CN116307299A - Photovoltaic power generation power short-term prediction method, system, equipment and storage medium - Google Patents

Photovoltaic power generation power short-term prediction method, system, equipment and storage medium Download PDF

Info

Publication number
CN116307299A
CN116307299A CN202310580304.9A CN202310580304A CN116307299A CN 116307299 A CN116307299 A CN 116307299A CN 202310580304 A CN202310580304 A CN 202310580304A CN 116307299 A CN116307299 A CN 116307299A
Authority
CN
China
Prior art keywords
power generation
photovoltaic power
lssvm
short
ibes
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
Application number
CN202310580304.9A
Other languages
Chinese (zh)
Inventor
孟庆霖
程泽
王议锋
许良
王瑞
刘�东
药炜
赵金
刘春雨
孙继科
孙京生
王一
陈浩
肖茂祥
孙宝平
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
State Grid Corp of China SGCC
State Grid Tianjin Electric Power Co Ltd
Marketing Service Center of State Grid Tianjin Electric Power Co Ltd
Original Assignee
State Grid Corp of China SGCC
State Grid Tianjin Electric Power Co Ltd
Marketing Service Center of State Grid Tianjin Electric Power Co Ltd
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 State Grid Corp of China SGCC, State Grid Tianjin Electric Power Co Ltd, Marketing Service Center of State Grid Tianjin Electric Power Co Ltd filed Critical State Grid Corp of China SGCC
Priority to CN202310580304.9A priority Critical patent/CN116307299A/en
Publication of CN116307299A publication Critical patent/CN116307299A/en
Pending legal-status Critical Current

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
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Economics (AREA)
  • General Physics & Mathematics (AREA)
  • Human Resources & Organizations (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Strategic Management (AREA)
  • Computing Systems (AREA)
  • Tourism & Hospitality (AREA)
  • Artificial Intelligence (AREA)
  • General Business, Economics & Management (AREA)
  • Mathematical Physics (AREA)
  • General Engineering & Computer Science (AREA)
  • Marketing (AREA)
  • Evolutionary Computation (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Operations Research (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Primary Health Care (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Water Supply & Treatment (AREA)
  • Public Health (AREA)
  • Molecular Biology (AREA)
  • Quality & Reliability (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Game Theory and Decision Science (AREA)
  • Development Economics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses a photovoltaic power generation power short-term prediction method, a system, equipment and a storage medium, and relates to the field of photovoltaic power generation power prediction, wherein the method comprises the following steps: acquiring photovoltaic power generation power of a photovoltaic electric field and meteorological factor data at the same time, dividing the photovoltaic power generation power into a training set and a testing set, and inputting the training set into an IBES-LSSVM for training; constructing an IBES-LSSVM model; the training process comprises an IBES model and an LSSVM prediction model; performing performance evaluation on the trained IBES-LSSVM model, and determining a photovoltaic power generation power short-term prediction model with optimal parameters; inputting weather factor data related to photovoltaic power generation power data to be predicted into an IBES-LSSVM model, and outputting a photovoltaic power generation power prediction result. The photovoltaic power generation system and the photovoltaic power generation method can predict photovoltaic power generation power in a short period, and can ensure stable operation of a power grid under the condition of photovoltaic grid connection.

Description

Photovoltaic power generation power short-term prediction method, system, equipment and storage medium
Technical Field
The invention belongs to the field of photovoltaic power generation power prediction, and particularly relates to a photovoltaic power generation power short-term prediction method, a system, equipment and a storage medium.
Background
With the continuous development of industrial level in China, the contradiction between the increasing energy demand and the energy-saving and carbon-reducing targets becomes one of the great challenges faced by China, and in order to ensure the energy supply and maintain the ecological environment stably, a large number of students at home and abroad develop and research on renewable energy sources, compared with traditional fossil energy sources, the renewable energy sources have the advantages of abundant reserves, greenness, cleanness and the like, and among a plurality of renewable energy sources, solar energy has the advantages of large reserves, wide distribution, convenient collection and the like. With the continuous expansion of the installed capacity of photovoltaic across the country, the application form of photovoltaic power generation has gradually evolved from an early off-grid power generation mode to a grid-connected form. Photovoltaic power generation has gradually become a major way of energy conservation and carbon reduction in the modern power industry.
The traditional power generation mode is mainly based on thermal power, thermal power generation has the characteristics of controllable electric energy, stable output electric energy and the like, but the photovoltaic power generation electrode is easily influenced by various weather and various factors due to the influence of power generation modes and principles, and the power generation performance shows obvious intermittence and discontinuity. Along with the continuous increase of grid-connected capacity of photovoltaic power generation, the variability of the photovoltaic power generation has a great influence on real-time scheduling and power quality of a power grid, so that the method for accurately predicting the generated energy of the photovoltaic power generation has important value for developing a photovoltaic power generation technology. The photovoltaic power prediction is classified into ultra-short term prediction, short term prediction and medium-long term prediction according to time scales, and is applied to real-time scheduling of a power grid, daily power generation planning and economic scheduling planning and maintenance and operation management of a photovoltaic electric field. The short-term photovoltaic power generation power prediction is crucial to the reasonable arrangement of a daily power generation plan, the realization of efficient economic dispatch and the promotion of clean energy development and utilization of an electric power department, so that the short-term photovoltaic power generation power prediction becomes a current research hotspot.
Disclosure of Invention
In order to achieve the above purpose, the invention adopts the following technical scheme:
the invention discloses a photovoltaic power generation power short-term prediction method, a system, equipment and a storage medium, which are used for solving the problem that the existing photovoltaic power generation power prediction accuracy is insufficient, so that the pressure of increasingly-growing distributed photovoltaic grid connection on a power grid is reduced.
Firstly, acquiring photovoltaic power generation power related data in a photovoltaic electric field acquired in real time, preprocessing the data, and eliminating missing data to obtain a photovoltaic power generation power related data sample; secondly, inputting a photovoltaic power generation power related data sample into a photovoltaic power generation power prediction model of a Least Squares Support Vector Machine (LSSVM) optimized through an improved balk search algorithm (IBES); and finally, optimizing parameters of the least square support vector machine by using an improved bald eagle search algorithm (IBES), and obtaining the photovoltaic power generation power prediction model with the optimal effect.
Further, the method specifically comprises the following steps:
step 1, acquiring collected photovoltaic power generation power related data, preprocessing the data, removing missing data to obtain a photovoltaic power generation power related data sample, and dividing the photovoltaic power generation power related data sample into a training data set and a test data set.
In the invention, the related data of the photovoltaic power generation power comprise air temperature, azimuth angle, cloud opacity, dew point temperature, solar scattering radiation index, solar direct radiation index, solar total horizontal radiation, fixed dip angle radiation, tracking dip angle radiation, atmospheric precipitation and relative humidity;
step 2, setting initial parameters of a model, including a bald eagle population, a fitness value and the maximum iteration times;
when the standard balying search algorithm selects the search space in the first stage, the random number is initialized so that the selected search space of the balying is simple and random, and the search space cannot be completely covered, which results in low algorithm solving precision and insufficient medium-term global searching capability. The invention is based on the improvement, introduces the idea of chaos optimization into the initialization of the bald eagle population, and the main idea of chaos optimization is to map variables into the value interval of the chaos variable space by utilizing the characteristics of ergodic property, randomness and the like, and finally linearly convert the obtained solution into the optimized variable space, thereby improving the algorithm performance. Therefore, the invention adopts Sine mapping in the chaotic mapping to initialize the population of the bald hawk searching algorithm, and the selected space mathematical model of the bald hawk searching algorithm after being initialized by the Sine chaotic mapping is as follows:
Figure SMS_1
wherein C is i Is [0,1 ]]Sine mapping of ub j And lb j The upper boundary and the lower boundary of the j-th dimension are respectively represented by a selected space mathematical model of a balying search algorithm after the initialization of the fine chaotic map of the fine, and population distribution map pairs of random initialization and the initialization of the fine chaotic map of the fine are shown in the figure 3.
Step 3, in the selection stage of the improved bald eagle searching algorithm: balding is used for determining the optimal searching position by randomly selecting searching area and judging the number of the hunting objects, so that hunting objects can be conveniently searched, and the balding position P at the stage i,n The update is determined by multiplying a priori information from a random search by a. The behavioral mathematical model is described as:
Figure SMS_2
wherein:
Figure SMS_3
for controlling the position-changing parameters>
Figure SMS_4
Is a random number between (0, 1)>
Figure SMS_5
Optimal value searched for the current bald eagle; ->
Figure SMS_6
For the average distribution position of bald hawk after the end of the previous search>
Figure SMS_7
Is the firstiOnly the location of bald hawks.
And 4, in a searching stage of the improved bald eagle searching algorithm, the improved bald eagle searching algorithm updates the bald eagle population through the following formula, the adaptability of the new bald eagle position is calculated, and the adaptability function is the Root Mean Square Error (RMSE) of the photovoltaic power generation power data training set.
Bal eagles fly in a spiral shape within the search space, looking for hunting. This flying approach can speed up the search process and help balding find the best dive capture location. For this purpose, we use polar equations to update the location of the bald eagle. The specific formula is as follows:
Figure SMS_8
Figure SMS_9
；
wherein:
Figure SMS_10
and->
Figure SMS_11
Polar angle and polar diameter of the spiral equation, respectively>
Figure SMS_12
And->
Figure SMS_13
Parameters for controlling the spiral track;
Figure SMS_14
is (0, 1) random number; ->
Figure SMS_15
And->
Figure SMS_16
Is the bald eagle position in polar coordinates.
Updating the bald eagle position through the searching process:
Figure SMS_17
wherein the method comprises the steps of
Figure SMS_18
Is->
Figure SMS_19
Only the bald eagle updates the position once.
And 5, in the dive stage of the improved bald hawk search algorithm, the third stage of the standard bald hawk search algorithm, namely the dive predation stage, is a key process for determining the convergence rate of the algorithm, and the standard bald hawk search algorithm is easy to deviate from the search direction and miss the optimal foraging area due to lack of effective control of step length, so that the algorithm falls into local optimum. Therefore, the invention provides a new non-inertial self-adaptive strategy, which can prevent the algorithm from sinking into local optimum while accelerating the algorithm convergence speed, so as to search for the hunting for quick reading, the bald eagle can start to dive from the optimum position in the search space, and other individuals can also add attack and move to the optimum position. Their state of motion can be described by polar equations.
Figure SMS_20
The improvement formula is as follows:
Figure SMS_21
wherein,,
Figure SMS_22
and->
Figure SMS_23
C, the variation of the movement of the bald hawk to the optimal center position in the polar coordinate space 1 And c 2 For the intensity of movement of the bald hawk to the optimal center position, ω represents the non-inertial weight factor, iter max Represents the maximum number of iterations, t represents the current number of iterations,P best for the optimal value searched for the current bald eagle,randis a random number in (0, 1).
P i,n Indicating the location of the bald eagle after updating.
And 5, judging whether a stopping condition is met, judging whether the fitness function is the Root Mean Square Error (RMSE) of the photovoltaic power generation power data test set is optimal, outputting a global optimal parameter if the fitness function is optimal, and otherwise, re-executing the step 4.
Step 6, regularization parameters under the condition that the fitness function reaches the optimal conditioncAnd the width of the center point from the kernel functionσAnd assigning values to the LSSVM, and predicting the photovoltaic power generation power by the LSSVM by utilizing the optimal parameters. The LSSVM method formula is as follows: LSSVM is an important improvement over standard Support Vector Machines (SVMs). It uses the sum of squares of the errors as an empirical loss and uses the equality constraint instead of the inequality constraint in the original algorithm, thereby eliminating the need to solve the complex quadratic programming problem. The improvement reduces the difficulty of solving, optimizes the overall convergence speed and improves the accuracy of the calculation result. The specific formula process is as follows: setting training setS={(x i ，y i ），x i ∈R n1 ，y i E R }, wherein,x i is an input vector;y i is an output vector;i=1,2,...,N；Ris a real number set;R n1 is thatn1A set of dimensional real numbers;Nis the number of training samples.
The optimized objective function of the LSSVM can be translated into:
Figure SMS_24
wherein:ωis a weight coefficient vector;
Figure SMS_25
is a nonlinear mapping function;bas a result of the bias term,Jas a loss function;e i is an error variable;cis a regularization parameter. By constructing Lagrange multipliersα i Solving an optimized objective function of an LSSVM：
Figure SMS_26
According to the Carlo-Coulomb-Tak condition, each variable in the above formula is biased: by solving, eliminateωAnde i ，finally, the LSSVM regression function is obtained as follows:
Figure SMS_27
wherein:α i in order to be a lagrange multiplier,
Figure SMS_28
as a kernel function of the LSSVM model, a gaussian kernel function (Radial Basis Function, RBF) is selected as a kernel function of the LSSVM model, and the specific expression of the RBF is as follows:
Figure SMS_29
Figure SMS_30
for inputting feature vectors, ++>
Figure SMS_31
Is the width of the kernel from the center point.
And 7, acquiring real-time data of the generated power of the photovoltaic electric field to be predicted and related influence factors.
And 8, taking the real-time data of the relevant influence factors of the generated power of the photovoltaic electric field to be predicted as the input of a pre-constructed photovoltaic generated power prediction model, and outputting the result of the photovoltaic generated power prediction of the photovoltaic electric field to be predicted.
Compared with the prior art, the invention has the beneficial effects that:
according to the photovoltaic power generation short-term prediction method, system, equipment and storage medium, firstly, aiming at the problems of low solving precision, low convergence speed and easiness in sinking into local optimum, a balying search algorithm (BES) is improved through sine mapping chaotic optimization (Sinemap) and a non-inertial weight self-adaptive strategy, when a standard balying search algorithm selects a search space in a first stage, a random number is initialized to enable the selected search space of the balying to be simple and random, the search space cannot be completely covered, and therefore the algorithm solving precision is low and the middle-term global search capability is insufficient. The main idea of the chaos optimization is to map the variable into the value interval of the chaos variable space by utilizing the characteristics of ergodic property, randomness and the like, and finally linearly convert the obtained solution into the optimized variable space, thereby improving the algorithm performance. Therefore, the invention adopts sine mapping in chaotic mapping to initialize the population of the bald eagle searching algorithm; the third stage of the standard bald hawk search algorithm, namely the diving predation stage, is a key process for determining the convergence rate of the algorithm, and the standard bald hawk search algorithm is easy to deviate from the search direction and miss the optimal foraging area due to lack of effective control of step length, so that the algorithm falls into local optimum. Therefore, the research provides a new non-inertial self-adaptive strategy, prevents the convergence rate of the algorithm from being in partial optimum while accelerating, obtains an improved baldness search algorithm (IBES), and optimizes the parameters of a least square support vector machine prediction model by adopting the method; and finally, inputting the related data of the photovoltaic power generation power into a prediction model, and adjusting the parameters of the IBES optimization algorithm through the result, thereby obtaining the photovoltaic power generation power prediction model with the optimal effect. The prediction accuracy of the photovoltaic power generation power is greatly improved, and the iteration speed of the model is improved.
Drawings
FIG. 1 is a process step diagram of the present invention;
FIG. 2 is a flow chart of a short-term photovoltaic power generation power prediction method of the present invention;
FIG. 3 is a population distribution comparison chart of random initialization and SIne chaotic map initialization according to the invention;
FIG. 4 is a graph comparing the fitness iteration curves of the BES-LSSVM method and the IBES-LSSVM of the present invention;
FIG. 5 is a graph comparing the predicted results of the training sets of LSSVM, BES-LSSVM and IBES-LSSVM of the present invention;
FIG. 6 is a graph of a comparison of the prediction results of the test set of the LSSVM, BES-LSSVM, and IBES-LSSVM of the present invention;
FIG. 7 is a graph comparing the predicted result evaluation indexes of the training sets of the LSSVM, BES-LSSVM and IBES-LSSVM of the invention;
FIG. 8 is a graph comparing the evaluation indexes of the prediction results of the test set of the LSSVM, BES-LSSVM and IBES-LSSVM according to the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely, and it is apparent that the described embodiments are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
As shown in fig. 1 to 8, the invention provides a method, a system, a device and a storage medium for short-term prediction of photovoltaic power generation power, which comprise the following steps:
step 1, acquiring collected photovoltaic power generation power related data, preprocessing the data, removing missing data to obtain a photovoltaic power generation power related data sample, and dividing the photovoltaic power generation power related data sample into a training data set and a test data set.
In the invention, the related data of the photovoltaic power generation power comprise air temperature, azimuth angle, cloud opacity, dew point temperature, DHI (solar scattering radiation index), DNI (direct solar radiation index), GHI (total solar horizontal radiation), GTI (fixed dip angle radiation), GTI (tracking dip angle radiation), atmospheric precipitation and relative humidity;
step 2, setting initial parameters of a model, including a bald eagle population, a fitness value and the maximum iteration times;
when the standard balying search algorithm selects the search space in the first stage, the random number is initialized so that the selected search space of the balying is simple and random, and the search space cannot be completely covered, which results in low algorithm solving precision and insufficient medium-term global searching capability. The invention is based on the improvement, introduces the idea of chaos optimization into the initialization of the bald eagle population, and the main idea of chaos optimization is to map variables into the value interval of the chaos variable space by utilizing the characteristics of ergodic property, randomness and the like, and finally linearly convert the obtained solution into the optimized variable space, thereby improving the algorithm performance. Therefore, the invention adopts Sine mapping in the chaotic mapping to initialize the population of the bald hawk searching algorithm, and the selected space mathematical model of the bald hawk searching algorithm after being initialized by the Sine chaotic mapping is as follows:
Figure SMS_32
wherein C is i Is [0,1 ]]Sine mapping of ub j And lb j The upper boundary and the lower boundary of the j-th dimension are respectively represented by a selected space mathematical model of a balying search algorithm after the initialization of the fine chaotic map of the fine, and population distribution map pairs of random initialization and the initialization of the fine chaotic map of the fine are shown in the figure 3.
Step 3, in the selection stage of the improved bald eagle searching algorithm: balding is used for determining the optimal searching position by randomly selecting searching area and judging the number of the hunting objects, so that hunting objects can be conveniently searched, and the balding position at the stage
Figure SMS_33
The update is determined by multiplying a priori information from a random search by a. The behavioral mathematical model is described as:
Figure SMS_34
wherein:
Figure SMS_35
for controlling the position-changing parameters>
Figure SMS_36
Is a random number between (0, 1)>
Figure SMS_37
Optimal value searched for the current bald eagle; ->
Figure SMS_38
For the average distribution position of bald hawk after the end of the previous search>
Figure SMS_39
The position of the ith bald eagle.
And 4, in a searching stage of the improved bald eagle searching algorithm, the improved bald eagle searching algorithm updates the bald eagle population through the following formula, the adaptability of the new bald eagle position is calculated, and the adaptability function is the Root Mean Square Error (RMSE) of the photovoltaic power generation power data training set.
Bal eagles fly in a spiral shape within the search space, looking for hunting. This flying approach can speed up the search process and help balding find the best dive capture location. For this purpose, we use polar equations to update the location of the bald eagle. The specific formula is as follows:
Figure SMS_40
Figure SMS_41
；
wherein:
Figure SMS_42
and->
Figure SMS_43
Polar angle and polar diameter of the spiral equation, respectively>
Figure SMS_44
And->
Figure SMS_45
Parameters for controlling the spiral track;
Figure SMS_46
is (0, 1) random number; ->
Figure SMS_47
And->
Figure SMS_48
Is the bald eagle position in polar coordinates.
Updating the bald eagle position through the searching process:
Figure SMS_49
wherein the method comprises the steps of
Figure SMS_50
Is->
Figure SMS_51
Only the bald eagle updates the position once.
And 5, in the dive stage of the improved bald hawk search algorithm, the third stage of the standard bald hawk search algorithm, namely the dive predation stage, is a key process for determining the convergence rate of the algorithm, and the standard bald hawk search algorithm is easy to deviate from the search direction and miss the optimal foraging area due to lack of effective control of step length, so that the algorithm falls into local optimum. Therefore, the invention provides a new non-inertial self-adaptive strategy, which can prevent the algorithm from sinking into local optimum while accelerating the algorithm convergence speed, so as to search for the hunting for quick reading, the bald eagle can start to dive from the optimum position in the search space, and other individuals can also add attack and move to the optimum position. Their state of motion can be described by polar equations.
Figure SMS_52
The improvement formula is as follows:
Figure SMS_53
wherein c 1 And c 2 For the intensity of movement of the bald hawk to the optimal center position, ω represents the non-inertial weight factor, iter max Represents the maximum number of iterations, t represents the current iterationNumber of generations, P i,n Indicating the location of the bald eagle after updating.
And 5, judging whether a stopping condition is met, judging whether the fitness function is the Root Mean Square Error (RMSE) of the photovoltaic power generation power data test set is optimal, outputting a global optimal parameter if the fitness function is optimal, and otherwise, re-executing the step 4.
Step 6, regularization parameters under the condition that the fitness function reaches the optimal conditioncAnd the width of the center point from the kernel functionσAnd assigning values to the LSSVM, and predicting the photovoltaic power generation power by the LSSVM by utilizing the optimal parameters. The LSSVM method formula is as follows: LSSVM is an important improvement over standard Support Vector Machines (SVMs). It uses the sum of squares of the errors as an empirical loss and uses the equality constraint instead of the inequality constraint in the original algorithm, thereby eliminating the need to solve the complex quadratic programming problem. The improvement reduces the difficulty of solving, optimizes the overall convergence speed and improves the accuracy of the calculation result. The specific formula process is as follows: setting training setS={(x i ，y i ），x i ∈R n1 ，y i E R }, wherein,x i is an input vector;y i is an output vector;i=1，2，…，N；Ris a real number set;R n1 is thatn1A set of dimensional real numbers;Nis the number of training samples.
The optimized objective function of the LSSVM can be translated into:
Figure SMS_54
Figure SMS_55
wherein:ωis a weight coefficient vector;φ( x)is a nonlinear mapping function;bas a result of the bias term,J as a loss function;e i is an error variable;cis a regularization parameter. General purpose medicineOver-structured Lagrangian multiplier α i Solving the above problems:
Figure SMS_56
according to the Carlo-Coulomb-Tak condition, each variable in the above formula is biased: by solving, eliminateωAnde i ，finally, the LSSVM regression function is obtained as follows:
Figure SMS_57
wherein:α i in order to be a lagrange multiplier,
Figure SMS_58
as a kernel function of the LSSVM model, a gaussian kernel function (Radial Basis Function, RBF) is selected as a kernel function of the LSSVM model, and the specific expression of the RBF is as follows:
Figure SMS_59
Figure SMS_60
for inputting feature vectors, ++>
Figure SMS_61
Is the width of the kernel from the center point.
And 7, acquiring real-time data of the generated power of the photovoltaic electric field to be predicted and related influence factors.
And 8, taking the real-time data of the relevant influence factors of the generated power of the photovoltaic electric field to be predicted as the input of a pre-constructed photovoltaic generated power prediction model, and outputting the result of the photovoltaic generated power prediction of the photovoltaic electric field to be predicted.
The invention also provides a photovoltaic power generation power short-term prediction system device, which comprises: a memory for storing a computer program; and the processor is used for realizing the photovoltaic power generation power short-term prediction method when the computer program is executed in the actual running process.
The processor when executing the computer program implements the steps of the above-mentioned photovoltaic power generation power short-term prediction method,
alternatively, the processor may implement functions of each module in the above system when executing the computer program, for example: the data acquisition module is used for acquiring the related data of the photovoltaic power generation power in the photovoltaic electric field acquired in real time; the data processing module is used for preprocessing the data and eliminating the missing data to obtain a photovoltaic power generation power related data sample; the sample input module is used for inputting the data samples related to the photovoltaic power generation power into the photovoltaic power generation power short-term prediction model; and the result output module is used for carrying out parameter optimization on the least square support vector machine through the IBES optimization algorithm so as to obtain a photovoltaic power generation power prediction model with the optimal effect and output a prediction result of the photovoltaic power generation power.
The photovoltaic power generation power short-term prediction device can adopt various computing devices such as a desktop computer, a notebook computer, a palm computer, a cloud server and the like. The apparatus is composed of multiple components of a processor, memory, etc., and is not limited thereto, and may include other components or different combinations. The device examples provided herein are for illustrative purposes only and are not limiting.
The processor may be various types of devices, such as CPU, DSP, ASIC, FPGA, etc., and may also include programmable logic devices, discrete gates, transistor logic, discrete hardware components, etc. The memory is used to store computer programs and modules, and the processor runs or executes the computer programs and/or modules stored in the memory and invokes data stored in the memory, thereby implementing various functions of the photovoltaic generation power short-term prediction model.
The memory may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program (such as a sound playing function, an image playing function, etc.) required for at least one function, and the like; the storage data area may store data (such as audio data, phonebook, etc.) created according to the use of the handset, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as a hard disk, memory, plug-in hard disk, smart memory card (SmartMediaCard, SMC), secure digital (SecureDigital, SD) card, flash card (FlashCard), at least one magnetic disk storage device, flash memory device, or other volatile solid state storage device.
The invention also provides a computer readable storage medium storing a computer program which when executed by a processor implements the steps of the photovoltaic power generation power short-term prediction method.
The photovoltaic power generation short-term prediction method, if implemented in the form of a software functional unit and sold or used as a separate product, may be stored in a computer-readable storage device.
Based on such understanding, the present invention may implement all or part of the above-mentioned short-term photovoltaic power generation power prediction method, or may be implemented by instructing relevant hardware through a computer program, where the computer program may be stored in a computer readable storage medium, and the computer program may implement the steps of the above-mentioned short-term photovoltaic power generation power prediction method when executed by a processor. The computer program comprises computer program code, and the computer program code can be in a source code form, an object code form, an executable file or a preset intermediate form and the like.
Computer readable storage medium may refer to any entity or device capable of carrying computer program code, including recording medium, USB flash disk, removable hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random Access Memory (RAM), software distribution medium, and so forth.
It should be noted that the computer readable storage medium may include content that is subject to appropriate increases and decreases as required by jurisdictions and by jurisdictions in which such computer readable storage medium does not include electrical carrier signals and telecommunications signals.
In this embodiment, the test is performed by using the data related to the photovoltaic power generation power of a certain photovoltaic electric field in Gansu province.
In order to evaluate the prediction effect, root mean square error (Root Mean Square Error, RMSE), mean deviation error MBE (Mean deviation error, MBE), mean absolute error (mean absoluteerror MAE) are selected as the main evaluation index of the model prediction accuracy.
As shown in fig. 3-8, fig. 3-8 respectively show comparison conditions of the photovoltaic power generation power prediction accuracy evaluation machine of the photovoltaic electric field. As can be seen from fig. 3 to 7, the least squares support vector machine (BES-LSSVM) is optimized compared to the Least Squares Support Vector Machine (LSSVM) and the conventional bald-Condition optimization algorithm; in the embodiment, the photovoltaic power generation power short-term prediction method based on the IBES-LSSVM has the advantages that the difference between the obtained photovoltaic power generation power prediction value and the true value is small, the iteration speed is higher, the practical application requirements are met, and the universality and the accuracy are higher; in the embodiment, the test set RMSE of the photovoltaic power generation power prediction method of the IBES-least square support vector machine is 0.26492, which is lower than other comparison methods, and has higher prediction precision; in addition, as can be seen from fig. 3-8, the global searching capability of the model is effectively improved by initializing the population of the bald eagle searching algorithm by sine mapping in the chaotic mapping, and the convergence speed of the algorithm is also effectively improved by fig. 4, compared with the Least Square Support Vector Machine (LSSVM) model and the Least Square Support Vector Machine (LSSVM), the prediction precision of the Least Square Support Vector Machine (LSSVM) model is improved, which indicates that the prediction effect by using the Least Square Support Vector Machine (LSSVM) is better.
The description of the relevant parts in the photovoltaic power generation short-term prediction system provided in this embodiment may refer to the detailed description of the corresponding specific implementation parts in the photovoltaic power generation short-term prediction described in this embodiment, which is not repeated here.
Compared with a single photovoltaic power generation power prediction model, the photovoltaic power generation power short-term prediction method and system provided by the invention effectively improve the prediction performance; the hawk search algorithm is improved through Sine mapping chaotic optimization (Sine map) and a non-inertial weight self-adaptive strategy, global search capacity and iteration speed of the algorithm are improved, a reliable basis is laid for improving prediction accuracy of a model, and a photovoltaic power generation power data set comprises the following data: air temperature, azimuth angle, cloud opacity, dew point temperature, DHI (solar scattering radiation index), DNI (solar direct radiation index), GHI (total solar horizontal radiation), GTI (fixed tilt angle radiation), GTI (tracking tilt angle radiation), atmospheric precipitation, and relative humidity.
The photovoltaic generating capacity and the meteorological features have strong correlation, the meteorological features are nonlinear, and the characteristics are contained in historical operation data, so that the least square support vector machine has good capacity for solving the problem, but the capacity is greatly reduced due to the problem of parameter setting, and therefore, the improved hawk searching algorithm is improved through Sine mapping chaotic optimization (Sine map) and a non-inertial weight self-adaptive strategy, the parameters of the least square support vector machine are optimized by the improved hawk searching algorithm, and the photovoltaic generating power short-term prediction model for obtaining optimal parameters can accurately predict photovoltaic generating power and has certain advantages compared with the existing machine learning model and group intelligent algorithm. The characteristics of an improved balying search algorithm (IBES) and a Least Square Support Vector Machine (LSSVM) are fully utilized, the problems of low prediction accuracy, low iteration speed and poor generalization performance of a single machine model are solved, and the model prediction capability is further improved.
The above embodiment is only one of the implementation manners capable of implementing the technical solution of the present invention, and the scope of the claimed invention is not limited to the embodiment, but also includes any changes, substitutions and other implementation manners easily recognized by those skilled in the art within the technical scope of the present invention.

Claims (14)

1. The short-term prediction method for the photovoltaic power generation power is characterized by comprising the following steps of:
acquiring photovoltaic power generation power data of a photovoltaic electric field and weather factor data at the same time;
constructing an IBES-LSSVM photovoltaic power generation short-term prediction model, wherein the IBES-LSSVM model comprises an improved baldness search algorithm and a least square support vector machine prediction model;
dividing the photovoltaic power generation power data of the obtained photovoltaic electric field and the meteorological factor data at the same time into a training data set and a test data set, and inputting the training data set into an IBES-LSSVM for training; obtaining a trained IBES-LSSVM model; the training process comprises optimizing parameters of a least square support vector machine prediction model and predicting a Least Square Support Vector Machine (LSSVM) through an improved baldness search algorithm, and meanwhile, obtaining an IBES-LSSVM photovoltaic power short-term prediction model with optimal parameters by taking root mean square error of a training data set as an fitness function;
inputting the test data set into a trained photovoltaic power generation short-term prediction model, performing performance evaluation on the trained IBES-LSSVM model, and determining a photovoltaic power generation short-term prediction model with optimal parameters;
and inputting meteorological factor data related to photovoltaic power generation power data to be predicted into an IBES-LSSVM photovoltaic power generation power short-term prediction model with optimal parameters, and outputting a predicted value of the photovoltaic power generation power.
2. The method for short-term prediction of photovoltaic power generation according to claim 1, wherein the obtaining of the photovoltaic power generation data of the photovoltaic electric field and the weather factor data at the same time specifically comprises: air temperature, azimuth angle, cloud opacity, dew point temperature, solar scattering radiation index, solar direct radiation index, solar total horizontal radiation, fixed tilt angle radiation, tracking tilt angle radiation, atmospheric precipitation, relative humidity.
3. The method for short-term prediction of photovoltaic power generation according to claim 1, wherein the improved process of the improved bald eagle search algorithm is specifically as follows:
when a search space is selected in the first stage aiming at a standard bald hawk search algorithm, introducing a chaos optimization idea into bald hawk population initialization;
the sinusoidal mapping in the chaotic mapping is adopted to initialize the population of the bald hawk searching algorithm, and the selected space mathematical model of the bald hawk searching algorithm after being initialized by the fine chaotic mapping is as follows:
Figure QLYQS_1
；/>
Figure QLYQS_2
；
wherein,,C i is [0,1 ]]Is used for the mapping of the sinusoids of (a),ub j andlb j respectively represent the firstjUpper and lower boundaries of dimensions;
based on the non-inertial adaptive strategy, the nose-down motion state of the bald eagle optimization algorithm can be described by a polar equation:
Figure QLYQS_3
；
Figure QLYQS_4
；
Figure QLYQS_5
；/>
Figure QLYQS_6
；
wherein,,
Figure QLYQS_7
and->
Figure QLYQS_8
For the change of the bald eagle to the optimal center position in the polar coordinate space, x 1 (i) And y 1 (i) Is the position of bald hawk in polar coordinates; p (P) best The most searched for the current bald eagleFigure of merit, P m Is the average distribution position of bald hawk after the previous search is finished, P i Is the position of the ith bald eagle; c 1 And c 2 The motion intensity of bald hawk to the optimal center position, and omega represents a non-inertial weight factor; ter (iter) max Representing a maximum number of iterations; t represents the current iteration number; p (P) i,n Indicating the location of the updated bald eagle; rand is a random number within (0, 1).
4. The method for short-term prediction of photovoltaic power generation according to claim 1, wherein the least squares support vector machine prediction model is specifically as follows:
the LSSVM uses the sum of squares of errors as an empirical loss and uses an equality constraint instead of an inequality constraint in the original algorithm, and the specific formula process is as follows: setting training setS={(x i ，y i ），x i ∈R n1 ，y i E R }, wherein,x i is an input vector;y i is an output vector;i=1，2, ..., N；Ris a real number set;R n1 is thatn1A set of dimensional real numbers;Nis the number of training samples;
the optimized objective function of the LSSVM can be translated into:
Figure QLYQS_9
；/>
Figure QLYQS_10
；
wherein:ωis a weight coefficient vector;
Figure QLYQS_11
is a nonlinear mapping function;bas a result of the bias term,J as a loss function;e i is an error variable;cfor regularization parameters, by constructing Lagrange multipliers α i Optimization for solving LSSVMObjective function:
Figure QLYQS_12
；
for the above formula
Figure QLYQS_13
The bias derivative of each variable in the (a): by solving, eliminateωAnde i ，finally, the LSSVM regression function is obtained as follows:
Figure QLYQS_14
；
wherein:
Figure QLYQS_15
as the kernel function of the LSSVM model, a Gaussian kernel function is selected as the kernel function of the LSSVM model, and the specific expression of the Gaussian kernel function is as follows:
Figure QLYQS_16
；
Figure QLYQS_17
for the input feature vector, σ is the width of the center point from the kernel function.
5. The photovoltaic power generation short-term prediction method according to claim 1, wherein the balding search algorithm optimizes parameters of a least square support vector machine prediction model, and the specific parameters are regularized parameters in the least square support vector machine prediction algorithmcAnd the width sigma of the center point from the kernel function.
6. The method for short-term prediction of photovoltaic power generation according to claim 1, wherein the fitness function of the IBES-LSSVM photovoltaic power generation short-term prediction model is the root mean square error of the training data set.
7. The method for short-term prediction of photovoltaic power generation according to claim 6, wherein the method for obtaining the optimal model parameters by judging the model parameters through the fitness function is as follows: and judging the model parameters to be optimal by minimizing the root mean square error of the training set.
8. The method for short-term prediction of photovoltaic power generation according to claim 1, wherein the process of determining the IBES-LSSVM photovoltaic power generation short-term prediction model is specifically as follows:
and comparing the evaluation indexes of the IBES-LSSVM model and other prediction models under the test set, and if the comparison result meets the set requirement, determining that the trained IBES-LSSVM model is the IBES-LSSVM photovoltaic power generation short-term prediction model.
9. The method of claim 8, wherein the evaluation index comprises a root mean square error, an average absolute error, and an average deviation error.
10. The photovoltaic power generation short-term prediction method according to claim 9, wherein the root mean square error, the average absolute error and the average deviation error are used as evaluation indexes to obtain the prediction results of the IBES-LSSVM model on the training set and the test set.
11. The photovoltaic power generation power short-term prediction system is characterized by comprising a data acquisition module, a data processing module, a sample input module and a result output module;
the data acquisition module is used for acquiring photovoltaic power generation power data of a photovoltaic electric field and weather factor data at the same time;
the data processing module is used for preprocessing the data and eliminating the missing data to obtain a photovoltaic power generation power related data sample;
the sample input module is used for inputting the photovoltaic power generation power related data samples into a photovoltaic power generation power short-term prediction model of the IBES-LSSVM least square support vector machine;
and the result output module is used for adjusting the parameters of the IBES optimization algorithm through the evaluation index result, so that the IBES-LSSVM photovoltaic power generation power short-term prediction model with the optimal effect is obtained, and the prediction result of the photovoltaic power generation power is output.
12. The photovoltaic power generation short-term prediction system according to claim 11, wherein the construction process of the IBES-LSSVM photovoltaic power generation short-term prediction model is as follows:
the hawk searching algorithm is improved through sine mapping chaotic optimization and a non-inertial weight self-adaptive strategy, an improved hawk searching algorithm is obtained, and the parameters of a least square support vector machine prediction model are optimized through the algorithm; and finally, inputting the data related to the photovoltaic power generation power into a prediction model, and adjusting the parameters of the least square support vector machine through a result IBES optimization algorithm, so as to obtain the photovoltaic power generation power prediction model with optimal parameters.
13. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method according to any one of claims 1 to 10 when the computer program is executed.
14. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method according to any one of claims 1 to 10.
CN202310580304.9A 2023-05-23 2023-05-23 Photovoltaic power generation power short-term prediction method, system, equipment and storage medium Pending CN116307299A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310580304.9A CN116307299A (en) 2023-05-23 2023-05-23 Photovoltaic power generation power short-term prediction method, system, equipment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310580304.9A CN116307299A (en) 2023-05-23 2023-05-23 Photovoltaic power generation power short-term prediction method, system, equipment and storage medium

Publications (1)

Publication Number Publication Date
CN116307299A true CN116307299A (en) 2023-06-23

Family

ID=86815313

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310580304.9A Pending CN116307299A (en) 2023-05-23 2023-05-23 Photovoltaic power generation power short-term prediction method, system, equipment and storage medium

Country Status (1)

Country Link
CN (1) CN116307299A (en)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115130375A (en) * 2022-06-17 2022-09-30 北京科技大学 Rock burst intensity prediction method
US20220373984A1 (en) * 2021-05-19 2022-11-24 Shandong University Hybrid photovoltaic power prediction method and system based on multi-source data fusion
CN115470987A (en) * 2022-09-14 2022-12-13 广西电网有限责任公司电力科学研究院 Short-term photovoltaic power generation prediction method based on improved long-term and short-term memory neural network

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20220373984A1 (en) * 2021-05-19 2022-11-24 Shandong University Hybrid photovoltaic power prediction method and system based on multi-source data fusion
CN115130375A (en) * 2022-06-17 2022-09-30 北京科技大学 Rock burst intensity prediction method
CN115470987A (en) * 2022-09-14 2022-12-13 广西电网有限责任公司电力科学研究院 Short-term photovoltaic power generation prediction method based on improved long-term and short-term memory neural network

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
周敏: "基于灰狼优化算法和最小二乘支持向量机的信用评估", 《成都理工大学学报》, vol. 46, no. 4, pages 507 - 512 *
李晟延: "基于PBES-LS-SVM的锂离子电池组SOC预测", 《电源技术》, vol. 46, no. 11, pages 1279 - 1283 *
石默涵: "基于混沌优化和自适应反向学习的秃鹰搜索算法", 《数学的实践与认识》, vol. 52, no. 6, pages 149 - 158 *

Similar Documents

Publication Publication Date Title
JP5888640B2 (en) Photovoltaic power generation prediction apparatus, solar power generation prediction method, and solar power generation prediction program
Lydia et al. Advanced algorithms for wind turbine power curve modeling
CN113128793A (en) Photovoltaic power combination prediction method and system based on multi-source data fusion
CN110942205B (en) Short-term photovoltaic power generation power prediction method based on HIMVO-SVM
CN111414717A (en) XGboost-L ightGBM-based unit power prediction method
Zhou et al. Transfer learning for photovoltaic power forecasting with long short-term memory neural network
CN113326969B (en) Short-term wind speed prediction method and system based on improved whale optimization algorithm and ELM
JP2017084360A (en) Method for operating photovoltaic power generation system
WO2021196743A1 (en) Tropical cyclone intensity forecast information generation method and system
Syama et al. A hybrid extreme learning machine model with lévy flight chaotic whale optimization algorithm for wind speed forecasting
Zhang et al. Short-term forecasting of wind power generation based on the similar day and Elman neural network
CN110866633A (en) Micro-grid ultra-short term load prediction method based on SVR support vector regression
CN115912502A (en) Intelligent power grid operation optimization method and device
Xiong et al. Research on a hybrid LSSVM intelligent algorithm in short term load forecasting
CN107844872B (en) Short-term wind speed forecasting method for wind power generation
Tahir et al. Enhanced solar photovoltaic power prediction using diverse machine learning algorithms with hyperparameter optimization
CN116307299A (en) Photovoltaic power generation power short-term prediction method, system, equipment and storage medium
CN113449466B (en) Solar radiation prediction method and system for optimizing RELM based on PCA and chaos GWO
CN115186865A (en) Mass distributed renewable energy power prediction method based on MapReduce and cloud picture
CN114444763A (en) Wind power prediction method based on AFSA-GNN
CN118171126A (en) Irradiance and photovoltaic power generation power combination prediction method and device
CN116776921B (en) Solar radiation prediction method and device based on improved patch-informer
Chen et al. Short-term PV power forecasting based on time-phased and error correction
CN117353302B (en) New energy power generation power prediction method, device, equipment and medium
CN117454456B (en) Offshore wind farm arrangement optimization simulation method and system based on ant colony algorithm

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
RJ01 Rejection of invention patent application after publication

Application publication date: 20230623

RJ01 Rejection of invention patent application after publication