respectively representing time t of the preceding dayPredicted load and actual load.
The result of the load prediction can be used as the input of the operation optimization configuration function, and the real-time operation condition is combined for optimization.
3) Social layer
The social layer mainly comprises two types of benefit subjects, namely power consumers and operators, wherein the load scene of the power consumers can be subdivided into residential users, industrial users, commercial intelligent buildings and charging loads of electric automobiles, the operators on the social layer are absolutely rational individuals, the decision is made by aiming at maximizing economic safety benefits, and the power consumers can generate uncertainty due to irrational factors such as social information and personal will and are expressed as limited rational individuals.
3.1 ) electric vehicle travel scene
For the construction of the charging load scene of the electric automobile, the charging load scene is divided into an industrial area, a commercial area and a residential area in space, the running characteristics of private automobiles are analyzed based on the historical charging data of a charging station and urban traffic development reports, the parking time, the arrival time and the like are obtained, and the space-time distribution of the urban multi-scene electric automobile is generated.
The key point of the user for considering whether the electric automobile is charged is the price and the demand, so that the membership function is set and the charging probability of the user is generated by taking the electricity price, the current SOC and the parking time as the input quantity of the fuzzy inference model for describing the charging behavior of the user based on the charging willingness of the user for a private automobile.
The joint Gaussian membership function calculation formula is as follows:
wherein, x is input quantity (i.e. electricity price, current SOC, parking time length), σ and c are shape coefficients of a joint gaussian membership function, and f (x, σ, c) is user charging probability under the corresponding input quantity.
For the real-time electricity price, a three-section fuzzy subset 'cheap, moderate and expensive' is adopted to represent the sensitivity degree of a user to the electricity price; for batteries whenThe front SOC adopts three fuzzy subsets of 'lack, moderate and abundant' to represent the SOC abundance; for the parking time, three sections of fuzzy subsets of 'short, medium and long' are adopted to describe the charging requirement of a user, generally, when alternating current slow charging is adopted, the parking time belongs to the fuzzy subset 'medium' or 'long', and when direct current fast charging is selected, the parking time belongs to the fuzzy subset 'short'. From this can get 3 3 And (4) performing a fuzzy charging decision, converting an output result into a charging probability, and obtaining a clear value by adopting a gravity center method of a fuzzy algorithm.
Defuzzification by a gravity center method:
in the formula (f) U (x) Is a fuzzy set on the continuous domain U, and C is a clear value after the fuzzy is solved.
And obtaining the charging probability of the user after ambiguity resolution, and obtaining the charging load prediction result of the electric automobile through Monte Carlo simulation.
3.2 User load scenario
The user load scene can be generally divided into a residential area, a business area and an industrial area, and the duty ratio of each load type in different areas is different. The industrial user and commercial building scenes are different from the residential user load scenes, the user load is mainly related to factors such as a work system, festivals and holidays, the residential user load is mainly related to factors such as user electricity utilization behaviors, weather and electricity prices, historical load data can be learned through a BP neural network, a key electricity utilization characteristic index training model is selected, and load prediction is carried out. However, the load of the user is often influenced by the electricity price, and a load prediction correction model based on consumer psychology is proposed here:
according to the consumer psychology model, the response degree of a user in a certain interval is in direct proportion to the peak-valley incentive electricity price, and the user changes the electricity utilization mode of the equipment according to the incentive. Due to the existence of the threshold effect and the saturation effect, the response degree of a user to the power price excitation can be reflected, namely, the peak-to-valley power price difference x has a corresponding functional relation with the load transfer rate lambda:
wherein, lambda is the load transfer rate under different peak-to-valley electrovalence differences x, k is the linear region proportionality coefficient, and lambda max The load transfer rate is the maximum load transfer rate of a saturation region, a and b are power price difference nodes of a dead zone, a linear region and a saturation region section, and when the power price difference is too small, the load amount of a power consumer is not reduced or transferred basically, namely, the power consumer is in the dead zone; when the electricity price difference is too large, the load transferring or reducing capacity of the consumer reaches the upper limit, namely, the consumer is in a saturation region; when the electricity price difference is between the dead zone and the saturation zone, the electricity price difference and the load transfer rate are basically in a linear relation.
In a demand response mechanism based on time-of-use electricity prices, 3 electricity price differences and negative peak-to-valley electricity prices are input as fuzzy membership degrees, and the peak-to-valley membership degree at each moment is determined, so that the basis of peak-to-valley period division is carried out 24 hours a day. Only the relative size of data needs to be compared when determining the peak-valley membership degree, so the probability that each moment on the load curve belongs to the peak time period is judged by adopting the linear part of the partial large-scale semi-trapezoidal membership function, and the probability that each moment on the load curve belongs to the valley time period is judged by adopting the linear part of the partial small-scale semi-trapezoidal membership function:
in the formula, x is the peak-to-valley electricity price difference, and m and n are parameters of the upper limit and the lower limit of the set function. The peak-valley position of the electricity price at each moment is divided into a peak time period, a valley time period and a flat time period.
After the time-of-use electricity price is implemented, the demand response model based on the psychological mechanism of consumers is as follows:
P L ＝λ p P L-p +λ f P L-f +λ g P L-g
in the formula, P L-p 、P L-f 、P L-g Is the load amount of the flat, peak and valley periods, λ p 、λ f 、λ g Load transfer rates for flat, peak and valley periods, P L The load quantity after load transfer in each time interval is integrated, namely the electric power curve of the user is used.
Example 3
A micro-grid CPSS (Cyber-Physical-Social Systems, CPSS) interaction architecture is shown in fig. 1.
The micro-grid under the CPSS framework is divided into three layers, namely a physical layer, an information layer and a social layer, information interaction is carried out between the physical layer and the information layer through the Internet of things, and information interaction is carried out between the social layer and the information layer through a communication network. Energy interaction among various energy individuals is mainly performed in the physical layer, servers and power generation/utilization users are included in the information layer, each server is responsible for managing users in the area, data are uploaded to a cloud end through a communication network for information interaction, the social layer divides space scenes into residential areas, business areas, industrial areas and operators, and social information interaction exists among the scenes.
The information layer multi-scene fusion load prediction function is shown in fig. 2.
The method is characterized in that the multi-scene of the information layer is developed from three dimensions of renewable energy output prediction, electric vehicle charging load prediction and user load prediction, and the specific process is as follows:
step1: selecting a node area to be predicted;
step2: selecting resource data of the node area, wherein the resource data comprises weather and output data of renewable energy sources, real-time electricity price of an electric vehicle, current battery charge SOC (state of charge), parking duration data, real-time electricity price of user load and SCADA (supervisory control and data acquisition) prediction data;
step3: establishing a multi-scene model of each part:
under the renewable energy output scene, performing Pearson correlation analysis on meteorological data and output data:
screening out significant characteristics such as solar irradiance, wind speed and the like as input of a neural network model, and predicting the output of the neural network of renewable energy;
under the charging scene of the electric automobile, historical traffic data and charging station data are analyzed to obtain the time-space characteristics of electric automobile charging, and a fuzzy membership function is combined
Calculating membership degree, and resolving ambiguity by gravity center method
Obtaining the charging probability, and carrying out Monte Carlo simulation charging load;
under the scene of user electricity consumption, three scenes of industrial load, commercial load and residential load are divided based on the types of basic users, and a peak-valley electricity price and load transfer rate model is established through a fuzzy membership rule, namely
Further modifying the load forecast based on the electricity price incentive response according to the historical SCADA load data forecast result, i.e. P L ＝λ p P L-p +λ f P L-f +λ g P L-g ；
Step4: scene fusion is carried out on three scenes of renewable energy output, electric vehicle charging and user electricity consumption, and the load prediction evaluation index w in the previous day is combined j Calculating the predicted load of the node area in the day ahead:
namely, the prediction results of the loads of the three scenes at the time t are calculated through the formula, and the comprehensive load prediction of the node area is obtained.
The electric vehicle load prediction flow is shown in fig. 3.
At a charging station terminal, predicting the charging behavior of a single trip chain of the electric vehicle charging user in the geographic space area, wherein the specific prediction process is as follows:
step1: historical traffic and charging station data are input, and the trip characteristics of users in the geographic area are obtained;
step2: acquiring the current SOC state and the real-time electricity price of a user, and inputting a fuzzy membership function by combining the travel characteristic of Step1
Step3: according to fuzzy membership rule
Performing ambiguity resolution to obtain the user charging probability, namely performing integral operation on the user charging probability according to the fuzzy membership function in Step 2;
step4: and carrying out Monte Carlo simulation to calculate the charging load of the user.
Step5: aggregating the predicted load of charged subscribers within a station, i.e.
And obtaining the regional predicted charging load of the charging station.
The preferred embodiments of the invention disclosed above are intended to be illustrative only. The preferred embodiments are not intended to be exhaustive or to limit the invention to the precise embodiments disclosed. Obviously, many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described in order to best explain the principles of the invention and the practical application, to thereby enable others skilled in the art to best utilize the invention. The invention is limited only by the claims and their full scope and equivalents.
Claims (10)
1. The CPSS architecture-based multi-scene low-voltage renewable resource load prediction method is characterized by comprising the following steps of:
coupling renewable energy output, electric vehicle charging load and user load, and predicting multi-scene power load spatial distribution;
establishing a fuzzy inference model of the user charging behavior;
establishing a relation model of peak-to-valley electricity price difference and load transfer rate, and establishing a demand response model based on a consumer psychology mechanism;
and (4) by combining the models, multi-scene fusion load prediction is carried out on the low-voltage renewable energy and the load under the CPSS framework.
2. The CPSS architecture-based multi-scenario low-voltage renewable resource load prediction method of claim 1, wherein the coupling is performed in combination with temporal-spatial characteristics to predict multi-scenario power load spatial distribution, and a multi-scenario load coupling formula is:
wherein λ is L 、λ EV 、λ WT 、λ PV Respectively representing various load prediction regulation coefficients, P i，t For multi-scene space load prediction at time t in the corresponding region of the node,
and the predicted power of the load individual, the electric automobile individual, the wind generating set individual and the photovoltaic set individual k at the time t is respectively represented.
3. The CPSS architecture-based multi-scenario low-voltage renewable resource load prediction method according to claim 2, wherein the load prediction adjustment coefficients of various energy individuals are set according to the deviation between the prediction data and the actual data of the previous day as follows:
in the formula, λ j Predicting the regulating coefficient for various loads, T is the number of rolling cycles per day,
and
the predicted load and the actual load at time t in the previous day are respectively indicated.
4. The CPSS architecture-based multi-scenario low-voltage renewable resource load prediction method of claim 3, wherein the energy source individuals include distributed photovoltaic, distributed wind turbine, diesel generator, distributed energy storage, electric vehicle, user load.
5. The CPSS architecture-based multi-scenario low-voltage renewable resource load prediction method of claim 4, wherein for energy individuals k of different kinds j, the power generation power is generated at time t according to access node i
Accumulated electric quantity
And rate of climbing
Unified individual physical characteristics are built for the standard and corresponding electrical constraints are established.
6. The CPSS architecture-based multi-scenario low-voltage renewable resource load prediction method of claim 1, wherein the fuzzy inference model includes three sections of fuzzy subsets, setting membership functions; and for the charging scene of the electric vehicle user, the electricity price, the current SOC and the parking time are used as input quantities of the fuzzy inference model of the charging behavior of the user, and the charging probability of the user is generated.
7. The CPSS architecture-based multi-scenario low-voltage renewable resource load prediction method of claim 6, wherein the membership function employs a joint Gaussian membership function:
wherein, x is input quantity including electricity price, current SOC and parking time, sigma and c are shape coefficients of a joint Gaussian membership function, and f (x, sigma, c) is user charging probability under the corresponding input quantity.
8. The CPSS architecture-based multi-scenario low-voltage renewable resource load prediction method of claim 6, wherein for an electric vehicle charging scenario, a gravity center method of a fuzzy algorithm is adopted to solve a fuzzy inference model, and a clear value of charging probability is obtained:
wherein f is U (x) And C is a clear value after the ambiguity is resolved.
9. The CPSS-architecture-based multi-scenario low-voltage renewable resource load prediction method as claimed in claim 1, wherein for a user electricity usage scenario, a relationship model between peak-to-valley electricity price difference x and load transfer rate λ is established based on consumer psychology:
wherein k is linear region proportionality coefficient, lambda max Maximum load transfer rate in saturation region, a and b are deadAnd the power price difference nodes of the zones, the linear zone and the saturation zone are segmented, and lambda is the load transfer rate under different peak-valley power price differences x.
10. The CPSS architecture-based multi-scenario low-voltage renewable resource load prediction method of any one of claims 1 to 9, wherein peak-valley intervals are divided by fuzzy membership degrees of a more large half trapezoid and a less small half trapezoid, and a demand response model based on a consumer psychology mechanism is established:
P L ＝λ p P L-p +λ f P L-f +λ g P L-g
in the formula, P L-p 、P L-f 、P L-g Is the load amount of the flat period, peak period and valley period, lambda p 、λ f 、λ g Load transfer rates for flat, peak and valley periods, P L The load quantity after load transfer in each time interval is integrated, namely the electric power curve of the user is used.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211147957.XA CN115545429A (en) | 2022-09-20 | 2022-09-20 | Multi-scene low-voltage renewable resource load prediction method based on CPSS framework |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211147957.XA CN115545429A (en) | 2022-09-20 | 2022-09-20 | Multi-scene low-voltage renewable resource load prediction method based on CPSS framework |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115545429A true CN115545429A (en) | 2022-12-30 |
Family
ID=84727193
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211147957.XA Pending CN115545429A (en) | 2022-09-20 | 2022-09-20 | Multi-scene low-voltage renewable resource load prediction method based on CPSS framework |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115545429A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117793008A (en) * | 2024-02-27 | 2024-03-29 | 广东电网有限责任公司中山供电局 | Security management method and device for Internet of things traffic and electric Internet of things system |
-
2022
- 2022-09-20 CN CN202211147957.XA patent/CN115545429A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117793008A (en) * | 2024-02-27 | 2024-03-29 | 广东电网有限责任公司中山供电局 | Security management method and device for Internet of things traffic and electric Internet of things system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109599856B (en) | Electric vehicle charging and discharging management optimization method and device in micro-grid multi-building | |
| CN113746144B (en) | Source-grid-load real-time interaction electric carbon control method and intelligent management system thereof | |
| Shang et al. | Internet of smart charging points with photovoltaic Integration: A high-efficiency scheme enabling optimal dispatching between electric vehicles and power grids | |
| Guo et al. | Multi-objective optimization design and multi-attribute decision-making method of a distributed energy system based on nearly zero-energy community load forecasting | |
| Alabi et al. | Improved hybrid inexact optimal scheduling of virtual powerplant (VPP) for zero-carbon multi-energy system (ZCMES) incorporating Electric Vehicle (EV) multi-flexible approach | |
| Zhang et al. | Grid-connected photovoltaic battery systems: A comprehensive review and perspectives | |
| CN111626527A (en) | Intelligent power grid deep learning scheduling method considering fast/slow charging/discharging form of schedulable electric vehicle | |
| CN113139711A (en) | Wisdom energy data center management system based on data integration | |
| CN114862252A (en) | Load-adjustable multi-layer aggregation scheduling potential analysis method, system, equipment and medium | |
| Li et al. | Energy management model of charging station micro-grid considering random arrival of electric vehicles | |
| CN116402307A (en) | Power grid planning capacity analysis method considering operation characteristics of schedulable flexible resources | |
| Hashemipour et al. | Uncertainty modeling for participation of electric vehicles in collaborative energy consumption | |
| CN115545429A (en) | Multi-scene low-voltage renewable resource load prediction method based on CPSS framework | |
| CN104112168B (en) | A kind of smart home optimization method based on multi-agent system | |
| Tong et al. | An intelligent scheduling control method for smart grid based on deep learning | |
| Zhu et al. | Dynamic multi-objective dispatch considering wind power and electric vehicles with probabilistic characteristics | |
| CN116054286A (en) | Residential area capacity optimal configuration method considering multiple elastic resources | |
| CN116307437A (en) | Virtual power plant optimal scheduling method and system based on carbon emission right exchange mechanism | |
| CN115939538A (en) | Comprehensive evaluation method and device for performance of battery energy storage system and computer equipment | |
| CN113052450B (en) | Urban energy Internet planning method suitable for electric energy substitution development strategy | |
| Yang et al. | Review on distribution network planning methods considering large-scale access of flexible load | |
| CN115330062A (en) | Scheduling optimization method for ordered charging service of new energy automobile in community scene | |
| CN114091958A (en) | Comprehensive energy efficiency evaluation method and power grid planning method for intelligent power utilization cell | |
| CN111130107A (en) | Power grid load prediction method and device | |
| Ren et al. | Decision-making approach in charging mode for electric vehicle based on cumulative prospect theory |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |