Photovoltaic power station power prediction method and system based on BP neural network

Abstract

The invention discloses a photovoltaic power station power prediction method and system based on a BP neural network, and relates to the technical field of operation and maintenance, and the method comprises the steps: collecting meteorological information and historical optical power data related to the power of a photovoltaic power station, and classifying the historical optical power data; constructing a BP neural network model, training the BP neural network by using an error back propagation algorithm, and minimizing a loss function; through defining a fitness function, optimization iteration is carried out on a particle swarm based on a prediction error of a BP neural network, and an optimal weight and bias combination is found. According to the method, the identification capability of the model on the power change of the photovoltaic power station under different meteorological conditions is improved, a foundation is laid for accurate prediction, the prediction accuracy is improved, due to the self-adaptive learning capability of the neural network, the model can better adapt to the environmental change, the prediction robustness is improved, and the prediction efficiency is improved. The generalization ability of the neural network model is improved, prediction errors are reduced, and the overall performance of photovoltaic power station power prediction is improved.

Description

Technical Field

[0001] The present invention relates to the field of operation and maintenance technology, and in particular to a BP neural network photovoltaic power station power prediction method and system. Background Art

[0002] In the context of today's global energy structure transformation, photovoltaic power generation, as a clean and renewable energy form, is becoming increasingly important. The construction and operation of photovoltaic power stations play an important role in promoting the optimization of energy structure and reducing greenhouse gas emissions. However, the intermittent and uncertain nature of photovoltaic power generation has brought challenges to grid dispatching and power station operation and maintenance. Therefore, accurate prediction of the power output of photovoltaic power stations has become a research focus in the field of operation and maintenance technology. For a long time, researchers have been committed to developing various prediction models in order to improve the accuracy and reliability of photovoltaic power station power prediction.

[0003] However, although the existing photovoltaic power station power prediction technology has achieved certain results, it still has many shortcomings in practical applications. The existing prediction methods often use simple preprocessing methods to process meteorological information and historical optical power data, lack the process of in-depth data mining and classification, and cannot fully extract effective information from the data, resulting in insufficient prediction accuracy and robustness of the prediction model under complex weather conditions. In the training process of the traditional BP neural network, the optimization of the loss function is not sufficient, and the adjustment of the network parameters fails to reach the global optimum, which affects the accuracy of the prediction results. The existing neural network optimization strategy mostly relies on the gradient descent algorithm, which is easy to fall into the local optimal solution, limiting the generalization ability of the prediction model. When applying the particle swarm optimization algorithm, the existing technology fails to fully consider the characteristics of the BP neural network, resulting in low efficiency in the optimization process of weights and biases, and cannot meet the actual needs of photovoltaic power station power prediction. In the process of model construction, the existing prediction methods often ignore the influence of nonlinear factors such as attenuation of photovoltaic components, dust accumulation and temperature changes on power output, resulting in large errors in long-term predictions. The existing prediction technology has limited processing capabilities in a big data environment. When faced with massive meteorological data and optical power data, the calculation efficiency is low and it is difficult to meet the needs of real-time prediction. Summary of the invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is: the existing operation and maintenance technology has low prediction accuracy, poor model generalization ability, insufficient optimization algorithm efficiency, and how to achieve high-precision, high-stability and high-efficiency photovoltaic power station power prediction.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: a BP neural network photovoltaic power station power prediction method, comprising collecting meteorological information and historical optical power data related to the photovoltaic power station power, and classifying the historical optical power data; constructing a BP neural network model, using an error back propagation algorithm to train the BP neural network, and minimizing the loss function; by defining a fitness function, optimizing and iterating the particle swarm based on the prediction error of the BP neural network to find the optimal weight and bias combination.

[0007] As a preferred solution of the photovoltaic power station power prediction method of the BP neural network described in the present invention, the meteorological information and historical light power data include irradiance G, temperature T, DC current I PV , DC voltage U PV , AC current I AC , AC voltage U AC The irradiance G determines the output power of the photovoltaic module. The irradiance G and the output power P of the photovoltaic array PV The relationship between them is expressed as:

[0008] P PV =G·A·η

[0009] Among them, P PV is the output power of the photovoltaic array, A is the area of ​​the photovoltaic array, and η is the conversion efficiency of the photovoltaic module; the temperature T affects the conversion efficiency of the photovoltaic module, and the working efficiency of the battery and the inverter is affected by the temperature; the output voltage U of the photovoltaic array PV and current I PV The relationship between and output power is expressed as:

[0010] P PV =U PV I PV

[0011] The inverter converts DC power into AC power, and the output power is expressed as:

[0012] P AC =U AC I AC

[0013] At the same time, the input power and output power of the inverter follow the principle of energy conservation, which can be expressed as:

[0014] P PV +P Battery =P AC ·η inv

[0015] Among them, P Battery is the output power of the battery, η invThe efficiency of the inverter, the inverter output power is affected by the load characteristics, input voltage and current.

[0016] As a preferred solution of the photovoltaic power station power prediction method of the BP neural network described in the present invention, the classification of historical optical power data includes classifying the historical optical power data according to the collected meteorological data using a fuzzy C-means clustering algorithm, and the fuzzy clustering classifies the historical data into different meteorologically similar days by analyzing the meteorological information, thereby forming a classification result with similar meteorological conditions.

[0017] As a preferred solution of the photovoltaic power station power prediction method of the BP neural network described in the present invention, wherein: the construction of the BP neural network model includes taking the historical data of meteorologically similar days and real-time meteorological information as input variables of the BP neural network model based on fuzzy clustering results; designing the input layer, including irradiance, temperature, and wind speed, as factors affecting the power output of the photovoltaic power station; designing the hidden layer, including the number of nodes in the hidden layer determined according to experimental results and cross-validation; designing the output layer, including using the output nodes as the power prediction value of the photovoltaic power station.

[0018] As a preferred solution of the photovoltaic power station power prediction method of the BP neural network described in the present invention, wherein: the construction of the BP neural network model also includes a training process using an error back propagation algorithm to train the neural network to minimize the loss function. In the prediction model, the nonlinear relationship of the photovoltaic power generation system is modeled and predicted through the BP neural network.

[0019] As a preferred solution of the photovoltaic power station power prediction method of the BP neural network described in the present invention, the optimization iteration of the particle swarm includes particle swarm initialization, each particle represents the weight and bias parameters of the BP neural network, and the particle swarm algorithm finds the global optimal solution by simulating the collaborative search process of different particles; the fitness function is defined as the prediction error of the BP neural network, and the performance of each particle is evaluated according to the fitness function.

[0020] As a preferred solution of the photovoltaic power station power prediction method of the BP neural network described in the present invention, wherein: the finding of the optimal weight and bias combination includes continuously updating the position of particles according to the fitness value, and finally converging to the global optimal solution to find the optimal weight and bias combination; the problem of the fitting effect of some points of the BP neural network is solved by PSO optimization.

[0021] Another object of the present invention is to provide a photovoltaic power station power prediction system based on a BP neural network, which can construct a BP neural network model, use an error back propagation algorithm to train the BP neural network, minimize the loss function, and solve the problem that the current operation and maintenance technology contains models with insufficient modeling capabilities for the nonlinear relationship of photovoltaic power station power output.

[0022] As a preferred solution of the photovoltaic power station power prediction system of the BP neural network described in the present invention, it includes: a data acquisition and classification module, a neural network construction module, and a neural network training module; the data acquisition and classification module is used to collect meteorological information and historical light power data related to the photovoltaic power station power, and classify the historical light power data; the neural network construction module is used to build a BP neural network model, use the error back propagation algorithm to train the BP neural network, and minimize the loss function; the neural network training module is used to define the fitness function, optimize and iterate the particle swarm based on the prediction error of the BP neural network, and find the optimal weight and bias combination.

[0023] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of a photovoltaic power station power prediction method using a BP neural network.

[0024] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a photovoltaic power station power prediction method using a BP neural network.

[0025] The invention has the beneficial effects as follows: the photovoltaic power station power prediction method of the BP neural network provided by the invention collects meteorological information and historical optical power data related to the photovoltaic power station power, and classifies the historical optical power data, thereby realizing data preprocessing and feature extraction, providing high-quality input data for the subsequent BP neural network model, improving the model's ability to identify changes in photovoltaic power station power under different meteorological conditions, and laying a foundation for accurate prediction; constructing a BP neural network model, and using an error back propagation algorithm for training to minimize the loss function, realizing modeling of the nonlinear relationship of photovoltaic power station power output, and improving the accuracy of prediction; due to the adaptive learning ability of the neural network, the model can better adapt to environmental changes and improve the robustness of prediction; by defining a fitness function, optimizing and iterating a particle swarm based on the prediction error of the BP neural network, finding the optimal weight and bias combination, and realizing the optimization of the neural network parameters; utilizing the global search ability of the particle swarm optimization algorithm, avoiding the problem of the traditional gradient descent algorithm falling into a local optimal solution, improving the generalization ability of the neural network model, and reducing the prediction error; finding a better weight and bias combination through global optimization, and improving the overall performance of photovoltaic power station power prediction; the invention achieves better results in terms of prediction accuracy, model generalization ability, and global search strategy. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0027] Figure 1 This is an overall flow chart of a photovoltaic power station power prediction method using a BP neural network provided in the first embodiment of the present invention.

[0028] Figure 2 This is an overall flow chart of a photovoltaic power station power prediction system using a BP neural network provided in the third embodiment of the present invention. DETAILED DESCRIPTION

[0029] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0030] Example 1, reference Figure 1 , as an embodiment of the present invention, provides a photovoltaic power station power prediction method using a BP neural network, comprising:

[0031] S1: Collect meteorological information and historical optical power data related to the power of photovoltaic power stations, and classify the historical optical power data.

[0032] Furthermore, meteorological information and historical optical power data include irradiance G, temperature T, DC current I PV , DC voltage U PV , AC current I AC , AC voltage U AC , the irradiance G determines the output power of the photovoltaic module, and the irradiance G and the output power U of the photovoltaic array PV The relationship between them is expressed as:

[0033] P PV =G·A·η

[0034] Among them, P PV is the output power of the photovoltaic array, A is the area of ​​the photovoltaic array, and η is the conversion efficiency of the photovoltaic module; the temperature T affects the conversion efficiency of the photovoltaic module, and the working efficiency of the battery and the inverter is affected by the temperature; the output voltage U of the photovoltaic array PV and current I PVThe relationship between and output power is expressed as:

[0035] P PV =U PV I PV

[0036] The inverter converts DC power into AC power, and the output power is expressed as:

[0037] P AC =U AC I AC

[0038] At the same time, the input power and output power of the inverter follow the principle of energy conservation, which can be expressed as:

[0039] P PV +P Battery =P AC ·η inv

[0040] Among them, P Battery is the output power of the battery, η inv The efficiency of the inverter, the inverter output power is affected by the load characteristics, input voltage and current.

[0041] It should be noted that the classification of historical optical power data includes classifying the historical optical power data based on the collected meteorological data using the fuzzy C-means clustering algorithm. Fuzzy clustering classifies historical data into different meteorologically similar days by analyzing meteorological information, thereby forming classification results with similar meteorological conditions.

[0042] It should also be noted that by collecting meteorological information and historical optical power data related to the power of photovoltaic power stations and classifying the historical optical power data, data preprocessing and feature extraction are achieved, providing high-quality input data for the subsequent BP neural network model, improving the model's ability to recognize changes in photovoltaic power station power under different meteorological conditions, and laying the foundation for accurate prediction.

[0043] S2: Build a BP neural network model, use the error back propagation algorithm to train the BP neural network, and minimize the loss function.

[0044] Furthermore, constructing the BP neural network model includes taking the historical data of similar meteorological days and real-time meteorological information as input variables of the BP neural network model based on the fuzzy clustering results; designing the input layer, including irradiance, temperature, and wind speed, as factors affecting the power output of the photovoltaic power station; designing the hidden layer, including determining the number of nodes in the hidden layer according to experimental results and cross-validation; and designing the output layer, including using the output nodes as the power prediction value of the photovoltaic power station.

[0045] It should be noted that constructing the BP neural network model also includes a training process in which the neural network is trained using an error back propagation algorithm to minimize the loss function. In the prediction model, the nonlinear relationship of the photovoltaic power generation system is modeled and predicted through the BP neural network.

[0046] It should also be noted that by constructing a BP neural network model and using the error back propagation algorithm for training to minimize the loss function, the nonlinear relationship of the photovoltaic power station power output is modeled. Through precise model training, the accuracy of the prediction is improved. Due to the adaptive learning ability of the neural network, the model can better adapt to environmental changes and improve the robustness of the prediction.

[0047] S3: By defining the fitness function, the particle swarm is optimized and iterated based on the prediction error of the BP neural network to find the optimal weight and bias combination.

[0048] Furthermore, the optimization iteration of the particle swarm includes particle swarm initialization. Each particle represents the weight and bias parameters of the BP neural network. The particle swarm algorithm simulates the collaborative search process of different particles to find the global optimal solution. The fitness function is defined as the prediction error of the BP neural network, and the performance of each particle is evaluated according to the fitness function.

[0049] It should be noted that finding the optimal weight and bias combination includes continuously updating the position of particles according to the fitness value, and finally converging to the global optimal solution to find the optimal weight and bias combination; solving the problem of fitting effect of some points of BP neural network through PSO optimization.

[0050] It should also be noted that by defining the fitness function, the particle swarm is optimized and iterated based on the prediction error of the BP neural network, the optimal weight and bias combination is found, and the optimization of the neural network parameters is achieved. The global search capability of the particle swarm optimization algorithm is utilized to avoid the problem of the traditional gradient descent algorithm falling into the local optimal solution, the generalization ability of the neural network model is improved, the prediction error is reduced, and a better weight and bias combination is found through global optimization, which improves the overall performance of photovoltaic power station power prediction. The present invention achieves better results in prediction accuracy, model generalization ability and global search strategy.

[0051] Example 2 is an embodiment of the present invention, which provides a photovoltaic power station power prediction method based on a BP neural network. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0052] Firstly, meteorological data and historical optical power data related to the power output of photovoltaic power stations are collected, including irradiance, temperature, direct current, direct voltage, alternating current and alternating voltage. The historical optical power data are classified by using fuzzy C-means clustering algorithm to form classification results with similar meteorological conditions. A BP neural network model is constructed. The input layer of the model contains three variables: irradiance, temperature and wind speed. The number of nodes in the hidden layer is determined to be 15 through experimental results and cross-validation. The output layer outputs the power prediction value of the photovoltaic power station through one node. During the model training process, an error back propagation algorithm is used to minimize the loss function. A particle swarm optimization (PSO) algorithm is used to optimize the weights and biases of the neural network. Each particle represents a set of possible weight and bias combinations. The global optimal solution is found by simulating the collaborative search process of particles. The fitness function is defined as the prediction error of the BP neural network to evaluate the performance of each particle. Experimental results show that the present invention performs better than traditional methods in terms of prediction accuracy and computational efficiency, effectively solves the problem of poor fitting effect of some points, and provides a more accurate and efficient solution for power prediction of photovoltaic power stations.

[0053] Example 3, reference Figure 2 , which is an embodiment of the present invention, provides a photovoltaic power station power prediction system based on a BP neural network, including a data acquisition and classification module, a neural network construction module, and a neural network training module.

[0054] The data acquisition and classification module is used to collect meteorological information and historical optical power data related to the power of photovoltaic power stations, and classify the historical optical power data; the neural network construction module is used to build a BP neural network model, and use the error back propagation algorithm to train the BP neural network to minimize the loss function; the neural network training module is used to define the fitness function, and optimize the particle swarm iteratively based on the prediction error of the BP neural network to find the optimal weight and bias combination.

[0055] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0056] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0057] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering, or processing in another suitable manner as necessary, and then stored in a computer memory.

[0058] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc. It should be noted that the above embodiments are only used to illustrate the technical solution of the present invention and are not limited. Although the present invention is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

[0059] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A photovoltaic power station power prediction method based on BP neural network, characterized in that: include: Collect meteorological information and historical optical power data related to the power of photovoltaic power stations, and classify the historical optical power data; Construct a BP neural network model, use the error back propagation algorithm to train the BP neural network, and minimize the loss function; By defining the fitness function, the particle swarm is optimized and iterated based on the prediction error of the BP neural network to find the optimal weight and bias combination.

2. The photovoltaic power station power prediction method using a BP neural network as claimed in claim 1, characterized in that: The meteorological information and historical optical power data include irradiance G, temperature T, DC current I PV , DC voltage U PV , AC current I AC , AC voltage U AC The irradiance G determines the output power of the photovoltaic module. The irradiance G and the output power P of the photovoltaic array PV The relationship between them is expressed as: P PV =G·A·h Among them, P PV is the output power of the photovoltaic array, A is the area of ​​the photovoltaic array, and η is the conversion efficiency of the photovoltaic module; Temperature T affects the conversion efficiency of photovoltaic modules, and the working efficiency of batteries and inverters is affected by temperature; The output voltage of the photovoltaic array is U PV and current I PV The relationship between and output power is expressed as: P PV =U PV ·AND PV The inverter converts DC power into AC power, and the output power is expressed as: P AC =U AC ·AND AC At the same time, the input power and output power of the inverter follow the principle of energy conservation, which can be expressed as: P PV +P Battery =P AC ·η inv Among them, P Battery is the output power of the battery, η inv The efficiency of the inverter, the inverter output power is affected by the load characteristics, input voltage and current.

3. The photovoltaic power station power prediction method using a BP neural network as claimed in claim 2, characterized in that: The classification of the historical optical power data includes classifying the historical optical power data according to the collected meteorological data using a fuzzy C-means clustering algorithm. The fuzzy clustering classifies the historical data into different meteorologically similar days by analyzing meteorological information to form a classification result with similar meteorological conditions.

4. The photovoltaic power station power prediction method using a BP neural network as claimed in claim 3, characterized in that: The construction of the BP neural network model includes taking the historical data of similar meteorological days and the real-time meteorological information as input variables of the BP neural network model based on the fuzzy clustering results; Design input layers, including irradiance, temperature, and wind speed as factors affecting the power output of the PV plant; The hidden layer is designed, including the number of nodes in the hidden layer is determined according to experimental results and cross-validation; the output layer is designed, including using the output nodes as power prediction values ​​of the photovoltaic power station.

5. The photovoltaic power station power prediction method using BP neural network as claimed in claim 4, characterized in that: The construction of the BP neural network model also includes a training process in which the neural network is trained using an error back propagation algorithm to minimize a loss function. In the prediction model, the nonlinear relationship of the photovoltaic power generation system is modeled and predicted through the BP neural network.

6. The photovoltaic power station power prediction method using BP neural network as claimed in claim 5, characterized in that: The optimization iteration of the particle swarm includes initialization of the particle swarm, each particle represents the weight and bias parameters of the BP neural network, and the particle swarm algorithm searches for the global optimal solution by simulating the collaborative search process of different particles; The fitness function is defined as the prediction error of the BP neural network, and the performance of each particle is evaluated based on the fitness function.

7. The photovoltaic power station power prediction method using BP neural network as claimed in claim 6, characterized in that: The finding of the optimal weight and bias combination includes continuously updating the position of the particle according to the fitness value, and finally converging to the global optimal solution to find the optimal weight and bias combination; The problem of fitting effect of some points in BP neural network is solved through PSO optimization.

8. A system for photovoltaic power station power prediction method using the BP neural network according to any one of claims 1 to 7, characterized in that: Including data collection and classification module, neural network construction module, neural network training module; The data collection and classification module is used to collect meteorological information and historical optical power data related to the power of the photovoltaic power station, and classify the historical optical power data; The neural network building module is used to build a BP neural network model, and use an error back propagation algorithm to train the BP neural network to minimize the loss function; The neural network training module is used to define the fitness function, optimize and iterate the particle swarm based on the prediction error of the BP neural network, and find the optimal weight and bias combination.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the photovoltaic power station power prediction method using the BP neural network according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the photovoltaic power station power prediction method using a BP neural network according to any one of claims 1 to 7 are implemented.

Images