Photovoltaic module PID detection method and device and electronic equipment

Abstract

The invention provides a photovoltaic module PID detection method and device and electronic equipment, and relates to the field of data processing. In the method, detection data for a target photovoltaic module are obtained, and the detection data comprise current, voltage, power output, environment temperature, relative humidity and solar irradiance; according to the detection data, a target feature group is determined, and the target feature group comprises a time sequence feature, a relative performance feature and an environment feature; inputting the time sequence feature, the relative performance feature and the environment feature into a support vector machine model to obtain a detection result; and if it is determined that the detection result indicates abnormity, determining that the target photovoltaic module has the PID effect. By implementing the technical scheme provided by the invention, the PID detection accuracy of the photovoltaic module can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of data processing, and in particular to a photovoltaic module PID detection method, device and electronic equipment. Background Art

[0002] During long-term use, photovoltaic modules may be affected by a variety of factors, resulting in degradation of their performance. Among them, the Potential Induced Degradation (PID) effect is a common fault phenomenon that significantly reduces the power generation efficiency of photovoltaic modules, thereby affecting the economy and reliability of the entire photovoltaic power generation system.

[0003] At present, traditional PID detection technology usually relies on empirical formulas and threshold judgments based on experience, and the accuracy and reliability of these methods are limited. However, due to the complex and changeable working environment of photovoltaic modules and the many factors that affect their performance, a single empirical formula lacks adaptability to real-time data and may fail under new environmental conditions or operating conditions, resulting in low accuracy in PID detection of photovoltaic modules. Summary of the invention

[0004] The present application provides a photovoltaic module PID detection method, device and electronic equipment, which are convenient for improving the PID detection accuracy of photovoltaic modules.

[0005] In a first aspect of the present application, a photovoltaic component PID detection method is provided, the method comprising: acquiring detection data for a target photovoltaic component, the detection data comprising current, voltage, power output, ambient temperature, relative humidity and solar irradiance; determining a target feature group based on the detection data, the target feature group comprising time series features, relative performance features and environmental features; inputting the time series features, relative performance features and environmental features into a support vector machine model to obtain a detection result; if it is determined that the detection result indicates an abnormality, determining that the target photovoltaic component has a PID effect.

[0006] By adopting the above technical solution, the method provides rich information for subsequent analysis by acquiring multiple detection data, namely current, voltage, power output, ambient temperature, relative humidity, and solar irradiance. This comprehensive data collection can more accurately reflect the actual working status of photovoltaic modules, making subsequent analysis more reliable. By determining the target feature group, the method integrates time series features, relative performance features, and environmental features. This systematic feature extraction can fully capture the performance changes of photovoltaic modules under different environmental conditions, help improve the sensitivity and accuracy of detection, and avoid the one-sidedness common in traditional methods. The use of support vector machine models for detection gives full play to the advantages of machine learning in data pattern recognition and classification. SVM can effectively process high-dimensional data and maintain good performance in complex nonlinear situations, thereby improving the accuracy and robustness of PID detection. By directly judging the test results, if the results indicate an abnormality, it can be quickly determined that the photovoltaic module has a PID effect. This clear feedback mechanism provides clear operational guidance for operation and maintenance personnel, so that they can take timely measures to reduce potential losses. Therefore, the above method can improve the accuracy of PID detection of photovoltaic modules.

[0007] Optionally, the acquisition of detection data for the target photovoltaic component specifically includes: receiving raw current data of the target photovoltaic component sent by a current sensor; receiving raw voltage data of the target photovoltaic component sent by a voltage sensor; receiving raw power data of the target photovoltaic component sent by a power sensor; receiving raw ambient temperature data of the target photovoltaic component sent by an ambient temperature sensor; receiving raw relative humidity data of the target photovoltaic component sent by a relative humidity sensor; receiving raw solar radiation data of the target photovoltaic component sent by a solar radiation sensor; and preprocessing the raw current data, the raw voltage data, the raw power data, the raw ambient temperature data, the raw relative humidity data and the raw solar radiation data to obtain the detection data, wherein the preprocessing includes denoising, filtering and normalization.

[0008] By adopting the above technical solution, the method can fully capture the working status of photovoltaic modules by receiving raw data from multiple sensors. This distributed multi-dimensional data collection provides rich information for subsequent analysis, which helps to more accurately evaluate the performance of the module and identify potential problems. Directly receiving the raw data sent by various sensors ensures the real-time and accuracy of the data. The sensor can provide real-time measurement values, thereby improving the timeliness of the data, so that the detection results can better reflect the current working status of the photovoltaic module. Denoising can significantly improve the reliability of the data and reduce interference with subsequent analysis by removing random noise in the measurement process. Filtering technology can smooth data fluctuations and eliminate misleading information caused by instantaneous fluctuations, thereby enhancing the interpretability of the data. Normalizing the data can standardize data of different dimensions and avoid analysis errors caused by dimensional differences. This is very important for subsequent feature extraction and model training. By preprocessing, the consistency of data from different sources is ensured, so that the comparability between each feature is enhanced. This consistency is crucial for the training and feature analysis of machine learning models, and helps to improve the accuracy and stability of the model. Preprocessed data can usually significantly reduce the complexity in subsequent analysis. Reducing noise and outliers can help the model focus more on key features, thereby improving training efficiency and reducing the risk of overfitting. The preprocessing step can lay a good foundation for subsequent data analysis and modeling. Cleaned and standardized data can be absorbed by the model faster, reducing the time analysts spend on data cleaning, thereby improving overall work efficiency.

[0009] Optionally, determining the target feature group based on the detection data specifically includes: using a sliding window to extract the target time points in the current and voltage, and the statistical features corresponding to the target time points, the statistical features including mean, standard deviation, maximum value, minimum value and time slope; generating the time series features based on the mean, standard deviation, maximum value, minimum value and time slope; obtaining the benchmark power output of the target photovoltaic component; calculating the relative power based on the benchmark power output and the power output at the current moment; obtaining the power output corresponding to the target time point, the target time point and the current moment are adjacent time points; calculating the relative power change rate based on the power output corresponding to the target time point and the power output at the current moment, the relative performance characteristics including the relative power and the relative power change rate.

[0010] By adopting the above technical solution, the sliding window technology is used to extract the target time points of current and voltage and their corresponding statistical features, which can capture the dynamic changes of time series data, allow real-time analysis of data trends, and enhance the understanding of the status of photovoltaic modules. By extracting statistical features such as mean and standard deviation, the distribution characteristics and fluctuations of the data can be fully reflected, which helps to identify potential abnormal patterns. Obtaining the benchmark power output of the target photovoltaic module and comparing it with the current power output provides a reference for the calculation of relative power, which can more effectively evaluate the performance of the current power output and help identify changes in the performance of photovoltaic modules. This comparison method based on benchmark values ​​can reduce errors caused by environmental changes and make the evaluation results more objective and reliable. By calculating the relative power, a quantitative indicator can be provided, which can effectively track and compare the performance of photovoltaic modules at different time points, which helps to identify trends and potential problems. Calculating the relative rate of change of power can provide a deeper understanding of the dynamic performance changes of photovoltaic modules. The calculation of the relative rate of change of power can capture smaller performance fluctuations, which is crucial for identifying early PID effects and helps to take maintenance measures early. By analyzing the relative rate of change, the trend of power output over time can be monitored to provide decision support for operation and maintenance personnel.

[0011] Optionally, determining the target feature group based on the detection data specifically also includes: calculating the average value, maximum value and minimum value corresponding to the environmental data at each current moment based on the ambient temperature, the relative humidity and the solar irradiance; obtaining the environmental data corresponding to the target time point; calculating the environmental relative change rate based on the environmental data corresponding to the target time point and the environmental data at the current moment, the environmental characteristics including the average value, the maximum value, the minimum value and the environmental relative change rate.

[0012] By adopting the above technical solutions, by calculating the average, maximum and minimum values ​​of ambient temperature, relative humidity and solar irradiance, comprehensive analysis of these environmental data helps to better understand the performance of components under different conditions. Statistical analysis of environmental data can more clearly reflect the potential impact of environmental factors on the output performance of photovoltaic modules, and provide a basis for analyzing the causes of performance fluctuations. The introduction of real-time environmental data makes feature extraction more in line with the actual situation, avoids the use of outdated data, and enhances the accuracy of photovoltaic module performance evaluation. Based on the latest environmental data, operation and maintenance strategies can be adjusted or maintenance measures can be taken in time to cope with sudden environmental changes. The relative rate of change can capture small environmental changes, which is particularly important for early identification of performance threats that photovoltaic modules may face, especially in the detection of PID effects. By monitoring the relative rate of change of the environment, long-term trends or periodic fluctuations can be discovered, which helps to develop more effective prediction models and operation and maintenance plans. This comprehensive feature group provides rich information for subsequent data analysis and helps to improve the ability of machine learning models to identify photovoltaic module performance. Multi-dimensional features enable models to better understand complex data relationships and enhance their generalization and accuracy, especially when faced with multiple environmental influences. Environmental factors are often related to the PID effect. By monitoring changes in environmental characteristics, the occurrence of the PID effect can be more accurately identified and diagnosed, and the detection accuracy can be improved. Timely identification of the impact of the environment on component performance can help develop a more effective early warning mechanism, reduce the risk of failure, and ensure the stable operation of the system. Through timely environmental monitoring and corresponding decision-making adjustments, the overall efficiency and sustainability of the photovoltaic power generation system can be improved.

[0013] Optionally, the time series characteristics, relative performance characteristics and environmental characteristics are input into a support vector machine model to obtain a detection result, specifically including: using a mean filling method to perform missing value processing on the time series characteristics, the relative performance characteristics and the environmental characteristics to obtain a framework feature group; using the support vector machine model to calculate the framework feature group to obtain a feature space position; determining the positional relationship between the feature space position and the hyperplane corresponding to the support vector machine model; and generating the detection result based on the positional relationship.

[0014] By adopting the above technical solutions, in actual detection, feature data may be missing due to sensor failure or other reasons. The mean filling method can effectively fill these missing values ​​and avoid incomplete model training caused by missing data. The mean filling method can retain the overall characteristics of the data, so that the impact on model training is small. The integration of multiple features enables the model to consider data in multiple dimensions, enhancing the sensitivity of the model to changes in the performance of photovoltaic modules. The unified feature input framework helps to simplify the subsequent data processing and model training process and improve efficiency. SVM maps low-dimensional features to high-dimensional space through kernel functions, so that complex nonlinear problems can be effectively classified in high-dimensional space, which helps to improve the detection accuracy of PID effects. SVM seeks to construct an optimal hyperplane to maximize the class spacing. This feature makes it perform well in handling complex classification tasks, especially when the samples are unbalanced. By judging the position relationship in the feature space, it can be clear whether the sample falls on one side or the other side of the hyperplane, simplifying the logic of anomaly detection. This spatial position-based judgment method makes the detection results more interpretable, and the operation and maintenance personnel can clearly understand the judgment basis of the model. Automated processing reduces the need for human intervention, improves the speed and efficiency of detection, and can promptly detect abnormal conditions of PV panels. Testing through a unified model can ensure the consistency and reliability of test results and reduce the possibility of human error.

[0015] Optionally, the method also includes: based on the position relationship, calculating a probability value, the probability value including a first probability value and a second probability value, the first probability value being used to represent the probability value corresponding to the feature space position on the first side of the hyperplane, and the second probability value being used to identify the probability value corresponding to the feature space position on the second side of the hyperplane, the first side and the second side being two opposite sides of the hyperplane; if it is determined that the feature space position is located on the first side and the first probability value is greater than or equal to a preset threshold, then it is determined that the detection result indicates normal, and it is determined that the target photovoltaic component does not have a PID effect; if it is determined that the feature space position is located on the second side and the second probability value is greater than or equal to the preset threshold, then it is determined that the detection result indicates abnormality, and it is determined that the target photovoltaic component has a PID effect.

[0016] By adopting the above technical solutions, the probability value can quantify the uncertainty of the classification results, making the decision more scientific. By setting a threshold to judge the status of the component, the decision-making process can be effectively optimized to ensure that maintenance or intervention decisions are made only under high confidence and reduce the false alarm rate. This division provides a clear classification standard, so that the model can quickly and accurately determine the health status of the photovoltaic component, which is convenient for the operation and maintenance personnel to take corresponding measures. The division of the hyperplane makes the results easy to understand, and the operation and maintenance personnel can intuitively know which category the current component status is in, so as to make quick decisions and responses. According to the operating environment and historical data of the photovoltaic component, the threshold can be adjusted to adapt to different scenarios and needs, and enhance the adaptability and accuracy of the model. By adjusting the threshold, a balance can be achieved between sensitivity and specificity to improve the overall detection effect. Clear judgment results enable operation and maintenance personnel to take quick action, reduce decision-making time, and improve operation and maintenance efficiency. Clear judgment results provide a reliable basis for operation and maintenance decisions, which can help operation and maintenance personnel formulate corresponding maintenance and repair strategies to ensure the stable operation of the system. Probability values ​​higher than the preset threshold ensure that PID effects are reported only at a higher confidence level, thereby reducing unnecessary maintenance costs and waste of resources. Through accurate detection and reporting, the overall reliability of the photovoltaic power generation system can be improved and long-term stable power generation performance can be ensured.

[0017] Optionally, the method also includes: acquiring training data, the training data including a training set and a test set; using the training set to train an initial model including a radial basis kernel and hyperparameters to obtain a correspondence between feature groups and labels, the labels including normal and abnormal; optimizing parameters of the initial model by constructing a Lagrangian function, and obtaining the support vector machine model that meets preset standards or converges through repeated iterations.

[0018] By adopting the above technical solution, dividing the data into a training set and a test set can effectively evaluate the generalization ability of the model and avoid the model merely memorizing the training data without the ability to process new data. The test set is used to independently evaluate the performance of the model, help understand the performance of the model on unknown data, and ensure that the model has good accuracy and reliability in practical applications. Through training, the model learns the corresponding relationship between the input features (and the output labels, laying the foundation for subsequent anomaly detection. The initial model can adapt to the different operating states of photovoltaic modules through continuous learning, thereby improving the accuracy of detection. The radial basis kernel function can map data from the original space to a higher-dimensional feature space, so that complex nonlinear relationships can be processed, which helps to improve the classification ability of the model. Secondly, the radial basis kernel function has good adaptability and can be optimized for different data distributions, which improves the performance of the model in various situations. Parameter optimization helps to find the optimal decision boundary and ensure that the hyperplane of the model in the feature space can Effectively distinguish normal and abnormal samples and improve classification accuracy. Through parameter adjustment, the model can better adapt to the training data, reduce the misclassification rate, and thus improve the reliability of detection. Through iterative optimization, the model will be fine-tuned at each step, gradually approaching the optimal model parameters and improving the overall performance. The iterative process can find a stable parameter combination in multiple attempts to ensure that the model can show good robustness under various data conditions. The convergence criterion helps to determine when to stop training, avoid unnecessary calculations and waste of resources, and improve training efficiency. Setting the convergence criterion can ensure that the model stops training after reaching a certain performance level, avoiding the risk of overfitting.

[0019] In a second aspect of the present application, a photovoltaic component PID detection device is provided, and the photovoltaic component PID detection device includes an acquisition module and a processing module, wherein the acquisition module is used to acquire detection data for a target photovoltaic component, and the detection data includes current, voltage, power output, ambient temperature, relative humidity and solar irradiance; the processing module is used to determine a target feature group based on the detection data, and the target feature group includes time series characteristics, relative performance characteristics and environmental characteristics; the processing module is also used to input the time series characteristics, relative performance characteristics and environmental characteristics into a support vector machine model to obtain a detection result; the processing module is also used to determine that the target photovoltaic component has a PID effect if it is determined that the detection result indicates an abnormality.

[0020] In the third aspect of the present application, an electronic device is provided, which includes a processor, a memory, a user interface and a network interface, the memory is used to store instructions, the user interface and the network interface are both used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device performs the method described above.

[0021] In a fourth aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores instructions, and when the instructions are executed, the method described above is executed.

[0022] In summary, one or more technical solutions provided in this application have at least the following technical effects or advantages: By acquiring a variety of test data, namely current, voltage, power output, ambient temperature, relative humidity, and solar irradiance, rich information is provided for subsequent analysis. This comprehensive data collection can more accurately reflect the actual working status of photovoltaic modules, making subsequent analysis more reliable. By determining the target feature group, the method integrates time series features, relative performance features, and environmental features. This systematic feature extraction can fully capture the performance changes of photovoltaic modules under different environmental conditions, help improve the sensitivity and accuracy of detection, and avoid the one-sidedness common in traditional methods. The use of support vector machine models for detection fully utilizes the advantages of machine learning in data pattern recognition and classification. SVM can effectively process high-dimensional data and maintain good performance in complex nonlinear situations, thereby improving the accuracy and robustness of PID detection. By directly judging the test results, if the results indicate an abnormality, it can be quickly determined that the photovoltaic module has a PID effect. This clear feedback mechanism provides clear operational guidance for operation and maintenance personnel, allowing them to take timely measures to reduce potential losses. Therefore, the above method can improve the accuracy of PID detection of photovoltaic modules. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 A schematic flow chart of a photovoltaic module PID detection method provided in an embodiment of the present application.

[0024] Figure 2 Another schematic diagram of a flow chart of a photovoltaic module PID detection method provided in an embodiment of the present application.

[0025] Figure 3 A module schematic diagram of a photovoltaic module PID detection device provided in an embodiment of the present application.

[0026] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0027] Explanation of the reference numerals: 31, acquisition module; 32, processing module; 41, processor; 42, communication bus; 43, user interface; 44, network interface; 45, memory. DETAILED DESCRIPTION

[0028] In order to enable technicians in this field to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments.

[0029] In the description of the embodiments of the present application, words such as "for example" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "for example" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "for example" or "for example" is intended to present related concepts in a specific way.

[0030] In the description of the embodiments of the present application, the meaning of the term "multiple" refers to two or more. For example, multiple systems refer to two or more systems, and multiple screen terminals refer to two or more screen terminals. In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. The terms "include", "comprise", "have" and their variations all mean "including but not limited to", unless otherwise specifically emphasized.

[0031] After a long period of use, photovoltaic modules may be affected by a variety of external factors, resulting in a decline in their overall performance. Potential Induced Degradation (PID) is the most common failure phenomenon in the operation of photovoltaic modules. It will significantly reduce the power generation efficiency of the modules, thus directly affecting the economy and reliability of the entire photovoltaic power generation system. This phenomenon not only causes damage to the photovoltaic modules themselves, but may also have a serious negative impact on the long-term operation of the power generation system.

[0032] At present, traditional PID detection technology often relies on empirical formulas and threshold judgments based on experience. Although these methods can provide certain guidance in some cases, their accuracy and reliability are largely limited. This is mainly because the working environment of photovoltaic modules is usually very complex and changeable, and there are relatively many factors that affect their performance. A single empirical formula lacks adaptability to real-time data, so it may fail under new environmental conditions or specific operating conditions, which leads to uncertainty in the accuracy of photovoltaic module PID detection.

[0033] In order to solve the above technical problems, the present application provides a photovoltaic module PID detection method, referring to Figure 1, Figure 1 A schematic flow chart of a photovoltaic module PID detection method provided in an embodiment of the present application. The detection method is applied to a server and includes steps S110 to S140, which are as follows: S110, acquiring detection data for a target photovoltaic module, the detection data including current, voltage, power output, ambient temperature, relative humidity, and solar irradiance.

[0034] Specifically, the main role of the server in this process is to collect and process data. It can be a local server or a cloud server, which is responsible for storing, analyzing and managing the relevant data of photovoltaic modules. For example, in a large photovoltaic power plant, multiple photovoltaic modules will be connected to a central server, and the server will regularly obtain real-time data from each photovoltaic module and process it centrally. The detection data is collected for a specific photovoltaic module, not the entire photovoltaic system. This targeting can help accurately diagnose the performance problems of a specific component. Assuming that a photovoltaic module performs abnormally during the power generation process, the operation and maintenance personnel can obtain all the detection data of the specific component through the server for analysis, instead of viewing all the component data of the entire power plant.

[0035] Among them, current is the current value generated by the PV module when it is working. Current is a key indicator reflecting the power generation capacity of the module. If a PV module should output 5 amperes of current under ideal conditions, but the actual measured value is only 2 amperes, this may indicate that the module is faulty. Voltage is the voltage value output by the PV module, which together with the current affects the power output of the module. A normal PV module may output 40 volts under standard test conditions. If the output voltage is found to be significantly lower than this value, it may indicate that there is an electrical fault in the module. Power output is the electrical energy generated by the PV module in a certain period of time, usually in watts. When a PV module should output 250 watts of power under good sunlight, but the actual output is only 100 watts, this may mean that the module is affected by the PID effect or other faults. Ambient temperature is the temperature of the environment around the PV module, and temperature affects the power generation efficiency of the PV module. If the PV module is operated in a high temperature environment, it may lead to reduced efficiency. Therefore, the server needs to monitor the ambient temperature to analyze its impact on the performance of the module. Relative humidity is the humidity level of the surrounding environment, which may affect the long-term performance and reliability of the module. In high humidity environments, photovoltaic modules are more susceptible to electrical short circuits or corrosion, so monitoring relative humidity is critical to assessing the health of modules. Solar irradiance is the intensity of sunlight received by photovoltaic modules, usually expressed in watts per square meter, and is an important indicator for evaluating the performance of photovoltaic modules. Changes in solar irradiance at different times and weather conditions will directly affect the power output of photovoltaic modules. Servers need to monitor this parameter in real time for comprehensive performance analysis.

[0036] In a possible implementation, obtaining detection data for a target photovoltaic component specifically includes: receiving raw current data of the target photovoltaic component sent by a current sensor; receiving raw voltage data of the target photovoltaic component sent by a voltage sensor; receiving raw power data of the target photovoltaic component sent by a power sensor; receiving raw ambient temperature data of the target photovoltaic component sent by an ambient temperature sensor; receiving raw relative humidity data of the target photovoltaic component sent by a relative humidity sensor; receiving raw solar radiation data of the target photovoltaic component sent by a solar radiation sensor; and preprocessing the raw current data, raw voltage data, raw power data, raw ambient temperature data, raw relative humidity data, and raw solar radiation data to obtain detection data, wherein the preprocessing includes denoising, filtering, and normalization.

[0037] Specifically, first, the server needs to obtain various detection data related to a specific photovoltaic module. These data are collected in a distributed manner through different types of sensors, ensuring comprehensive monitoring of the working status of the photovoltaic module. The current sensor is responsible for measuring the output current of the photovoltaic module. This is an important indicator for evaluating the power generation capacity of the module. The voltage sensor is used to measure the output voltage of the photovoltaic module, and combined with the current, the power output can be calculated. The power sensor directly measures the power output of the photovoltaic module. For example, the power sensor shows that the actual power output of the module is 105 watts, which is crucial for evaluating its efficiency. The ambient temperature sensor measures the ambient temperature around the photovoltaic module. The ambient temperature directly affects the power generation efficiency of the module. The relative humidity sensor monitors the humidity level of the environment, which may affect the long-term performance of the module. The solar radiation sensor measures the intensity of sunlight received by the photovoltaic module, usually expressed in watts per square meter (W / m²). Among them, the above-mentioned multiple sensor devices are pre-installed at the corresponding measurement positions of the photovoltaic module and maintain communication connection with the server.

[0038] Among them, after the data collection is completed, the server needs to preprocess the raw data to improve the data quality and the accuracy of the analysis. Preprocessing includes the following aspects: Eliminate random noise that may be introduced during the data collection process. For example, the current sensor may produce instantaneous current fluctuations due to electromagnetic interference. Through denoising, these unreal fluctuations can be eliminated to make the data more stable. Apply filtering technology to smooth the raw data to reduce the volatility and error of the data. This can help highlight the main trends of the data, such as using a low-pass filter to eliminate high-frequency noise and retain the main signal changes. Convert data of different dimensions to the same standard range, usually between 0 and 1. This step is especially important for comparing different types of data, such as current, voltage, power, etc., because the dimensions and ranges of these data may vary greatly. If the current range is 0 to 10 amperes, the voltage range is 0 to 60 volts, and the power range is 0 to 300 watts, they can be converted to the same scale through normalization, making it easier to compare in subsequent analysis.

[0039] S120. Determine a target feature group according to the detection data, where the target feature group includes time series features, relative performance features, and environmental features.

[0040] Specifically, the feature group refers to the key indicators extracted from the detection data, which are used to reflect the operating status and performance of the photovoltaic module. The selection of features is crucial to the accuracy and effectiveness of the model. In this example, the target feature group contains three types of features: Time series features refer to the changing trend of photovoltaic module detection data over a period of time. These features can reveal the performance changes of photovoltaic modules at different time points. For example: Assume that during the monitoring process, the current data is recorded as [3A, 3.2A, 3.5A, 2.8A, 3.1A] in different time periods. From these data, the mean, standard deviation, maximum value, minimum value and time slope can be calculated. For example, a negative slope indicates a downward trend in current, thereby obtaining time series features. These features help to detect early signs of performance degradation. Relative performance features are indicators obtained by comparing current performance with benchmark performance, usually performance under ideal conditions. These features can help determine whether the actual power generation capacity of the module meets the standard. Environmental features refer to external environmental factors that affect the performance of photovoltaic modules within a specific time, including temperature, humidity and solar radiation. For example, assume that during the recording period, the ambient temperature is 25°C, the relative humidity is 60%, and the solar radiation is 800W / m². These data can be used to analyze the difference between the actual performance and the expected performance of the PV module under the current environmental conditions, thereby helping to assess the health status of the module. For example, high temperature and high humidity may cause the performance of the module to deteriorate, which requires special attention.

[0041] In a possible implementation, a target feature group is determined based on the detection data, specifically including: using a sliding window to extract the target time point in the current and voltage, and the statistical features corresponding to the target time point, the statistical features including the mean, standard deviation, maximum value, minimum value and time slope; generating time series features based on the mean, standard deviation, maximum value, minimum value and time slope; obtaining the reference power output of the target photovoltaic module; calculating the relative power based on the reference power output and the power output at the current moment; obtaining the power output corresponding to the target time point, the target time point and the current moment are adjacent time points; calculating the relative power change rate based on the power output corresponding to the target time point and the power output at the current moment, the relative performance features including the relative power and the relative power change rate.

[0042] Specifically, sliding window is a method for analyzing time series data. It sets a fixed-length window, moves on the data, and gradually extracts the data segments within the window. In the detection data of photovoltaic modules, the current and voltage signals can be analyzed through sliding windows to analyze their characteristics of changing over time.

[0043] For example, suppose there is the following current data: [2.5A, 2.6A, 2.4A, 2.7A, 2.9A, 2.5A]. If a window size is set to 3, the server can extract the following data segments: Window 1: [2.5A, 2.6A, 2.4A]; Window 2: [2.6A, 2.4A, 2.7A]; Window 3: [2.4A, 2.7A, 2.9A]; Window 4: [2.7A, 2.9A, 2.5A]; For each extracted window, a series of statistical features such as mean, standard deviation, maximum, minimum and time slope are calculated to capture the changing trend and volatility of the data. The mean is the average value of the data in the window. The standard deviation is the degree of dispersion of the data. The maximum value is the maximum value of the data in the window. The minimum value is the minimum value of the data in the window. The time slope can be calculated by linear regression to reflect the changing trend of the data over time. For example, for window 1 [2.5A, 2.6A, 2.4A], the mean can be calculated to be 2.5A, the standard deviation is 0.0816A, the maximum value is 2.6A, the minimum value is 2.4A, and the time slope is 0.1. The above statistical features are used to generate time series features, which are used to describe the performance changes of photovoltaic modules over a period of time. The benchmark power output refers to the output power that the photovoltaic module should have under ideal conditions. This indicator is usually based on the design parameters or historical performance data of the module. The relative power is generated by comparing the current power output with the benchmark power output, which can reflect the actual power generation capacity of the photovoltaic module. The relative rate of change of power is used to reflect the degree of change in power output. It is calculated based on the power output at the current moment and the power output at adjacent time points. The relative performance feature combines relative power and relative rate of change of power to provide a performance evaluation of PV modules under specific conditions. In summary, this passage details how to extract features from the detection data of PV modules for subsequent performance analysis and fault diagnosis. This feature extraction process can improve the ability to monitor and evaluate the health status of PV modules, thereby providing data support for maintenance decisions.

[0044] In a possible implementation, a target feature group is determined based on the detection data, which specifically includes: calculating the average value, maximum value and minimum value corresponding to the environmental data at each current moment based on the ambient temperature, relative humidity and solar irradiance; obtaining the environmental data corresponding to the target time point; and calculating the relative environmental change rate based on the environmental data corresponding to the target time point and the environmental data at the current moment, wherein the environmental features include the average value, maximum value, minimum value and relative environmental change rate.

[0045] Specifically, for environmental data (such as temperature, humidity, and irradiance) collected within a certain period of time, the overall characteristics of the environmental conditions can be obtained by calculating the average, maximum, and minimum values ​​of these data. These statistical characteristics can reflect the trend and magnitude of environmental changes, which helps to analyze the impact of the environment on the performance of photovoltaic modules. During the analysis process, environmental data at a specific time point is required for comparison with the data at the current moment. The target time point usually refers to a time point related to the power output or other performance indicators of the photovoltaic module. Assume that the environmental data recorded at a specific time point, such as 10:00AM, is: temperature: 26°C, humidity: 48%, irradiance: 800W / m 2The relative rate of change of the environment is used to indicate the degree of change of the environmental conditions between a specific time point and the current moment. This rate of change can reflect the possible impact of environmental conditions on the performance of photovoltaic modules. For example: Assume that at the current moment, for example, 11:00 AM, the environmental data recorded are: temperature: 27°C, humidity: 50%, irradiance: 900W / m 2 The calculated relative change rate of ambient temperature is 3.85%, the relative change rate of ambient humidity is 4.17%, and the relative change rate of ambient radiation is 125%.

[0046] S130, inputting the time series characteristics, relative performance characteristics and environmental characteristics into a support vector machine model to obtain a detection result.

[0047] Specifically, when performing PID detection, it is first necessary to input the features extracted from the operating data of the PV module, including time series features, relative performance features, and environmental features, into the trained support vector machine model. The SVM model is a powerful classification and regression tool that can make predictions by learning the relationship between features and labels. After receiving the input features, the support vector machine model will perform calculations in its feature space, determine the positional relationship corresponding to these features, and output a detection result. The result is usually a classification result (such as "normal" or "abnormal"), which is used to indicate whether the PV module has a PID effect.

[0048] In one possible implementation, time series features, relative performance features, and environmental features are input into a support vector machine model to obtain detection results, specifically including: using a mean filling method to process missing values ​​for time series features, relative performance features, and environmental features to obtain a framework feature group; using a support vector machine model to calculate the framework feature group to obtain a feature space position; determining a positional relationship between the feature space position and a hyperplane corresponding to the support vector machine model; and generating a detection result based on the positional relationship.

[0049] Specifically, when performing data analysis, there may be missing values ​​in the feature data, that is, data that has not been collected. Mean filling is a commonly used missing value processing technique that replaces missing values ​​with the mean of the feature to ensure data integrity and availability. This can prevent calculation errors caused by missing values ​​during model training. Suppose that in the time series feature, the current data at a certain time point is missing, and the mean of other current data is known to be 5.0A. Then, the missing current value will be filled with 5.0A. After missing value processing, all time series features, relative performance features, and environmental features will be combined into a complete feature group, called the framework feature group. This group of features will be used for subsequent model calculations. The framework feature group is input into the trained support vector machine model, and the model will calculate the position of the feature group in the feature space based on the input features. The feature space is a multidimensional space in which each dimension corresponds to a feature. Once the position in the feature space is obtained, the model will determine the relationship between the position and the defined hyperplane. The hyperplane is used to divide data points into different categories (such as normal and abnormal). If the position is on one side of the hyperplane, it usually represents one category (such as normal), and on the other side it represents another category (such as abnormal). Finally, based on the relationship between the feature space position and the hyperplane, the final detection result is generated, indicating whether the photovoltaic module has a PID effect.

[0050] S140: If it is determined that the detection result indicates an abnormality, it is determined that the target photovoltaic component has a PID effect.

[0051] Specifically, after the input features are processed by the support vector machine model, the model will generate a test result. This result will be divided into two categories: normal and abnormal. If the test result is judged to be "abnormal", it means that the performance of the photovoltaic module is abnormal and there may be some kind of fault. Once the test result indicates an abnormality, the system can infer that the photovoltaic module may be affected by the PID effect. This is because the PID effect will cause the current, voltage and power output of the component to decrease, causing the test result to deviate from the normal range. Through the above process, the abnormal indication of the test result is an important basis for judgment, which can effectively identify the risk of PID effect in photovoltaic modules. This timely feedback can not only help maintenance personnel respond quickly, but also effectively maintain the overall performance and economy of the photovoltaic power generation system.

[0052] In a possible implementation, based on the positional relationship, a probability value is calculated, the probability value including a first probability value and a second probability value, the first probability value being used to indicate the probability value corresponding to the feature space position on the first side of the hyperplane, and the second probability value being used to identify the probability value corresponding to the feature space position on the second side of the hyperplane, the first side and the second side being two opposite sides of the hyperplane; if it is determined that the feature space position is located on the first side, and the first probability value is greater than or equal to a preset threshold, then it is determined that the detection result indicates normal, and it is determined that the target photovoltaic component does not have a PID effect; if it is determined that the feature space position is located on the second side, and the second probability value is greater than or equal to the preset threshold, then it is determined that the detection result indicates abnormality, and it is determined that the target photovoltaic component has a PID effect.

[0053] Specifically, the hyperplane is the key structure used by the support vector machine for classification, which divides the feature space into two parts. The data points in the feature space, that is, the feature positions, will be classified as normal or abnormal according to their relationship with the hyperplane. The support vector machine model calculates two probability values. The first probability value represents the probability that the feature space position is on the first side of the hyperplane, corresponding to the normal state. The second probability value represents the probability that the feature space position is on the second side of the hyperplane, corresponding to the abnormal state. If the feature space position is on the first side of the hyperplane and the first probability value is greater than or equal to the preset threshold, such as 0.7, the server determines it as normal and believes that the photovoltaic module has no PID effect. If the feature space position is on the second side of the hyperplane and the second probability value is greater than or equal to the set 0.7, the server determines it as abnormal and believes that the photovoltaic module has a PID effect. The server systematically judges the operating status of the photovoltaic module by analyzing the relationship between the feature space position and the hyperplane and the relevant probability values. Through the set threshold judgment, the normal and abnormal states can be effectively distinguished, so as to timely discover the potential PID effect, which provides a basis for maintenance and repair. This method improves the accuracy and reliability of photovoltaic module detection.

[0054] In one possible implementation, refer to Figure 2 , Figure 2 Another flow chart of a photovoltaic module PID detection method provided in an embodiment of the present application includes steps S210 to S230, and the above steps are as follows: S210, obtaining training data, the training data including a training set and a test set; S220, using the training set to train an initial model including a radial basis kernel and hyperparameters to obtain a correspondence between feature groups and labels, and the labels include normal and abnormal; S230, using a constructed Lagrangian function to optimize the parameters of the initial model, and through repeated iterations, obtaining a support vector machine model that meets preset standards or converges.

[0055] Specifically, in machine learning, training data is used to train the model to help the model learn the relationship between input features and output labels. The test set is usually used after the model is trained to verify the generalization ability of the model. The radial basis kernel function is a kernel function that can map the input data to a high-dimensional space so that classification can be performed more easily in this space. Through this mapping, the support vector machine can process nonlinearly separable data. During the model training process, hyperparameters are parameters that need to be set manually, such as the parameters and penalty parameters of the kernel function. Hyperparameters have an important impact on the performance of the model. The input data of the model is a feature group, that is, the detection data of the photovoltaic module, such as current, voltage and other features. The label is the output result of the model, which is a classification result, such as "normal" or "abnormal" in the embodiment of the present application. Through training, the model learns how to predict the corresponding label based on the input features. In optimization problems, the Lagrangian function is a mathematical tool for solving constrained optimization problems. In the support vector machine, the Lagrangian function is used to construct the objective function so that the model can find the best classification boundary. By constructing the Lagrangian function, the model parameters are adjusted through repeated iterations during the training process of the model to optimize the classification boundary. This process can be achieved through algorithms, such as gradient descent. During the training process, the algorithm continuously adjusts parameters through iterations until the performance of the model no longer changes significantly and reaches a convergence state, such as the loss function. The results of the training need to meet certain preset standards, such as classification accuracy or error rate, in order to consider the model effective.

[0056] The present application also provides a photovoltaic module PID detection device, referring to Figure 3 , Figure 3 A module schematic diagram of a photovoltaic component PID detection device provided in an embodiment of the present application, wherein the detection device is a server, and the server includes an acquisition module 31 and a processing module 32, wherein the acquisition module 31 acquires detection data for a target photovoltaic component, the detection data including current, voltage, power output, ambient temperature, relative humidity and solar irradiance; the processing module 32 determines a target feature group based on the detection data, the target feature group including time series features, relative performance features and environmental features; the processing module 32 inputs the time series features, relative performance features and environmental features into a support vector machine model to obtain a detection result; if the processing module 32 determines that the detection result indicates an abnormality, it determines that a PID effect exists in the target photovoltaic component.

[0057] In a possible implementation, the acquisition module 31 acquires detection data for the target photovoltaic component, specifically including: the acquisition module 31 receives the original current data of the target photovoltaic component sent by the current sensor; the acquisition module 31 receives the original voltage data of the target photovoltaic component sent by the voltage sensor; the acquisition module 31 receives the original power data of the target photovoltaic component sent by the power sensor; the acquisition module 31 receives the original ambient temperature data of the target photovoltaic component sent by the ambient temperature sensor; the acquisition module 31 receives the original relative humidity data of the target photovoltaic component sent by the relative humidity sensor; the acquisition module 31 receives the original solar radiation data of the target photovoltaic component sent by the solar radiation sensor; the processing module 32 preprocesses the original current data, original voltage data, original power data, original ambient temperature data, original relative humidity data and original solar radiation data to obtain detection data, and the preprocessing includes denoising, filtering and normalization.

[0058] In a possible implementation, the processing module 32 determines the target feature group based on the detection data, specifically including: the processing module 32 uses a sliding window to extract the target time point in the current and voltage, and the statistical features corresponding to the target time point, the statistical features include mean, standard deviation, maximum value, minimum value and time slope; the processing module 32 generates time series features based on the mean, standard deviation, maximum value, minimum value and time slope; the acquisition module 31 acquires the reference power output of the target photovoltaic component; the processing module 32 calculates the relative power based on the reference power output and the power output at the current moment; the acquisition module 31 acquires the power output corresponding to the target time point, the target time point and the current moment are adjacent time points; the processing module 32 calculates the relative power change rate based on the power output corresponding to the target time point and the power output at the current moment, and the relative performance characteristics include relative power and relative power change rate.

[0059] In a possible implementation, the processing module 32 determines the target feature group based on the detection data, and specifically also includes: the processing module 32 calculates the average value, maximum value and minimum value corresponding to the environmental data at each current moment based on the ambient temperature, relative humidity and solar irradiance; the acquisition module 31 acquires the environmental data corresponding to the target time point; the processing module 32 calculates the relative change rate of the environment based on the environmental data corresponding to the target time point and the environmental data at the current moment, and the environmental characteristics include the average value, maximum value, minimum value and the relative change rate of the environment.

[0060] In one possible implementation, the processing module 32 inputs the time series features, relative performance features and environmental features into a support vector machine model to obtain a detection result, which specifically includes: the processing module 32 uses a mean filling method to process missing values ​​of the time series features, relative performance features and environmental features to obtain a framework feature group; the processing module 32 calculates the framework feature group through a support vector machine model to obtain a feature space position; the processing module 32 determines the positional relationship between the feature space position and the hyperplane corresponding to the support vector machine model; the processing module 32 generates a detection result based on the positional relationship.

[0061] In a possible implementation, the processing module 32 calculates a probability value based on the position relationship, and the probability value includes a first probability value and a second probability value, the first probability value is used to indicate the probability value corresponding to the feature space position on the first side of the hyperplane, and the second probability value is used to identify the probability value corresponding to the feature space position on the second side of the hyperplane, and the first side and the second side are two opposite sides of the hyperplane; if the processing module 32 determines that the feature space position is located on the first side and the first probability value is greater than or equal to a preset threshold, it is determined that the detection result indicates normal, and it is determined that the target photovoltaic component does not have a PID effect; if the processing module 32 determines that the feature space position is located on the second side and the second probability value is greater than or equal to the preset threshold, it is determined that the detection result indicates abnormality, and it is determined that the target photovoltaic component has a PID effect.

[0062] In one possible implementation, the acquisition module 31 acquires training data, and the training data includes a training set and a test set; the processing module 32 uses the training set to train an initial model including a radial basis kernel and hyperparameters to obtain a correspondence between feature groups and labels, and the labels include normal and abnormal; the processing module 32 uses a constructed Lagrangian function to optimize the parameters of the initial model, and through repeated iterations, obtains a support vector machine model that meets preset standards or converges.

[0063] It should be noted that: when the device provided in the above embodiment realizes its function, only the division of the above functional modules is used as an example. In actual application, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the device and method embodiments provided in the above embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0064] The present application also provides an electronic device, referring to Figure 4 , Figure 4 The electronic device may include: at least one processor 41 , at least one network interface 44 , a user interface 43 , a memory 45 , and at least one communication bus 42 .

[0065] The communication bus 42 is used to realize the connection and communication between these components.

[0066] The user interface 43 may include a display screen (Display) and a camera (Camera). Optionally, the user interface 43 may also include a standard wired interface and a wireless interface.

[0067] The network interface 44 may optionally include a standard wired interface or a wireless interface (such as a WI-FI interface).

[0068] Among them, the processor 41 may include one or more processing cores. The processor 41 uses various interfaces and lines to connect various parts in the entire server, and executes various functions of the server and processes data by running or executing instructions, programs, code sets or instruction sets stored in the memory 45, and calling data stored in the memory 45. Optionally, the processor 41 can be implemented in at least one hardware form of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), and programmable logic array (Programmable Logic Array, PLA). The processor 41 can integrate one or a combination of a central processing unit (Central Processing Unit, CPU), a graphics processing unit (Graphics Processing Unit, GPU) and a modem. Among them, the CPU mainly processes the operating system, user interface and application programs; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 41, and it can be implemented separately through a chip.

[0069] Among them, the memory 45 may include a random access memory (RAM) or a read-only memory (Read-Only Memory). Optionally, the memory 45 includes a non-transitory computer-readable storage medium. The memory 45 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 45 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned method embodiments, etc.; the data storage area may store data involved in the above-mentioned method embodiments, etc. The memory 45 may also be optionally at least one storage device located away from the aforementioned processor 41. As Figure 4 As shown, the memory 45 as a computer storage medium may include an operating system, a network communication module, a user interface module and an application program of a photovoltaic component PID detection method.

[0070] exist Figure 4 In the electronic device shown, the user interface 43 is mainly used to provide an input interface for the user and obtain data input by the user; and the processor 41 can be used to call an application program storing a photovoltaic component PID detection method in the memory 45. When executed by one or more processors, the electronic device executes one or more methods in the above-mentioned embodiments.

[0071] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the order of the actions described, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required for the present application.

[0072] The present application also provides a computer-readable storage medium, which stores instructions. When executed by one or more processors, the electronic device executes one or more of the methods described in the above embodiments.

[0073] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0074] In the several embodiments provided in this application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0075] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0076] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0077] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a memory and includes several instructions for a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned memory includes: various media that can store program codes, such as USB flash drives, mobile hard drives, magnetic disks or optical disks.

[0078] The above is only an exemplary embodiment of the present disclosure, and the scope of the present disclosure cannot be limited thereto. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the disclosure of the specification and the truth of practice, it will be easy for those skilled in the art to think of other embodiments of the present disclosure. This application is intended to cover any modification, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the technical field that are not recorded in the present disclosure. The description and examples are regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. A photovoltaic module PID detection method, characterized in that: The method comprises: Acquire detection data for a target photovoltaic module, wherein the detection data includes current, voltage, power output, ambient temperature, relative humidity, and solar irradiance; Determining a target feature group according to the detection data, the target feature group including time series features, relative performance features, and environmental features; Inputting the time series characteristics, relative performance characteristics and environmental characteristics into a support vector machine model to obtain a detection result; If it is determined that the detection result indicates an abnormality, it is determined that the target photovoltaic component has a PID effect.

2. The photovoltaic module PID detection method according to claim 1, characterized in that: The obtaining of detection data for the target photovoltaic module specifically includes: Receiving raw current data of the target photovoltaic component sent by a current sensor; Receiving raw voltage data of the target photovoltaic component sent by a voltage sensor; Receiving raw power data of the target photovoltaic component sent by a power sensor; Receiving raw ambient temperature data of the target photovoltaic module sent by an ambient temperature sensor; Receiving raw relative humidity data of the target photovoltaic module sent by a relative humidity sensor; Receiving raw solar radiation data of the target photovoltaic module sent by a solar radiation sensor; The original current data, the original voltage data, the original power data, the original ambient temperature data, the original relative humidity data and the original solar radiation data are preprocessed to obtain the detection data, and the preprocessing includes denoising, filtering and normalization processing.

3. The photovoltaic module PID detection method according to claim 1, characterized in that: Determining the target feature group according to the detection data specifically includes: A sliding window is used to extract the target time points in the current and voltage, and the statistical features corresponding to the target time points, wherein the statistical features include mean, standard deviation, maximum value, minimum value and time slope; Generate the time series feature according to the mean, the standard deviation, the maximum value, the minimum value and the time slope; Obtaining a reference power output of the target photovoltaic module; Calculating the relative power according to the reference power output and the power output at the current moment; Obtaining the power output corresponding to the target time point, where the target time point and the current moment are adjacent time points; The power relative change rate is calculated based on the power output corresponding to the target time point and the power output at the current moment, and the relative performance characteristics include the relative power and the power relative change rate.

4. The photovoltaic module PID detection method according to claim 3, characterized in that: Determining the target feature group according to the detection data specifically includes: Calculate the average value, maximum value and minimum value of the environmental data at each current moment according to the ambient temperature, the relative humidity and the solar irradiance; Acquire environmental data corresponding to the target time point; The relative rate of change of the environment is calculated based on the environmental data corresponding to the target time point and the environmental data at the current moment, and the environmental characteristics include the average value, the maximum value, the minimum value and the relative rate of change of the environment.

5. The photovoltaic module PID detection method according to claim 1, characterized in that: The time series characteristics, relative performance characteristics and environmental characteristics are input into the support vector machine model to obtain the detection results, specifically including: Using the mean filling method to process missing values ​​of the time series characteristics, the relative performance characteristics and the environmental characteristics, to obtain a framework feature group; The framework feature group is calculated by the support vector machine model to obtain a feature space position; Determining the positional relationship between the feature space position and the hyperplane corresponding to the support vector machine model; The detection result is generated according to the positional relationship.

6. The photovoltaic module PID detection method according to claim 5, characterized in that: The method further comprises: Based on the positional relationship, a probability value is calculated, the probability value comprising a first probability value and a second probability value, the first probability value being used to represent a probability value corresponding to the first side of the hyperplane of the feature space position, the second probability value being used to identify a probability value corresponding to the second side of the hyperplane of the feature space position, the first side and the second side being two opposite sides of the hyperplane; If it is determined that the characteristic spatial position is located on the first side, and the first probability value is greater than or equal to a preset threshold, it is determined that the detection result indicates normal, and it is determined that the target photovoltaic component does not have a PID effect; If it is determined that the characteristic space position is located on the second side and the second probability value is greater than or equal to a preset threshold, it is determined that the detection result indicates an abnormality, and it is determined that the target photovoltaic component has a PID effect.

7. The photovoltaic module PID detection method according to claim 1, characterized in that: The method further comprises: Acquire training data, where the training data includes a training set and a test set; The training set is used to train an initial model including a radial basis kernel and hyperparameters to obtain a corresponding relationship between feature groups and labels, wherein the labels include normal and abnormal; The initial model is optimized by constructing a Lagrangian function, and the support vector machine model that meets a preset standard or converges is obtained through repeated iterations.

8. A photovoltaic module PID detection device, characterized in that: The photovoltaic module PID detection device comprises an acquisition module (31) and a processing module (32), wherein: The acquisition module (31) is used to acquire detection data for a target photovoltaic component, wherein the detection data includes current, voltage, power output, ambient temperature, relative humidity and solar irradiance; The processing module (32) is used to determine a target feature group based on the detection data, wherein the target feature group includes a time series feature, a relative performance feature and an environmental feature; The processing module (32) is also used to input the time series characteristics, relative performance characteristics and environmental characteristics into a support vector machine model to obtain a detection result; The processing module (32) is further configured to determine that a PID effect exists in the target photovoltaic component if it is determined that the detection result indicates an abnormality.

9. An electronic device, characterized in that: The electronic device comprises a processor (41), a memory (45), a user interface (43) and a network interface (44), wherein the memory (45) is used to store instructions, the user interface (43) and the network interface (44) are both used to communicate with other devices, and the processor (41) is used to execute the instructions stored in the memory (45) so that the electronic device executes the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed, the method according to any one of claims 1 to 7 is performed.

Images