Photovoltaic string fault diagnosis state evaluation method, system, equipment and medium

Abstract

The invention discloses a photovoltaic string fault diagnosis state evaluation method, system, equipment and medium, and belongs to the field of photovoltaic stations, and the method comprises the steps: obtaining the IV scanning data of each photovoltaic string, unifying the voltage scale through a linear difference method, generating an IV curve of each string, and calculating the fault diagnosis state of each string; open-circuit voltage, short-circuit current, maximum power point parameters, slope parameters, fitting deviation, curve fluctuation characteristics and corresponding characteristic values are selected and solved, for each characteristic value, a preset parameter is taken, each state evaluation value is obtained by using a corresponding calculation method, and a decision tree fault diagnosis model is constructed by combining the IV curve characteristic value and the evaluation value. According to the method, the fault of each group of string is marked according to the comparison result of the evaluation value and the preset threshold value, finally, the preset photovoltaic string fault coding library is associated, the fault diagnosis result and the corresponding fault reason are output, and the photovoltaic string fault can be pre-judged timely and accurately with low cost and high efficiency.

Description

Technical Field

[0001] This invention belongs to the field of photovoltaic power plants, specifically relating to methods, systems, equipment, and media for assessing the fault diagnosis status of photovoltaic strings. Background Technology

[0002] Faced with the rapid growth of photovoltaic (PV) installed capacity, unmanned or minimally manned operation and maintenance (O&M) models will become an important way to solve the problem of high O&M pressure in PV power plants. IV (Current-Voltage) characteristic curve scanning, as a core technology for PV string performance diagnosis, accurately identifies various typical faults such as module degradation, hot spot effects, and PID (Potential Induced Degradation) effects by measuring the dynamic relationship curves of current and voltage in real time and relying on recognition algorithms. Currently, PV string fault diagnosis methods based on IV scanning are mainly divided into the following two types: One approach is a photovoltaic string fault diagnosis method based on artificial intelligence algorithms. This method utilizes classification algorithms, image recognition algorithms, etc., to collect sample data and train a model to diagnose faults in photovoltaic strings. However, this method has high requirements for the quantity and quality of training data and computer hardware, and its generalization ability is poor in practical engineering applications.

[0003] The second method is based on mathematical statistics. This method extracts trend features from the IV curve and relies on statistical methods or expert experience to set thresholds to diagnose faults in photovoltaic strings. Because this method has many parameters and depends on feature values ​​and thresholds, the selection of feature values, the calculation method, and the setting of thresholds have a significant impact on the accuracy of photovoltaic string fault identification.

[0004] Currently, if we can predict photovoltaic string failures in a low-cost, efficient, timely and accurate manner, it will be of great significance to the efficient operation and maintenance and quality improvement of photovoltaic power plants. Summary of the Invention

[0005] The purpose of this invention is to overcome the problem of being unable to predict photovoltaic string faults in a low-cost, efficient, timely and accurate manner, and to propose a method, system, equipment and medium for photovoltaic string fault diagnosis and status assessment.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a method for assessing the fault status of photovoltaic strings, comprising the following steps: Obtain the current and voltage IV scan results data for each photovoltaic string; The voltage dimension scale of the IV scan results was uniformly formatted using the linear interpolation method to obtain the IV curves of each photovoltaic string. Based on the IV curves of each photovoltaic string, the characteristics of the IV curves are selected, including open-circuit voltage, short-circuit current, maximum power point parameter, slope parameter, fitting deviation, and curve fluctuation characteristics. The characteristic values ​​of each IV curve feature are then solved. Among them, the open-circuit voltage and short-circuit current are solved using the local linear fitting method. The state assessment of the feature values ​​is specifically carried out by selecting the feature values ​​within the preset quantile interval for the same feature value of all photovoltaic strings, calculating the average value, and calculating the state assessment value based on the feature values ​​and average values ​​of each photovoltaic string. Using the IV curves, eigenvalues, and state assessment values ​​of each photovoltaic string, a decision tree fault diagnosis model is constructed. Based on the decision tree fault diagnosis model, each photovoltaic string is marked for faults by comparing the state assessment values ​​with preset thresholds. Based on the fault label association with the preset photovoltaic string fault code library, the fault diagnosis results and fault causes are output.

[0007] Furthermore, the process of uniformly formatting the voltage dimension scale of the IV scan results data using the linear interpolation method includes: Obtain the current photovoltaic string IV scan result data, sort it in ascending order by voltage value, and obtain the sorted data; Extract the maximum and minimum voltage values ​​from the sorted data; Set voltage intervals, determine voltage ranges based on minimum voltage, maximum voltage and voltage intervals, generate equally spaced sequences within voltage ranges, and multiply equally spaced sequences by voltage intervals to obtain standardized voltage sequences; The linear interpolation method is used to solve for the current value corresponding to each voltage in the standardized voltage sequence, and the current IV curve of the current photovoltaic string is obtained. Repeat the above steps to complete the formatting of all photovoltaic string IV scan data and obtain the IV curves for each photovoltaic string.

[0008] Furthermore, the calculation formula for the linear interpolation method is shown below:

[0009] in, , Indicates the first In the IV scan results of each photovoltaic string, all voltages less than the standard voltage... The maximum voltage and current values, , Indicates the first In the IV scan results of each photovoltaic string, all voltages are greater than the standard voltage. The minimum voltage and current values.

[0010] Furthermore, the process of solving for the eigenvalues ​​of each IV curve characteristic includes: Select the last few data points from the IV curve of each photovoltaic string, perform linear fitting with current as the independent variable and voltage as the dependent variable, and solve for the voltage value when the current is zero as the open circuit voltage. Select the first few data points from the IV curve of each photovoltaic string, perform linear fitting with voltage as the independent variable and current as the dependent variable, and solve for the current value when the voltage is zero as the short-circuit current. The power is obtained by multiplying the voltage and current in the IV curve of each photovoltaic string, and the voltage and current corresponding to the maximum power are selected as the maximum power point parameters. The slope parameters include horizontal slope and vertical slope. Based on open-circuit voltage, short-circuit current and maximum power point parameters, the horizontal slope and vertical slope are calculated using the horizontal slope formula and the vertical slope formula. The voltage in the IV curve of each photovoltaic string is taken as the independent variable and the current is taken as the dependent variable. The polynomial fitting algorithm is input to obtain the fitting equation and output the predicted current. The absolute error between the actual current and the predicted current is calculated, and the average value of the error value within the preset quantile interval of the absolute error is selected as the fitting deviation. The slope set is obtained by calculating the slope between two consecutive points in the IV curve of each photovoltaic string. The hysteresis difference value with a preset hysteresis interval is calculated based on the slope set. After binarizing the hysteresis difference value, the number of the largest consecutive non-zero values ​​is extracted as the curve fluctuation feature.

[0011] Furthermore, the formula for calculating the lag difference value of the curve fluctuation characteristics is as follows:

[0012] in, Indicates the first The hysteresis interval of each photovoltaic string is The The hysteresis difference at each measurement point For the current measuring point, , The lag interval, .

[0013] Furthermore, the construction and fault labeling process of the decision tree fault diagnosis model includes: Set preset thresholds for each IV curve feature; The state evaluation value of the fitting deviation is compared with the corresponding preset threshold. Values ​​higher than the threshold are marked as zero values, and values ​​lower than the threshold are marked as first values. For photovoltaic strings marked as zero, the state evaluation value of the curve fluctuation characteristics is compared with the corresponding preset threshold. If it is higher than the threshold, it is marked as zero; if it is lower than the threshold, it is marked as second value. For photovoltaic strings marked as zero, the open-circuit voltage state assessment value is compared with the corresponding preset threshold. If it is higher than the threshold, it is marked as zero; if it is lower than the threshold, it is marked as the third value. For photovoltaic strings marked as zero, the short-circuit current state assessment value is compared with the corresponding preset threshold. If it is higher than the threshold, it is marked as zero; if it is lower than the threshold, it is marked as the fourth value. For photovoltaic strings marked as zero, the state assessment values ​​of the maximum power point current and the horizontal slope are compared with the corresponding preset thresholds. If at least zero values ​​are below the threshold, they are marked as the fifth value; if all values ​​are above the threshold, they are marked as the zero value. For photovoltaic strings marked as zero, the state assessment values ​​of the maximum power point voltage and vertical slope are compared with the corresponding preset thresholds. If at least zero values ​​are below the threshold, they are marked as the sixth value; if all values ​​are above the threshold, they are marked as the zero value.

[0014] Furthermore, the state assessment value is the ratio of the characteristic value of each photovoltaic string to its average value.

[0015] Secondly, the present invention provides a photovoltaic string fault diagnosis and status assessment system, comprising: The IV scan data acquisition module is used to acquire the IV scan result data of each photovoltaic string; The formatting module is used to perform uniform formatting of the voltage dimension scale of the IV scan results data using the linear interpolation method to obtain the IV curves of each photovoltaic string. The eigenvalue solving module is used to select IV curve features, including open-circuit voltage, short-circuit current, maximum power point parameter, slope parameter, fitting deviation, and curve fluctuation characteristics, based on the IV curves of each photovoltaic string, and solve for the eigenvalues ​​of each IV curve feature; among which, the open-circuit voltage and short-circuit current are solved using a local linear fitting method; The feature value state assessment module is used to assess the state of feature values. Specifically, for the same feature value of all photovoltaic strings, the feature values ​​within the preset quantile interval are selected to calculate the average value. Based on the feature values ​​and average values ​​of each photovoltaic string, the state assessment value is calculated. The fault labeling module is used to construct a decision tree fault diagnosis model using the IV curves, eigenvalues, and state evaluation values ​​of each photovoltaic string. Based on the decision tree fault diagnosis model, each photovoltaic string is labeled with a fault by comparing the state evaluation value with a preset threshold. The fault diagnosis result output module is used to output fault diagnosis results and fault causes based on the fault marker association with the preset photovoltaic string fault code library.

[0016] Thirdly, the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement a photovoltaic string fault diagnosis status assessment method.

[0017] Fourthly, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements a method for assessing the fault diagnosis status of photovoltaic strings.

[0018] Compared with the prior art, the present invention has the following beneficial technical effects: The photovoltaic string fault diagnosis and status assessment method proposed in this invention uses the linear interpolation method to format IV scan data, avoiding the problem of inconsistent voltage intervals in the IV scan results data, which leads to deviations in the extraction of IV scan curve fluctuation features. It uses a local linear fitting method to solve the open-circuit voltage and short-circuit current of IV scan data, avoiding the overfitting problem at the beginning and end of existing overall high-order polynomial fitting. It proposes fitting deviation and curve fluctuation feature indicators, which can quickly and accurately detect problems such as obvious and slight steps or depressions in the IV scan curve. Based on the feature value status assessment results, a photovoltaic string fault diagnosis model is constructed. Using the decision tree idea, string faults are quickly diagnosed according to the primary and secondary relationships of photovoltaic string faults, enabling low-cost, high-efficiency, timely and accurate prediction of photovoltaic string faults. Attached Figure Description

[0019] The accompanying drawings described herein are for illustrative purposes only and are not intended to limit the scope of the invention in any way. Furthermore, the shapes and proportions of the components in the drawings are merely illustrative to aid in understanding the invention and do not specifically limit the shapes and proportions of the components of the invention.

[0020] In the attached diagram: Figure 1 This is a flowchart of the photovoltaic string fault diagnosis and status assessment method of the present invention.

[0021] Figure 2 This is a simplified structural diagram of the photovoltaic string fault diagnosis and status assessment system of the present invention.

[0022] Figure 3 This is a schematic diagram of the electronic device used in the photovoltaic string fault diagnosis and status assessment method of the present invention.

[0023] Figure 4 This is a flowchart of the photovoltaic string fault diagnosis status assessment method in an embodiment of the present invention.

[0024] Figure 5 This is a specific logic diagram for evaluating the characteristic value state of photovoltaic strings in an embodiment of the present invention. Detailed Implementation

[0025] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0026] Example 1 See Figure 1 A method for diagnosing and assessing the condition of photovoltaic strings faults includes the following steps: The process involves acquiring the IV scan results of each photovoltaic (PV) string; then, using the linear interpolation method to standardize the voltage dimension scale of the IV scan results, resulting in the IV curves for each PV string. The standardization process includes: acquiring the IV scan results of the current PV string, sorting them in ascending order of voltage value to obtain the sorted data; extracting the maximum and minimum voltage values ​​from the sorted data; setting voltage intervals, determining voltage ranges based on the minimum, maximum, and voltage intervals, generating equally spaced sequences within each voltage range, and multiplying these sequences by the voltage intervals to obtain a standardized voltage sequence; using the linear interpolation method to calculate the current value corresponding to each voltage in the standardized voltage sequence, thus obtaining the IV curve for the current PV string; repeating the above steps to complete the formatting of the IV scan data for all PV strings, resulting in the IV curves for each PV string. This standardization of the voltage dimension of the IV data for each PV string eliminates format differences, ensures the comparability of IV curves, and provides accurate and consistent data support for subsequent feature extraction and fault diagnosis. The calculation formula for the linear interpolation method is shown below: ;in, , Indicates the first In the IV scan results of each photovoltaic string, all voltages less than the standard voltage... The maximum voltage and current values, , Indicates the first In the IV scan results of each photovoltaic string, all voltages are greater than the standard voltage. The minimum voltage and current values ​​are determined. Based on the IV curves of each photovoltaic string, IV curve features including open-circuit voltage, short-circuit current, maximum power point parameter, slope parameter, fitting deviation, and curve fluctuation characteristics are selected, and the eigenvalues ​​of each IV curve feature are solved. The open-circuit voltage and short-circuit current are solved using a local linear fitting method. Solving for the eigenvalues ​​of each IV curve feature involves: selecting the last few data points from each photovoltaic string's IV curve, performing a linear fit with current as the independent variable and voltage as the dependent variable, and determining the voltage value at which the current is zero as the open-circuit voltage; selecting the first few data points from each photovoltaic string's IV curve, performing a linear fit with voltage as the independent variable and current as the dependent variable, and determining the current value at which the voltage is zero as the short-circuit current; calculating the product of each voltage and current in each photovoltaic string's IV curve to obtain the power, and selecting the maximum power... The corresponding voltage and current are used as maximum power point parameters; the slope parameters include horizontal and vertical slopes. Based on the open-circuit voltage, short-circuit current, and maximum power point parameters, the horizontal and vertical slopes are calculated using the horizontal and vertical slope formulas. The voltage in the IV curve of each photovoltaic string is used as the independent variable, and the current as the dependent variable. A polynomial fitting algorithm is input to obtain the fitting equation and output the predicted current. The absolute error between the actual current and the predicted current is calculated, and the average value of the error within a preset quantile interval is selected as the fitting deviation. The slope between two consecutive points in the IV curve of each photovoltaic string is calculated to obtain a slope set. Based on the slope set, the hysteresis difference value of the preset hysteresis interval is calculated. After binarization of the hysteresis difference value, the number of the largest consecutive non-zero values ​​is extracted as the curve fluctuation feature. Through precise regional fitting and multi-dimensional feature extraction, key parameters such as open-circuit voltage and features such as slope and fluctuation can be accurately obtained. The data reliability is high, providing comprehensive and high-quality feature support for subsequent condition assessment and accurate fault diagnosis. The formula for calculating the hysteresis difference value of the curve fluctuation feature is: ,in, Indicates the first The hysteresis interval of each photovoltaic string is The The hysteresis difference at each measurement point For the current measuring point, , The lag interval, The process involves state evaluation of eigenvalues. Specifically, for the same eigenvalue across all photovoltaic (PV) strings, the average value of eigenvalues ​​within a preset quantile interval is calculated. Based on the eigenvalues ​​and average values ​​of each PV string, a state evaluation value is calculated. A decision tree fault diagnosis model is constructed using the IV curves, eigenvalues, and state evaluation values ​​of each PV string. Based on this model, each PV string is fault-marked by comparing the state evaluation value with a preset threshold. The construction and fault marking process of the decision tree fault diagnosis model includes: setting preset thresholds for each IV curve feature; comparing the state evaluation value with the corresponding preset threshold, marking values ​​above the threshold as zero and values ​​below as first; for PV strings marked as zero, comparing the state evaluation value of the curve fluctuation feature with the corresponding preset threshold, marking values ​​above the threshold as first; and for PV strings marked as zero, comparing the state evaluation value of the curve fluctuation feature with the corresponding preset threshold, keeping the value above the threshold as the first. For photovoltaic strings marked with a zero value, those below a threshold are marked with a second value. For strings marked with a zero value, the open-circuit voltage state assessment value is compared with the corresponding preset threshold; if it is higher than the threshold, it remains marked with a zero value; if it is lower, it is marked with a third value. Similarly, for strings marked with a zero value, the short-circuit current state assessment value is compared with the corresponding preset threshold; if it is higher than the threshold, it remains marked with a zero value; if it is lower, it is marked with a fourth value. For strings marked with a zero value, the maximum power point current and horizontal slope state assessment values ​​are compared with their corresponding preset thresholds; if at least zero values ​​are lower than the threshold, they are marked with a fifth value; if both are higher, they remain marked with a zero value. Finally, for strings marked with a zero value, the maximum power point voltage and vertical slope state assessment values ​​are compared with their corresponding preset thresholds; if at least zero values ​​are lower than the threshold, they are marked with a sixth value; if both are higher, they remain marked with a zero value. This hierarchical, progressive threshold comparison logic provides clear criteria and orderly filtering, accurately marking different types of faults, reducing misjudgments, improving the accuracy and efficiency of fault identification, and providing reliable marking support for subsequent association with the fault coding library. Based on the fault label association with the preset photovoltaic string fault code library, the fault diagnosis results and fault causes are output.

[0027] This embodiment efficiently addresses the need for precise diagnosis in photovoltaic (PV) operation and maintenance. By standardizing the voltage dimension scale through linear interpolation, it ensures standardized IV curve data, laying a solid foundation for accurate feature extraction and avoiding diagnostic biases caused by data format differences. The selected multi-dimensional features, such as open-circuit voltage and maximum power point parameters, combined with local linear fitting, comprehensively capture subtle changes in the IV curve, accurately identify potential fault signals, and significantly improve the fault detection rate. The state assessment method based on preset quantile intervals eliminates outlier interference, making the assessment results more objective and providing a reliable basis for fault diagnosis. The application of a decision tree model enables rapid fault classification and labeling. Combined with a preset fault coding library, it can directly output the fault cause, simplifying the diagnostic process and shortening fault investigation time. The entire process balances data accuracy and diagnostic efficiency, reducing manual operation and maintenance costs and improving the operational stability and power generation efficiency of the PV system.

[0028] Example 2 See Figure 2 A photovoltaic string fault diagnosis and status assessment system, including: The IV scan data acquisition module is used to acquire the IV scan result data of each photovoltaic string; The formatting module is used to perform uniform formatting of the voltage dimension scale of the IV scan results data using the linear interpolation method to obtain the IV curves of each photovoltaic string. The eigenvalue solving module is used to select IV curve features, including open-circuit voltage, short-circuit current, maximum power point parameter, slope parameter, fitting deviation, and curve fluctuation characteristics, based on the IV curves of each photovoltaic string, and solve for the eigenvalues ​​of each IV curve feature; among which, the open-circuit voltage and short-circuit current are solved using a local linear fitting method; The feature value state assessment module is used to assess the state of feature values. Specifically, for the same feature value of all photovoltaic strings, the feature values ​​within the preset quantile interval are selected to calculate the average value. Based on the feature values ​​and average values ​​of each photovoltaic string, the state assessment value is calculated. The fault labeling module is used to construct a decision tree fault diagnosis model using the IV curves, eigenvalues, and state evaluation values ​​of each photovoltaic string. Based on the decision tree fault diagnosis model, each photovoltaic string is labeled with a fault by comparing the state evaluation value with a preset threshold. The fault diagnosis result output module is used to output fault diagnosis results and fault causes based on the fault marker association with the preset photovoltaic string fault code library.

[0029] Example 3 See Figure 3An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements a photovoltaic string fault diagnosis and status assessment method: acquiring IV scan result data for each photovoltaic string; performing voltage dimension scaling uniform formatting on the IV scan result data using a linear interpolation method to obtain IV curves for each photovoltaic string; based on the IV curves of each photovoltaic string, selecting IV curve features including open-circuit voltage, short-circuit current, maximum power point parameter, slope parameter, fitting deviation, and curve fluctuation characteristics, and solving for the eigenvalues ​​of each IV curve feature; wherein the open-circuit... Voltage and short-circuit current are solved using a local linear fitting method. A state assessment is performed on the eigenvalues. Specifically, for the same eigenvalue across all photovoltaic strings, the average value is calculated for eigenvalues ​​falling within a preset quantile interval. Based on the eigenvalues ​​and average values ​​of each photovoltaic string, a state assessment value is calculated. A decision tree fault diagnosis model is constructed using the IV curves, eigenvalues, and state assessment values ​​of each photovoltaic string. Based on this model, faults are marked for each photovoltaic string by comparing the state assessment values ​​with preset thresholds. The fault diagnosis results and causes are output based on the fault markings and a preset photovoltaic string fault coding library.

[0030] Example 4 A computer-readable storage medium stores a computer program that, when executed by a processor, implements a photovoltaic string fault diagnosis and state assessment method: acquiring IV scan result data for each photovoltaic string; performing voltage dimension scaling uniform formatting on the IV scan result data using a linear interpolation method to obtain IV curves for each photovoltaic string; based on the IV curves of each photovoltaic string, selecting IV curve features including open-circuit voltage, short-circuit current, maximum power point parameter, slope parameter, fitting deviation, and curve fluctuation characteristics, and solving for the eigenvalues ​​of each IV curve feature; wherein the open-circuit voltage and short-circuit current are... The current is solved using a local linear fitting method. A state assessment is performed on the eigenvalues. Specifically, for the same eigenvalue across all photovoltaic strings, the average value is calculated for eigenvalues ​​within a preset quantile interval. Based on the eigenvalues ​​and average values ​​of each photovoltaic string, a state assessment value is calculated. A decision tree fault diagnosis model is constructed using the IV curves, eigenvalues, and state assessment values ​​of each photovoltaic string. Based on the decision tree fault diagnosis model, each photovoltaic string is fault-marked by comparing the state assessment value with a preset threshold. The fault diagnosis result and fault cause are output based on the fault markings and a preset photovoltaic string fault coding library.

[0031] Example 5 See Figure 4 The photovoltaic string fault diagnosis and status assessment method specifically includes the following steps: Step 1: Acquire and collect the IV scan results data of the group sequence.

[0032] Specifically, by sending an IV scan request to the controller, the IV scan program is driven to execute the scan task, and after the scan is completed, the IV scan result data of each string is returned. .

[0033] The IV scan results mainly consist of data on multiple sets of voltages and their corresponding currents collected during the scan mission, namely: .

[0034] In the above expression, Indicates the number of strings. Indicates the string number. Indicates the first The string number Group current value, Indicates the first The string number Group voltage value.

[0035] Step 2: IV scan data formatting processing.

[0036] To address the issue of inconsistent voltage intervals in the IV scan results of different groups, which leads to deviations in the amount of data in the fluctuation range of the IV scan curve, an IV scan data formatting method is proposed. Its main function is to ensure that all IV curves have the same scale in the voltage dimension, thus providing a basic condition for the extraction of IV curve feature values.

[0037] The specific processing method is as follows: Step 2.1: Obtain any set of string IV scan result data And sort them in ascending order of voltage value; Step 2.2: Obtain the IV scan result data for this group of strings. Maximum value of medium voltage and minimum value ; Step 2.3: Set the voltage interval value And respectively the minimum voltage value in 2.2 and maximum value With voltage interval value By dividing the values ​​and rounding, we obtain the voltage interval range. , in: .

[0038] Based on this, in the interval Generate a sequence with intervals of 1. And multiply the sequence value by the voltage interval value. Thus, the standardized voltage sequence values ​​are obtained. The calculation formula is as follows:

[0039] In the above formula, Indicates the first The string corresponds to the first A standardized voltage value.

[0040] Step 2.4: Using the linear interpolation method, solve for the standardized voltage sequence values ​​sequentially. Corresponding current value This yields the formatted string IV scan result data. .

[0041]

[0042] The formula for the linear interpolation method is as follows:

[0043] In the above formula, , Indicates the first In the IV scan results of each string, all voltages less than the standard voltage... The maximum voltage value and its corresponding current value, , Indicates the first In the IV scan results of each group of strings, all voltages are greater than the standard voltage. The minimum voltage value and its corresponding current value.

[0044] Step 2.5: Following steps 2.1 to 2.4, format all IV scan data for each group to obtain the formatted IV scan data result. .

[0045] Step 3: Feature selection and eigenvalue solving.

[0046] Common IV curve anomalies include: low current, steep slope on the horizontal leg (i.e., increased slope), stepped / dipping curves, shallow slope on the vertical leg (i.e., decreased slope), and low voltage. Considering the severity of the fault, the IV scan curve features of this invention include eight features: open-circuit voltage, short-circuit current, mpp current, mpp voltage, horizontal slope, vertical slope, fitting bias, and curve hysteresis difference. Open-circuit voltage... Short-circuit current A local linear fitting scheme is adopted; mpp current , mpp voltage Power is obtained by multiplying voltage by current. The current and voltage corresponding to the maximum power are then selected, and the current (mpp) is the output voltage. and MPP voltage Horizontal slope Vertical slope By obtaining the open-circuit voltage of this string Short-circuit current , mpp current and MPP voltage The solution is as follows. The fitting bias is determined by calculating the deviation between the actual current and the predicted current, reflecting the obvious steps and dips in the photovoltaic string. The curve hysteresis difference is calculated using a curve fluctuation feature extraction method based on hysteresis difference.

[0047] The specific calculation methods for each feature are as follows: (1) Open circuit voltage Short-circuit current .

[0048] Because the adjustable range of the device voltage is limited during IV curve scanning, and it is impossible or impossible to adjust it to the voltage limit value, it is necessary to predict the open-circuit voltage and short-circuit current using the IV curve trend. Traditional methods for calculating open-circuit voltage and short-circuit current employ a high-order polynomial fitting algorithm for the entire IV scan data. This method is prone to overfitting at the beginning and end of the IV curve. To address this issue, this invention employs a local linear fitting scheme, thereby avoiding this problem. Specifically: Open circuit voltage Select any set of formatted IV scan data. After selecting this data The data is used with current as the independent variable and voltage as the dependent variable. A linear fitting algorithm is then applied to fit the data, and the voltage corresponding to a current of 0 is calculated from the fitted function. This voltage value is the open-circuit voltage. ; short circuit current Select any set of formatted IV scan data. After selecting this data Data points were used, with voltage as the independent variable and current as the dependent variable. A linear fitting algorithm was then applied to fit the data, and the current corresponding to a voltage of 0 was calculated from the fitted function. This current value is then considered the short-circuit current. ; (2) mpp current , mpp voltage : Select any set of formatted IV scan data The voltage is then multiplied by the current to obtain the power. Based on this, the current and voltage corresponding to the maximum power are selected as the mpp current. and MPP voltage .

[0049] (3) Horizontal slope Vertical slope : Select any string and obtain the open-circuit voltage of that string. Short-circuit current , mpp current and MPP voltage Then, the horizontal slope of the string is calculated according to the following formula. Vertical slope :

[0050] (4) Fitting bias .

[0051] Steps and dips are prominent characteristics of photovoltaic (PV) string faults on the IV scan curve. To address this, this invention proposes an IV scan curve fitting deviation evaluation method based on a polynomial fitting algorithm. This method calculates the deviation between the actual current and the predicted current to reflect the obvious steps and dips in the PV string. The specific calculation process is as follows: Step 3.4.1: Select any set of formatted IV scan data. With voltage For independent variable, current The dependent variable is input into a polynomial fitting algorithm of order 4 for model training, resulting in the fitted equation:

[0052] In the above formula, , , , , These are the model parameters; specific parameter values ​​can be obtained through model training.

[0053] Step 3.4.2, will Medium voltage data The input is then fed into the fitted equation to obtain the predicted current. .

[0054] Step 3.4.3: Transfer IV scan data The absolute error between the current in the current and the predicted current is calculated. The calculation formula is as follows: .

[0055] Step 3.4.4: Calculate the absolute error The average of the error values ​​in the 90%-95% quantile range is used to obtain the fitting deviation of the string. .

[0056] Step 3.4.5: Calculate the fitting deviation values ​​for all strings according to steps 3.4.1 to 3.4.4. .

[0057] (5) Curve lag difference.

[0058] To address the issue of minor faults in photovoltaic strings causing small dips or steps in the IV scan curve, this invention proposes a curve fluctuation feature extraction method based on hysteresis difference. The specific steps of this method are as follows: Step 3.5.1: Select any set of formatted IV scan data. And calculate the slope between two consecutive points. Thus, the set of slopes of the IV curves of this series is obtained. For any string middle, slope at point The calculation formula is as follows:

[0059] Step 3.5.2: For the set of IV curve slopes of this series... Calculate the lag difference values ​​at intervals from 1 to 10 sequentially, and form the set of lag difference values ​​for the slope of the IV curve. ,in:

[0060] In the above formula, Indicates the first The lag interval of each string is The Lag difference value at each measuring point For the current measuring point, , The lag interval, .

[0061] Step 3.5.3: Set the slope lag difference. In this context, values ​​greater than 0 are set to 1, and values ​​less than or null are set to 0. Step 3.5.4: Obtain the slope lag difference set In the text, the maximum number of consecutive values ​​of 1 in each row. .

[0062] Step 3.5.5, Obtain maximum value .

[0063] Step 3.5.6: Calculate the lag differences of all string curves according to steps 3.5.1 to 3.5.5. .

[0064] Step 4: Evaluation of photovoltaic string characteristic values.

[0065] Since IV scan data is significantly affected by irradiance, ambient temperature, and module temperature, directly using feature values ​​for photovoltaic string fault identification yields poor accuracy. This invention proposes a photovoltaic string feature value state assessment method to provide data support for photovoltaic string fault identification, as detailed below: (1) Open-circuit voltage status assessment method.

[0066] Obtain the open-circuit voltage of all strings , and select Open-circuit voltages with median quantiles between 85% and 95% and their average values ​​were calculated. Then calculate the open-circuit voltage of all strings. and The ratio of these values ​​yields the open-circuit voltage state assessment value for each group of strings. .

[0067] (2) Short-circuit current status assessment method.

[0068] Get the short-circuit current of all strings , and select The median of the short-circuit currents between 85% and 95% is calculated, and their average value is also calculated. Then calculate the short-circuit current of all strings. and The ratio of the two values ​​is used to obtain the short-circuit current state assessment value for each group of strings. .

[0069] (3) MPP current state assessment method.

[0070] Get the mpp current of all strings , and select The median current in mpp is between 85% and 95%, and its average value is calculated. Then calculate the mpp current of all strings. and The ratio of the two values ​​is used to obtain the current state assessment value of each string mpp. .

[0071] (4) MPP voltage state assessment method.

[0072] Get the MPP voltage of all strings , and select The median voltage of mpp between 85% and 95% is calculated and its average value is calculated. Then calculate the MPP voltage of all strings. and The ratio of these values ​​yields the voltage state assessment value (mpp) for each string group. .

[0073] (5) Horizontal slope status assessment method.

[0074] Get the horizontal slope of all strings , and select The median is the horizontal slope between 85% and 95% and its average is calculated. Then calculate the horizontal slope of all strings. and The ratio of the values ​​yields the horizontal slope state evaluation value for each group of strings. .

[0075] (6) Vertical slope status assessment method.

[0076] Get the vertical slope of all strings , and select The vertical slope of the median between 85% and 95% is calculated and its average value is obtained. Then calculate the vertical slope of all strings. and The ratio of these ratios yields the vertical slope state evaluation value for each group of strings. .

[0077] (7) Method for evaluating the state of fitting deviation.

[0078] Get the fitting deviation of all strings , and select The fitting deviation of the median between 85% and 95% was calculated and its average value was obtained. Then calculate the fitting deviation of all strings. and The ratio of the values ​​yields the state evaluation value of the fitting deviation for each group of strings. .

[0079] (8) Curve lag difference state assessment method.

[0080] Obtain the lag difference of all string curves And obtain the lag difference state evaluation value of any set of series curves according to the following formula. Then, the lag difference state values ​​of all cascade curves are calculated sequentially. .

[0081]

[0082] Step 5: Based on the eigenvalue state assessment results, construct a photovoltaic string fault diagnosis model.

[0083] Photovoltaic string fault diagnosis model obtains photovoltaic string feature value state threshold The photovoltaic string characteristic value and state evaluation value are used to mark all strings by constructing a decision tree to determine the size of the state value and the threshold.

[0084] The specific logic of the photovoltaic string fault diagnosis model is as follows: Step 5.1: Set and obtain the characteristic value state threshold of the photovoltaic string. .

[0085] Step 5.2: Obtain the fitting deviation state evaluation value. The fitting deviation state evaluation value of each string is marked with reference to the fitting deviation state threshold. The value above the threshold is marked as 0, and the value below the threshold is marked as 1.

[0086] Step 5.3: Obtain the lag difference state evaluation value of the string curve marked as 0 in Step 5.2. The lag difference state evaluation values ​​of each string curve are marked according to the lag difference threshold. Values ​​above the threshold are marked as 0, and values ​​below the threshold are marked as 2.

[0087] Step 5.4: Obtain the string open-circuit voltage state assessment value marked as 0 in Step 5.3. The open-circuit voltage state evaluation values ​​of each string are marked with reference to the open-circuit voltage state threshold. Values ​​higher than the threshold are marked as 0, and values ​​lower than the threshold are marked as 3.

[0088] Step 5.5: Obtain the string short-circuit current state assessment value marked as 0 in Step 5.4. The short-circuit current state evaluation values ​​of each string are marked with reference to the short-circuit current state threshold. Values ​​higher than the threshold are marked as 0, and values ​​lower than the threshold are marked as 4.

[0089] Step 5.6: Obtain the string MPP current state assessment value marked as 0 in Step 5.5. and string horizontal slope status assessment value The evaluation values ​​of the MPP current state and the horizontal slope state of each string are marked with reference to the MPP current state threshold and the horizontal slope state threshold. If at least one of them is lower than the threshold, it is marked as 5, and if both are higher than the threshold, it is marked as 0.

[0090] Step 5.7: Obtain the string MPP voltage state assessment value marked as 0 in Step 5.6. and string vertical slope status assessment The MPP voltage state evaluation value and vertical slope state evaluation value of each string are marked according to the MPP voltage state threshold and vertical slope state threshold, respectively. Values ​​with at least one value below the threshold are marked as 6, and values ​​with both values ​​above the threshold are marked as 0. The specific logic is as follows: Figure 5 As shown.

[0091] Step 6: Constructing the photovoltaic string fault coding library.

[0092] Based on the photovoltaic string fault diagnosis model, a photovoltaic string fault coding library was constructed, the details of which are as follows:

[0093] Step 7: Application of photovoltaic string fault diagnosis model.

[0094] First, a photovoltaic string fault diagnosis task is created for the photovoltaic string to be diagnosed; then, IV scan data of the photovoltaic string to be diagnosed is obtained, and the characteristic value state threshold of the photovoltaic string is manually set. Based on this, the photovoltaic strings are assessed for status and diagnosed for faults according to steps 1-5, thereby outputting photovoltaic string fault marking codes. Finally, the output photovoltaic string fault marking codes are associated with the photovoltaic string fault coding library constructed in step 6 to obtain fault diagnosis results and fault causes, thereby guiding photovoltaic power station operation and maintenance personnel to conduct inspections and troubleshooting for faulty strings.

[0095] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0096] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0097] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0098] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0099] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for diagnosing and assessing the condition of photovoltaic strings, characterized in that, Includes the following steps: Obtain the current and voltage IV scan results data for each photovoltaic string; The voltage dimension scale of the IV scan results data was uniformly formatted using the linear interpolation method to obtain the IV curves of each photovoltaic string. Based on the IV curves of each photovoltaic string, the IV curve features are selected, including open-circuit voltage, short-circuit current, maximum power point parameter, slope parameter, fitting deviation and curve fluctuation characteristics, and the characteristic values ​​of each IV curve feature are solved. The open-circuit voltage and the short-circuit current are solved using a local linear fitting method. For the same characteristic value of all photovoltaic strings, the average value of the characteristic value within the preset quantile interval is calculated. Based on the characteristic value and the average value of each photovoltaic string, the state evaluation value is calculated. Using the IV curves of each photovoltaic string, the feature values, and the state evaluation values, a decision tree fault diagnosis model is constructed. Based on the decision tree fault diagnosis model, each photovoltaic string is marked for faults by comparing the state evaluation values ​​with preset thresholds. Based on the fault marker associated with the preset photovoltaic string fault code library, the fault diagnosis results and fault causes are output.

2. The photovoltaic string fault diagnosis and status assessment method according to claim 1, characterized in that, The process of uniformly formatting the voltage dimension scale of the IV scan result data using the linear interpolation method includes: Obtain the current photovoltaic string IV scan result data, sort it in ascending order by voltage value, and obtain the sorted data; Extract the maximum and minimum voltage values ​​from the sorted data; Set a voltage interval, determine a voltage range based on the minimum voltage value, the maximum voltage value and the voltage interval, generate an equally spaced sequence within the voltage range, and multiply the equally spaced sequence with the voltage interval to obtain a standardized voltage sequence; The linear interpolation method is used to solve for the current value corresponding to each voltage in the standardized voltage sequence, and the current IV curve of the current photovoltaic string is obtained. Repeat the above steps to complete the formatting of all photovoltaic string IV scan data and obtain the IV curves for each photovoltaic string.

3. The photovoltaic string fault diagnosis and status assessment method according to claim 1, characterized in that, The calculation formula for the linear interpolation method is shown below: in, , Indicates the first In the IV scan results of each photovoltaic string, all voltages less than the standard voltage... The maximum voltage and current values, , Indicates the first In the IV scan results of each photovoltaic string, all voltages are greater than the standard voltage. The minimum voltage and current values.

4. The photovoltaic string fault diagnosis and status assessment method according to claim 1, characterized in that, The process of solving for the eigenvalues ​​of each of the IV curve features includes: Select the last few data points from the IV curves of each photovoltaic string, perform linear fitting with current as the independent variable and voltage as the dependent variable, and solve for the voltage value when the current is zero as the open circuit voltage. Select the first few data points from the IV curves of each photovoltaic string, perform linear fitting with voltage as the independent variable and current as the dependent variable, and solve for the current value when the voltage is zero as the short-circuit current. The power is obtained by multiplying the voltage and current in the IV curve of each photovoltaic string, and the voltage and current corresponding to the maximum power are selected as the maximum power point parameters. The slope parameters include horizontal slope and vertical slope. Based on the open-circuit voltage, the short-circuit current, and the maximum power point parameter, the horizontal slope and the vertical slope are calculated using the horizontal slope formula and the vertical slope formula. Using the voltage as the independent variable and the current as the dependent variable in the IV curve of each photovoltaic string, the multinomial fitting algorithm is input to obtain the fitting equation and output the predicted current. The absolute error between the actual current and the predicted current is calculated, and the average value of the error value within the preset quantile interval of the absolute error is selected as the fitting deviation. The slope set is obtained by calculating the slope between two consecutive points in the IV curve of each photovoltaic string. The hysteresis difference value with a preset hysteresis interval is calculated based on the slope set. After binarizing the hysteresis difference value, the number of the largest consecutive non-zero values ​​is extracted as the curve fluctuation feature.

5. The photovoltaic string fault diagnosis and status assessment method according to claim 4, characterized in that, The formula for calculating the hysteresis difference value of the curve fluctuation characteristics is as follows: in, Indicates the first The hysteresis interval of each photovoltaic string is The The hysteresis difference at each measurement point For the current measuring point, , The lag interval, .

6. The photovoltaic string fault diagnosis and status assessment method according to claim 1, characterized in that, The process of constructing and labeling the decision tree fault diagnosis model includes: Set a preset threshold corresponding to each of the IV curve features; The state evaluation value of the fitting deviation is compared with the corresponding preset threshold. Values ​​higher than the threshold are marked as zero values, and values ​​lower than the threshold are marked as first values. For the photovoltaic string marked with the zeroth value, the state evaluation value of the curve fluctuation characteristic is compared with the corresponding preset threshold. If it is higher than the threshold, it is marked with the zeroth value; if it is lower than the threshold, it is marked with the second value. For the photovoltaic string marked with the zeroth value, the open-circuit voltage state assessment value is compared with the corresponding preset threshold. If it is higher than the threshold, it is marked with the zeroth value; if it is lower than the threshold, it is marked with the third value. For the photovoltaic string marked with the zeroth value, the state assessment value of the short-circuit current is compared with the corresponding preset threshold. If it is higher than the threshold, it is marked with the zeroth value; if it is lower than the threshold, it is marked with the fourth value. For the photovoltaic string marked with the zeroth value, the state evaluation values ​​of the maximum power point current and the horizontal slope are compared with the corresponding preset thresholds. If at least zero values ​​are lower than the threshold, they are marked as the fifth value. If all values ​​are higher than the threshold, the zeroth value is retained. For the photovoltaic string marked with the zeroth value, the state evaluation values ​​of the maximum power point voltage and the vertical slope are compared with the corresponding preset thresholds. If at least zero values ​​are below the threshold, they are marked as the sixth value; if all values ​​are above the threshold, the zeroth value is retained.

7. The photovoltaic string fault diagnosis and status assessment method according to claim 1, characterized in that, The state assessment value is the ratio of the characteristic value of each photovoltaic string to the average value.

8. A photovoltaic string fault diagnosis and status assessment system, characterized in that, include: The IV scan data acquisition module is used to acquire the IV scan result data of each photovoltaic string; The formatting module is used to perform voltage dimension scale uniform formatting on the IV scan result data using the linear interpolation method to obtain the IV curve of each photovoltaic string. The eigenvalue solving module is used to select IV curve features, including open-circuit voltage, short-circuit current, maximum power point parameter, slope parameter, fitting deviation, and curve fluctuation characteristics, based on the IV curves of each photovoltaic string, and solve for the eigenvalues ​​of each IV curve feature; wherein the open-circuit voltage and the short-circuit current are solved using a local linear fitting method; The feature value state assessment module is used to assess the state of the feature value. Specifically, for the same feature value of all photovoltaic strings, the feature value within a preset quantile interval is selected to calculate the average value. Based on the feature value of each photovoltaic string and the average value, a state assessment value is calculated. The fault labeling module is used to construct a decision tree fault diagnosis model using the IV curves of each photovoltaic string, the feature values, and the state evaluation values. Based on the decision tree fault diagnosis model, each photovoltaic string is labeled with a fault by comparing the state evaluation values ​​with a preset threshold. The fault diagnosis result output module is used to output fault diagnosis results and fault causes based on the fault mark associated with the preset photovoltaic string fault code library.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the photovoltaic string fault diagnosis status assessment method according to any one of claims 1-7.

10. A computer-readable storage medium storing a computer program that, when executed by a processor, implements the photovoltaic string fault diagnosis status assessment method according to any one of claims 1-7.

Images