Self-adaptive fault diagnosis method of distributed photovoltaic system and related device

Abstract

The invention discloses a self-adaptive fault diagnosis method of a distributed photovoltaic system and a related device, and belongs to the technical field of fault diagnosis, and the method comprises the following steps: collecting operation state data of the distributed photovoltaic system, and carrying out the preprocessing of the operation state data; constructing a multi-dimensional environment parameter mapping table to dynamically adjust the reference parameters of the equipment; inputting the preprocessed operation state data and the adjusted equipment reference parameters into an adaptive fault diagnosis model to obtain a fault diagnosis result; according to the adaptive fault diagnosis model, preprocessed operation state data and adjusted equipment reference parameters are input into a convolutional neural network, optimization is carried out through an adaptive learning rate adjustment algorithm, and a fault category is used as a fault diagnosis result to be output. Based on a fault diagnosis result, a fault judgment threshold and a feature extraction weight are optimized through a DQN algorithm and then training is carried out. According to the invention, the problem of low fault identification accuracy under different working conditions due to a fixed diagnosis threshold in the prior art can be solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of fault diagnosis, and particularly relates to an adaptive fault diagnosis method and related device for a distributed photovoltaic system. Background Art

[0002] With the increasing demand for energy transformation and sustainable development, distributed photovoltaic systems have been widely used in residential buildings, commercial buildings, industrial parks and other fields due to their advantages such as flexibility, high efficiency, and environmental protection. However, distributed photovoltaic systems are long-term in a complex outdoor environment and are affected by various environmental factors such as light intensity, temperature, and humidity. Various components such as photovoltaic modules, inverters, and electrical connection parts are prone to various faults. These faults will not only affect the power generation efficiency of the system, but may also cause safety hazards such as fires and electric shocks.

[0003] At present, although the fault diagnosis methods for distributed photovoltaic systems can achieve real-time monitoring, most of them use fixed fault feature extraction algorithms and diagnostic thresholds, which cannot adapt to the complex and changeable working conditions of distributed photovoltaic systems, resulting in low accuracy of fault identification. The environment where the distributed photovoltaic system is located is complex and changeable. For example, the light intensity, temperature, humidity, etc. in different regions vary greatly. Existing methods often do not fully consider the influence of environmental factors on the system performance, resulting in poor generality of the fault diagnosis model in different environments. Summary of the Invention

[0004] The purpose of the present invention is to provide an adaptive fault diagnosis method and related device for a distributed photovoltaic system to solve the problem of low accuracy of fault identification under different working conditions caused by fixed diagnostic thresholds in the prior art.

[0005] To achieve the above object, the present invention adopts the following technical solutions: In a first aspect, an adaptive fault diagnosis method for a distributed photovoltaic system includes the following steps: Collect the operation status data of the distributed photovoltaic system and perform preprocessing; Construct a multi-dimensional environmental parameter mapping table to dynamically adjust the device reference parameters; Input the preprocessed operation status data and the adjusted device reference parameters into an adaptive fault diagnosis model to obtain a fault diagnosis result and achieve adaptive fault diagnosis; The adaptive fault diagnosis model is obtained by inputting the preprocessed operation status data and the adjusted device reference parameters into a convolutional neural network, optimizing through an adaptive learning rate adjustment algorithm, outputting the fault category as the fault diagnosis result, and training after optimizing the fault judgment threshold and feature extraction weight through the DQN algorithm based on the fault diagnosis result.

[0006] In some embodiments, the operating state data includes: light intensity, temperature, current, and voltage. The multi-dimensional environmental parameter mapping table represents the correspondence between temperature, humidity, and device reference parameters.

[0007] In some embodiments, the convolutional layer of the convolutional neural network uses a 3×3 convolutional kernel to extract features from the light intensity, a 5×5 convolutional kernel to extract features from the temperature, and a 1D temporal convolutional layer to extract features from the current and voltage. Dropout is added after the convolutional layer of the convolutional neural network, and L2 regularization is used in the fully connected layer of the convolutional neural network.

[0008] In some embodiments, the adaptive learning rate adjustment algorithm is the Adam algorithm, its initial learning rate is set to 0.001, and a cosine annealing strategy is used for adjustment.

[0009] In some embodiments, a BN layer is added before the activation function of the convolutional neural network, and the Focal Loss is used as the loss function of the convolutional neural network.

[0010] In some embodiments, the step of collecting the operating state data of the distributed photovoltaic system and performing preprocessing specifically includes: Performing Kalman filtering on the light intensity, detecting outliers in the operating state data through the Isolation Forest algorithm and removing the outliers, and finally performing normalization processing on the operating state data.

[0011] In a second aspect, a distributed photovoltaic system adaptive fault diagnosis system includes: A data preprocessing module, configured to collect the operating state data of the distributed photovoltaic system and perform preprocessing; A device reference parameter dynamic adjustment module, configured to construct a multi-dimensional environmental parameter mapping table to dynamically adjust the device reference parameters; An adaptive fault diagnosis module, configured to input the preprocessed operating state data and the adjusted device reference parameters into an adaptive fault diagnosis model to obtain a fault diagnosis result and implement adaptive fault diagnosis; The adaptive fault diagnosis model inputs the preprocessed operating state data and the adjusted device reference parameters into a convolutional neural network, is optimized through an adaptive learning rate adjustment algorithm, outputs the fault category as the fault diagnosis result, and is trained based on the fault diagnosis result by optimizing the fault judgment threshold and feature extraction weights through the DQN algorithm.

[0012] In a third aspect, an electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the adaptive fault diagnosis method for the distributed photovoltaic system are implemented.

[0013] In a fourth aspect, a computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the steps of the adaptive fault diagnosis method for the distributed photovoltaic system are implemented.

[0014] In a fifth aspect, a computer program product includes a computer program, characterized in that when the computer program is executed by a processor, the steps of the adaptive fault diagnosis method for the distributed photovoltaic system are implemented.

[0015] Compared with the prior art, the present invention has the following beneficial effects: The present invention dynamically associates temperature, humidity with device reference parameters through a multi-dimensional parameter mapping table, solves the misjudgment problem caused by environmental fluctuations in the traditional fixed threshold method, automatically adjusts the current reference value in high-temperature and high-humidity environments, and avoids misjudging normal temperature rise as a fault. The DQN algorithm optimizes the fault judgment threshold and feature extraction weights, and can adjust the fault judgment threshold in real time through reinforcement learning, so that the system maintains stable diagnostic performance in different seasons and climate conditions and adapts to the performance decay during the long-term operation of the photovoltaic array. Finally, combined with a convolutional neural network for fault identification, the diagnostic threshold can be dynamically adjusted to improve the fault identification accuracy of the distributed photovoltaic system under different working conditions.

[0016] Further, the present invention uses a 3×3 convolutional kernel to extract features of light intensity, a 5×5 convolutional kernel to extract features of temperature, and a 1D temporal convolutional to extract features of current and voltage, and performs differential feature extraction for different physical quantities. The light intensity captures local mutations through a small convolutional kernel, the temperature field analyzes the hot spot diffusion pattern using a large convolutional kernel, and the current and voltage use temporal convolution to identify intermittent faults. In addition, the present invention adds Dropout after the convolutional layer of the convolutional neural network, and the fully connected layer of the convolutional neural network uses L2 regularization, which can enable the adaptive fault diagnosis model to maintain a high accuracy on complex environmental data and improve the generalization ability.

[0017] Further, the present invention adds a BN layer before the activation function of the convolutional neural network, and the loss function of the convolutional neural network uses Focal Loss, which can solve the problem of data imbalance and improve the small-sample fault identification ability.

[0018] Furthermore, the present invention optimizes the convolutional neural network through the Adam algorithm and adjusts it in combination with the cosine annealing strategy, enabling it to converge rapidly in the initial stage of the adaptive fault diagnosis model training and fine-tuning in the later stage, and can shorten more training time compared with the fixed learning rate scheme. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 FIG. [X] is a flowchart of an adaptive fault diagnosis method for a distributed photovoltaic system provided in this embodiment; Figure 2 FIG. [Y] is a structural diagram of an adaptive fault diagnosis system for a distributed photovoltaic system provided in this embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solution of the present invention will be further described in detail below in conjunction with the drawings. The content is an explanation of the present invention rather than a limitation.

[0021] It should be noted that the terms "comprising" and "having" and any variations thereof in the specification and claims of the present invention are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device including a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, systems, products or devices.

[0022] As Figure 1 shown, this embodiment provides an adaptive fault diagnosis method for a distributed photovoltaic system, including the following steps: S1. Data acquisition and preprocessing. Install light intensity sensors, temperature sensors, current sensors, and voltage sensors in the distributed photovoltaic system to collect the operation status data of the system in real time. The operation status data includes: light intensity, temperature, current, and voltage. At the same time, collect the fault record data of the system, including fault type, fault occurrence time, and fault repair time. Sensor network deployment; Among them, the light intensity sensor adopts a silicon-based photodiode array, which is installed on the surface of the photovoltaic module, with a sampling frequency of 50 Hz and a dynamic range of 0 - 2000 W / m²; the temperature sensor adopts a fiber Bragg grating (FBG) sensor, which is distributedly deployed at key nodes of the busbar box and inverter, with a temperature resolution of 0.1 °C; the current and voltage adopt Hall effect sensors to achieve non-contact measurement of current (0 - 500 A) and voltage (0 - 1000 V); Environmental sensor: Integrate temperature, humidity, and air pressure sensors, with a sampling period of 1 second and data accuracy of RH ± 2% and air pressure ± 0.3 hPa.

[0023] Adopt the IEEE1588 precise time protocol to achieve timestamp alignment of multi-sensor data, with a synchronization error < 1μs. Data fusion is performed through the edge gateway to generate a standardized data frame containing timestamps, device IDs, and parameter types.

[0024] Noise suppression: Apply Kalman filtering to the light intensity to effectively suppress the instantaneous fluctuations caused by cloud movement; Data cleaning: Detect outliers based on the Isolation Forest algorithm, set the outlier score threshold to 0.65, and process the outliers generated by sensor drift; Normalization processing: Use Z-score standardization to calculate the mean and standard deviation by grouping according to device ID, eliminating the dimension difference. The formula is as follows:

[0025] where is the original data, is the mean, is the standard deviation.

[0026] After the above processing, the normalized operation state data of the system is obtained.

[0027] S2. Construct and train an adaptive fault diagnosis model; S2.1 Construction of the adaptive fault diagnosis model: An adaptive fault diagnosis model is obtained by training a convolutional neural network. The convolutional neural network includes an input layer, a convolutional layer, a pooling layer, a fully connected layer, and an output layer; the convolutional layer extracts local features of the input data through convolutional kernels.

[0028] The input layer is the normalized system operation state data; In the convolutional layer, for the light intensity data, a 3×3 convolutional kernel is used for feature extraction; for the temperature data, a 5×5 convolutional kernel can be used for feature extraction, and for the voltage and current data, 1D temporal convolution is used for feature extraction.

[0029] The pooling layer uses the max pooling method for downsampling to reduce the data dimension and computational complexity. The pooling window size is adjusted according to the size of the input data.

[0030] The fully connected layer is used for the final classification decision. The number of neurons in the fully connected layer is set according to the number of fault types. Currently, there are 11 common fault types in the distributed photovoltaic system. Among them, there are 4 common faults at the photovoltaic module level, including cell aging, hot spot effect, hidden crack, and encapsulation material damage; there are 3 faults in the inverter, including power device failure, control circuit failure, and poor heat dissipation; there are 4 faults in the system electrical connection, including line aging, poor contact, short circuit, and open circuit; therefore, the number of neurons in the fully connected layer can be set to 11.

[0031] The output layer is the fault recognition result.

[0032] S2.2 Training of the adaptive fault diagnosis model. The normalized operating state data is divided into a training set (70% of the data), a validation set (15% of the data), and a test set (15%).

[0033] The Adam (Adaptive Moment Estimation) algorithm is a commonly used adaptive learning rate adjustment algorithm. It combines the advantages of the momentum method and the RMSProp algorithm and can dynamically adjust the learning rate according to the gradient changes during training. When training a CNN model, the Adam algorithm can be used as an optimizer to improve the convergence speed and generalization ability of the model. Therefore, the training set is used to train the CNN model, and the weights and biases of the model are continuously updated through the backpropagation algorithm and the Adam optimizer to minimize the error between the model output and the true labels.

[0034] The AdamW optimizer is adopted with an initial learning rate of 0.001, and the cosine annealing strategy is used for adjustment; Dropout (rate = 0.5) is added after the convolutional layer, and L2 regularization (λ = 0.0001) is adopted for the fully connected layer; a BN (BatchNormalization) layer is added before the activation function to accelerate the training convergence. The focal loss is used as the loss function of this model, with the balance factor set to 0.25 and the focus factor set to 2.

[0035] Construct a multi-dimensional environmental parameter mapping table to establish the corresponding relationship between temperature, humidity, and the device reference parameters. When the ambient temperature > 45°C, the maximum allowable temperature rise of the inverter is reduced by 20%; when the relative humidity > 85%, the component insulation resistance threshold is increased by 30%. During the data acquisition stage, according to the environmental parameter mapping table, the device reference parameters are dynamically adjusted, and the adjusted device reference parameters are used as new input features and input into the CNN together with the normalized operating state data; then the Deep Q-Network (DQN) algorithm is adopted, with the fault diagnosis accuracy rate as the reward function, to dynamically adjust: the fault judgment threshold (voltage deviation and temperature threshold), the feature extraction weight, and the model update frequency. Specifically, after the adaptive fault diagnosis model outputs the fault diagnosis result, the DQN algorithm adjusts the fault judgment threshold and the feature extraction weight according to the fault diagnosis result, and inputs the adjusted parameters back into the CNN for model optimization.

[0036] S3. Input the normalized operating state data in S1 and the adjusted device reference parameters in S2 into a Convolutional Neural Network (CNN), and then obtain the fault diagnosis result.

[0037] Experimental verification: 5-fold cross-validation is adopted to ensure the generalization ability of the model. The performance metrics for verifying the adaptive fault diagnosis model are accuracy: the proportion of correctly classified samples, recall rate: the proportion of actually faulty samples that are correctly identified, F1-score: the harmonic mean of precision and recall rate, and average diagnosis time: the delay from data acquisition to result output.

[0038] Replicate typical working conditions in a climate test chamber: sudden change in light intensity: 1000 W / m² → 200 W / m² (simulating cloud cover), temperature shock: -40°C → 85°C cyclic test, humidity loading: RH95% for 72 hours.

[0039] Test results of the adaptive fault diagnosis model: Fault identification accuracy: 98.2% (a 21.5% improvement compared to the accuracy of 76.7% of the traditional method), average diagnosis time: 42 ms (a 168 ms improvement compared to 210 ms of the traditional method), false alarm rate: 0.3%.

[0040] As Figure 2 shown, this embodiment provides an adaptive fault diagnosis system for a distributed photovoltaic system, including: A data preprocessing module for collecting and preprocessing the operation status data of the distributed photovoltaic system; A device reference parameter dynamic adjustment module for constructing a multi-dimensional environmental parameter mapping table to dynamically adjust the device reference parameters; An adaptive fault diagnosis module for inputting the preprocessed operation status data and the adjusted device reference parameters into an adaptive fault diagnosis model to obtain a fault diagnosis result and achieve adaptive fault diagnosis; The adaptive fault diagnosis model inputs the preprocessed operation status data and the adjusted device reference parameters into a convolutional neural network, is optimized through an adaptive learning rate adjustment algorithm, outputs the fault category as the fault diagnosis result, and is trained based on the fault diagnosis result by optimizing the fault judgment threshold and feature extraction weights through the DQN algorithm.

[0041] The division of modules in the embodiments of the present invention is illustrative, merely a logical function division. In actual implementation, there may be other division methods. In addition, in each embodiment of the present invention, the functional modules can be integrated in one processor, or can exist separately physically, or two or more modules can be integrated in one module. The above integrated modules can be implemented in the form of hardware or in the form of software functional modules.

[0042] In this embodiment, a computer device is further provided. The computer device includes a processor and a memory. The memory is used to store a computer program (in this embodiment, the computer program includes a calculation component and an iteration component, and can perform model calculation and model update). The computer program includes program instructions, and the processor is used to execute the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or may also be other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, and is suitable for implementing one or more instructions. Specifically, it is suitable for loading and executing one or more instructions in the computer storage medium to implement the corresponding method flow or corresponding function. The processor described in the embodiment of the present invention can be used for the operation of an adaptive fault diagnosis method for a distributed photovoltaic system.

[0043] This embodiment also provides a storage medium, specifically a computer-readable storage medium (Memory). The computer-readable storage medium is a memory device in a computer device and is used to store programs and data. It can be understood that the computer-readable storage medium here can include both the built-in storage medium in the computer device and, of course, the extended storage medium supported by the computer device. The computer-readable storage medium provides a storage space, and this storage space stores the operating system of the terminal. And, one or more instructions suitable for being loaded and executed by the processor are also stored in this storage space. These instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. One or more instructions stored in the computer-readable storage medium can be loaded and executed by the processor to implement the corresponding steps of an adaptive fault diagnosis method for a distributed photovoltaic system in the above embodiment.

[0044] This embodiment also provides a computer program product. The computer program product includes a computer program. When the computer program is executed by the processor, the corresponding steps of an adaptive fault diagnosis method for a distributed photovoltaic system in the above embodiment are implemented.

[0045] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0046] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, as well as the combination of flows and / or blocks in the flowchart and / or block diagram, can be realized by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in one Figure 1 flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0047] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing devices to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in one Figure 1 flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0048] These computer program instructions can also be loaded onto a computer or other programmable data processing devices, such that a series of operation steps are executed on the computer or other programmable devices to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable devices provide steps for implementing the functions specified in one Figure 1 flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0049] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that it is still possible to modify the specific implementation manners of the present invention or make equivalent substitutions. Any modification or equivalent substitution that does not depart from the spirit and scope of the present invention shall be covered by the protection scope of the claims of the present invention.

Claims

1. An adaptive fault diagnosis method for a distributed photovoltaic system, characterized in that It includes the following steps: Collect the operation status data of the distributed photovoltaic system and preprocess it; Construct a multi-dimensional environmental parameter mapping table to dynamically adjust the device reference parameters; Input the preprocessed operation status data and the adjusted device reference parameters into the adaptive fault diagnosis model to obtain the fault diagnosis result and realize adaptive fault diagnosis; The adaptive fault diagnosis model inputs the preprocessed operation status data and the adjusted device reference parameters into a convolutional neural network, optimizes it through an adaptive learning rate adjustment algorithm, outputs the fault category as the fault diagnosis result, and is trained based on the fault diagnosis result by optimizing the fault judgment threshold and feature extraction weight through the DQN algorithm.

2. The adaptive fault diagnosis method for a distributed photovoltaic system according to claim 1, characterized in that The operation status data includes: light intensity, temperature, current, and voltage. The multi-dimensional environmental parameter mapping table represents the corresponding relationship between temperature, humidity, and device reference parameters.

3. The adaptive fault diagnosis method for a distributed photovoltaic system according to claim 1, characterized in that The convolutional layer of the convolutional neural network uses a 3×3 convolutional kernel to extract features from the light intensity, a 5×5 convolutional kernel to extract features from the temperature, and a 1D temporal convolutional layer to extract features from the current and voltage; Dropout is added after the convolutional layer of the convolutional neural network, and L2 regularization is used in the fully connected layer of the convolutional neural network.

4. The adaptive fault diagnosis method for a distributed photovoltaic system according to claim 1, characterized in that The adaptive learning rate adjustment algorithm is the Adam algorithm, its initial learning rate is set to 0.001, and a cosine annealing strategy is used for adjustment.

5. The adaptive fault diagnosis method for a distributed photovoltaic system according to claim 1, wherein A BN layer is added before the activation function of the convolutional neural network, and the focal loss is used as the loss function of the convolutional neural network.

6. The adaptive fault diagnosis method for a distributed photovoltaic system according to claim 1, wherein The step of collecting the operation status data of the distributed photovoltaic system and preprocessing it specifically includes: Perform Kalman filtering on the light intensity, detect outliers in the operation status data through the isolation forest algorithm and remove the outliers, and finally normalize the operation status data.

7. A distributed photovoltaic system adaptive fault diagnosis system, characterized in that, It includes: A data preprocessing module for collecting the operation status data of the distributed photovoltaic system and preprocessing it; A device reference parameter dynamic adjustment module for constructing a multi-dimensional environmental parameter mapping table to dynamically adjust the device reference parameters; An adaptive fault diagnosis module for inputting the preprocessed operation status data and the adjusted device reference parameters into the adaptive fault diagnosis model to obtain the fault diagnosis result and realize adaptive fault diagnosis; The adaptive fault diagnosis model inputs the preprocessed operation status data and the adjusted device reference parameters into a convolutional neural network, optimizes it through an adaptive learning rate adjustment algorithm, outputs the fault category as the fault diagnosis result, and is trained based on the fault diagnosis result by optimizing the fault judgment threshold and feature extraction weight through the DQN algorithm.

8. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable in the processor. When the processor executes the computer program, it implements the steps of the adaptive fault diagnosis method for a distributed photovoltaic system according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the adaptive fault diagnosis method for a distributed photovoltaic system according to any one of claims 1 to 6 are implemented.

10. A computer program product, the computer program product comprising a computer program, characterized in that, When the computer program is executed by a processor, the steps of the adaptive fault diagnosis method for a distributed photovoltaic system according to any one of claims 1 to 6 are implemented.

Images