State evaluation method and device, terminal equipment and storage medium

Abstract

The application discloses a state evaluation method and device, a terminal equipment and a storage medium. The state evaluation method comprises the following steps: obtaining target data of a photovoltaic inverter; inputting the target data into a trained state evaluation model to obtain state evaluation information of the photovoltaic inverter, wherein the trained state evaluation model is obtained by training an initial support vector machine model based on sample data of the photovoltaic inverter and preset optimization parameters, the sample data is obtained based on an index system of the photovoltaic inverter, and the index system is constructed based on quantitative parameters and qualitative parameters of the photovoltaic inverter. According to the application, the index system is constructed based on the quantitative parameters and the qualitative parameters, so that the related data can be fully mined and utilized, the shortage of a single data type is made up, the trained state evaluation model has more comprehensive state evaluation capability, and the accuracy and reliability of the state evaluation of the photovoltaic inverter are improved.

Description

Technical Field

[0001] This application relates to the field of photovoltaic power generation technology, and in particular to a condition assessment method, apparatus, terminal equipment, and storage medium. Background Technology

[0002] As a key component of solar power generation systems, the health status of photovoltaic inverters directly affects the safety and stability of the power generation system. Therefore, the need for status assessment of photovoltaic inverter data has emerged.

[0003] Data related to the condition assessment of photovoltaic (PV) inverters exhibits characteristics of large-scale, multi-source, and multi-dimensional big data. However, current condition assessment methods for PV inverters typically only involve measuring the inverter's inputs and outputs to evaluate its condition. This approach fails to fully utilize the relevant data, resulting in inaccurate and unreliable condition assessment results for PV inverters. Summary of the Invention

[0004] The main purpose of this application is to provide a condition assessment method, apparatus, terminal equipment, and storage medium, which aims to solve the problem that the condition results of photovoltaic inverters are not accurate and reliable enough.

[0005] To achieve the above objectives, this application provides a state assessment method, the state assessment method comprising:

[0006] Obtain target data for the photovoltaic inverter;

[0007] The target data is input into the trained state evaluation model to obtain the state evaluation information of the photovoltaic inverter. The trained state evaluation model is obtained by training an initial support vector machine model based on the sample data of the photovoltaic inverter and preset optimization parameters. The sample data is obtained based on the index system of the photovoltaic inverter. The index system is constructed based on the quantitative and qualitative parameters of the photovoltaic inverter.

[0008] Optionally, before the step of acquiring the target data of the photovoltaic inverter, the method further includes:

[0009] Obtain at least one quantitative parameter and at least one qualitative parameter of the photovoltaic inverter;

[0010] Based on at least one quantitative parameter and at least one qualitative parameter of the photovoltaic inverter, a basic index system for the photovoltaic inverter is constructed.

[0011] Based on the basic index system of the photovoltaic inverter, the key index system of the photovoltaic inverter is constructed.

[0012] The sample data was obtained based on the key indicator system of the photovoltaic inverter.

[0013] The initial support vector machine model is trained based on the sample data and the preset optimization parameters to obtain the trained state evaluation model.

[0014] Optionally, the step of obtaining the target data of the photovoltaic inverter includes:

[0015] Obtain the raw data from the photovoltaic inverter;

[0016] The raw data is filtered according to the key parameter system of the photovoltaic inverter to obtain the filtered raw data;

[0017] The filtered raw data is preprocessed to obtain the target data.

[0018] Optionally, the step of constructing the basic index system of the photovoltaic inverter based on at least one quantitative parameter and at least one qualitative parameter of the photovoltaic inverter includes:

[0019] Based on a preset first analysis rule, at least one quantitative parameter of the photovoltaic inverter is analyzed to obtain the sensitivity factor corresponding to each of the at least one quantitative parameter.

[0020] Based on a preset second analysis rule, at least one qualitative parameter of the photovoltaic inverter is analyzed to obtain the degradation factor corresponding to each of the at least one qualitative parameter.

[0021] Based on the at least one quantitative parameter, the corresponding sensitivity factor of the at least one quantitative parameter, the at least one qualitative parameter, and the corresponding degradation factor of the at least one qualitative parameter, the basic index system of the photovoltaic inverter is constructed.

[0022] Optionally, the at least one quantitative parameter belongs to its corresponding feature type, the number of feature types is at least one, and each of the at least one quantitative parameter includes at least one quantitative parameter sample. The step of analyzing the at least one quantitative parameter of the photovoltaic inverter based on a preset first analysis rule to obtain the sensitivity factor corresponding to each of the at least one quantitative parameter includes:

[0023] Based on the quantitative parameter samples corresponding to at least one quantitative parameter under each feature type, the average intra-class distance and average inter-class distance corresponding to at least one quantitative parameter under each feature type are calculated.

[0024] Based on the average intra-class distance and the average inter-class distance, the sensitivity factor corresponding to at least one quantitative parameter under each feature type is calculated.

[0025] Optionally, the at least one qualitative parameter belongs to its corresponding feature type, and each of the at least one qualitative parameter includes an actual value, a threshold value, and a factory value. The step of analyzing the at least one qualitative parameter of the photovoltaic inverter based on a preset second analysis rule to obtain the degradation factor corresponding to each of the at least one qualitative parameter includes:

[0026] Based on the actual value, threshold, and factory value of at least one qualitative parameter under each feature type, and based on a preset offset calculation rule, the degradation factor corresponding to at least one qualitative parameter under each feature type is calculated.

[0027] Optionally, the step of constructing the key indicator system of the photovoltaic inverter based on the basic indicator system of the photovoltaic inverter includes:

[0028] Based on the basic index system of the photovoltaic inverter, at least one quantitative parameter and at least one qualitative parameter of the photovoltaic inverter are screened to obtain at least one preliminary quantitative parameter and at least one preliminary qualitative parameter.

[0029] Based on the preset Gaussian kernel function, the quantitative parameter samples corresponding to each of the at least one initial screening quantitative parameter, and the qualitative parameter samples corresponding to each of the at least one initial screening qualitative parameter, a normalized kernel matrix is ​​constructed.

[0030] Based on the normalized kernel matrix, the corresponding feature matrix and feature vector matrix are constructed.

[0031] Based on the normalized kernel matrix, the feature matrix, and the feature vector matrix, the target matrix is ​​constructed.

[0032] Based on the target matrix, a key indicator system for the photovoltaic inverter is constructed.

[0033] Optionally, the step of training the initial support vector machine model based on the sample data and the preset optimization parameters to obtain the trained state evaluation model includes:

[0034] The sample data is preprocessed to obtain training samples and test samples;

[0035] The initial support vector machine model is trained based on the training samples;

[0036] Determine whether the initial support vector machine model meets the preset cross-validation conditions;

[0037] If the initial support vector machine model satisfies the cross-validation condition, then training of the initial support vector machine model is stopped, and the state evaluation model to be tested is obtained.

[0038] If the initial support vector machine model does not meet the cross-validation condition, the initial support vector machine model is trained based on the training samples and preset optimization parameters until the initial support vector machine model meets the cross-validation condition. Then, the training of the initial support vector machine model is stopped, and the state evaluation model to be tested is obtained.

[0039] The state evaluation model to be tested is tested based on the test samples.

[0040] If the state evaluation model to be tested passes the test, then the trained state evaluation model is obtained.

[0041] Optionally, before the step of training the initial support vector machine model based on the training samples and preset optimization parameters until the initial support vector machine model meets the cross-validation condition, and then stopping the training of the initial support vector machine model to obtain the state evaluation model to be tested, the method further includes:

[0042] Based on the preset Grey Wolf algorithm, at least one optimization parameter is initialized.

[0043] Optionally, after the step of inputting the target data into the trained state assessment model to obtain the state assessment information of the photovoltaic inverter, the method further includes:

[0044] If the status assessment information of the photovoltaic inverter meets the preset alarm conditions, then the alarm category corresponding to the status assessment information of the photovoltaic inverter is determined.

[0045] Based on the alarm category, the corresponding alarm information is pushed.

[0046] This application also proposes a state assessment device, the state assessment device comprising:

[0047] The acquisition module is used to acquire target data for the photovoltaic inverter;

[0048] The evaluation module is used to input the target data into the trained state evaluation model to obtain the state evaluation information of the photovoltaic inverter. The trained state evaluation model is obtained by training an initial support vector machine model based on the sample data of the photovoltaic inverter and preset optimization parameters. The sample data is obtained based on the index system of the photovoltaic inverter. The index system is constructed based on the quantitative and qualitative parameters of the photovoltaic inverter.

[0049] This application also proposes a terminal device, which includes a memory, a processor, and a state evaluation program stored in the memory and executable on the processor. When the state evaluation program is executed by the processor, it implements the steps of the state evaluation method described above.

[0050] This application also proposes a computer-readable storage medium storing a state evaluation program, which, when executed by a processor, implements the steps of the state evaluation method described above.

[0051] The state assessment method, apparatus, terminal device, and storage medium proposed in this application acquire target data of a photovoltaic inverter; input the target data into a trained state assessment model to obtain state assessment information of the photovoltaic inverter. The trained state assessment model is obtained by training an initial support vector machine model based on sample data of the photovoltaic inverter and preset optimization parameters. The sample data is obtained based on the indicator system of the photovoltaic inverter, which is constructed based on the quantitative and qualitative parameters of the photovoltaic inverter. Based on this application's solution, combining quantitative and qualitative parameters to construct the indicator system can fully mine and utilize relevant data, compensate for the shortcomings of single data types, and enable the trained state assessment model to have a more comprehensive state assessment capability, improving the accuracy and reliability of photovoltaic inverter state assessment. Attached Figure Description

[0052] Figure 1 This is a schematic diagram of the functional modules of the terminal equipment to which the status assessment device of this application belongs;

[0053] Figure 2 This is a schematic diagram of the first exemplary embodiment of the status assessment method of this application;

[0054] Figure 3 This is a schematic diagram of the second exemplary embodiment of the status assessment method of this application;

[0055] Figure 4 This is a schematic diagram of the third exemplary embodiment of the status assessment method of this application;

[0056] Figure 5 This is a schematic diagram of the fourth exemplary embodiment of the status assessment method of this application;

[0057] Figure 6 This is a schematic diagram of the fifth exemplary embodiment of the status assessment method of this application;

[0058] Figure 7 This is a schematic diagram of the sixth exemplary embodiment of the status assessment method of this application;

[0059] Figure 8 This is a schematic diagram of the seventh exemplary embodiment of the status assessment method of this application;

[0060] Figure 9 This is a schematic diagram of the eighth exemplary embodiment of the status assessment method of this application;

[0061] Figure 10 This is a schematic diagram of the model training involved in the state assessment method of this application;

[0062] Figure 11 This is a schematic diagram of the ninth exemplary embodiment of the status assessment method of this application;

[0063] Figure 12 This is a schematic diagram of the tenth exemplary embodiment of the status assessment method of this application.

[0064] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0065] It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit this application.

[0066] The main solution of this application embodiment is: acquiring target data of a photovoltaic inverter; inputting the target data into a trained state assessment model to obtain state assessment information of the photovoltaic inverter. The trained state assessment model is obtained by training an initial support vector machine model based on sample data of the photovoltaic inverter and preset optimization parameters. The sample data is obtained based on the indicator system of the photovoltaic inverter, and the indicator system is constructed based on the quantitative and qualitative parameters of the photovoltaic inverter. Based on this application solution, combining quantitative and qualitative parameters to construct the indicator system can fully mine and utilize relevant data, compensate for the shortcomings of single data types, and enable the trained state assessment model to have a more comprehensive state assessment capability, thereby improving the accuracy and reliability of photovoltaic inverter state assessment.

[0067] Specifically, refer to Figure 1 , Figure 1 This is a schematic diagram of the functional modules of the terminal device to which the status assessment device belongs in this application. The status assessment device can be an independent device capable of performing status assessments, and it can be implemented on the terminal device in hardware or software form. The terminal device can be a smart mobile terminal with data processing capabilities, such as a mobile phone or tablet computer, or it can be a fixed terminal device or server with data processing capabilities.

[0068] In this embodiment, the terminal device to which the status assessment device belongs includes at least an output module 110, a processor 120, a memory 130, and a communication module 140.

[0069] The memory 130 stores the operating system and the state assessment program. The state assessment device can acquire target data of the photovoltaic inverter; input the target data into the trained state assessment model, and store the obtained state assessment information of the photovoltaic inverter in the memory 130. The output module 110 can be a display screen, etc. The communication module 140 can include a WIFI module, a mobile communication module, and a Bluetooth module, etc., and communicates with external devices or servers through the communication module 140.

[0070] When the state evaluation program in memory 130 is executed by the processor, it performs the following steps:

[0071] Obtain target data for the photovoltaic inverter;

[0072] The target data is input into the trained state evaluation model to obtain the state evaluation information of the photovoltaic inverter. The trained state evaluation model is obtained by training an initial support vector machine model based on the sample data of the photovoltaic inverter and preset optimization parameters. The sample data is obtained based on the index system of the photovoltaic inverter. The index system is constructed based on the quantitative and qualitative parameters of the photovoltaic inverter.

[0073] Furthermore, when the state evaluation program in memory 130 is executed by the processor, it also performs the following steps:

[0074] Obtain at least one quantitative parameter and at least one qualitative parameter of the photovoltaic inverter;

[0075] Based on at least one quantitative parameter and at least one qualitative parameter of the photovoltaic inverter, a basic index system for the photovoltaic inverter is constructed.

[0076] Based on the basic index system of the photovoltaic inverter, the key index system of the photovoltaic inverter is constructed.

[0077] The sample data was obtained based on the key indicator system of the photovoltaic inverter.

[0078] The initial support vector machine model is trained based on the sample data and the preset optimization parameters to obtain the trained state evaluation model.

[0079] Furthermore, when the state evaluation program in memory 130 is executed by the processor, it also performs the following steps:

[0080] Obtain the raw data from the photovoltaic inverter;

[0081] The raw data is filtered according to the key parameter system of the photovoltaic inverter to obtain the filtered raw data;

[0082] The filtered raw data is preprocessed to obtain the target data.

[0083] Furthermore, when the state evaluation program in memory 130 is executed by the processor, it also performs the following steps:

[0084] Based on a preset first analysis rule, at least one quantitative parameter of the photovoltaic inverter is analyzed to obtain the sensitivity factor corresponding to each of the at least one quantitative parameter.

[0085] Based on a preset second analysis rule, at least one qualitative parameter of the photovoltaic inverter is analyzed to obtain the degradation factor corresponding to each of the at least one qualitative parameter.

[0086] Based on the at least one quantitative parameter, the corresponding sensitivity factor of the at least one quantitative parameter, the at least one qualitative parameter, and the corresponding degradation factor of the at least one qualitative parameter, the basic index system of the photovoltaic inverter is constructed.

[0087] Furthermore, when the state evaluation program in memory 130 is executed by the processor, it also performs the following steps:

[0088] Based on the quantitative parameter samples corresponding to at least one quantitative parameter under each feature type, the average intra-class distance and average inter-class distance corresponding to at least one quantitative parameter under each feature type are calculated.

[0089] Based on the average intra-class distance and the average inter-class distance, the sensitivity factor corresponding to at least one quantitative parameter under each feature type is calculated.

[0090] Furthermore, when the state evaluation program in memory 130 is executed by the processor, it also performs the following steps:

[0091] Based on the actual value, threshold, and factory value of at least one qualitative parameter under each feature type, and based on a preset offset calculation rule, the degradation factor corresponding to at least one qualitative parameter under each feature type is calculated.

[0092] Furthermore, when the state evaluation program in memory 130 is executed by the processor, it also performs the following steps:

[0093] Based on the basic index system of the photovoltaic inverter, at least one quantitative parameter and at least one qualitative parameter of the photovoltaic inverter are screened to obtain at least one preliminary quantitative parameter and at least one preliminary qualitative parameter.

[0094] Based on the preset Gaussian kernel function, the quantitative parameter samples corresponding to each of the at least one initial screening quantitative parameter, and the qualitative parameter samples corresponding to each of the at least one initial screening qualitative parameter, a normalized kernel matrix is ​​constructed.

[0095] Based on the normalized kernel matrix, the corresponding feature matrix and feature vector matrix are constructed.

[0096] Based on the normalized kernel matrix, the feature matrix, and the feature vector matrix, the target matrix is ​​constructed.

[0097] Based on the target matrix, a key indicator system for the photovoltaic inverter is constructed.

[0098] Furthermore, when the state evaluation program in memory 130 is executed by the processor, it also performs the following steps:

[0099] The sample data is preprocessed to obtain training samples and test samples;

[0100] The initial support vector machine model is trained based on the training samples;

[0101] Determine whether the initial support vector machine model meets the preset cross-validation conditions;

[0102] If the initial support vector machine model satisfies the cross-validation condition, then training of the initial support vector machine model is stopped, and the state evaluation model to be tested is obtained.

[0103] If the initial support vector machine model does not meet the cross-validation condition, the initial support vector machine model is trained based on the training samples and preset optimization parameters until the initial support vector machine model meets the cross-validation condition. Then, the training of the initial support vector machine model is stopped, and the state evaluation model to be tested is obtained.

[0104] The state evaluation model to be tested is tested based on the test samples.

[0105] If the state evaluation model to be tested passes the test, then the trained state evaluation model is obtained.

[0106] Furthermore, when the state evaluation program in memory 130 is executed by the processor, it also performs the following steps:

[0107] Based on the preset Grey Wolf algorithm, at least one optimization parameter is initialized.

[0108] Furthermore, when the state evaluation program in memory 130 is executed by the processor, it also performs the following steps:

[0109] If the status assessment information of the photovoltaic inverter meets the preset alarm conditions, then the alarm category corresponding to the status assessment information of the photovoltaic inverter is determined.

[0110] Based on the alarm category, the corresponding alarm information is pushed.

[0111] This embodiment, through the above-described scheme, specifically acquires target data of the photovoltaic inverter; inputs the target data into a trained state assessment model to obtain the state assessment information of the photovoltaic inverter. The trained state assessment model is obtained by training an initial support vector machine model based on sample data of the photovoltaic inverter and preset optimization parameters. The sample data is obtained based on the indicator system of the photovoltaic inverter, which is constructed based on the quantitative and qualitative parameters of the photovoltaic inverter. In this embodiment, combining quantitative and qualitative parameters to construct the indicator system can fully mine and utilize relevant data, compensate for the shortcomings of a single data type, and enable the trained state assessment model to have a more comprehensive state assessment capability, improving the accuracy and reliability of photovoltaic inverter state assessment.

[0112] Reference Figure 2 The first embodiment of the status assessment method of this application provides a flowchart, the status assessment method including:

[0113] Step S10: Obtain the target data for the photovoltaic inverter.

[0114] Specifically, as a key component of a solar power generation system, the health of the photovoltaic (PV) inverter directly affects the safety and stability of the system. Therefore, the need for condition assessment of PV inverter data has emerged. During long-term operation, PV inverters often experience abnormal operating conditions. Traditional maintenance methods mainly include regular manual inspections and post-incident troubleshooting. However, these methods are not only costly but also waste maintenance resources and negatively impact the stability of the entire power generation system.

[0115] With the continuous development of IoT and big data technologies, new detection equipment and sensors are widely used, resulting in photovoltaic inverter data exhibiting large-scale, multi-source, and multi-dimensional big data characteristics. Although the industry has begun to conduct simple condition monitoring and assessment of photovoltaic inverter data, the sheer volume of data makes collection and mining difficult. This leads to inaccuracies in equipment data feature mining, coarse condition assessments, and insufficient real-time effectiveness of fault diagnosis. These shortcomings hinder the efficient implementation of condition-based maintenance for photovoltaic inverters and limit the development of reasonable maintenance strategies. Currently, in photovoltaic inverter condition assessment methods, external characteristic data can be obtained by measuring the inputs and outputs of the photovoltaic inverter system. However, to understand the dynamic laws of photovoltaic inverters, internal state variables are needed, which are usually not directly measurable.

[0116] Therefore, this embodiment proposes a new condition assessment method for photovoltaic inverters. First, it is necessary to obtain the target data required for condition assessment, which may include quantitative parameter data and qualitative parameter data.

[0117] In photovoltaic (PV) inverters, quantitative parameters typically refer to parameters that can be expressed with specific numerical values, such as voltage, current, and power, used to describe the inverter's operating performance and status. The target data may include one or more of the following quantitative parameters: ① Input DC voltage and current: The DC voltage and current of the PV cell string input to the inverter. ② Output AC voltage and current: The AC voltage and current output by the inverter. ③ Maximum power point (MPP) voltage and current: The optimal operating point of the inverter, corresponding to the maximum output power. ④ Efficiency: The inverter's energy conversion efficiency, representing the conversion efficiency from DC to AC. ⑤ Temperature: The internal temperature of the inverter, including the temperature of the heat sink, electronic components, etc. ⑥ Frequency: The frequency of the AC output from the inverter. ⑦ Operating time: The duration of inverter operation.

[0118] In addition, the qualitative parameters of photovoltaic inverters are descriptive parameters that cannot be directly represented by numerical values. They are usually expressed in words or categories, such as fault type and operating status. The target data includes one or more of the following qualitative parameters: ① Fault type: Indicates potential inverter faults, such as overvoltage, overcurrent, and short circuit. ② Operating status: Whether the inverter is currently in normal operation, standby, or fault state. ③ Shutdown record: Records the inverter's downtime and cause. ④ Warning information: Warning information indicating potential inverter faults. ⑤ Operating environment: The environmental conditions of the inverter, such as weather, temperature, and humidity. ⑥ Maintenance record: The inverter's maintenance history, including repair and maintenance records. ⑦ Instruction manual: The instruction manual may describe various operating modes of the inverter, wiring methods, button and display function descriptions, and the meaning of alarm codes. A comprehensive analysis of both quantitative and qualitative parameters provides a more complete understanding of the inverter's status and performance, helping to assess the inverter's health and operating condition.

[0119] It is worth noting that the target data should be acquired in a timely manner to ensure that the final condition assessment information is as synchronized as possible with the actual operating state of the photovoltaic inverter. Furthermore, the target data should conform to the input specifications of the condition assessment model.

[0120] Step S20: Input the target data into the trained state evaluation model to obtain the state evaluation information of the photovoltaic inverter. The trained state evaluation model is obtained by training an initial support vector machine model based on the sample data of the photovoltaic inverter and preset optimization parameters. The sample data is obtained based on the index system of the photovoltaic inverter. The index system is constructed based on the quantitative and qualitative parameters of the photovoltaic inverter.

[0121] The target data requiring state assessment is input into the trained state assessment model. During this process, the trained model analyzes and processes the input target data based on its previous training experience and learned patterns, thereby generating state assessment information about the photovoltaic inverter's condition.

[0122] Quantitative parameters directly reflect the performance and operation of the photovoltaic inverter, while qualitative parameters describe its various attributes. Combining these parameters constructs a suitable index system for photovoltaic inverters, which provides sample data closely related to inverter condition assessment. Furthermore, selecting appropriate optimization parameters ensures the model better fits the data and has higher predictive power. Considering the advantages of support vector machine (SVM) models—high robustness, suitability for nonlinear data, insensitivity to outliers, strong generalization performance, and good interpretability—an initial SVM model can be selected as the training object. Thus, with the advantages of the index system, sample data, optimization parameters, and SVM model, a well-trained condition assessment model can more comprehensively understand the state of the photovoltaic inverter and provide high-quality condition assessment information.

[0123] Understandably, the condition assessment information reflects the internal and external conditions of the photovoltaic inverter, providing a more comprehensive reference for operation and maintenance decisions regarding the photovoltaic inverter.

[0124] This embodiment, through the above-described scheme, specifically acquires target data of the photovoltaic inverter; inputs the target data into a trained state assessment model to obtain the state assessment information of the photovoltaic inverter. The trained state assessment model is obtained by training an initial support vector machine model based on sample data of the photovoltaic inverter and preset optimization parameters. The sample data is obtained based on the indicator system of the photovoltaic inverter, which is constructed based on the quantitative and qualitative parameters of the photovoltaic inverter. In this embodiment, combining quantitative and qualitative parameters to construct the indicator system can fully mine and utilize relevant data, compensate for the shortcomings of a single data type, and enable the trained state assessment model to have a more comprehensive state assessment capability, improving the accuracy and reliability of photovoltaic inverter state assessment.

[0125] Furthermore, referring to Figure 3 The second embodiment of the status assessment method of this application provides a flowchart, based on the above. Figure 2 In the embodiment shown, before obtaining the target data of the photovoltaic inverter in step S10, the method further includes:

[0126] Step S01: Obtain at least one quantitative parameter and at least one qualitative parameter of the photovoltaic inverter.

[0127] Specifically, to enable the trained state assessment model to possess a more comprehensive state assessment capability, it is necessary to obtain at least one quantitative parameter and at least one qualitative parameter of the photovoltaic inverter in advance. It is understandable that the relevant quantitative or qualitative parameters will differ for different characteristics or indicators of the photovoltaic inverter; therefore, not all of the acquired quantitative or qualitative parameters may be used for model training.

[0128] Step S02: Based on at least one quantitative parameter and at least one qualitative parameter of the photovoltaic inverter, construct the basic index system of the photovoltaic inverter.

[0129] Specifically, a basic index system for photovoltaic inverters is constructed based on at least one quantitative parameter and at least one qualitative parameter obtained from the photovoltaic inverter. This basic index system includes, on the one hand, at least one quantitative parameter and its corresponding sensitivity factor, and on the other hand, at least one qualitative parameter and its corresponding degradation factor. The sensitivity factor is used to strengthen key quantitative parameters and weaken non-key quantitative parameters; the degradation factor is used to strengthen key qualitative parameters and weaken non-key qualitative parameters.

[0130] Understandably, under the basic indicator system, the impact of both quantitative and qualitative parameters on condition assessment has been quantified, which can provide a basis for condition assessment and monitoring of photovoltaic inverters, and help to achieve more accurate and effective operation analysis and maintenance decisions.

[0131] Step S03: Based on the basic index system of the photovoltaic inverter, construct the key index system of the photovoltaic inverter.

[0132] Specifically, since the basic index system of photovoltaic inverters quantifies the impact of various quantitative and qualitative parameters on the state assessment, based on the basic index system, some key quantitative and qualitative parameters can be selected to construct the key index system of photovoltaic inverters.

[0133] More specifically, the selection of key quantitative and qualitative parameters can be referenced using sensitivity factors and degradation factors. The magnitude of the sensitivity factor is positively correlated with the sensitivity of the quantitative parameter; a larger sensitivity factor value indicates greater sensitivity, and a smaller value indicates less sensitivity. The magnitude of the degradation factor is positively correlated with the degree of degradation of the qualitative parameter and negatively correlated with the operating state of the photovoltaic inverter; a smaller degradation factor value indicates lower degradation of the qualitative parameter and better operating state of the photovoltaic inverter, while a larger degradation factor value indicates higher degradation of the qualitative parameter and worse operating state of the photovoltaic inverter.

[0134] Step S04: Obtain the sample data according to the key indicator system of the photovoltaic inverter.

[0135] Specifically, after the selection process described above, the key indicator system includes at least one key quantitative parameter and at least one key qualitative parameter. At least one key quantitative parameter and at least one key qualitative parameter are crucial for the condition assessment of photovoltaic inverters.

[0136] Based on the acquisition of multiple internal and external operating data of the photovoltaic inverter, the multiple internal and external operating data are screened according to the key indicator system to obtain sample data related to at least one key quantitative parameter and sample data related to at least one key qualitative parameter.

[0137] Step S05: Train the initial support vector machine model based on the sample data and the preset optimization parameters to obtain the trained state evaluation model.

[0138] Specifically, based on the sample data related to at least one key quantitative parameter and at least one key qualitative parameter, the initial support vector machine model is continuously trained on the sample data related to at least one key quantitative parameter and at least one key qualitative parameter until the training is completed, and the trained state evaluation model can be obtained.

[0139] The initial support vector machine (SVM) model is built based on a pre-defined SVM algorithm. SVM is a machine learning algorithm used for classification and regression tasks. When building the initial SVM model, the SVM algorithm is selected, and the parameters are set according to the characteristics of the sample data and the problem requirements. Then, the model parameters are adjusted using sample data to enable the initial SVM model to better classify or regress the sample data. After training, a trained state evaluation model is obtained. This trained state evaluation model is essentially an improved SVM model.

[0140] This embodiment, through the above-described scheme, specifically involves obtaining at least one quantitative parameter and at least one qualitative parameter of the photovoltaic inverter; constructing a basic index system for the photovoltaic inverter based on these parameters; constructing a key index system for the photovoltaic inverter based on the basic index system; obtaining sample data according to the key index system; and training the initial support vector machine model based on the sample data and preset optimization parameters to obtain the trained state evaluation model. In this embodiment, by obtaining the quantitative and qualitative parameters of the photovoltaic inverter, a basic index system is constructed, which in turn forms a key index system. Sample data is obtained through the key index system, and the initial support vector machine model is trained based on the sample data and optimization parameters to finally obtain a trained state evaluation model. Thus, the generalization ability of the trained state evaluation model can be improved based on the above steps, giving it a more comprehensive state evaluation capability and achieving accurate state evaluation of the photovoltaic inverter.

[0141] Furthermore, referring to Figure 4 The third embodiment of the status assessment method of this application provides a flowchart, based on the above. Figure 3 In the embodiment shown, step S10, further refining the target data of the photovoltaic inverter, includes:

[0142] Step S101: Obtain the raw data of the photovoltaic inverter.

[0143] Specifically, before conducting a condition assessment of a photovoltaic (PV) inverter, it is necessary to obtain the inverter's raw data. This raw data can consist of various internal and external operational data points. Typically, the raw data is relatively large in scale, and some of it will not actually be used in the PV inverter's condition assessment process.

[0144] For example, raw data may include: 1. Current and voltage data: including DC input current and voltage and AC output current and voltage, reflecting the conversion and output of electrical energy. 2. Power data: including DC input power and AC output power, used to evaluate the power conversion efficiency of the photovoltaic inverter. 3. Temperature data: internal and external temperature data of the photovoltaic inverter, used to monitor the thermal management of the equipment and potential overheating. 4. Frequency data: AC output frequency data, used to ensure that the electrical energy output by the photovoltaic inverter conforms to the standard frequency. 5. Operating status data: including the switching status, operating mode, fault status, etc. of the photovoltaic inverter, used to evaluate the operating status and performance of the equipment. 6. DC input and AC output waveform data: current and voltage waveform data, used to analyze power quality and detect possible waveform distortion. 7. Communication data: communication data with other equipment or monitoring systems, used for remote monitoring and control of the photovoltaic inverter. 8. Event log: records equipment events and fault information for fault diagnosis and maintenance. 9. Environmental Condition Data: Includes environmental condition data such as light intensity, temperature, and humidity, which can affect the performance of the photovoltaic inverter. 10. Timestamp Data: Records the timestamps of data acquisition for time series analysis and data alignment. 11. Operating Mode: Describes the current operating mode of the photovoltaic inverter, such as normal operation, fault state, or shutdown. 12. Fault Codes: If the photovoltaic inverter is in a fault state, includes the corresponding fault code and description for problem diagnosis. 13. Maintenance / Inspection Records: Records the date and time of maintenance, inspection, or repair activities; describes the type of maintenance activity, such as periodic inspection, fault repair, or preventative maintenance; records the details of the maintenance or inspection performed, including the items inspected, problems found, and measures taken; describes the results of the maintenance activities, including whether the problem was resolved and whether performance improved; if the photovoltaic inverter has previously experienced a fault, records the relevant fault code, repair measures, and repair date.

[0145] The examples of raw data above are only for the purpose of understanding the content of raw data and are not intended to limit the scope of raw data. Raw data may include one or more of the examples above, as well as data other than those listed above. It is understood that raw data may include both quantitative and qualitative data.

[0146] Step S102: The original data is filtered according to the key parameter system of the photovoltaic inverter to obtain the filtered original data.

[0147] Specifically, the key indicator system includes at least one key quantitative parameter and at least one key qualitative parameter. At least one key quantitative parameter and at least one key qualitative parameter are crucial for the condition assessment of photovoltaic inverters.

[0148] Based on the raw data obtained from the photovoltaic inverter, the raw data is filtered according to a key indicator system to obtain the filtered raw data. The filtered raw data includes data related to at least one key quantitative parameter and data related to at least one key qualitative parameter.

[0149] Step S103: Preprocess the filtered raw data to obtain the target data.

[0150] Specifically, in order to ensure that the filtered raw data meets the input specifications of the trained state evaluation model, the filtered raw data needs to be preprocessed.

[0151] Preprocessing can include one or more of the following steps: ① Data cleaning: Checking for missing values, outliers, or other incomplete or erroneous data, and then repairing, deleting, or imputing them as needed. ② Data transformation: Transforming the data to meet the requirements of the model. For example, standardizing (making the mean 0 and the standard deviation 1) or normalizing (scaling the data to a specific range) continuous data. ③ Feature selection: Selecting appropriate features or attributes according to the needs of the problem. Selection can be based on the importance and relevance of features to reduce unnecessary dimensionality. ④ Feature extraction: Extracting more meaningful features from the original data. For example, extracting statistical features from time series data or texture features from images. ⑤ Data balancing: When dealing with classification problems, it may be necessary to handle class imbalance, such as balancing the number of samples from different classes through undersampling or oversampling. ⑥ Data encoding: Converting categorical data into a numerical representation so that the model can process it. For example, one-hot encoding of class labels. ⑦ Data dimensionality reduction: Reducing the dimensionality of the data to reduce computational complexity and the risk of model overfitting. Common methods include Principal Component Analysis (PCA).

[0152] After preprocessing the filtered raw data, the target data can be obtained. Understandably, the goal of preprocessing is to clean, transform, and normalize the data. The preprocessing process helps improve the performance and stability of the model, enabling it to better adapt to real-world data and perform accurate state assessments.

[0153] This embodiment, through the above-described scheme, specifically involves acquiring the raw data of the photovoltaic inverter; filtering the raw data according to the key parameter system of the photovoltaic inverter to obtain filtered raw data; and preprocessing the filtered raw data to obtain the target data. In this embodiment, filtering and preprocessing the raw data based on the key parameter system can yield target data that conforms to the input specifications of the trained state assessment model, thereby obtaining more reliable state assessment results based on the target data.

[0154] Furthermore, referring to Figure 5 The fourth embodiment of the status assessment method of this application provides a flowchart, based on the above. Figure 3 In the illustrated embodiment, step S02, based on at least one quantitative parameter and at least one qualitative parameter of the photovoltaic inverter, further refines the basic index system of the photovoltaic inverter, including:

[0155] Step S021: Analyze at least one quantitative parameter of the photovoltaic inverter based on a preset first analysis rule to obtain the sensitivity factor corresponding to each of the at least one quantitative parameter.

[0156] Specifically, to reveal the importance of a certain quantitative parameter for the condition assessment of a photovoltaic inverter, a sensitivity factor calculation method can be used. Based on a preset first analysis rule, at least one quantitative parameter of the photovoltaic inverter can be analyzed to obtain the sensitivity factor corresponding to each of the at least one quantitative parameter. The sensitivity factor is expressed in numerical form.

[0157] Step S022: Analyze at least one qualitative parameter of the photovoltaic inverter based on the preset second analysis rule to obtain the degradation factor corresponding to each of the at least one qualitative parameter.

[0158] Specifically, to reveal the importance of a certain qualitative parameter for the condition assessment of a photovoltaic inverter, a degradation factor can be calculated. Based on a preset second analysis rule, at least one qualitative parameter of the photovoltaic inverter can be analyzed to obtain the degradation factor corresponding to each of the at least one qualitative parameter. The degradation factor is expressed in numerical form.

[0159] Step S023: Based on the at least one quantitative parameter, the sensitivity factor corresponding to each of the at least one quantitative parameter, the at least one qualitative parameter, and the degradation factor corresponding to each of the at least one qualitative parameter, the basic index system of the photovoltaic inverter is constructed.

[0160] Specifically, a basic index system for photovoltaic inverters can be constructed based on at least one quantitative parameter, at least one corresponding sensitivity factor for each quantitative parameter, at least one qualitative parameter, and at least one corresponding degradation factor for each qualitative parameter. The magnitude of the sensitivity factor is positively correlated with the sensitivity of the quantitative parameter, and is used to strengthen key quantitative parameters while weakening non-key quantitative parameters. Similarly, the magnitude of the degradation factor is positively correlated with the degree of degradation of the qualitative parameter, and is used to strengthen key qualitative parameters while weakening non-key qualitative parameters.

[0161] This embodiment, through the above-described scheme, specifically analyzes at least one quantitative parameter of the photovoltaic inverter based on a preset first analysis rule to obtain the sensitivity factor corresponding to each of the at least one quantitative parameter; analyzes at least one qualitative parameter of the photovoltaic inverter based on a preset second analysis rule to obtain the degradation factor corresponding to each of the at least one qualitative parameter; and constructs a basic index system for the photovoltaic inverter based on the at least one quantitative parameter, the corresponding sensitivity factor, the at least one qualitative parameter, and the corresponding degradation factor. In this embodiment, sensitivity factors and degradation factors are introduced to construct the basic index system for the photovoltaic inverter. Under this basic index system, the influence of both quantitative and qualitative parameters on condition assessment is quantified, providing a basis for condition assessment and monitoring of the photovoltaic inverter, and helping to achieve more accurate and effective operation analysis and maintenance decisions.

[0162] Furthermore, referring to Figure 6 The fifth embodiment of the status assessment method of this application provides a flowchart, based on the above. Figure 5 In the illustrated embodiment, the at least one quantitative parameter belongs to its corresponding feature type, the number of feature types is at least one, and each of the at least one quantitative parameter includes at least one quantitative parameter sample. Step S021 involves analyzing the at least one quantitative parameter of the photovoltaic inverter based on a preset first analysis rule to obtain further refined sensitivity factors corresponding to each of the at least one quantitative parameter, including:

[0163] Step S0211: Based on the quantitative parameter samples corresponding to at least one quantitative parameter under each feature type, calculate the average intra-class distance and average inter-class distance corresponding to at least one quantitative parameter under each feature type.

[0164] Specifically, feature types refer to different types or categories of features or attributes. These feature types can be selected and defined according to specific circumstances to better describe the state characteristics of photovoltaic inverters. Each feature type may affect the state of the photovoltaic inverter, therefore, the corresponding quantitative parameters need to be analyzed during state assessment.

[0165] The intra-class distance of the j-th quantitative parameter for the i-th characteristic type of the photovoltaic inverter and the average intra-class distance of the j-th quantitative parameter.

[0166]

[0167]

[0168] Among them, D ijN is the intra-class distance of the j-th quantitative parameter of the i-th feature type; i f is the number of samples for the quantitative parameter of the i-th feature type in a certain fault; ij (m), f ij (n) represent the m-th and n-th samples in the j-th quantitative parameter of the i-th feature type, respectively; d 2 f ij (m) and f ij The square of the distance d between (n); N is the number of feature types;

[0169] Calculate the average inter-class distance D of the j-th quantitative parameter of the i-th characteristic type of the inverter. j :

[0170]

[0171]

[0172] in, These are the m-th and n-th sample averages of the j-th quantitative parameter for the i-th feature type, respectively.

[0173] Step S0212: Based on the average intra-class distance and the average inter-class distance, calculate the sensitivity factor corresponding to at least one quantitative parameter under each feature type.

[0174] Specifically, after calculating the above average intra-class distance... and the average inter-class distance D j Then, the sensitivity factor α of the j-th quantitative parameter can be further calculated. j :

[0175]

[0176] Where, α j The larger the value, the more sensitive the j-th quantitative parameter is, and the more reasonable it is to select the j-th quantitative parameter for photovoltaic inverter condition assessment.

[0177] This embodiment, through the above-described scheme, specifically calculates the average intra-class distance and average inter-class distance corresponding to at least one quantitative parameter under each feature type based on the corresponding quantitative parameter samples; and calculates the sensitivity factor corresponding to at least one quantitative parameter under each feature type based on the average intra-class distance and the average inter-class distance. In this embodiment, the sensitivity factor corresponding to the quantitative parameter is calculated based on the average intra-class distance and the average inter-class distance. The sensitivity factor is used to strengthen key quantitative parameters and weaken non-key quantitative parameters, helping to achieve more accurate and effective operation analysis and maintenance decisions.

[0178] Furthermore, referring to Figure 7 The sixth embodiment of the status assessment method of this application provides a flowchart, based on the above. Figure 6 In the illustrated embodiment, the at least one qualitative parameter belongs to its corresponding feature type, and each of the at least one qualitative parameter includes an actual value, a threshold value, and a factory value. Step S022 involves analyzing the at least one qualitative parameter of the photovoltaic inverter based on a preset second analysis rule to obtain further refined degradation factors corresponding to each of the at least one qualitative parameter, including:

[0179] Step S0221: Based on the actual value, threshold, and factory value of the qualitative parameter corresponding to at least one qualitative parameter under each feature type, and based on the preset offset calculation rules, calculate the degradation factor corresponding to at least one qualitative parameter under each feature type.

[0180] Specifically, feature types refer to different types or categories of features or attributes. These feature types can be selected and defined according to specific circumstances to better describe the state characteristics of photovoltaic inverters. Each feature type may affect the state of the photovoltaic inverter, therefore, the corresponding qualitative parameters need to be analyzed during state assessment.

[0181] For a certain characteristic parameter of a photovoltaic inverter, the formula for calculating the degradation factor β under the preset offset calculation rules is as follows:

[0182]

[0183] Where, x i It is the actual value of the qualitative parameter, x k It is a qualitative parameter threshold, x b These are qualitative parameters and their factory values. x i -x k x represents the first-type offset of the actual value of a qualitative parameter relative to its threshold value. b -x k The second type of deviation is the qualitative parameter's factory value relative to its threshold. Therefore, the degradation factor β obtained by dividing the first type of deviation by the second type of deviation characterizes the relative degree of degradation between the actual state and the factory state related to this qualitative parameter. The smaller the degradation factor β of the qualitative parameter, the better the equipment condition of the photovoltaic inverter.

[0184] This embodiment, through the above-described scheme, specifically calculates the degradation factor corresponding to at least one qualitative parameter under each feature type based on the actual value, threshold value, and factory value of the qualitative parameter, and according to a preset offset calculation rule. In this embodiment, the degradation factor corresponding to the qualitative parameter is calculated based on the actual value, threshold value, and factory value of the qualitative parameter. The degradation factor is used to strengthen critical qualitative parameters and weaken non-critical qualitative parameters, helping to achieve more accurate and effective operation analysis and maintenance decisions.

[0185] Furthermore, referring to Figure 8 The seventh embodiment of the status assessment method of this application provides a flowchart, based on the above. Figure 3 In the embodiment shown, step S03 involves further refining the key indicator system of the photovoltaic inverter based on its fundamental indicator system, including:

[0186] Step S031: Based on the basic index system of the photovoltaic inverter, at least one quantitative parameter and at least one qualitative parameter of the photovoltaic inverter are screened to obtain at least one preliminary quantitative parameter and at least one preliminary qualitative parameter.

[0187] Specifically, under the basic index system of photovoltaic inverters, the impact of both quantitative and qualitative parameters on condition assessment has been quantified. Therefore, based on sensitivity factors and degradation factors, some quantitative and qualitative parameters can be selected as initial screening quantitative and qualitative parameters. At least one initial screening quantitative parameter and one initial screening qualitative parameter are required.

[0188] Step S032: Based on the preset Gaussian kernel function, the quantitative parameter samples corresponding to each of the at least one initial screening quantitative parameter, and the qualitative parameter samples corresponding to each of the at least one initial screening qualitative parameter, a normalized kernel matrix is ​​constructed.

[0189] Specifically, based on the quantitative parameter samples corresponding to at least one initial screening quantitative parameter and the qualitative parameter samples corresponding to at least one initial screening qualitative parameter, the original kernel matrix K is constructed using a preset Gaussian kernel function.

[0190] Then, by normalizing the original kernel matrix K, we can obtain the normalized kernel matrix K'.

[0191] Step S033: Based on the normalized kernel matrix, construct the corresponding feature matrix and feature vector matrix.

[0192] Specifically, based on the normalized kernel matrix K', the eigenma matrix K1 and the eigenvector matrix K2 in diagonal matrix form can be constructed.

[0193] Step S034: Based on the normalized kernel matrix, the feature matrix, and the feature vector matrix, the target matrix is ​​constructed.

[0194] Specifically, based on the feature matrix K1, several corresponding eigenvalues ​​are calculated, as well as the contribution rate corresponding to each eigenvalue, where the contribution rate is a certain eigenvalue divided by the sum of all eigenvalues.

[0195] Then, the eigenvalues ​​are sorted in descending order of their contribution rate. The eigenvalues ​​at the top of the list are selected, and the number of eigenvalues ​​to retain can be determined based on a preset threshold or cumulative contribution rate. A new matrix K11 is then constructed based on the selected eigenvalues.

[0196] Then, the eigenvectors of the eigenvector matrix K2 are sorted according to the eigenvalue order of matrix K11 to obtain matrix K21. Dividing each row of matrix K21 by the corresponding eigenvalue of matrix K11 yields a new matrix K3.

[0197] Finally, the target matrix KK can be obtained from the normalized kernel matrix K' and matrix K3:

[0198] KK=K3*K'

[0199] Step S035: Based on the target matrix, construct the key indicator system of the photovoltaic inverter.

[0200] Specifically, after obtaining the target matrix KK, each column (or dimension) of the target matrix KK can be sorted. Further, key indicators can be selected based on several columns of the sorted target matrix KK. For example, the first and second columns of the target matrix KK will correspond to the first and second key indicators, respectively.

[0201] In actual application scenarios, key indicators can be selected by combining the target matrix KK. These key indicators will help to more accurately assess the status of photovoltaic inverters and play an important role in practical applications.

[0202] Understandably, the key indicator system is built upon key indicators, and key indicators have corresponding quantitative and qualitative parameters. In other words, the key indicator system includes key quantitative parameters and key qualitative parameters.

[0203] This embodiment, through the above-described scheme, specifically uses the basic index system of the photovoltaic inverter to screen at least one quantitative parameter and at least one qualitative parameter of the photovoltaic inverter, obtaining at least one initial screening quantitative parameter and at least one initial screening qualitative parameter. Based on a preset Gaussian kernel function, the quantitative parameter samples corresponding to each of the at least one initial screening quantitative parameter, and the qualitative parameter samples corresponding to each of the at least one initial screening qualitative parameter, a normalized kernel matrix is ​​constructed. Based on the normalized kernel matrix, a corresponding feature matrix and eigenvector matrix are constructed. Based on the normalized kernel matrix, the feature matrix, and the eigenvector matrix, a target matrix is ​​constructed. Based on the target matrix, the key index system of the photovoltaic inverter is constructed. In this embodiment, by using the improved kernel nonlinear mapping dimensionality reduction method described above, key indicators are selected from the basic index system, and a key index system is further constructed, which can provide an effective reference for accurately evaluating the state of the photovoltaic inverter.

[0204] Furthermore, referring to Figure 9 The eighth embodiment of the status assessment method of this application provides a flowchart, based on the above. Figure 3 In the embodiment shown, step S05 involves training the initial support vector machine model based on the sample data and the preset optimization parameters to obtain a further refined state evaluation model, including:

[0205] Step S051: Preprocess the sample data to obtain training samples and test samples.

[0206] Specifically, in combination Figure 10 , Figure 10 This is a schematic diagram of model training involved in the state assessment method of this application. After obtaining the sample data, the sample data can be preprocessed. The preprocessing process is similar to step S103 of the third embodiment of the state assessment method of this application. Unlike step S103, the preprocessing process of this embodiment adds a data segmentation step. The data segmentation step refers to dividing the sample data into training samples and test samples so that independent data can be used when training and testing the model.

[0207] Furthermore, the training and test samples can be normalized to eliminate feature scale differences between them during model training, thereby accelerating convergence, improving stability and performance, and maintaining data consistency.

[0208] Step S052: Train the initial support vector machine model based on the training samples.

[0209] Specifically, training samples are used as input to the initial support vector machine model to adjust and train the model so that it can better adapt to the actual data, thereby improving the model's prediction accuracy and performance.

[0210] Step S053: Determine whether the initial support vector machine model meets the preset cross-validation conditions.

[0211] Specifically, after the above training steps are completed, it can be determined whether the initial support vector machine model meets the preset cross-validation conditions. The determination process involves cross-validation.

[0212] Cross-validation is a method for evaluating the performance of machine learning models. It involves dividing an existing dataset into multiple subsets and repeatedly using one subset as the validation set and the remaining subsets as the training set, alternating between them to assess the model's performance on different data subsets. This helps to test the model's generalization ability under different data conditions and avoids overfitting or underfitting. The most common cross-validation method is k-fold cross-validation, which divides the dataset into k subsets, selecting one subset as the validation set each time and using the remaining k-1 subsets as the training set. This process is repeated multiple times, and the average performance metric is calculated to obtain a more accurate evaluation of the model's performance.

[0213] It is understandable that if the initial support vector machine model passes cross-validation, then the cross-validation condition is met; if the initial support vector machine model fails cross-validation, then the cross-validation condition is not met.

[0214] Step S054: If the initial support vector machine model does not meet the cross-validation condition, then the initial support vector machine model is trained based on the training samples and preset optimization parameters until the initial support vector machine model meets the cross-validation condition. Then, the training of the initial support vector machine model is stopped, and the state evaluation model to be tested is obtained.

[0215] Specifically, if the initial support vector machine (SVM) model does not meet the cross-validation criteria, training of the initial SVM model continues. In subsequent training, optimization parameters can be added as input to the initial SVM model based on the training samples. These optimization parameters provide more adjustment options, influencing the learning process and results of the initial SVM model, enabling it to better adapt to the training samples, and improving its generalization ability, accuracy, stability, and reliability.

[0216] Thus, after repeated training until the initial support vector machine model meets the cross-validation conditions, training of the initial support vector machine model can be stopped, and the state evaluation model to be tested can be obtained.

[0217] In some cases, after step S053, if the initial support vector machine model satisfies the cross-validation condition, training of the initial support vector machine model is stopped, resulting in the state evaluation model to be tested. At this point, the state evaluation model to be tested already possesses good generalization performance, but further testing is still required.

[0218] Step S055: Test the state evaluation model to be tested based on the test sample.

[0219] Specifically, test samples are used as input to the state evaluation model under test, and the model is then tested. Correspondingly, the state evaluation model under test outputs state evaluation information from the test samples based on the relationships and patterns it has learned.

[0220] The purpose of test samples is to verify the performance and generalization ability of the state assessment model under test. By observing its performance on test samples, the predictive ability of the state assessment model under test on unseen data can be evaluated, and it can be determined whether the model has good generalization ability, thereby judging the accuracy and reliability of the state assessment model under test in practical applications.

[0221] Step S056: If the state evaluation model to be tested passes the test, the trained state evaluation model is obtained.

[0222] Specifically, if the state evaluation information of the test samples output by the state evaluation model under test meets the preset requirements and can accurately simulate and evaluate the state of the photovoltaic inverter, then the state evaluation model under test can be determined to have passed the test. Furthermore, the state evaluation model that has passed the test is used as the trained state evaluation model. It can be understood that the trained state evaluation model is an improved SVM model, and its optimal parameter set can be expressed as (C, σ... 2 Here, C is the regularization parameter in the trained state evaluation model, often called the soft margin parameter, used to control the model's complexity and tolerance. A smaller C value leads to a looser margin, tolerating some classification errors on the training data and making the model more generalizable. A larger C value leads to a tighter margin, forcing the model to fit the training data better, but may lead to overfitting. σ 2 This is the bandwidth parameter of the kernel function, typically used in conjunction with radial basis function (RBF) kernels. The bandwidth parameter determines the shape of the kernel function and affects the mapping of data in high-dimensional space. A smaller σ... 2 A larger σ value will result in a steeper kernel function, a more concentrated mapping of data points in high-dimensional space, and may make the model more sensitive to noise. 2The value will result in a smoother kernel function and a more dispersed mapping of data points in high-dimensional space, which may improve the generalization ability of the model.

[0223] If the state evaluation information of the test sample output by the state evaluation model under test does not meet the preset requirements, it may be necessary to continue training the state evaluation model under test, or take other measures to change the training steps to make it converge.

[0224] This embodiment, through the above-described scheme, specifically involves preprocessing the sample data to obtain training and test samples; training the initial support vector machine (SVM) model based on the training samples; determining whether the initial SVM model meets preset cross-validation conditions; if the initial SVM model meets the cross-validation conditions, training of the initial SVM model is stopped, resulting in a state evaluation model to be tested; if the initial SVM model does not meet the cross-validation conditions, training of the initial SVM model is continued based on the training samples and preset optimization parameters until the initial SVM model meets the cross-validation conditions, at which point training of the initial SVM model is stopped, resulting in the state evaluation model to be tested; testing the state evaluation model to be tested based on the test samples; if the state evaluation model to be tested passes the test, the trained state evaluation model is obtained. In this embodiment, through steps such as data preprocessing, model training, cross-validation, parameter optimization, and model testing, the trained state evaluation model is ensured to have strong generalization ability, improving the accuracy and reliability of prediction results, and providing a reliable tool for the state evaluation of photovoltaic inverters.

[0225] Furthermore, referring to Figure 11 The ninth embodiment of the status assessment method of this application provides a flowchart, based on the above. Figure 9 In the embodiment shown, before step S054, which involves training the initial support vector machine model based on the training samples and preset optimization parameters until the initial support vector machine model meets the cross-validation conditions, and then stopping the training of the initial support vector machine model to obtain the state evaluation model to be tested, the method further includes:

[0226] Step S057: Based on the preset Grey Wolf algorithm, initialize at least one optimization parameter.

[0227] Specifically, the optimization parameters play a role in adjusting and optimizing model performance during model training, which can improve the model's generalization ability on new data and enhance its robustness and reliability.

[0228] Therefore, the Grey Wolf algorithm is an optimization algorithm that can be used to initialize at least one optimization parameter based on a pre-defined Grey Wolf algorithm. The optimization parameter can include one or more parameters such as scale, scaling factor, crossover probability, and number of iterations in parameter optimization.

[0229] The process of initializing optimization parameters can be as follows: ① Initialize the parameter space: Define the value range of each parameter, which will affect the model's performance. ② Initialize the wolf pack: Set an initial group of gray wolves, each representing a set of parameter values. ③ Calculate fitness: Use an objective function to calculate the fitness of each gray wolf, which represents the model's performance on the training data. ④ Update gray wolf positions: Update the gray wolf positions using algorithmic rules based on the relationship between fitness values ​​and individual gray wolves to better explore the parameter space. ⑤ Find the optimal individual: Determine the optimal individual in the wolf pack based on the fitness value, i.e., the parameter combination with the best performance. ⑥ Update parameters: Use the parameters of the current optimal individual as new parameter values ​​to continue optimizing the model. ⑦ Iterative optimization: Repeat steps 3 to 6 until the preset number of iterations or convergence conditions are reached. ⑧ Obtain the optimal optimization parameters: After the iteration is complete, obtain the optimal combination of optimization parameters for training the model.

[0230] Understandably, in addition to the Grey Wolf algorithm, other algorithms such as grid search, random search, Bayesian optimization, genetic algorithm, particle swarm optimization, simulated annealing, and local optimization can be used to set the initial optimization parameters for the support vector machine model.

[0231] Compared to other algorithms that initialize optimization parameters, the Grey Wolf Algorithm has the following advantages: ① Compared to some other complex algorithms, the Grey Wolf Algorithm has fewer parameters, typically requiring only setting a few parameters such as population size and number of iterations, making parameter tuning easier. ② The Grey Wolf Algorithm achieves a global search of the parameter space by simulating the cooperative behavior of a wolf pack, giving it an advantage in finding the global optimum, especially in high-dimensional parameter spaces. ③ The Grey Wolf Algorithm is adaptive, automatically adjusting its search strategy based on the nature of the problem, making it suitable for various problem types, including continuous parameter optimization and discrete parameter optimization. ④ The Grey Wolf Algorithm has good parallelism, easily running on multi-core processors or in distributed computing environments, accelerating the search process. ⑤ The Grey Wolf Algorithm is generally less sensitive to the choice of initial solution, thus exhibiting robustness to a certain extent and handling the diversity of different problem domains.

[0232] This embodiment uses the above-described scheme, specifically by initializing at least one optimization parameter based on a preset Grey Wolf algorithm. In this embodiment, the Grey Wolf algorithm automatically searches the parameter space, optimizes model performance, improves model accuracy, and accelerates convergence, effectively avoiding the tediousness and uncertainty of manually adjusting parameters.

[0233] Furthermore, referring to Figure 12 The tenth embodiment of the status assessment method of this application provides a flowchart, based on the above. Figure 2 In the embodiment shown, after step S20, which involves inputting the target data into the trained state assessment model to obtain the state assessment information of the photovoltaic inverter, the method further includes:

[0234] Step S30: If the status assessment information of the photovoltaic inverter meets the preset alarm conditions, then determine the alarm category corresponding to the status assessment information of the photovoltaic inverter.

[0235] Specifically, in order to promptly identify faults in photovoltaic inverters and take measures to prevent potential escalation and damage, an alarm can be issued when the condition assessment information of the photovoltaic inverter meets preset alarm conditions.

[0236] Specifically, this embodiment adopts a multi-category alarm mode. First, it is necessary to determine the alarm category corresponding to the status assessment information of the photovoltaic inverter.

[0237] For example, the alarm categories corresponding to the status assessment information are divided into four types: Z1, Z2, Z3, and Z4. Z1 represents the normal operating state of the photovoltaic inverter, with corresponding photovoltaic inverter operating indicators being normal. Z2 represents the general operating state of the photovoltaic inverter, with corresponding photovoltaic inverter operating indicator data showing fluctuations, requiring maintenance personnel to continuously monitor the fluctuations of various indicator data. Z3 represents the unstable operating state of the photovoltaic inverter, with corresponding photovoltaic inverter operating indicator data deviating from normal values, requiring maintenance personnel to analyze the abnormal data and develop corresponding maintenance plans to prevent faults. Z4 represents the abnormal operating state of the photovoltaic inverter, with corresponding photovoltaic inverter operating indicator key data being abnormal, requiring immediate analysis and timely rectification of the corresponding indicators.

[0238] Step S40: Push the corresponding alarm information according to the alarm category.

[0239] Specifically, corresponding alarm information can be pushed to different alarm categories.

[0240] For example, regarding the four alarm categories Z1, Z2, Z3, and Z4 mentioned above, when the current status assessment information of the photovoltaic inverter corresponds to alarm category Z1, the corresponding alarm icon can be displayed in the background; when the current status assessment information of the photovoltaic inverter corresponds to alarm category Z2, the corresponding alarm icon can be displayed in the background and an alarm sound can be emitted; when the current status assessment information of the photovoltaic inverter corresponds to alarm category Z3, the corresponding alarm icon can be displayed in the background, an alarm sound can be emitted, and alarm information can be pushed to the management personnel; when the current status assessment information of the photovoltaic inverter corresponds to alarm category Z4, the corresponding alarm icon can be displayed in the background, an alarm sound can be emitted, and alarm information can be pushed to the management personnel and maintenance personnel near the abnormal photovoltaic inverter. The alarm information can include the location information of the abnormal photovoltaic inverter so that the management personnel or maintenance personnel can go to the site to handle the abnormality in a timely manner.

[0241] The above examples are only to illustrate that alarm information is pushed according to alarm category. The core is that the content of the alarm information is associated with the alarm category, and different alarm categories correspond to different alarm information.

[0242] This embodiment, through the above-described scheme, specifically determines the alarm category corresponding to the photovoltaic inverter's status assessment information if the inverter's status assessment information meets preset alarm conditions; and pushes corresponding alarm information based on the alarm category. In this embodiment, by selectively pushing corresponding alarm information based on the determined alarm category, the problem with the photovoltaic inverter can be accurately indicated, accelerating maintenance response and repair, and effectively reducing photovoltaic inverter failure losses.

[0243] Furthermore, embodiments of this application also propose a state assessment device, the state assessment device comprising:

[0244] The acquisition module is used to acquire target data for the photovoltaic inverter;

[0245] The evaluation module is used to input the target data into the trained state evaluation model to obtain the state evaluation information of the photovoltaic inverter. The trained state evaluation model is obtained by training an initial support vector machine model based on the sample data of the photovoltaic inverter and preset optimization parameters. The sample data is obtained based on the index system of the photovoltaic inverter. The index system is constructed based on the quantitative and qualitative parameters of the photovoltaic inverter.

[0246] The principle and implementation process of state assessment in this embodiment are described in the above embodiments and will not be repeated here.

[0247] Furthermore, this application also proposes a terminal device, which includes a memory, a processor, and a state evaluation program stored in the memory and executable on the processor. When the state evaluation program is executed by the processor, it implements the steps of the state evaluation method described above.

[0248] Since this status assessment program employs all the technical solutions of all the foregoing embodiments when executed by the processor, it has at least all the beneficial effects brought about by all the technical solutions of all the foregoing embodiments, which will not be elaborated here.

[0249] Furthermore, embodiments of this application also propose a computer-readable storage medium storing a state evaluation program, which, when executed by a processor, implements the steps of the state evaluation method as described above.

[0250] Since this status assessment program employs all the technical solutions of all the foregoing embodiments when executed by the processor, it has at least all the beneficial effects brought about by all the technical solutions of all the foregoing embodiments, which will not be elaborated here.

[0251] Compared to existing technologies, the state assessment method, apparatus, terminal device, and storage medium proposed in this application obtain target data of a photovoltaic inverter; input the target data into a trained state assessment model to obtain state assessment information of the photovoltaic inverter. The trained state assessment model is obtained by training an initial support vector machine model based on sample data of the photovoltaic inverter and preset optimization parameters. The sample data is obtained based on the indicator system of the photovoltaic inverter, which is constructed based on the quantitative and qualitative parameters of the photovoltaic inverter. Based on this application's solution, combining quantitative and qualitative parameters to construct the indicator system can fully mine and utilize relevant data, compensate for the shortcomings of single data types, and enable the trained state assessment model to have a more comprehensive state assessment capability, improving the accuracy and reliability of photovoltaic inverter state assessment.

[0252] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0253] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0254] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, controlled terminal, or network device, etc.) to execute the methods of each embodiment of this application.

[0255] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A state assessment method, characterized in that, The state assessment method includes: Obtain target data for the photovoltaic inverter; The target data is input into the trained state evaluation model to obtain the state evaluation information of the photovoltaic inverter. The trained state evaluation model is obtained by training an initial support vector machine model based on the sample data of the photovoltaic inverter and preset optimization parameters. The sample data is obtained based on the key index system of the photovoltaic inverter. The key index system is constructed based on the basic index system of the photovoltaic inverter. The basic index system is constructed based on at least one quantitative parameter, at least one sensitivity factor corresponding to each quantitative parameter, at least one qualitative parameter, and at least one degradation factor corresponding to each qualitative parameter. The value of the sensitivity factor is positively correlated with the sensitivity of the quantitative parameter; the value of the degradation factor is positively correlated with the degradation degree of the qualitative parameter.

2. The state assessment method as described in claim 1, characterized in that, Before the step of acquiring the target data of the photovoltaic inverter, the method further includes: Obtain at least one quantitative parameter and at least one qualitative parameter of the photovoltaic inverter; Based on at least one quantitative parameter and at least one qualitative parameter of the photovoltaic inverter, a basic index system for the photovoltaic inverter is constructed. Based on the basic index system of the photovoltaic inverter, the key index system of the photovoltaic inverter is constructed. The sample data was obtained based on the key indicator system of the photovoltaic inverter. The initial support vector machine model is trained based on the sample data and the preset optimization parameters to obtain the trained state evaluation model. The steps for constructing the basic index system of the photovoltaic inverter based on at least one quantitative parameter and at least one qualitative parameter of the photovoltaic inverter include: Based on a preset first analysis rule, at least one quantitative parameter of the photovoltaic inverter is analyzed to obtain the sensitivity factor corresponding to each of the at least one quantitative parameter. Based on a preset second analysis rule, at least one qualitative parameter of the photovoltaic inverter is analyzed to obtain the degradation factor corresponding to each of the at least one qualitative parameter. Based on the at least one quantitative parameter, the corresponding sensitivity factor of the at least one quantitative parameter, the at least one qualitative parameter, and the corresponding degradation factor of the at least one qualitative parameter, the basic index system of the photovoltaic inverter is constructed.

3. The state assessment method as described in claim 2, characterized in that, The steps for obtaining the target data of the photovoltaic inverter include: Obtain the raw data from the photovoltaic inverter; The raw data is filtered according to the key parameter system of the photovoltaic inverter to obtain the filtered raw data; The filtered raw data is preprocessed to obtain the target data.

4. The state assessment method as described in claim 2, characterized in that, The at least one quantitative parameter belongs to its corresponding feature type, the number of feature types is at least one, and each of the at least one quantitative parameter includes at least one quantitative parameter sample. The step of analyzing the at least one quantitative parameter of the photovoltaic inverter based on a preset first analysis rule to obtain the sensitivity factor corresponding to each of the at least one quantitative parameter includes: Based on the quantitative parameter samples corresponding to at least one quantitative parameter under each feature type, the average intra-class distance and average inter-class distance corresponding to at least one quantitative parameter under each feature type are calculated. Based on the average intra-class distance and the average inter-class distance, the sensitivity factor corresponding to at least one quantitative parameter under each feature type is calculated.

5. The condition assessment method as described in claim 4, characterized in that, The at least one qualitative parameter belongs to its corresponding characteristic type, and each of the at least one qualitative parameter includes an actual value, a threshold value, and a factory value. The step of analyzing the at least one qualitative parameter of the photovoltaic inverter based on a preset second analysis rule to obtain the degradation factor corresponding to each of the at least one qualitative parameter includes: Based on the actual value, threshold, and factory value of at least one qualitative parameter under each feature type, and based on a preset offset calculation rule, the degradation factor corresponding to at least one qualitative parameter under each feature type is calculated.

6. The state assessment method as described in claim 2, characterized in that, The steps for constructing the key indicator system of the photovoltaic inverter based on the basic indicator system of the photovoltaic inverter include: Based on the basic index system of the photovoltaic inverter, at least one quantitative parameter and at least one qualitative parameter of the photovoltaic inverter are screened to obtain at least one preliminary quantitative parameter and at least one preliminary qualitative parameter. Based on the preset Gaussian kernel function, the quantitative parameter samples corresponding to each of the at least one initial screening quantitative parameter, and the qualitative parameter samples corresponding to each of the at least one initial screening qualitative parameter, a normalized kernel matrix is ​​constructed. Based on the normalized kernel matrix, the corresponding feature matrix and feature vector matrix are constructed. Based on the normalized kernel matrix, the feature matrix, and the feature vector matrix, the target matrix is ​​constructed. Based on the target matrix, a key indicator system for the photovoltaic inverter is constructed.

7. The state assessment method as described in claim 2, characterized in that, The step of training the initial support vector machine model based on the sample data and the preset optimization parameters to obtain the trained state evaluation model includes: The sample data is preprocessed to obtain training samples and test samples; The initial support vector machine model is trained based on the training samples; Determine whether the initial support vector machine model meets the preset cross-validation conditions; If the initial support vector machine model does not meet the cross-validation condition, the initial support vector machine model is trained based on the training samples and preset optimization parameters until the initial support vector machine model meets the cross-validation condition. Then, the training of the initial support vector machine model is stopped, and the state evaluation model to be tested is obtained. The state evaluation model to be tested is tested based on the test samples. If the state evaluation model to be tested passes the test, then the trained state evaluation model is obtained.

8. The state assessment method as described in claim 7, characterized in that, Before the step of training the initial support vector machine model based on the training samples and preset optimization parameters until the initial support vector machine model meets the cross-validation conditions, and then stopping the training of the initial support vector machine model to obtain the state evaluation model to be tested, the method further includes: Based on the preset Grey Wolf algorithm, at least one optimization parameter is initialized.

9. The condition assessment method as described in claim 1, characterized in that, After the step of inputting the target data into the trained state assessment model to obtain the state assessment information of the photovoltaic inverter, the method further includes: If the status assessment information of the photovoltaic inverter meets the preset alarm conditions, then the alarm category corresponding to the status assessment information of the photovoltaic inverter is determined. Based on the alarm category, the corresponding alarm information is pushed.

10. A condition assessment device, characterized in that, The condition assessment device includes: The acquisition module is used to acquire target data for the photovoltaic inverter; An evaluation module is used to input the target data into a trained state evaluation model to obtain the state evaluation information of the photovoltaic inverter. The trained state evaluation model is obtained by training an initial support vector machine model based on sample data of the photovoltaic inverter and preset optimization parameters. The sample data is obtained based on the key indicator system of the photovoltaic inverter. The key indicator system is constructed based on the basic indicator system of the photovoltaic inverter. The basic indicator system is constructed based on at least one quantitative parameter, at least one corresponding sensitivity factor for each quantitative parameter, at least one qualitative parameter, and at least one corresponding degradation factor for each qualitative parameter. The magnitude of the sensitivity factor is positively correlated with the sensitivity of the quantitative parameter; the magnitude of the degradation factor is positively correlated with the degree of degradation of the qualitative parameter.

11. A terminal device, characterized in that, The terminal device includes a memory, a processor, and a state assessment program stored in the memory and executable on the processor. When the state assessment program is executed by the processor, it implements the steps of the state assessment method as described in any one of claims 1-9.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a state assessment program, which, when executed by a processor, implements the steps of the state assessment method as described in any one of claims 1-9.

Images