Photovoltaic power generation equipment fault diagnosis method and system based on artificial intelligence

Abstract

The invention belongs to the technical field of fault diagnosis, and discloses a photovoltaic power generation equipment fault diagnosis method and system based on artificial intelligence. The method comprises the following steps: acquiring a plurality of historical comprehensive monitoring data based on a cloud data center, and preprocessing the historical comprehensive monitoring data; carrying out clustering processing; using an artificial intelligence algorithm and a photovoltaic power generation equipment fault diagnosis model; using a reinforcement learning algorithm; acquiring real-time comprehensive monitoring data based on an equipment monitoring device; dividing real-time power generation time periods; performing fault diagnosis by using the matched photovoltaic power generation equipment fault diagnosis model; generating a real-time alarm signal according to the real-time fault diagnosis result; and generating an operation optimization strategy by using the photovoltaic power generation equipment operation optimization model, and returning the operation optimization strategy to the equipment control device of the corresponding photovoltaic power generation equipment. According to the method, the problems of low fault diagnosis efficiency and accuracy, insufficient real-time response capability and poor analysis comprehensiveness in the prior art are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of fault diagnosis, and in particular relates to a photovoltaic power generation equipment fault diagnosis method and system based on artificial intelligence. Background Art

[0002] With the development and popularization of photovoltaic power generation technology, more and more photovoltaic power generation equipment has been put into operation. Photovoltaic power generation equipment, also known as solar photovoltaic equipment or solar panels, is a new energy power generation equipment that directly converts sunlight into electrical energy. However, photovoltaic power generation equipment may have various faults during long-term operation, affecting the performance and life of the equipment. Online monitoring and fault diagnosis of photovoltaic power generation equipment is an important measure to ensure the stable and reliable operation of photovoltaic power generation equipment.

[0003] The existing fault diagnosis technology has the following defects: 1) Low fault diagnosis efficiency and accuracy: Existing fault diagnosis methods rely on manual experience and require professionals to conduct on-site inspection and analysis, which is time-consuming and inefficient. In addition, manual diagnosis is easily affected by subjective factors, resulting in insufficient accuracy and consistency of diagnostic results.

[0004] 2) Insufficient real-time response capability: Existing fault diagnosis methods often lack the ability to quickly respond to real-time monitoring data, and are unable to detect and handle faults in a timely manner, affecting the operating efficiency and reliability of photovoltaic power generation equipment.

[0005] 3) Poor analysis comprehensiveness: Existing fault diagnosis methods often perform fault analysis based on equipment operation data, ignoring the different fault manifestation characteristics in different power generation time periods, and ignoring other factors that have a close impact on fault diagnosis. Summary of the invention

[0006] In order to solve the problems of low fault diagnosis efficiency and accuracy, insufficient real-time response capability and poor analysis comprehensiveness in the prior art, the present invention aims to provide a photovoltaic power generation equipment fault diagnosis method and system based on artificial intelligence.

[0007] The technical solution adopted by the present invention is: A photovoltaic power generation equipment fault diagnosis method based on artificial intelligence comprises the following steps: Based on the cloud data center, the historical comprehensive monitoring data of different historical power generation moments of a number of photovoltaic power generation equipment are collected, and the historical comprehensive monitoring data are preprocessed to obtain a number of preprocessed historical comprehensive monitoring data; Taking the power generation difference as the clustering purpose, several pre-processed historical comprehensive monitoring data at different historical power generation moments are clustered to obtain several historical cluster centers and corresponding historical data sets at different historical power generation time periods; Based on the historical data set of each historical power generation time period, an artificial intelligence algorithm is used to construct a photovoltaic power generation equipment fault diagnosis model corresponding to the historical power generation time period, and generate several historical fault diagnosis results for each historical power generation time period; Based on several historical fault diagnosis results of all historical power generation time periods, a reinforcement learning algorithm is used to build an optimization model for the operation of photovoltaic power generation equipment; Based on the equipment monitoring device, the real-time comprehensive monitoring data of the current real-time power generation time of the photovoltaic power generation equipment is collected, and the real-time comprehensive monitoring data is uploaded to the cloud data center; Based on the cloud data center, the Euclidean distance between the real-time comprehensive monitoring data and several historical cluster centers is obtained, and the historical power generation time period corresponding to the historical cluster center with the closest Euclidean distance is used as the real-time power generation time period to which the current real-time power generation moment belongs; According to the real-time power generation time period, the corresponding matching photovoltaic power generation equipment fault diagnosis model is extracted, and the matching photovoltaic power generation equipment fault diagnosis model is used to perform fault diagnosis on the real-time comprehensive monitoring data to obtain real-time fault diagnosis results; If the real-time fault diagnosis result is that there is a fault, a corresponding real-time alarm signal is generated, the real-time alarm signal is visualized, and the next step is entered; otherwise, the real-time comprehensive monitoring data of the next real-time power generation moment is continuously collected, and the cluster center matching step is returned; According to the real-time fault diagnosis results, the operation optimization model of the photovoltaic power generation equipment is used to generate an operation optimization strategy, obtain the real-time operation optimization strategy, and return the real-time operation optimization strategy to the equipment control device of the corresponding photovoltaic power generation equipment.

[0008] Furthermore, the historical / real-time comprehensive monitoring data includes historical / real-time control signal data, historical / real-time equipment operation data and historical / real-time environmental monitoring data of the photovoltaic power generation equipment collected at the same historical / real-time power generation moment.

[0009] Furthermore, the preprocessing includes format conversion, denoising, data cleaning and normalization processing of a number of historical comprehensive monitoring data in sequence.

[0010] Furthermore, taking the power generation difference as the clustering purpose, the FCM clustering algorithm is used to cluster several pre-processed historical comprehensive monitoring data at different historical power generation moments, and several historical clustering centers and corresponding historical data sets of different historical power generation time periods are obtained.

[0011] Furthermore, according to the historical data set of each historical power generation time period, an artificial intelligence algorithm is used to construct a photovoltaic power generation equipment fault diagnosis model corresponding to the historical power generation time period, and generate a number of historical fault diagnosis results for each historical power generation time period, including the following steps: Use deep learning algorithms to build an initial photovoltaic power generation equipment fault diagnosis model, and use the initial network parameters of the photovoltaic power generation equipment fault diagnosis model as the optimization target. Use swarm intelligence optimization algorithms to search and optimize, and obtain the optimal initial network parameters of the initial photovoltaic power generation equipment fault diagnosis model. Labels are added to the historical data set of the current historical power generation period to obtain a photovoltaic power generation equipment fault diagnosis sample set with real labels, and the photovoltaic power generation equipment fault diagnosis sample set is divided into a model training set and a model test set in a ratio of 7:3; Based on the optimal initial network parameters, the initial photovoltaic power generation equipment fault diagnosis model is optimized to obtain an optimized photovoltaic power generation equipment fault diagnosis model, and the optimized photovoltaic power generation equipment fault diagnosis model is trained and optimized according to the model training set to obtain a trained photovoltaic power generation equipment fault diagnosis model; According to the model test set, the trained photovoltaic power generation equipment fault diagnosis model is tested to obtain several historical prediction labels, which are compared and counted with several real labels of the model test set to obtain the model test accuracy; If the model test accuracy is greater than the accuracy threshold, the optimal photovoltaic power generation equipment fault diagnosis model for the current historical power generation period is output, and a number of corresponding historical fault diagnosis results are generated; otherwise, the training optimization of the photovoltaic power generation equipment fault diagnosis model is continued; The historical data sets of all historical power generation time periods are traversed to build a fault diagnosis model for photovoltaic power generation equipment in each historical power generation time period.

[0012] Furthermore, the deep learning algorithm is the LSTM-Attention-cGAN algorithm, and the photovoltaic power generation equipment fault diagnosis model includes an input layer, a feature extraction module based on the LSTM algorithm, an attention weight module based on the Attention mechanism, an adversarial training module based on the cGAN algorithm, a classification layer, and an output layer; The swarm intelligence optimization algorithm is the ICPO algorithm.

[0013] Furthermore, based on several historical fault diagnosis results of all historical power generation time periods, a photovoltaic power generation equipment operation optimization model is constructed using a reinforcement learning algorithm, including the following steps: The photovoltaic power generation equipment operation optimization problem is used as the simulation environment of the reinforcement learning algorithm to build an intelligent agent and experience replay pool; According to several historical photovoltaic power generation equipment failure conditions included in several historical fault diagnosis results, several historical photovoltaic power generation equipment failure states are set, and all historical photovoltaic power generation equipment failure states are integrated to define the state space of the reinforcement learning algorithm; Based on several photovoltaic power generation equipment control signals included in the preset operation optimization strategy of the photovoltaic power generation equipment, the action space of the reinforcement learning algorithm is defined, and the reward function is set according to the interaction between the intelligent agent and the simulation environment; According to the spatial structure of the state space and the spatial structure of the action space, a neural network is constructed, and according to the intelligent agent of the reinforcement learning algorithm, the experience replay pool and the neural network, the model architecture of the photovoltaic power generation equipment operation optimization model is constructed; Set the state space, action space and reward function for the model architecture to obtain the initial photovoltaic power generation equipment operation optimization model; According to several historical fault diagnosis results of all historical power generation time periods, the initial photovoltaic power generation equipment operation optimization model is optimized and trained to obtain the final photovoltaic power generation equipment operation optimization model, and generate several historical operation optimization strategy generation experiences; Several historical operation optimization strategy generation experiences and corresponding historical fault diagnosis results are stored in the experience playback pool of the final photovoltaic power generation equipment operation optimization model.

[0014] Furthermore, the reinforcement learning algorithm is a DQN algorithm, and the neural network is a deep Q network.

[0015] Further, according to the real-time fault diagnosis result, the operation optimization model of the photovoltaic power generation equipment is used to generate an operation optimization strategy, obtain the real-time operation optimization strategy, and return the real-time operation optimization strategy to the equipment control device of the corresponding photovoltaic power generation equipment, including the following steps: Obtain the Euclidean distance between the real-time fault diagnosis result and several historical fault diagnosis results in the experience playback pool of the photovoltaic power generation equipment operation optimization model, and use the historical fault diagnosis result whose Euclidean distance exceeds the Euclidean distance threshold as the matching historical fault diagnosis result of the real-time fault diagnosis result; Extracting a number of matching historical operation optimization strategy generation experiences corresponding to the matching historical fault diagnosis results, and updating the action space of the photovoltaic power generation equipment operation optimization model according to the number of matching historical operation optimization strategy generation experiences to obtain an updated action space; According to several real-time photovoltaic power generation equipment fault conditions included in the real-time fault diagnosis results, several real-time photovoltaic power generation equipment fault states are set, and all real-time photovoltaic power generation equipment fault states are integrated to update the state space of the photovoltaic power generation equipment operation optimization model to obtain an updated state space; Input the updated state space into the deep Q network of the photovoltaic power generation equipment operation optimization model, and use the deep Q network to generate the Q value of each possible action in the updated action space based on the updated state space; Using the agent, according to the reward function, obtain the reward value of each possible action in the updated action space, and according to the reward value, update the Q value of the possible action to obtain the updated Q value of each possible action; Repeat the above Q-value update steps until the number of iterations reaches the iteration threshold, and use the greedy strategy to select the possible action with the highest updated Q-value in each iteration as the execution action; Using an intelligent agent, several execution actions are integrated to obtain a real-time operation optimization strategy, and the real-time operation optimization strategy is returned to the equipment control device of the corresponding photovoltaic power generation equipment.

[0016] A photovoltaic power generation equipment fault diagnosis system based on artificial intelligence is used to implement a photovoltaic power generation equipment fault diagnosis method. The system includes a cloud data center, a plurality of equipment monitoring devices, a plurality of equipment control devices, and a plurality of photovoltaic power generation equipment. The cloud data center is respectively connected to the plurality of equipment monitoring devices and the plurality of equipment control devices in communication. Each equipment monitoring device and each equipment control device are arranged at a photovoltaic power generation equipment in a one-to-one correspondence, and both the equipment monitoring device and the equipment control device are connected to the corresponding photovoltaic power generation equipment in communication. The cloud data center is provided with a data preprocessing unit, a clustering processing unit, a fault diagnosis model building unit, an operation optimization model building unit, a data partitioning unit, a fault diagnosis unit, an alarm signal generating unit and an operation optimization unit which are connected in sequence.

[0017] The beneficial effects of the present invention are: The invention discloses a photovoltaic power generation equipment fault diagnosis method and system based on artificial intelligence. The method adopts an artificial intelligence algorithm to construct a photovoltaic power generation equipment fault diagnosis model, performs automated and intelligent photovoltaic power generation equipment fault diagnosis, avoids manual inspection and analysis, improves fault diagnosis efficiency, eliminates the influence of subjective factors, and improves fault diagnosis accuracy and consistency. Real-time data acquisition and cloud data processing are performed. After uploading real-time comprehensive monitoring data to a cloud data center, the matching fault diagnosis model is extracted by matching historical clustering centers, and fault diagnosis is performed on real-time data, thereby achieving rapid response to real-time data, timely discovering and processing faults, and improving the operation efficiency and reliability of photovoltaic power generation equipment. The method comprehensively considers the potential influence of signal data, equipment operation data and environmental monitoring data on photovoltaic power generation equipment, and considers that the fault manifestation characteristics in different power generation time periods are different, establishes photovoltaic power generation equipment fault diagnosis models in different power generation time periods, and improves the comprehensiveness and accuracy of model analysis. A fault diagnosis result feedback mechanism is established, and the operation state of photovoltaic power generation equipment is adaptively adjusted and optimized through a photovoltaic power generation equipment operation optimization model of a forced learning algorithm. When a potential fault is first discovered, the photovoltaic power generation equipment is controlled in real time to ensure stable and reliable operation of the photovoltaic power generation equipment.

[0018] Other beneficial effects of the present invention will be further described in the specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a flowchart of the photovoltaic power generation equipment fault diagnosis method based on artificial intelligence in the present invention.

[0020] Figure 2 It is a structural block diagram of the photovoltaic power generation equipment fault diagnosis system based on artificial intelligence in the present invention. DETAILED DESCRIPTION

[0021] The present invention will be further explained below in conjunction with the accompanying drawings and specific embodiments.

[0022] Embodiment 1: like Figure 1 As shown, this embodiment provides a photovoltaic power generation equipment fault diagnosis method based on artificial intelligence, comprising the following steps: S1: Based on the cloud data center, historical comprehensive monitoring data of several photovoltaic power generation equipment at different historical power generation times are collected, and several historical comprehensive monitoring data are preprocessed to obtain several preprocessed historical comprehensive monitoring data; The historical comprehensive monitoring data includes the historical control signal data, historical equipment operation data and historical environmental monitoring data of the photovoltaic power generation equipment collected at the same historical power generation time; The control signal data includes the maximum power point tracking control signal of the photovoltaic power generation equipment, the inverter control signal, the grid synchronization control signal, the fault detection and protection control signal, the energy storage system control signal, and the remote communication and control signal; The maximum power point tracking control signal is used to adjust the voltage and current of the photovoltaic array to maximize the output power. Common control signals include the output voltage and current regulation instructions of the maximum power point tracking controller, etc. The inverter control signal includes the switching frequency, output voltage and current regulation instructions, etc. The grid synchronization control signal includes the output frequency and phase adjustment instructions of the grid synchronization controller, etc. The fault detection and protection control signal includes the fault detection threshold and fault response instructions, etc. The remote communication and control signal includes the data transmission instructions and control commands of the remote communication interface, etc. The energy storage system control signal includes the charge and discharge control and energy management strategy, etc. Equipment operation data includes the power generation power, operating current, operating voltage, operating temperature, battery status, inverter status, grid interaction data, system operation time and grid load data of the photovoltaic power generation equipment; Environmental monitoring data includes ambient temperature, ambient humidity, light intensity, wind speed, weather data, altitude, and atmospheric pressure; Preprocessing includes format conversion, denoising, data cleaning and normalization of several historical comprehensive monitoring data in sequence; By converting the data format of heterogeneous data into homogeneous data, data support is provided for subsequent model establishment and analysis and prediction; Denoising: using Gaussian denoising algorithm to remove noise from data, improve data quality, and avoid noise interfering with model accuracy; Data cleaning includes: missing value processing, filling or deleting missing data, outlier detection and processing, identifying and processing abnormal or outlier values, duplicate data processing, deleting duplicate records, and improving data quality; Normalization processing unifies the magnitude of different data to eliminate the accuracy impact caused by magnitude differences; S2: Taking power generation difference as the clustering purpose, the fuzzy C-mean (FCM) clustering algorithm is used to cluster several pre-processed historical comprehensive monitoring data at different historical power generation moments to obtain several historical cluster centers and corresponding historical data sets at different historical power generation time periods, including the following steps: S2-1: Select clustering parameters of the FCM clustering algorithm, including the fuzzy factor and the total number of cluster centers. In this embodiment, the power generation time period is divided into a valley power generation time period, a normal power generation time period, and a peak power generation time period based on the power generation difference as the clustering purpose. Therefore, the total number of cluster centers is 3; The comprehensive monitoring data of photovoltaic power generation equipment in different power generation time periods (valley, normal, peak) usually have significant distribution differences. The valley power generation time period has lower power output, lower operating current and operating voltage, the battery charge state may be low, and the equipment is in standby or partial operation. The normal power generation time period has a medium level of power output, relatively stable operating current and operating voltage, the battery charge state remains stable, and the equipment operates in the standard operation mode. The peak power generation time period has high power output, high operating current and high operating voltage, the battery is fully charged, and the equipment operates at the maximum power generation efficiency. The corresponding control signal data and environmental monitoring data also have large differences. When analyzing and processing the comprehensive monitoring data, it is necessary to take into account the characteristics of the time period in order to more accurately understand the operating status of the equipment; S2-2: Based on the clustering parameters, the fuzzy C-mean (FCM) clustering algorithm is used to cluster a number of pre-processed historical comprehensive monitoring data to obtain a number of initial cluster centers; S2-3: According to the Euclidean distance between each pre-processed historical comprehensive monitoring data and a number of initial cluster centers, a corresponding fuzzy membership degree is set for each initial cluster center; The formula is:

[0023] In the formula, For the i The historical comprehensive monitoring data after preprocessing and the j Euclidean distance of cluster centers; For the Preprocessed historical comprehensive monitoring data; For the j Cluster center; is the data indicator; is the cluster center indicator; S2-4: Update the cluster center according to the fuzzy membership degree to obtain a number of corresponding updated cluster centers; The formula for updating fuzzy membership is:

[0024] In the formula, For the Preprocessed historical comprehensive monitoring data; is the data indicator; All are cluster center indicators; is the total number of cluster centers; For the i After preprocessing, historical comprehensive monitoring data is j , k The distance to the cluster center; For the i After preprocessing, the historical comprehensive monitoring data belongs to Updated fuzzy membership of cluster centers; The formula for updating the cluster center is:

[0025] In the formula, For the j Updated cluster centers; is the fuzzy factor; is the data indicator; is the total number of data; is the cluster center indicator; For the Preprocessed historical comprehensive monitoring data; For the i After preprocessing, the historical comprehensive monitoring data belongs to Fuzzy membership of cluster centers; S2-5: Use the Lagrange multiplier method to calculate the merge function and obtain the merge function value and merge function change value. The formula is:

[0026] In the formula, For the t , t -1 iteration of the combined function value of the Lagrange multiplier method; is the corresponding change value; For the i Characteristic parameters; t is the indicator of the number of iterations; For the i After preprocessing, the historical comprehensive monitoring data belongs to The membership degree of the cluster center; is the fuzzy factor; is the data indicator; is the total number of data; is the cluster center indicator; is the total number of cluster centers; For the i After preprocessing, historical comprehensive monitoring data is j Euclidean distance of cluster centers; S2-6: Repeat the above steps. If the merge function value is greater than the function threshold, or the merge function change value is greater than the change value threshold, continue to update the cluster center. Otherwise, use the current cluster center as the final cluster center. S2-7: Divide the pre-processed historical comprehensive monitoring data into the final cluster center with the closest Euclidean distance to obtain the corresponding historical data set; S2-8: setting a corresponding historical power generation time period label for each final cluster center, and spreading the historical power generation time period label to the corresponding historical data set, to obtain several historical cluster centers and corresponding historical data sets of different historical power generation time periods; S3: Based on the historical data set of each historical power generation time period, an artificial intelligence algorithm is used to construct a photovoltaic power generation equipment fault diagnosis model corresponding to the historical power generation time period, and generate several historical fault diagnosis results for each historical power generation time period, including the following steps: S3-1: Use the deep learning algorithm to build an initial photovoltaic power generation equipment fault diagnosis model, and use the initial network parameters of the photovoltaic power generation equipment fault diagnosis model as the optimization target. Use the swarm intelligence optimization algorithm to search and optimize to obtain the optimal initial network parameters of the initial photovoltaic power generation equipment fault diagnosis model, including the following steps: S3-1-1: Use the Long Short-Term Memory (LSTM)-Attention-Conditional Generative Adversarial Network (cGAN) algorithm to build an initial photovoltaic power generation equipment fault diagnosis model; The photovoltaic power generation equipment fault diagnosis model includes an input layer, a feature extraction module based on the LSTM algorithm, an attention weight module based on the Attention mechanism, an adversarial training module based on the cGAN algorithm, a classification layer, and an output layer. The input layer is used to receive comprehensive monitoring data. The LSTM network realizes feature extraction of data through multiple gating mechanisms and has a memory function, which improves the accuracy and stability of the model. It is used to extract key features from time series data and convert the processed data into a hidden state sequence. The Attention mechanism realizes the splicing of key features through preset attention weights, integrates scattered features, enhances the influence of key features on label prediction, and increases the flexibility and explanatory power of the model. The cGAN network generates data based on conditional input, simulates different operating scenarios or generates anomaly detection samples, and is used to input the attention weights and LSTM network outputs into cGAN as conditions for adversarial training. Adversarial training helps the model learn more complex distributions and improve generalization capabilities. The cGAN generator tries to generate data similar to the real data distribution, while the discriminator tries to distinguish between the real data and the data generated by the generator. The softmax classification function is used according to the output of the adversarial training module to perform feature classification and obtain the corresponding prediction code. The output layer converts the prediction code into the corresponding fault diagnosis result. S3-1-2: Taking the initial network parameters of the photovoltaic power generation equipment fault diagnosis model as the optimization target, the Improved Crested Porcupine Optimizer (ICPO) optimization algorithm is used to search and optimize to obtain the optimal initial network parameters of the initial photovoltaic power generation equipment fault diagnosis model, including the following steps: S3-1-2-1: According to the optimization goal, set the encoding vector, fitness function, ICPO population parameters and maximum number of iterations of the ICPO individual; The formula for the fitness value is:

[0027] In the formula, is the fitness function; is the mean square error function; For the The true output value of the neurons in the classification layer; For the The ideal output value of the classification layer neurons; is the indicator of the classification layer neurons; N is the total number of neurons in the classification layer; S3-1-2-2: Based on the coding vector of the ICPO individual and the ICPO population parameters, the Circle chaotic mapping sequence is used for initialization to generate several initial ICPO individuals of the initial ICPO population; The formula is:

[0028] In the formula, is the initial ICPO individual of the Circle chaos map, i.e. the initial solution; is the randomly generated initial ICPO individual; It is the individual indicator of ICPO; S3-1-2-3: Introduce a cyclic population reduction mechanism to limit the number of individuals in the ICPO population parameters and obtain updated ICPO population parameters for the next iteration; The formula is:

[0029] In the formula, For the The number of individuals in the ICPO population parameter of the iteration; For the The number of individuals in the ICPO population parameter of the iteration; is the minimum number of individuals in the ICPO population parameter; Evaluate arguments for functions; Evaluate loop parameters for a function; Evaluate loop parameters for the maximum function; t is the indicator of the number of iterations; S3-1-2-4: Calculate the initial fitness value of the initial ICPO individuals in the initial ICPO population according to the fitness function; S3-1-2-5: According to the initial fitness value and the updated ICPO population parameters, the first defense strategy, the second defense strategy, the third defense strategy and the fourth defense strategy are used to update the initial ICPO population to obtain an updated ICPO population; The formula for the first defense strategy is:

[0030] In the formula, For the updated ICPO individuals within the first defense range; The initial ICPO individual within the first defense range; is a random number based on normal distribution; is a random value in the interval [0,1]; It is the optimal solution within the first defense range; is the vector generated between the true optimal solution within the first defense range and the optimal solution randomly selected from the ICPO population; It is the individual indicator of ICPO; is the iteration indicator; The formula for the second defense strategy is:

[0031] In the formula, For the updated ICPO individuals within the second defense range; The initial ICPO individual within the second defense range; is the search upper limit vector of the second defense range; is a random value in the interval [0,1]; Respectively Initial ICPO individuals; All are [1, ]; is the vector generated between the true optimal solution within the second defense range and the optimal solution randomly selected from the ICPO population; The formula for the third defense strategy is:

[0032] In the formula, For the updated ICPO individuals within the third defense range; The initial ICPO individual within the third defense range; is the search upper limit vector of the third defense range; Respectively Initial ICPO individuals; is [1, ]; The odor diffusion factor defined for the fitness function; It is a defense factor; Control parameters for search direction; The formula for the fourth defense strategy is:

[0033] In the formula, For the updated ICPO individuals within the fourth defense range; The initial ICPO individual within the fourth defense range; It is the optimal solution within the fourth defense range; All are random values ​​in the interval [0,1]; It is a defense factor; Control parameters for search direction; is the average force affecting the search direction; is the convergence speed factor; S3-1-2-6: Use the dynamic reverse learning algorithm to perform dynamic reverse learning on the updated ICPO population to generate a dynamic reverse ICPO population; The formula is:

[0034] In the formula, It is a dynamically reversed ICPO individual; is the decreasing inertia coefficient; are the maximum and minimum values ​​of the vector space respectively; For the updated ICPO individuals; S3-1-2-7: According to the fitness function, calculate the fitness values ​​of all ICPO individuals in the updated ICPO population and the dynamically reversed ICPO population, take the ICPO individual with the minimum fitness value as the optimal individual, and retain the optimal individual; S3-1-2-8: If the number of iterations of iterative optimization reaches the maximum number of iterations or the fitness value of the optimal individual meets the requirements, the optimal solution corresponding to the optimal individual is output to obtain the optimal initial network parameters of the initial photovoltaic power generation equipment fault diagnosis model. The initial network parameters include the number of neurons of LSTM, the initial weight of the gate structure, the initial bias, the first learning rate, and the initial generator network structure parameters, the initial discriminator network structure parameters, the second initial learning rate and the initial noise dimension of the adversarial training network; Photovoltaic power generation equipment fault diagnosis models based on deep learning algorithms generally have defects such as being sensitive to initial values, being easily trapped in local optimal values ​​and unable to jump out, and being prone to premature maturity. By using the ICPO optimization algorithm to optimize the initial network parameters, the above defects are solved and the accuracy, efficiency and stability of the model are improved. S3-2: adding labels to the historical data set of the current historical power generation time period to obtain a photovoltaic power generation equipment fault diagnosis sample set with real labels, and dividing the photovoltaic power generation equipment fault diagnosis sample set into a model training set and a model test set in a ratio of 7:3; S3-3: Based on the optimal initial network parameters, the initial photovoltaic power generation equipment fault diagnosis model is optimized to obtain an optimized photovoltaic power generation equipment fault diagnosis model, and the optimized photovoltaic power generation equipment fault diagnosis model is trained and optimized according to the model training set to obtain a trained photovoltaic power generation equipment fault diagnosis model; S3-4: According to the model test set, the trained photovoltaic power generation equipment fault diagnosis model is tested to obtain several historical prediction labels, and the labels are compared and counted with several real labels of the model test set to obtain the model test accuracy; S3-5: If the model test accuracy is greater than the accuracy threshold, the optimal photovoltaic power generation equipment fault diagnosis model for the current historical power generation time period is output, and a number of corresponding historical fault diagnosis results are generated; otherwise, the training optimization of the photovoltaic power generation equipment fault diagnosis model is continued; S3-6: Traverse the historical data set of all historical power generation time periods and build a photovoltaic power generation equipment fault diagnosis model for each historical power generation time period; The fault types of photovoltaic power generation equipment in different time periods (off-peak, normal, and peak) are quite different. Off-peak power generation time periods include battery charging failures, inverter efficiency problems, sensor failures or false alarms, and temporary failures caused by maintenance operations. Normal power generation time periods include routine failures during normal equipment operation, such as component aging, electrical connection problems, and system stability problems, such as voltage fluctuations and frequency deviations. Peak power generation time periods include overheating failures, such as overheating of batteries, inverters, or photovoltaic panels, overload protection triggering, cooling system failures, and transmission line or transformer failures. The comprehensive monitoring data of different power generation time periods (low, normal, and peak) show significantly different fault characteristics or operation modes. It is necessary to train a special fault diagnosis model for each power generation time period to improve the adaptability, prediction accuracy, and stability of the fault diagnosis model. S4: Based on several historical fault diagnosis results of all historical power generation time periods, a photovoltaic power generation equipment operation optimization model is constructed using the Deep Q Network (DQN) algorithm, including the following steps: S4-1: Use the photovoltaic power generation equipment operation optimization problem as the simulation environment of the DQN algorithm to build an intelligent agent and experience replay pool; S4-2: according to several historical photovoltaic power generation equipment failure conditions included in several historical fault diagnosis results, several historical photovoltaic power generation equipment failure states are set, and all historical photovoltaic power generation equipment failure states are integrated to define the state space of the DQN algorithm; S4-3: Define the action space of the DQN algorithm according to the photovoltaic power generation equipment control signals included in the preset operation optimization strategy of the photovoltaic power generation equipment, and set the reward function according to the interaction between the intelligent agent and the simulation environment; S4-4: Based on the spatial structure of the state space and the spatial structure of the action space, a deep Q network is constructed, and based on the intelligent agent, experience replay pool and neural network of the DQN algorithm, the model architecture of the photovoltaic power generation equipment operation optimization model is constructed; S4-5: Set the state space, action space and reward function for the model architecture to obtain the initial photovoltaic power generation equipment operation optimization model; S4-6: Based on several historical fault diagnosis results of all historical power generation time periods, the initial photovoltaic power generation equipment operation optimization model is optimized and trained to obtain the final photovoltaic power generation equipment operation optimization model, and several historical operation optimization strategy generation experiences are generated; S4-7: storing several historical operation optimization strategy generation experiences and corresponding historical fault diagnosis results into the experience playback pool of the final photovoltaic power generation equipment operation optimization model; S5: Based on the equipment monitoring device, collect the real-time comprehensive monitoring data of the current real-time power generation time of the photovoltaic power generation equipment, and upload the real-time comprehensive monitoring data to the cloud data center; The real-time comprehensive monitoring data includes the real-time control signal data of the photovoltaic power generation equipment, the real-time equipment operation data and the real-time environmental monitoring data collected at the same real-time power generation time; S6: Based on the cloud data center, obtain the Euclidean distance between the real-time comprehensive monitoring data and several historical cluster centers, and use the historical power generation time period corresponding to the historical cluster center with the closest Euclidean distance as the real-time power generation time period to which the current real-time power generation moment belongs; S7: extracting the corresponding matching photovoltaic power generation equipment fault diagnosis model according to the real-time power generation time period, and using the matching photovoltaic power generation equipment fault diagnosis model to perform fault diagnosis on the real-time comprehensive monitoring data to obtain a real-time fault diagnosis result; S8: If the real-time fault diagnosis result is that there is a fault, a corresponding real-time alarm signal is generated, the real-time alarm signal is visualized, and the next step is entered; otherwise, the real-time comprehensive monitoring data of the next real-time power generation moment is continuously collected, and the cluster center matching step is returned; S9: Based on the real-time fault diagnosis result, using the photovoltaic power generation equipment operation optimization model, generating an operation optimization strategy, obtaining a real-time operation optimization strategy, and returning the real-time operation optimization strategy to the equipment control device of the corresponding photovoltaic power generation equipment, including the following steps: S9-1: obtaining the Euclidean distance between the real-time fault diagnosis result and several historical fault diagnosis results in the experience playback pool of the photovoltaic power generation equipment operation optimization model, and taking the historical fault diagnosis result whose Euclidean distance exceeds the Euclidean distance threshold as the matching historical fault diagnosis result of the real-time fault diagnosis result; S9-2: extracting the matching historical operation optimization strategy generation experience corresponding to several matching historical fault diagnosis results, and updating the action space of the photovoltaic power generation equipment operation optimization model based on the matching historical operation optimization strategy generation experience, and obtaining the updated action space ,in, For the updated Action value, is the action indicator, is the total number of action space dimensions; S9-3: According to the real-time fault diagnosis results including several real-time photovoltaic power generation equipment fault conditions, set several real-time photovoltaic power generation equipment fault states, integrate all real-time photovoltaic power generation equipment fault states, update the state space of the photovoltaic power generation equipment operation optimization model, and obtain the updated state space ,in, For the updated Status value, is the state indicator, is the total number of state space dimensions; S9-4: inputting the updated state space into the deep Q network of the photovoltaic power generation equipment operation optimization model, and using the deep Q network to generate a Q value for each possible action in the updated action space based on the updated state space; S9-5: Using the agent, according to the reward function, obtain the reward value of each possible action in the updated action space, and according to the reward value, update the Q value of the possible action to obtain the updated Q value of each possible action; The formula is:

[0035] In the formula, The updated status value and the updated action value The corresponding updated Q value; Status value and action value The corresponding predicted Q value; is the learning rate; is the highest predicted Q value; It is a comprehensive indicator; To update the parameters; S9-6: Repeat the above Q value update steps until the number of iterations reaches the iteration number threshold, and use the greedy strategy to select the possible action with the highest updated Q value in each iteration as the execution action; S9-7: using an intelligent agent, integrating a number of execution actions to obtain a real-time operation optimization strategy, and returning the real-time operation optimization strategy to a device control device of a corresponding photovoltaic power generation device; Real-time operation optimization strategies include, but are not limited to, maximum power point tracking adjustment of photovoltaic power generation equipment, inverter control parameter adjustment (adjusting the switching frequency, output voltage and current settings of the inverter to adapt to grid requirements and reduce power loss), environmental condition response (adjusting the angle of the photovoltaic array, the position of the sunshade, etc. to optimize power generation efficiency), energy storage system management (controlling the charging and discharging strategy of the energy storage system to smooth power output and improve the system's adaptability to the grid), grid interaction adjustment (adjusting the power generation of the photovoltaic power generation system to support the stable operation of the grid), fault mitigation measures (when a potential fault is diagnosed, taking measures to mitigate the impact of the fault, such as limiting power output, starting backup equipment, etc.), system status adjustment (adjusting the system's operating mode according to the equipment's operating status, such as switching from normal operation to maintenance mode or fault mode), remote communication and control (adjusting the communication method with the remote monitoring center to improve the efficiency and reliability of data transmission).

[0036] Embodiment 2: like Figure 2 As shown, this embodiment provides a photovoltaic power generation equipment fault diagnosis system based on artificial intelligence, which is used to implement a photovoltaic power generation equipment fault diagnosis method. The system includes a cloud data center, a plurality of equipment monitoring devices, a plurality of equipment control devices, and a plurality of photovoltaic power generation equipment. The cloud data center is respectively connected to the plurality of equipment monitoring devices and the plurality of equipment control devices in communication. Each equipment monitoring device and each equipment control device are arranged at a photovoltaic power generation equipment in a one-to-one correspondence, and the equipment monitoring device and the equipment control device are both connected to the corresponding photovoltaic power generation equipment in communication. Equipment monitoring device, used to collect real-time comprehensive monitoring data of the current real-time power generation moment of photovoltaic power generation equipment, and upload the real-time comprehensive monitoring data to the cloud data center; An equipment control device, used to receive the real-time operation optimization strategy sent by the cloud data center, and control the corresponding photovoltaic power generation equipment according to the real-time operation optimization strategy; The cloud data center is provided with a data preprocessing unit, a clustering processing unit, a fault diagnosis model building unit, an operation optimization model building unit, a data partitioning unit, a fault diagnosis unit, an alarm signal generating unit and an operation optimization unit which are connected in sequence; A data preprocessing unit is used to collect historical comprehensive monitoring data of different historical power generation moments of a number of photovoltaic power generation equipment, and preprocess the historical comprehensive monitoring data to obtain a number of preprocessed historical comprehensive monitoring data; A cluster processing unit is used to cluster a number of pre-processed historical comprehensive monitoring data at different historical power generation moments based on power generation difference as the clustering purpose, and obtain a number of historical cluster centers and corresponding historical data sets at different historical power generation time periods; A fault diagnosis model building unit is used to build a photovoltaic power generation equipment fault diagnosis model corresponding to the historical power generation time period based on the historical data set of each historical power generation time period using an artificial intelligence algorithm, and generate a number of historical fault diagnosis results for each historical power generation time period; An operation optimization model building unit is used to build an operation optimization model of photovoltaic power generation equipment using a reinforcement learning algorithm based on a number of historical fault diagnosis results in all historical power generation time periods; A data partitioning unit is used to obtain the Euclidean distance between the real-time comprehensive monitoring data and several historical cluster centers, and use the historical power generation time period corresponding to the historical cluster center with the closest Euclidean distance as the real-time power generation time period to which the current real-time power generation moment belongs; A fault diagnosis unit is used to extract a corresponding matching photovoltaic power generation equipment fault diagnosis model according to the real-time power generation time period, and use the matching photovoltaic power generation equipment fault diagnosis model to perform fault diagnosis on the real-time comprehensive monitoring data to obtain a real-time fault diagnosis result; An alarm signal generating unit, used to generate a corresponding real-time alarm signal when the real-time fault diagnosis result indicates that a fault exists, and to visually display the real-time alarm signal; The operation optimization unit is used to generate an operation optimization strategy based on the real-time fault diagnosis result and the operation optimization model of the photovoltaic power generation equipment, obtain the real-time operation optimization strategy, and return the real-time operation optimization strategy to the equipment control device of the corresponding photovoltaic power generation equipment.

[0037] The invention discloses a photovoltaic power generation equipment fault diagnosis method and system based on artificial intelligence. The method adopts an artificial intelligence algorithm to construct a photovoltaic power generation equipment fault diagnosis model, performs automated and intelligent photovoltaic power generation equipment fault diagnosis, avoids manual inspection and analysis, improves fault diagnosis efficiency, eliminates the influence of subjective factors, and improves fault diagnosis accuracy and consistency. Real-time data acquisition and cloud data processing are performed. After uploading real-time comprehensive monitoring data to a cloud data center, the matching fault diagnosis model is extracted by matching historical clustering centers, and fault diagnosis is performed on real-time data, thereby achieving rapid response to real-time data, timely discovering and processing faults, and improving the operation efficiency and reliability of photovoltaic power generation equipment. The method comprehensively considers the potential influence of signal data, equipment operation data and environmental monitoring data on photovoltaic power generation equipment, and considers that the fault manifestation characteristics in different power generation time periods are different, establishes photovoltaic power generation equipment fault diagnosis models in different power generation time periods, and improves the comprehensiveness and accuracy of model analysis. A fault diagnosis result feedback mechanism is established, and the operation state of photovoltaic power generation equipment is adaptively adjusted and optimized through a photovoltaic power generation equipment operation optimization model of a forced learning algorithm. When a potential fault is first discovered, the photovoltaic power generation equipment is controlled in real time to ensure stable and reliable operation of the photovoltaic power generation equipment.

[0038] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A photovoltaic power generation equipment fault diagnosis method based on artificial intelligence, characterized in that: The steps include: Based on the cloud data center, the historical comprehensive monitoring data of different historical power generation moments of a number of photovoltaic power generation equipment are collected, and the historical comprehensive monitoring data are preprocessed to obtain a number of preprocessed historical comprehensive monitoring data; Taking the power generation difference as the clustering purpose, several pre-processed historical comprehensive monitoring data at different historical power generation moments are clustered to obtain several historical cluster centers and corresponding historical data sets at different historical power generation time periods; Based on the historical data set of each historical power generation time period, an artificial intelligence algorithm is used to construct a photovoltaic power generation equipment fault diagnosis model corresponding to the historical power generation time period, and generate several historical fault diagnosis results for each historical power generation time period; Based on several historical fault diagnosis results of all historical power generation time periods, a reinforcement learning algorithm is used to build an optimization model for the operation of photovoltaic power generation equipment; Based on the equipment monitoring device, the real-time comprehensive monitoring data of the current real-time power generation time of the photovoltaic power generation equipment is collected, and the real-time comprehensive monitoring data is uploaded to the cloud data center; Based on the cloud data center, the Euclidean distance between the real-time comprehensive monitoring data and several historical cluster centers is obtained, and the historical power generation time period corresponding to the historical cluster center with the closest Euclidean distance is used as the real-time power generation time period to which the current real-time power generation moment belongs; According to the real-time power generation time period, the corresponding matching photovoltaic power generation equipment fault diagnosis model is extracted, and the matching photovoltaic power generation equipment fault diagnosis model is used to perform fault diagnosis on the real-time comprehensive monitoring data to obtain real-time fault diagnosis results; If the real-time fault diagnosis result is that there is a fault, a corresponding real-time alarm signal is generated, the real-time alarm signal is visualized, and the next step is entered; otherwise, the real-time comprehensive monitoring data of the next real-time power generation moment is continuously collected, and the cluster center matching step is returned; According to the real-time fault diagnosis results, the operation optimization model of the photovoltaic power generation equipment is used to generate an operation optimization strategy, obtain the real-time operation optimization strategy, and return the real-time operation optimization strategy to the equipment control device of the corresponding photovoltaic power generation equipment.

2. The photovoltaic power generation equipment fault diagnosis method based on artificial intelligence according to claim 1, characterized in that: The historical / real-time comprehensive monitoring data includes historical / real-time control signal data, historical / real-time equipment operation data and historical / real-time environmental monitoring data of the photovoltaic power generation equipment collected at the same historical / real-time power generation moment.

3. The photovoltaic power generation equipment fault diagnosis method based on artificial intelligence according to claim 1, characterized in that: The preprocessing includes format conversion, denoising, data cleaning and normalization processing performed on a number of historical comprehensive monitoring data in sequence.

4. The photovoltaic power generation equipment fault diagnosis method based on artificial intelligence according to claim 1, characterized in that: Taking power generation difference as the clustering purpose, the FCM clustering algorithm is used to cluster several pre-processed historical comprehensive monitoring data at different historical power generation moments, and several historical cluster centers and corresponding historical data sets of different historical power generation time periods are obtained.

5. The photovoltaic power generation equipment fault diagnosis method based on artificial intelligence according to claim 1, characterized in that: According to the historical data set of each historical power generation time period, an artificial intelligence algorithm is used to construct a photovoltaic power generation equipment fault diagnosis model corresponding to the historical power generation time period, and generate several historical fault diagnosis results for each historical power generation time period, including the following steps: Use deep learning algorithms to build an initial photovoltaic power generation equipment fault diagnosis model, and use the initial network parameters of the photovoltaic power generation equipment fault diagnosis model as the optimization target. Use swarm intelligence optimization algorithms to search and optimize, and obtain the optimal initial network parameters of the initial photovoltaic power generation equipment fault diagnosis model. Labels are added to the historical data set of the current historical power generation period to obtain a photovoltaic power generation equipment fault diagnosis sample set with real labels, and the photovoltaic power generation equipment fault diagnosis sample set is divided into a model training set and a model test set in a ratio of 7:3; Based on the optimal initial network parameters, the initial photovoltaic power generation equipment fault diagnosis model is optimized to obtain an optimized photovoltaic power generation equipment fault diagnosis model, and the optimized photovoltaic power generation equipment fault diagnosis model is trained and optimized according to the model training set to obtain a trained photovoltaic power generation equipment fault diagnosis model; According to the model test set, the trained photovoltaic power generation equipment fault diagnosis model is tested to obtain several historical prediction labels, which are compared and counted with several real labels of the model test set to obtain the model test accuracy; If the model test accuracy is greater than the accuracy threshold, the optimal photovoltaic power generation equipment fault diagnosis model for the current historical power generation period is output, and a number of corresponding historical fault diagnosis results are generated; otherwise, the training optimization of the photovoltaic power generation equipment fault diagnosis model is continued; The historical data sets of all historical power generation time periods are traversed to build a fault diagnosis model for photovoltaic power generation equipment in each historical power generation time period.

6. The photovoltaic power generation equipment fault diagnosis method based on artificial intelligence according to claim 5 is characterized in that: The deep learning algorithm is the LSTM-Attention-cGAN algorithm, and the photovoltaic power generation equipment fault diagnosis model includes an input layer, a feature extraction module based on the LSTM algorithm, an attention weight module based on the Attention mechanism, an adversarial training module based on the cGAN algorithm, a classification layer, and an output layer; The swarm intelligence optimization algorithm is the ICPO algorithm.

7. The photovoltaic power generation equipment fault diagnosis method based on artificial intelligence according to claim 1, characterized in that: Based on several historical fault diagnosis results of all historical power generation time periods, a photovoltaic power generation equipment operation optimization model is constructed using a reinforcement learning algorithm, including the following steps: The photovoltaic power generation equipment operation optimization problem is used as the simulation environment of the reinforcement learning algorithm to build an intelligent agent and experience replay pool; According to several historical photovoltaic power generation equipment failure conditions included in several historical fault diagnosis results, several historical photovoltaic power generation equipment failure states are set, and all historical photovoltaic power generation equipment failure states are integrated to define the state space of the reinforcement learning algorithm; Based on several photovoltaic power generation equipment control signals included in the preset operation optimization strategy of the photovoltaic power generation equipment, the action space of the reinforcement learning algorithm is defined, and the reward function is set according to the interaction between the intelligent agent and the simulation environment; According to the spatial structure of the state space and the spatial structure of the action space, a neural network is constructed, and according to the intelligent agent of the reinforcement learning algorithm, the experience replay pool and the neural network, the model architecture of the photovoltaic power generation equipment operation optimization model is constructed; Set the state space, action space and reward function for the model architecture to obtain the initial photovoltaic power generation equipment operation optimization model; According to several historical fault diagnosis results of all historical power generation time periods, the initial photovoltaic power generation equipment operation optimization model is optimized and trained to obtain the final photovoltaic power generation equipment operation optimization model, and generate several historical operation optimization strategy generation experiences; Several historical operation optimization strategy generation experiences and corresponding historical fault diagnosis results are stored in the experience playback pool of the final photovoltaic power generation equipment operation optimization model.

8. The photovoltaic power generation equipment fault diagnosis method based on artificial intelligence according to claim 7 is characterized in that: The reinforcement learning algorithm is the DQN algorithm, and the neural network is a deep Q network.

9. The photovoltaic power generation equipment fault diagnosis method based on artificial intelligence according to claim 8, characterized in that: According to the real-time fault diagnosis result, the operation optimization model of the photovoltaic power generation equipment is used to generate the operation optimization strategy, obtain the real-time operation optimization strategy, and return the real-time operation optimization strategy to the equipment control device of the corresponding photovoltaic power generation equipment, including the following steps: Obtain the Euclidean distance between the real-time fault diagnosis result and several historical fault diagnosis results in the experience playback pool of the photovoltaic power generation equipment operation optimization model, and use the historical fault diagnosis result whose Euclidean distance exceeds the Euclidean distance threshold as the matching historical fault diagnosis result of the real-time fault diagnosis result; Extracting a number of matching historical operation optimization strategy generation experiences corresponding to the matching historical fault diagnosis results, and updating the action space of the photovoltaic power generation equipment operation optimization model according to the number of matching historical operation optimization strategy generation experiences to obtain an updated action space; According to several real-time photovoltaic power generation equipment fault conditions included in the real-time fault diagnosis results, several real-time photovoltaic power generation equipment fault states are set, and all real-time photovoltaic power generation equipment fault states are integrated to update the state space of the photovoltaic power generation equipment operation optimization model to obtain an updated state space; Input the updated state space into the deep Q network of the photovoltaic power generation equipment operation optimization model, and use the deep Q network to generate the Q value of each possible action in the updated action space based on the updated state space; Using the agent, according to the reward function, obtain the reward value of each possible action in the updated action space, and according to the reward value, update the Q value of the possible action to obtain the updated Q value of each possible action; Repeat the above Q-value update steps until the number of iterations reaches the iteration number threshold, and use the greedy strategy to select the possible action with the highest updated Q-value in each iteration as the execution action; Using an intelligent agent, several execution actions are integrated to obtain a real-time operation optimization strategy, and the real-time operation optimization strategy is returned to the equipment control device of the corresponding photovoltaic power generation equipment.

10. A photovoltaic power generation equipment fault diagnosis system based on artificial intelligence, used to implement the photovoltaic power generation equipment fault diagnosis method according to any one of claims 1 to 9, characterized in that: The system includes a cloud data center, a plurality of equipment monitoring devices, a plurality of equipment control devices, and a plurality of photovoltaic power generation equipment. The cloud data center is respectively connected to the plurality of equipment monitoring devices and the plurality of equipment control devices in communication. Each of the equipment monitoring devices and each of the equipment control devices are arranged at a photovoltaic power generation equipment in a one-to-one correspondence, and the equipment monitoring devices and the equipment control devices are both connected to the corresponding photovoltaic power generation equipment in communication. The cloud data center is provided with a data preprocessing unit, a clustering processing unit, a fault diagnosis model building unit, an operation optimization model building unit, a data partitioning unit, a fault diagnosis unit, an alarm signal generating unit and an operation optimization unit which are connected in sequence.

Images