An intelligent control method based on distributed photovoltaic online diagnosis

Abstract

The present invention discloses an intelligent management and control method based on distributed photovoltaic online diagnosis, which specifically includes the following steps: first, using a collection device to collect the operation data of a photovoltaic power station in real time, and transmitting the collected data to a data control terminal, and completing the processing operation of the collected data, and building a distributed photovoltaic operation fault diagnosis decision forest based on the collected data. The present invention relates to the field of distributed photovoltaic online diagnosis technology. The intelligent management and control method based on distributed photovoltaic online diagnosis combines a recurrent neural network and a deep neural network for the intelligent management and control method of distributed photovoltaic online diagnosis, which can fully utilize the advantages of both. The recurrent neural network is responsible for processing time series data and capturing long-term memory, while the deep neural network is responsible for extracting deep features and processing complex nonlinear relationships. This combination makes the fault diagnosis model more accurate, efficient and reliable.

Description

Technical Field

[0001] The present invention relates to the technical field of distributed photovoltaic online diagnosis, and in particular to an intelligent management and control method based on distributed photovoltaic online diagnosis. Background Art

[0002] A distributed photovoltaic power station specifically refers to a distributed power generation system that uses photovoltaic modules to directly convert solar energy into electrical energy. It is generally connected to a power grid with a voltage level of less than 35 kV or lower and is connected to the power system near the user.

[0003] According to the patent document, the name is: A method for diagnosing operating faults of distributed photovoltaic power stations (patent publication number: CN116317937A, patent publication date: 2023-06-23), including the following steps: S1, collecting historical meteorological data of the distributed photovoltaic power station grid connection point and the operating characteristic parameter values ​​of the distributed photovoltaic power station, as well as the operating characteristic parameters and corresponding fault type information under the faulty operating state, and preprocessing the collected data; S2, building a distributed photovoltaic operation fault diagnosis decision tree, and completing the training of the decision tree based on historical meteorological data and photovoltaic power station operation data; S3, introducing adaptive reinforcement learning technology, constructing a distributed photovoltaic operation fault diagnosis AdaBoost-DT model, and training the AdaBoost-DT model based on historical meteorological data and photovoltaic power station operation data; S4, by inputting meteorological data and photovoltaic power station operation characteristic parameters of any period of time into the trained AdaBoost-DT model, it is determined whether the power station has a fault and outputs the corresponding fault type. The scheme realizes the rapid and accurate diagnosis of distributed photovoltaic power station operation faults.

[0004] Based on the description in the above documents, the existing online diagnosis methods of distributed photovoltaic systems mostly rely on decision tree technology. However, this method generally has the problem of poor diagnostic effect and is unable to achieve intelligent management and control. Therefore, the present invention provides an intelligent management and control method based on distributed photovoltaic online diagnosis. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention provides an intelligent management and control method based on distributed photovoltaic online diagnosis, which solves the problem that the existing online diagnosis methods of distributed photovoltaic systems mostly rely on decision tree technology. However, this method generally has the problem of poor diagnostic effect and is unable to achieve intelligent management and control.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: an intelligent management and control method based on distributed photovoltaic online diagnosis, specifically comprising the following steps:

[0007] A1. First, use the collection equipment to collect the operation data of the photovoltaic power station in real time;

[0008] A2, based on the collected data, the data is transmitted to the data control terminal, and the processing operation of the collected data is completed;

[0009] A3. Build a distributed photovoltaic operation fault diagnosis decision forest based on the collected data, and complete the training of the decision forest based on the collected data;

[0010] A4. Input the preprocessed real-time data into the trained recurrent neural network and deep neural network fault diagnosis model. The model determines whether the photovoltaic power station has a fault based on the input data and outputs the fault type and fault location.

[0011] A5. Based on the fault diagnosis results, the intelligent management and control system automatically triggers the corresponding early warning and processing mechanism. At the same time, the intelligent management and control system feeds back the fault diagnosis results and processing measures to the operation and maintenance personnel.

[0012] Preferably, the collection equipment includes a temperature sensor, a humidity sensor, a light intensity sensor, a voltage sensor, and a current sensor, and the collected photovoltaic component temperature, humidity, light intensity, voltage, current data and the inverter output voltage, output current, and power data are transmitted to the data control terminal to facilitate data processing operations.

[0013] Preferably, the data processing operation of the control terminal in A2 on the collected data is:

[0014] B1: First, the collected raw data is preprocessed to fill missing values, delete duplicate data, and identify and process outliers;

[0015] B2: Convert the preprocessed data into numerical values, and store the required data based on the integration and screening of the numerical values.

[0016] Preferably, the decision forest in A3 is constructed as follows:

[0017] C1; collect relevant operating data from each monitoring point of the distributed photovoltaic power generation system, pre-process the collected data, and extract characteristic variables related to fault diagnosis from the original data;

[0018] C2: Generate a decision tree using the extracted data alone, and form a decision forest using multiple decision trees;

[0019] C3: Use the data set to train the decision forest so that each decision tree can perform classification or regression prediction on the input feature variables.

[0020] Preferably, the deep neural network in A4 is used to process high-dimensional data and nonlinear relationships, while the recurrent neural network is used to process time series data and can capture the time dependency in the data. The deep neural network is combined with the recurrent neural network for real-time fault diagnosis of photovoltaic power stations.

[0021] Preferably, the recurrent neural network fault diagnosis model in A4 is established in the following manner:

[0022] D1: Based on the collected real-time operation data, clean and normalize the data to make it meet the model input requirements;

[0023] D2: Design the structure of input layer, hidden layer and output layer. The input layer receives the real-time operation data of the photovoltaic system, the hidden layer captures the timing information, and the output layer outputs the fault diagnosis results.

[0024] D3: Divide the dataset into training set, validation set and test set for model training, validation and testing respectively to train the model.

[0025] Preferably, the deep neural network fault diagnosis model in A4 is established in the following manner:

[0026] E1: Based on the collected real-time operation data, clean and normalize the data to make it meet the model input requirements;

[0027] E2: Determine the number of network layers, the number of neurons in each layer, and the activation function;

[0028] E3: Divide the dataset into training set, validation set, and test set for model training, validation, and testing, respectively, to train the model.

[0029] Preferably, the intelligent management and control steps in A5 are:

[0030] F1: Based on the diagnosis results, the intelligent management and control system accurately locates the location and cause of the fault;

[0031] F2: After that, the intelligent management and control system automatically triggers the corresponding early warning and processing mechanism;

[0032] F3: When the performance of the PV modules is detected to be degraded, the system can automatically adjust the working status of the PV modules or issue an early warning to replace the modules; when an inverter failure is detected, the system can automatically switch to a backup inverter or issue a maintenance early warning.

[0033] Beneficial Effects

[0034] The present invention provides an intelligent management and control method based on distributed photovoltaic online diagnosis. Compared with the prior art, it has the following beneficial effects:

[0035] 1. The intelligent management and control method based on distributed photovoltaic online diagnosis combines recurrent neural networks and deep neural networks for distributed photovoltaic online diagnosis. It can make full use of the advantages of both. The recurrent neural network is responsible for processing time series data and capturing long-term memory, while the deep neural network is responsible for extracting deep features and processing complex nonlinear relationships. This combination makes the fault diagnosis model more accurate, efficient and reliable.

[0036] 2. The intelligent management and control method based on distributed photovoltaic online diagnosis can significantly improve the accuracy of fault diagnosis by setting up a decision forest. The decision forest builds multiple decision trees and integrates their output results. Each decision tree can learn different features from the data. Through integrated learning, the prediction results of multiple decision trees can be combined to obtain more accurate fault diagnosis. The decision forest can adaptively handle these complex situations without frequent adjustment of model parameters, thereby enhancing the robustness and stability of the system.

[0037] 3. The intelligent management and control method based on distributed photovoltaic online diagnosis can monitor the operating status of the distributed photovoltaic system in real time through the intelligent management and control system. Once an abnormality or fault is found, it can respond quickly and issue an alarm. This real-time monitoring and rapid response capability helps to discover and solve problems in a timely manner and reduce the impact of faults on the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 is a flow chart of the present invention;

[0039] Figure 2 is a data processing flow chart of the present invention;

[0040] Figure 3 A flowchart for constructing a decision forest of the present invention;

[0041] Figure 4 Establishing a flow chart for the recurrent neural network fault diagnosis model of the present invention;

[0042] Figure 5 It is a flow chart of the deep neural network fault diagnosis model of the present invention;

[0043] Figure 6 This is a flow chart of the intelligent management and control of the present invention. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] See also Figure 1-Figure 6 , the present invention provides a technical solution:

[0046] An intelligent management and control method based on distributed photovoltaic online diagnosis specifically includes the following steps:

[0047] A1. First, use the collection equipment to collect the operation data of the photovoltaic power station in real time;

[0048] A2, based on the collected data, the data is transmitted to the data control terminal, and the processing operation of the collected data is completed;

[0049] A3. Build a distributed photovoltaic operation fault diagnosis decision forest based on the collected data, and complete the training of the decision forest based on the collected data;

[0050] A4. Input the preprocessed real-time data into the trained recurrent neural network and deep neural network fault diagnosis model. The model determines whether the photovoltaic power station has a fault based on the input data and outputs the fault type and fault location.

[0051] A5. Based on the fault diagnosis results, the intelligent management and control system automatically triggers the corresponding early warning and processing mechanism. At the same time, the intelligent management and control system feeds back the fault diagnosis results and processing measures to the operation and maintenance personnel.

[0052] In this embodiment, the collection equipment includes a temperature sensor, a humidity sensor, a light intensity sensor, a voltage sensor, and a current sensor, and the collected photovoltaic component temperature, humidity, light intensity, voltage, current data and the inverter output voltage, output current, and power data are transmitted to the data control terminal to facilitate data processing operations.

[0053] In this embodiment, the data processing operation of the control terminal in A2 for the collected data is:

[0054] B1: First, the collected raw data is preprocessed to fill missing values, delete duplicate data, and identify and process outliers;

[0055] B2: Convert the preprocessed data into numerical values, and store the required data based on the integration and screening of the numerical values.

[0056] In this embodiment, the decision forest in A3 is constructed as follows:

[0057] C1; collect relevant operating data from each monitoring point of the distributed photovoltaic power generation system, pre-process the collected data, and extract characteristic variables related to fault diagnosis from the original data;

[0058] C2: Generate a decision tree using the extracted data alone, and form a decision forest using multiple decision trees;

[0059] C3: Use the data set to train the decision forest so that each decision tree can perform classification or regression prediction on the input feature variables.

[0060] In a decision forest, each decision tree is independently generated based on the training data set. During the generation process, a bootstrap sampling method is used to randomly extract samples from the original training data set to construct a training subset for each tree. At the same time, when a node is split, a portion of the features are randomly selected for consideration to increase the diversity of the model.

[0061] Decision forests can significantly improve the accuracy of fault diagnosis by constructing multiple decision trees and integrating their output results. Each decision tree can learn different features from the data. Through ensemble learning, the prediction results of multiple decision trees can be combined to obtain more accurate fault diagnosis. Decision forests can adaptively handle these complex situations without frequent adjustments to model parameters, thereby enhancing the robustness and stability of the system.

[0062] In this embodiment, the deep neural network in A4 is used to process high-dimensional data and nonlinear relationships, while the recurrent neural network is used to process time series data and can capture the time dependency in the data. The deep neural network is combined with the recurrent neural network for real-time fault diagnosis of photovoltaic power stations.

[0063] In this embodiment, the recurrent neural network fault diagnosis model in A4 is established in the following manner:

[0064] D1: Based on the collected real-time operation data, clean and normalize the data to make it meet the model input requirements;

[0065] D2: Design the structure of input layer, hidden layer and output layer. The input layer receives the real-time operation data of the photovoltaic system, the hidden layer captures the timing information, and the output layer outputs the fault diagnosis results.

[0066] D3: Divide the dataset into training set, validation set and test set for model training, validation and testing respectively to train the model.

[0067] In this embodiment, the deep neural network fault diagnosis model in A4 is established as follows:

[0068] E1: Based on the collected real-time operation data, clean and normalize the data to make it meet the model input requirements;

[0069] E2: Determine the number of network layers, the number of neurons in each layer, and the activation function;

[0070] E3: Divide the dataset into training set, validation set, and test set for model training, validation, and testing, respectively, to train the model.

[0071] The intelligent management and control method that combines recurrent neural networks and deep neural networks for distributed photovoltaic online diagnosis can fully utilize the advantages of both. The recurrent neural network is responsible for processing time series data and capturing long-term memory, while the deep neural network is responsible for extracting deep features and processing complex nonlinear relationships. This combination makes the fault diagnosis model more accurate, efficient and reliable.

[0072] In this embodiment, the intelligent management and control steps in A5 are:

[0073] F1: Based on the diagnosis results, the intelligent management and control system accurately locates the location and cause of the fault;

[0074] F2: After that, the intelligent management and control system automatically triggers the corresponding early warning and processing mechanism;

[0075] F3: When the performance of the PV modules is detected to be degraded, the system can automatically adjust the working status of the PV modules or issue an early warning to replace the modules; when an inverter failure is detected, the system can automatically switch to a backup inverter or issue a maintenance early warning.

[0076] The intelligent management and control system can monitor the operating status of the distributed photovoltaic system in real time. Once an abnormality or fault is found, it can respond quickly and issue an alarm. This real-time monitoring and rapid response capability helps to detect and solve problems in a timely manner and reduce the impact of faults on the system.

[0077] In summary, firstly, the operation data of the photovoltaic power station is collected in real time by using the collection equipment, and the collected data is transmitted to the data control terminal, and the processing operation of the collected data is completed. Based on the collected data, a distributed photovoltaic operation fault diagnosis decision forest is built, and the training of the decision forest is completed according to the collected data. The pre-processed real-time data is input into the trained recurrent neural network and deep neural network fault diagnosis model. The model determines whether the photovoltaic power station has a fault according to the input data, and outputs the fault type and fault location. According to the fault diagnosis results, the intelligent management and control system automatically triggers the corresponding early warning and processing mechanism. At the same time, the intelligent management and control system feeds back the fault diagnosis results and processing measures to the operation and maintenance personnel.

[0078] Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0079] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0080] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent management and control method based on distributed photovoltaic online diagnosis, characterized by: The specific steps include: A1. First, use the collection equipment to collect the operation data of the photovoltaic power station in real time; A2, based on the collected data, the data is transmitted to the data control terminal, and the processing operation of the collected data is completed; A3. Build a distributed photovoltaic operation fault diagnosis decision forest based on the collected data, and complete the training of the decision forest based on the collected data; A4. Input the preprocessed real-time data into the trained recurrent neural network and deep neural network fault diagnosis model. The model determines whether the photovoltaic power station has a fault based on the input data and outputs the fault type and fault location. A5. Based on the fault diagnosis results, the intelligent management and control system automatically triggers the corresponding early warning and processing mechanism. At the same time, the intelligent management and control system feeds back the fault diagnosis results and processing measures to the operation and maintenance personnel.

2. According to claim 1, an intelligent management and control method based on distributed photovoltaic online diagnosis is characterized in that: The collection equipment includes temperature sensors, humidity sensors, light intensity sensors, voltage sensors, and current sensors, and transmits the collected photovoltaic module temperature, humidity, light intensity, voltage, current data and the inverter output voltage, output current, and power data to the data control terminal for data processing operations.

3. According to claim 1, an intelligent management and control method based on distributed photovoltaic online diagnosis is characterized in that: The data processing operation of the control terminal in A2 for the collected data is: B1: First, the collected raw data is preprocessed to fill missing values, delete duplicate data, and identify and process outliers; B2: Convert the preprocessed data into numerical values, and store the required data based on the integration and screening of the numerical values.

4. According to claim 1, an intelligent management and control method based on distributed photovoltaic online diagnosis is characterized in that: The decision forest in A3 is constructed as follows: C1; collect relevant operating data from each monitoring point of the distributed photovoltaic power generation system, pre-process the collected data, and extract characteristic variables related to fault diagnosis from the original data; C2: Generate a decision tree using the extracted data alone, and form a decision forest using multiple decision trees; C3: Use the data set to train the decision forest so that each decision tree can perform classification or regression prediction on the input feature variables.

5. According to claim 1, an intelligent management and control method based on distributed photovoltaic online diagnosis is characterized in that: The deep neural network in A4 is used to process high-dimensional data and nonlinear relationships, while the recurrent neural network is used to process time series data and can capture the time dependency in the data. The deep neural network is combined with the recurrent neural network for real-time fault diagnosis of photovoltaic power plants.

6. The intelligent management and control method based on distributed photovoltaic online diagnosis according to claim 1 is characterized by: The method for establishing the recurrent neural network fault diagnosis model in A4 is as follows: D1: Based on the collected real-time operation data, clean and normalize the data to make it meet the model input requirements; D2: Design the structure of input layer, hidden layer and output layer. The input layer receives the real-time operation data of the photovoltaic system, the hidden layer captures the timing information, and the output layer outputs the fault diagnosis results. D3: Divide the dataset into training set, validation set and test set for model training, validation and testing respectively to train the model.

7. The intelligent management and control method based on distributed photovoltaic online diagnosis according to claim 1 is characterized by: The deep neural network fault diagnosis model in A4 is established as follows: E1: Based on the collected real-time operation data, clean and normalize the data to make it meet the model input requirements; E2: Determine the number of network layers, the number of neurons in each layer, and the activation function; E3: Divide the dataset into training set, validation set, and test set for model training, validation, and testing, respectively, to train the model.

8. The intelligent management and control method based on distributed photovoltaic online diagnosis according to claim 1 is characterized by: The intelligent control steps in A5 are: F1: Based on the diagnosis results, the intelligent management and control system accurately locates the location and cause of the fault; F2: After that, the intelligent management and control system automatically triggers the corresponding early warning and processing mechanism; F3: When the performance of the PV modules is detected to be degraded, the system can automatically adjust the working status of the PV modules or issue an early warning to replace the modules; when an inverter failure is detected, the system can automatically switch to a backup inverter or issue a maintenance early warning.

Images