High and low voltage control monitoring system and method for photovoltaic substation

Abstract

The invention discloses a high-low voltage control monitoring system and method for a photovoltaic substation, and relates to the technical field of photovoltaic substation monitoring, fused dynamic data is input into a fault prediction model, and fault prediction results of a high-voltage side and a low-voltage side are output in real time; constructing a protection strategy library, dynamically matching a protection strategy according to a real-time fault prediction result, generating a protection instruction, carrying out dynamic balance regulation and control on the high and low voltage side voltage of the photovoltaic substation according to the protection instruction, and outputting an optimal prediction voltage regulation and control reference scheme; and monitoring actual high and low voltage side voltage data in real time, and comparing the actual high and low voltage side voltage data with the optimal prediction voltage regulation and control reference scheme to generate a fault detection report. By inputting the fused dynamic data into the fault prediction model, the fault prediction result is output in real time, the fault prediction precision is improved, the fault influence is reduced, stable operation of the photovoltaic substation is guaranteed, and the reliability and safety are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic substation monitoring, and in particular to a high and low voltage control monitoring system and method for a photovoltaic substation. Background Art

[0002] In recent years, with the rapid development of renewable energy, photovoltaic power generation has gained widespread application as a clean and sustainable energy source. Within photovoltaic power generation systems, photovoltaic substations are a critical link, and the stability and reliability of their operation directly impact the efficiency and safety of the entire system. In photovoltaic substations, real-time monitoring and control of parameters such as voltage and current on both the high-voltage and low-voltage sides, leveraging industrial data acquisition technologies and sensor networks, is crucial for ensuring stable operation.

[0003] While existing technologies have made progress in PV substation monitoring and control, such as the introduction of fault prediction techniques based on data mining and machine learning, these technologies are often limited to static data analysis and fail to fully utilize dynamic data characteristics, resulting in low accuracy in fault prediction for both high- and low-voltage sides. While real-time fusion analysis of dynamic data can better reflect grid status, existing technologies lack sufficient data collection, processing, and fusion, making it difficult to capture changing trends in operating conditions. Furthermore, due to the significant differences in operating characteristics between the high- and low-voltage sides, existing technologies struggle to address both, further reducing prediction accuracy. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a high and low voltage control and monitoring method for a photovoltaic substation to solve the problem of low accuracy in predicting high and low voltage faults in a photovoltaic substation.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In the first aspect, the present invention provides a high and low voltage control and monitoring method for a photovoltaic substation, which includes real-time collection of multi-source data on the high-voltage side and the low-voltage side of the photovoltaic substation, and preprocessing, fusing through a dynamic weighted fusion method, and generating fused dynamic data; inputting the fused dynamic data into a fault prediction model, and outputting the fault prediction results of the high-voltage side and the low-voltage side in real time; constructing a protection strategy library, dynamically matching the protection strategy according to the real-time fault prediction results, generating protection instructions, and dynamically balancing and regulating the high and low voltage side voltages of the photovoltaic substation according to the protection instructions, and outputting the optimal predicted voltage control benchmark scheme; real-time monitoring of the actual high and low voltage side voltage data, and comparing with the optimal predicted voltage control benchmark scheme to generate a fault detection report.

[0008] As a preferred solution of the high and low voltage control and monitoring method for a photovoltaic substation according to the present invention, the multi-source data includes voltage data, environmental data, and operation data of the high and low voltage sides of the photovoltaic substation;

[0009] The preprocessing includes a sliding window detection method for integrity check, a cubic spline interpolation method for repairing missing values, a Db4 wavelet threshold denoising method for eliminating electromagnetic interference signals, and a Min-Max normalization method for normalization processing.

[0010] As a preferred solution of the high and low voltage control and monitoring method for photovoltaic substations described in the present invention, wherein: the fusion is performed by a dynamic weighted fusion method to generate fused dynamic data, and the specific steps are as follows:

[0011] The analytic hierarchy process is used to extract the operating status features from the pre-processed multi-source data of the high-voltage and low-voltage sides of the photovoltaic substation and assign different weights.

[0012] Through the dynamic time warping method, the extracted operating status features are aligned in chronological order, and weighted calculation is performed according to the operating status feature weights to generate fused dynamic data.

[0013] As a preferred solution of the high and low voltage control and monitoring method for photovoltaic substations described in the present invention, wherein: the fusion dynamic data is input into the fault prediction model, and the fault prediction results of the high voltage side and the low voltage side are output in real time. The specific steps are as follows:

[0014] Based on historical fault data, the ST-GCN model is trained through supervised learning to obtain a fault prediction model;

[0015] The fused dynamic data is input into the fault prediction model to predict and output the comprehensive risk scores of the high-voltage and low-voltage sides in real time;

[0016] Based on historical fault data, the fault judgment threshold K is defined, and the comprehensive risk scores of the high-voltage side and the low-voltage side are judged to generate the fault prediction results.

[0017] As a preferred solution of the high and low voltage control and monitoring method for photovoltaic substations described in the present invention, wherein: the construction of a protection strategy library, dynamic matching of protection strategies according to real-time fault prediction results, and generation of protection instructions are performed in the following specific steps:

[0018] Generate corresponding processing actions based on historical fault data to form an initial protection strategy, optimize it through rule induction, and store it in a structured database to generate a protection strategy library;

[0019] Match the real-time fault prediction results with the protection strategies in the protection strategy library to generate protection instructions.

[0020] As a preferred solution of the high and low voltage control and monitoring method for photovoltaic substations described in the present invention, wherein: according to the protection instruction, the high and low voltage side voltages of the photovoltaic substation are dynamically balanced and regulated, and the optimal predicted voltage control reference scheme is output. The specific steps are as follows:

[0021] Define the control targets and the task list of high and low voltage side adjustments in priority according to the protection instructions;

[0022] Based on the task list, the high and low voltage sides of the photovoltaic substation are dynamically balanced and regulated through the coordinated control of the photovoltaic inverter, reactive power compensation and transformer tap;

[0023] Record the voltage regulation results of the high and low voltage sides, output the voltage stability target values ​​of the high and low voltage sides, the voltage adjustment actions of the high and low voltage sides of the photovoltaic substation, and the voltage stability status of the high and low voltage sides after regulation, and generate the optimal predicted voltage regulation benchmark plan.

[0024] As a preferred solution of the high and low voltage control monitoring method for photovoltaic substations described in the present invention, wherein: the real-time monitoring of actual high and low voltage side voltage data and comparison with the optimal predicted voltage control benchmark solution to generate a fault detection report, the specific steps are as follows:

[0025] Real-time monitoring of high and low voltage side voltage data of photovoltaic substations, point-by-point comparison with the optimal predicted voltage control benchmark solution, and output of comparison results;

[0026] Analyze and compare the distribution characteristics of abnormal high and low voltage side voltage data points in the results, and determine whether the fault judgment conditions are met. If so, generate a fault detection report in combination with the protection instructions.

[0027] In a second aspect, the present invention provides a high and low voltage control and monitoring system for a photovoltaic substation, comprising:

[0028] Multi-source data fusion module, fault prediction module, voltage control module and fault monitoring module; Multi-source data fusion module is used to collect multi-source data on the high-voltage side and low-voltage side of the photovoltaic substation in real time, and pre-process it, and fuse it through the dynamic weighted fusion method to generate fused dynamic data; Fault prediction module is used to input the fused dynamic data into the fault prediction model and output the fault prediction results of the high-voltage side and low-voltage side in real time; Voltage control module is used to build a protection strategy library, dynamically match the protection strategy according to the real-time fault prediction results, and dynamically balance the high and low voltage side voltages of the photovoltaic substation according to the protection instructions, and output the optimal predicted voltage control benchmark solution; Fault monitoring module is used to monitor the actual high and low voltage side voltage data in real time, and compare it with the optimal predicted voltage control benchmark solution to generate a fault detection report.

[0029] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the high and low voltage control and monitoring method for a photovoltaic substation as described in the first aspect of the present invention is implemented.

[0030] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the high and low voltage control and monitoring method for a photovoltaic substation as described in the first aspect of the present invention is implemented.

[0031] The beneficial effects of the present invention are as follows: fusion dynamic data is generated by a dynamic weighted fusion method, operation status features are extracted from pre-processed multi-source data using the hierarchical analysis method and weighted, and then the features are aligned and weighted in chronological order using the dynamic time warping method to solve the problem of inconsistency in the time series of multi-source data, comprehensively and accurately reflect the operation status of the photovoltaic substation, and provide rich and accurate data support for analysis; the fusion dynamic data is input into the fault prediction model to output the fault prediction results in real time, and the ST-GCN model is combined with historical fault data training to learn the relationship between the fault mode and the characteristics of multi-source data, and the fault judgment threshold is defined to determine the results, and the possibility of high and low voltage side faults is monitored in real time, providing operation and maintenance personnel with a clear risk assessment basis so that preventive measures can be taken in advance, improving the accuracy of fault prediction and reducing the impact of faults, ensuring the stable operation of photovoltaic substations, and improving reliability and safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0033] Figure 1 This is a flow chart of the high and low voltage control and monitoring method for a photovoltaic substation in Example 1.

[0034] Figure 2 This is a flowchart for multi-source data preprocessing in Example 1.

[0035] Figure 3 This is a flowchart for fault prediction model training and prediction in Example 1.

[0036] Figure 4 Schematic diagram of the high and low voltage control and monitoring system for photovoltaic substations in Example 1. DETAILED DESCRIPTION

[0037] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0038] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0039] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0040] Example 1, with reference to Figure 1~Figure 4 This embodiment provides a high and low voltage control and monitoring method for a photovoltaic substation, comprising the following steps:

[0041] S1. Real-time collection of multi-source data on the high-voltage and low-voltage sides of photovoltaic substations and pre-processing

[0042] S1.1. Multi-source data includes voltage data, environmental data, and operational data on the high-voltage and low-voltage sides of photovoltaic substations.

[0043] It should be noted that voltage data is collected by voltage sensors arranged on the high-voltage side of the photovoltaic substation. These sensors are distributed in key locations such as the high-voltage bus, circuit breakers, and the high-voltage side of the transformer to monitor parameters such as voltage amplitude and frequency in real time; temperature sensors, humidity sensors, light intensity sensors, and wind speed sensors are used to monitor environmental parameters such as temperature, humidity, light intensity, and wind speed in the photovoltaic power station; operation data is collected by operation monitoring sensors, and inverter monitoring sensors, transformer load sensors, and circuit breaker status sensors are used to collect operating status parameters such as inverter output power conversion efficiency, transformer load rate, and circuit breaker switch status; voltage data, environmental data, and operation data are summarized and preliminarily organized through communication interfaces and data acquisition terminals; a timestamp is added to each set of data through a timestamp marking device to ensure the time consistency of the data; the marked voltage data, environmental data, and operation data are integrated to form multi-source data on the high-voltage side of the photovoltaic substation; multi-source data on the low-voltage side of the photovoltaic substation is collected in the same way; and finally, multi-source data on the high-voltage and low-voltage sides of the photovoltaic substation are obtained.

[0044] S1.2. The preprocessing includes a sliding window detection method for integrity check, a cubic spline interpolation method for repairing missing values, a Db4 wavelet threshold denoising method for eliminating electromagnetic interference signals, and a Min-Max normalization method for normalization.

[0045] It should be noted that when preprocessing the multi-source data on the high-voltage and low-voltage sides of the photovoltaic substation, the sliding window detection method is used to perform integrity check. The sliding window gradually scans the data sequence at the set time interval to detect whether there are missing or incomplete data segments with abnormal point markings; then the cubic spline interpolation method is used to repair the marked missing values, and the cubic spline function is constructed according to the data points adjacent to the missing values ​​to calculate the value of the missing position to complete the data; then the Db4 wavelet threshold denoising method is used to denoise the multi-source data on the high-voltage and low-voltage sides of the photovoltaic substation, and the data is decomposed into different An appropriate threshold is set for the wavelet component of the frequency to remove the noise signal in the high-frequency component and retain the effective signal component; the multi-source data on the high-voltage side and low-voltage side of the photovoltaic substation are normalized through the Min-Max normalization method, and the value range of the multi-source data on the high-voltage side and low-voltage side of the photovoltaic substation is mapped to between 0 and 1, eliminating the influence of different dimensions on the data and improving the comparability of the multi-source data on the high-voltage side and low-voltage side of the photovoltaic substation; after integrity check, missing value repair, denoising and normalization processing, clean and standardized multi-source data on the high-voltage side and low-voltage side of the photovoltaic substation are obtained.

[0046] S2. Fusion is performed through a dynamic weighted fusion method to generate fused dynamic data.

[0047] S2.1. Using the analytic hierarchy process, the operating status features are extracted from the pre-processed multi-source data on the high-voltage and low-voltage sides of the photovoltaic substation and assigned different weights.

[0048] It should be noted that the operating status assessment of the high-voltage and low-voltage sides of the photovoltaic substation is taken as the target layer, the voltage characteristics, environmental characteristics and operating characteristics are classified as the criterion layer, and the voltage amplitude, frequency fluctuation, temperature, humidity, light intensity, inverter efficiency, transformer load rate, etc. are determined as the solution layer; through expert evaluation or historical data analysis methods, for each sub-feature of the criterion layer, the specific characteristic indicators of the solution layer are compared pairwise to determine the relative importance of each characteristic indicator and construct a judgment matrix; the eigenvalue and eigenvector of the constructed judgment matrix are calculated, and the weight vector of each characteristic indicator is obtained through normalization processing to determine the importance of each characteristic indicator in the overall evaluation; the consistency index and consistency ratio of the judgment matrix are calculated to check whether the consistency of the judgment matrix is ​​within an acceptable range. If the consistency ratio exceeds the range, the judgment matrix is ​​adjusted and the weight vector is recalculated; according to the weight vector, the operating status characteristics corresponding to each characteristic indicator are extracted from the multi-source data, and the corresponding weight is assigned to each extracted operating status feature.

[0049] S2.2. Using the dynamic time warping method, the extracted operating status features are aligned in chronological order, and weighted calculation is performed based on the operating status feature weights to generate fused dynamic data.

[0050] It should be noted that the expression for generating fusion dynamic data is:

[0051] X = ∑ (m,j) ∈ π w t(m,j) ⋅ [ x H,m x L,j ] ;

[0052] in, It is the fusion of dynamic data, is the index of the high-voltage side time series feature, is the index of the time series characteristics of the low-pressure side, High-voltage side The time series characteristics The value of the time step, It is the high-voltage side. It is the low pressure side The time series characteristics The value of the time step, It is the low pressure side. is the optimal alignment of the high-pressure side and low-pressure side time steps, is the time step The weight of

[0053] It should be noted that the extracted operating status features of the high-pressure side and the low-pressure side are sorted in time series respectively to ensure that the time points of each feature correspond to each other; a time series set of the operating status features of the high-pressure side and a time series set of the operating status features of the low-pressure side are constructed, and dynamic time warping calculations are performed on the time series sets of the high-pressure side and the low-pressure side respectively, and the optimal alignment path between the two time series is found through the dynamic programming method, so that the aligned time series are consistent in the time dimension; according to the weights assigned by the hierarchical analysis method, the aligned operating status features of the high-pressure side and the low-pressure side are weighted, and the characteristic value at each time point is multiplied by the corresponding weight to obtain the weighted time series; the weighted high-pressure side and low-pressure side time series are added or fused point by point to generate fused dynamic data, ensuring that the fusion result reflects the operating status features and weights of the high-pressure side and the low-pressure side; the generated fused dynamic data is smoothed to eliminate noise or discontinuity that may be introduced by alignment or weighted calculation; and the final fused dynamic data is generated.

[0054] S3. Input the fused dynamic data into the fault prediction model and output the fault prediction results of the high-voltage side and the low-voltage side in real time.

[0055] S3.1. Based on historical fault data, the ST-GCN model is trained through supervised learning to obtain a fault prediction model.

[0056] It should be noted that the ST-GCN model is trained through supervised learning based on historical fault data to obtain the fault prediction model. The specific steps include: collecting historical fault data on the high-voltage and low-voltage sides of the photovoltaic substation, including voltage data, environmental data, operating data and corresponding fault labels, and dividing the data into training and test sets; organizing the historical fault data in a time series format, and constructing a spatiotemporal graph structure suitable for ST-GCN model input, where nodes represent the high-voltage and low-voltage side equipment of the photovoltaic substation, edges represent the connection relationship between devices, and node features and edge features correspond to the operating status and environmental parameters of the equipment respectively; initializing the network parameters of the ST-GCN model, and setting hyperparameters such as the number of layers, convolution kernel size and learning rate of the ST-GCN model; calculating the loss function to measure the difference between the prediction results of the ST-GCN model and the actual fault labels;

[0057] It should be noted that the loss function expression for calculating and training the ST-GCN model is:

[0058] ;

[0059] in, is the loss function, is the total number of training samples, is the index variable of the training sample, is the total number of categories for fault classification, is the index variable of the category of fault classification, It is The samples belong to The true label of the class, The ST-GCN model predicts The samples belong to The probability of the class, is the weight moment of the ST-GCN model, The regularization hyperparameter value range is 0.1-1;

[0060] The training set data is input into the ST-GCN model, and the prediction results of the ST-GCN model are calculated through forward propagation. The loss value between the prediction result and the true fault label is calculated according to the loss function. The ST-GCN model parameters are updated using the backpropagation algorithm to optimize the ST-GCN model performance. The training process is repeated multiple times, and the parameters of the ST-GCN model are gradually adjusted to make the loss value converge to the minimum value. The performance of the trained ST-GCN model is evaluated on the test set. Evaluation indicators such as the accuracy, recall rate and F1 score of the ST-GCN model are calculated to verify the prediction ability of the ST-GCN model. Finally, a fault prediction model is generated.

[0061] S3.2. Input the fused dynamic data into the fault prediction model to predict and output the comprehensive risk scores of the high-voltage side and the low-voltage side in real time.

[0062] It should be noted that the expression for the comprehensive risk score of the high-pressure side and the low-pressure side output in real time is:

[0063] S H =σ( ∑ τ=1 U v t ⋅ g H ( X [t])), S L =σ( ∑ τ=1 U v t ⋅ g L ( X [t])) ;

[0064] in, is the comprehensive risk score on the high-pressure side, is the comprehensive risk score of the low-pressure side, is the activation function, is the number of time steps for fusing dynamic data, is the time step index of the fused dynamic data, is the time step is the weight, is the prediction function on the high-pressure side, is the prediction function on the low-pressure side, X [t ] is the fusion of dynamic data at time step The eigenvector of

[0065] It should be noted that the fused dynamic data generated by the dynamic weighted fusion method is obtained to ensure that the data contains the operating status characteristics and weight information of the high-voltage side and the low-voltage side; the fused dynamic data is organized in a time series format to make the fused dynamic data consistent with the input requirements of the fault prediction model, and a spatiotemporal graph structure containing the high-voltage side and low-voltage side equipment nodes and their connection relationships is constructed, and the node characteristics and edge characteristics correspond to the operating status and environmental parameters of the equipment respectively; the organized fused dynamic data is input into the fault prediction model, and the prediction results of the fault prediction model are calculated through forward propagation; according to the loss function set in the fault prediction model, the difference between the prediction result and the internal reference value of the fault prediction model is calculated to complete the prediction process; and the comprehensive risk scores of the high-voltage side and the low-voltage side are output in real time.

[0066] S3.3. Based on historical fault data, define the fault judgment threshold K, and judge the comprehensive risk scores of the high-voltage side and the low-voltage side to generate the fault prediction results.

[0067] When the comprehensive risk score of the high-voltage side is greater than K, it is considered that the high-voltage side voltage has failed;

[0068] When the comprehensive risk score of the low-voltage side is greater than K, it is considered that the low-voltage side voltage has a fault;

[0069] When the comprehensive risk score of the high-voltage side is less than K, the voltage on the high-voltage side is considered normal;

[0070] When the comprehensive risk score of the low-voltage side is less than K, the voltage on the low-voltage side is considered normal.

[0071] It should be noted that historical fault data on the high-voltage and low-voltage sides of the photovoltaic substation are collected, and voltage-related fault features are extracted, including data such as voltage amplitude, frequency fluctuation, temperature, humidity, light intensity, inverter efficiency, and transformer load rate; the historical fault data are preprocessed, and the sliding window detection method is used for integrity check, the cubic spline interpolation method is used to repair missing values, the Db4 wavelet threshold denoising method is used to eliminate electromagnetic interference signals, and the Min-Max normalization method is used to normalize the data; sample data of faults on the high-voltage and low-voltage sides are screened out from the preprocessed historical fault data, and the characteristic distribution of the high-voltage and low-voltage side fault samples is statistically analyzed respectively; the upper limit value of the characteristic distribution of the high-voltage side fault sample and the upper limit value of the characteristic distribution of the low-voltage side fault sample are calculated, for example, the 95% quantile of the fault sample characteristic distribution is selected to determine a reasonable upper limit value; the upper limit values ​​of the characteristic distribution of the high-voltage and low-voltage sides are compared, and a value that can simultaneously cover the fault characteristics of the high-voltage and low-voltage sides is selected as the fault judgment threshold K.

[0072] S4. Build a protection strategy library, dynamically match protection strategies based on real-time fault prediction results, and generate protection instructions.

[0073] S4.1. Generate corresponding processing actions based on historical fault data to form an initial protection strategy, optimize it through rule induction, and store it in a structured database to generate a protection strategy library.

[0074] It should be noted that historical fault data on the high-voltage and low-voltage sides of the photovoltaic substation are collected, and the operating status characteristics, environmental data and processing measures when the fault occurs are extracted; each historical fault data is classified into high-voltage side faults and low-voltage side faults according to the fault type, and the corresponding processing actions are recorded; the processing actions are associated with the fault type to form an initial protection strategy, including processing actions for high-voltage side faults and processing actions for low-voltage side faults; the initial protection strategy is summarized according to rules, the common characteristics of different fault types are analyzed, similar processing actions are merged, and the strategy content is optimized; the optimized protection strategy is checked to ensure that each strategy is clear and covers common fault types; the optimized protection strategy is classified according to the high-voltage side and low-voltage side, and stored in a structured database to form a protection strategy library.

[0075] S4.2. Match the real-time fault prediction results with the protection strategies in the protection strategy library to generate protection instructions.

[0076] It should be explained that the expression for generating protection instructions is;

[0077] ;

[0078] in, It is a specific protection instruction. It is the first A protective action, It is the protection strategy library, is the fault prediction result, It is the first The judgment conditions of the protection strategy, It indicates protection action The weight of is the index variable of the protection strategy library;

[0079] It should be noted that the real-time fault prediction results are obtained, and the fault type, fault location and comprehensive risk score in the prediction results are extracted; according to the fault type, the real-time fault prediction results are classified as high-voltage side faults or low-voltage side faults; the protection strategies corresponding to the high-voltage side faults or low-voltage side faults are retrieved in the protection strategy library, and compared one by one according to the fault type; the applicable conditions of each protection strategy in the protection strategy library are checked, including fault characteristics, risk score range and equipment status, to find the protection strategy that matches the real-time fault prediction results; if a matching protection strategy is found, the corresponding processing action is extracted to form a preliminary protection instruction; if a completely matching protection strategy is not found, the closest protection strategy is selected based on the similarity of the fault type, and the processing action is adjusted to adapt to the real-time fault prediction results; the preliminary protection instruction is verified with the real-time fault prediction results to ensure the feasibility and rationality of the protection instruction; and the final protection instruction is generated.

[0080] S5. According to the protection instructions, the high and low voltage sides of the photovoltaic substation are dynamically balanced and regulated, and the optimal predicted voltage control benchmark solution is output.

[0081] S5.1. Define the control targets and the task list of high and low voltage side voltages to be adjusted first according to the protection instructions.

[0082] It should be noted that the protection instruction is parsed to extract the fault type, fault location and operational steps contained in the protection instruction; the voltage control requirements of the high-voltage side or low-voltage side are judged according to the fault type, and the control target is determined to restore voltage stability or adjust the voltage amplitude; check whether the protection instruction clearly specifies the high-voltage side or low-voltage side voltage to be adjusted first. If it is not clearly specified, the priority is evaluated according to the scope and severity of the fault impact; the voltage on the side that needs to be adjusted first is taken as the primary control target and recorded as high-voltage side priority or low-voltage side priority; according to the control target, a list of specific tasks that need to be performed is listed, including adjusting the output power of the photovoltaic inverter, starting the reactive compensation device or adjusting the transformer tap; assign a priority to each task to ensure that the priority adjustment tasks on the high-voltage side or low-voltage side are ranked first; check whether each task in the task list conflicts with other tasks to avoid executing operations that may affect each other at the same time; combine the final task list with the control target to form a complete control task plan, clarify the execution order and target value of each task; generate a task list for the control target and the high-voltage and low-voltage side voltages that need to be adjusted first.

[0083] S5.2. Based on the task list, dynamically balance and regulate the high and low voltage sides of the photovoltaic substation through the coordinated control of the photovoltaic inverter, reactive power compensation and transformer tap.

[0084] It should be noted that the task list should be read to clearly define the control targets and the priority adjustment of the high and low voltage side voltages; if the task list prioritizes adjusting the high voltage side voltage, check the high voltage side voltage status, and adjust the output power of the photovoltaic inverter according to the control capability of the photovoltaic inverter, reduce or increase the output power to affect the high voltage side voltage; at the same time, check the status of the reactive compensation device, and start or adjust the reactive compensation device according to the high voltage side voltage situation, increase or decrease the reactive power output, and stabilize the high voltage side voltage; if further adjustment is required, operate the transformer tap, and adjust the tap position according to the target value in the task list to change the high voltage side voltage amplitude; if the task list prioritizes adjusting the high voltage side voltage, check the status of the reactive compensation device, and adjust the output power of the photovoltaic inverter according to the control capability of the photovoltaic inverter, and reduce or increase the output power to affect the high voltage side voltage; at the same time, check the status of the reactive compensation device, and start or adjust the reactive compensation device according to the high voltage side voltage situation, increase or decrease the reactive power output, and stabilize the high voltage side voltage; if further adjustment is required, operate the transformer tap, and adjust the tap position according to the target value in the task list to change the high voltage side voltage amplitude; Adjust the low-voltage side voltage, check the low-voltage side voltage status, adjust the output power through the photovoltaic inverter, and control the low-voltage side voltage fluctuation; enable the reactive power compensation device, adjust the reactive power output according to the low-voltage side voltage situation, and improve the low-voltage side voltage stability; adjust the transformer tap when necessary to change the low-voltage side voltage amplitude; during the regulation process, monitor the high and low voltage side voltage changes in real time, compare the target values ​​in the task list, and determine whether the expected regulation effect is achieved; if the target value is not reached, re-evaluate the task list and adjust the control strategy of the photovoltaic inverter, reactive power compensation device and transformer tap until the dynamic balance regulation of the high and low voltage sides of the photovoltaic substation is completed.

[0085] S5.3. Record the voltage control results of the high and low voltage sides, output the voltage stability target values ​​of the high and low voltage sides, the voltage adjustment actions of the high and low voltage sides of the photovoltaic substation, and the voltage stability status of the high and low voltage sides after control, and generate the optimal predicted voltage control benchmark plan.

[0086] It should be noted that the high-voltage side and low-voltage side voltage data after regulation are read, and the current voltage value is recorded; the high-voltage side voltage value after regulation is compared with the high-voltage side voltage stability target value in the task list to confirm whether the target value is reached, and the comparison result is recorded; the low-voltage side voltage value after regulation is compared with the low-voltage side voltage stability target value in the task list to confirm whether the target value is reached, and the comparison result is recorded; the operation records of the photovoltaic inverter during the regulation process are extracted, including the specific values ​​and time points of the output power adjustment, to form a detailed list of high-voltage and low-voltage side voltage adjustment actions; the operation records of the reactive compensation device during the regulation process are sorted out, including the start time, the adjusted reactive power size and the adjustment time point, and added to the voltage adjustment action list; the operation of the transformer tap during the regulation process is recorded, including the position change and adjustment time point before and after the tap adjustment, and added to the voltage adjustment action list; the voltage adjustment action list of the high-voltage side and low-voltage side is summarized to form a complete record of the high- and low-voltage side voltage adjustment actions of the photovoltaic substation;

[0087] Check the voltage data of the high-voltage and low-voltage sides after regulation to confirm whether they are in a stable state, and record the stable state of the voltages on the high-voltage and low-voltage sides after regulation; integrate the high-voltage side voltage stability target value, the low-voltage side voltage stability target value, the high- and low-voltage side voltage adjustment action records of the photovoltaic substation, and the stable state of the high- and low-voltage sides after regulation to form a complete regulation result report; analyze the operational effects and voltage change trends during the regulation process based on the data and records in the regulation result report, and evaluate the effectiveness of the regulation strategy; based on the regulation results and analysis, generate the optimal predicted voltage regulation benchmark plan.

[0088] S6. Monitor the actual high and low voltage side voltage data in real time, and compare it with the optimal predicted voltage control benchmark solution to generate a fault detection report.

[0089] S6.1. Monitor the voltage data on the high and low voltage sides of the photovoltaic substation in real time, compare them point by point with the optimal predicted voltage control benchmark solution, and output the comparison results.

[0090] What is needed is to start the real-time monitoring program, continuously collect voltage data on the high-voltage and low-voltage sides of the photovoltaic substation, and ensure that the frequency of data collection is aligned with the time point of the optimal prediction benchmark scheme; arrange the collected high-voltage side voltage data in chronological order, and compare them one by one with the high-voltage side voltage target value in the optimal prediction benchmark scheme, and record the difference between the actual voltage value and the target value at each time point; arrange the collected low-voltage side voltage data in chronological order, and compare them one by one with the low-voltage side voltage target value in the optimal prediction benchmark scheme, and record the difference between the actual voltage value and the target value at each time point; mark the comparison results at each time point, distinguish whether the actual voltage value is higher or lower than the target value, and record the specific value of the difference; check whether there are differences beyond the allowable range in the comparison results, mark abnormal data points, and record the time when the abnormality occurred and the specific difference value; summarize the comparison results of the high-voltage and low-voltage sides to form a complete point-by-point comparison result of the high- and low-voltage side voltage data with the optimal prediction benchmark scheme.

[0091] S6.2. Analyze the distribution characteristics of abnormal high and low voltage side voltage data points in the comparison results, and determine whether the fault judgment conditions are met. If so, generate a fault detection report in combination with the protection instructions.

[0092] It should be noted that the abnormal high and low voltage side voltage data points in the real-time monitoring comparison results are extracted, and the time, high-voltage side voltage value, low-voltage side voltage value and the difference between the abnormal data points and the optimal prediction benchmark solution are recorded; the time distribution characteristics of the abnormal data points are analyzed, and it is checked whether the abnormal data points are concentrated in a specific time period to form a time distribution characteristic description; the voltage value distribution characteristics of the abnormal data points are checked, and it is analyzed whether the high-voltage side and low-voltage side voltage values ​​show a continuous upward or downward trend, or whether there are drastic fluctuations, and the voltage value change trend is recorded; the number of abnormal data points is counted, the proportion of abnormal data points in the total monitoring data points is calculated, and the severity of the abnormal situation is evaluated; according to the time distribution, voltage value change trend and quantity ratio of the abnormal data points, the fault judgment conditions are compared to determine whether any fault judgment condition is met; if the distribution characteristics of the abnormal data points meet the fault judgment conditions, the protection instructions at the corresponding time point are extracted, including the protection action records of the high-voltage side and the low-voltage side; according to the protection instructions and the characteristics of the abnormal data points, a fault detection report is generated.

[0093] This embodiment further provides a high and low voltage control and monitoring system for a photovoltaic substation, comprising:

[0094] Multi-source data fusion module, fault prediction module, voltage control module and fault monitoring module; Multi-source data fusion module is used to collect multi-source data on the high-voltage side and low-voltage side of the photovoltaic substation in real time, and pre-process it, and fuse it through the dynamic weighted fusion method to generate fused dynamic data; Fault prediction module is used to input the fused dynamic data into the fault prediction model and output the fault prediction results of the high-voltage side and low-voltage side in real time; Voltage control module is used to build a protection strategy library, dynamically match the protection strategy according to the real-time fault prediction results, and dynamically balance the high and low voltage side voltages of the photovoltaic substation according to the protection instructions, and output the optimal predicted voltage control benchmark solution; Fault monitoring module is used to monitor the actual high and low voltage side voltage data in real time, and compare it with the optimal predicted voltage control benchmark solution to generate a fault detection report.

[0095] This embodiment also provides a computer device suitable for use in a high and low voltage control and monitoring method for a photovoltaic substation, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the high and low voltage control and monitoring method for a photovoltaic substation proposed in the above embodiment.

[0096] The computer device may be a terminal, comprising a processor, memory, a communication interface, a display, and an input device connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system and computer programs. The internal memory provides an environment for the operating system and computer programs stored in the non-volatile storage media. The communication interface of the computer device is used to communicate with external terminals via wired or wireless communication. Wireless communication may be achieved via Wi-Fi, a carrier network, NFC (near-field communication), or other technologies. The display of the computer device may be a liquid crystal display or an electronic ink display. The input device may be a touchscreen overlay on the display, buttons, a trackball, or a touchpad on the computer device housing, or an external keyboard, touchpad, or mouse.

[0097] This embodiment further provides a storage medium having a computer program stored thereon. When the program is executed by a processor, the processor implements the high and low voltage control and monitoring method for a photovoltaic substation as proposed in the above embodiment. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0098] In summary, the present invention uses: a dynamic weighted fusion method to generate fused dynamic data, uses the hierarchical analysis method to extract operating status features from the preprocessed multi-source data and assign weights, and then uses the dynamic time warping method to align the features in chronological order and perform weighted calculations to solve the problem of inconsistency in the time series of multi-source data, comprehensively and accurately reflect the operating status of the photovoltaic substation, and provide rich and accurate data support for analysis; inputs the fused dynamic data into the fault prediction model to output the fault prediction results in real time, uses the ST-GCN model combined with historical fault data training to learn the relationship between fault modes and multi-source data features, and defines the fault judgment threshold judgment results, monitors the possibility of high and low voltage side faults in real time, and provides operation and maintenance personnel with a clear risk assessment basis so that preventive measures can be taken in advance, improves fault prediction accuracy and reduces fault impact, ensures stable operation of photovoltaic substations, and improves reliability and safety.

[0099] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A high and low voltage control and monitoring method for a photovoltaic substation, characterized by: include, Real-time collection of multi-source data from the high-voltage and low-voltage sides of photovoltaic substations, pre-processing, and fusion through dynamic weighted fusion method to generate fused dynamic data; The fused dynamic data is input into the fault prediction model, and the fault prediction results of the high-voltage side and the low-voltage side are output in real time; Build a protection strategy library, dynamically match protection strategies based on real-time fault prediction results, generate protection instructions, dynamically balance and regulate the high and low voltage sides of the photovoltaic substation based on the protection instructions, and output the optimal predicted voltage regulation benchmark solution; Monitor actual high and low voltage side voltage data in real time, and compare it with the optimal predicted voltage control benchmark solution to generate a fault detection report.

2. The high and low voltage control and monitoring method for a photovoltaic substation according to claim 1, wherein: The multi-source data includes voltage data, environmental data and operation data of the high-voltage side and the low-voltage side of the photovoltaic substation; The preprocessing includes a sliding window detection method for integrity check, a cubic spline interpolation method for repairing missing values, a Db4 wavelet threshold denoising method for eliminating electromagnetic interference signals, and a Min-Max normalization method for normalization processing.

3. The high and low voltage control and monitoring method for a photovoltaic substation according to claim 2, characterized in that: The dynamic weighted fusion method is used to perform fusion to generate fused dynamic data. The specific steps are as follows: The analytic hierarchy process is used to extract the operating status features from the pre-processed multi-source data of the high-voltage and low-voltage sides of the photovoltaic substation and assign different weights. Through the dynamic time warping method, the extracted operating status features are aligned in chronological order, and weighted calculation is performed according to the operating status feature weights to generate fused dynamic data.

4. The high and low voltage control and monitoring method for a photovoltaic substation according to claim 3, characterized in that: The fusion dynamic data is input into the fault prediction model, and the fault prediction results of the high-voltage side and the low-voltage side are output in real time. The specific steps are as follows: Based on historical fault data, the ST-GCN model is trained through supervised learning to obtain a fault prediction model; The fused dynamic data is input into the fault prediction model to predict and output the comprehensive risk scores of the high-voltage and low-voltage sides in real time; Based on historical fault data, the fault judgment threshold K is defined, and the comprehensive risk scores of the high-voltage side and the low-voltage side are judged to generate the fault prediction results.

5. The high and low voltage control and monitoring method for a photovoltaic substation according to claim 4, characterized in that: The construction of the protection strategy library dynamically matches the protection strategy according to the real-time fault prediction results and generates protection instructions. The specific steps are as follows: Generate corresponding processing actions based on historical fault data to form an initial protection strategy, optimize it through rule induction, and store it in a structured database to generate a protection strategy library; Match the real-time fault prediction results with the protection strategies in the protection strategy library to generate protection instructions.

6. The high and low voltage control and monitoring method for a photovoltaic substation according to claim 5, characterized in that: According to the protection instructions, the high and low voltage sides of the photovoltaic substation are dynamically balanced and regulated to output the optimal predicted voltage regulation benchmark solution. The specific steps are as follows: Define the control targets and the task list of high and low voltage side adjustments in priority according to the protection instructions; Based on the task list, the high and low voltage sides of the photovoltaic substation are dynamically balanced and regulated through the coordinated control of the photovoltaic inverter, reactive power compensation and transformer tap; Record the voltage regulation results of the high and low voltage sides, output the voltage stability target values ​​of the high and low voltage sides, the voltage adjustment actions of the high and low voltage sides of the photovoltaic substation, and the voltage stability status of the high and low voltage sides after regulation, and generate the optimal predicted voltage regulation benchmark plan.

7. The high and low voltage control and monitoring method for a photovoltaic substation according to claim 6, characterized in that: The real-time monitoring of the actual high and low voltage side voltage data and the comparison with the optimal predicted voltage control benchmark solution to generate a fault detection report are as follows: Real-time monitoring of high and low voltage side voltage data of photovoltaic substations, point-by-point comparison with the optimal predicted voltage control benchmark solution, and output of comparison results; Analyze and compare the distribution characteristics of abnormal high and low voltage side voltage data points in the results, and determine whether the fault judgment conditions are met. If so, generate a fault detection report in combination with the protection instructions.

8. A high and low voltage control and monitoring system for a photovoltaic substation, based on the high and low voltage control and monitoring method for a photovoltaic substation according to any one of claims 1 to 7, characterized in that: Including multi-source data fusion module, fault prediction module, voltage regulation module and fault monitoring module; The multi-source data fusion module is used to collect multi-source data from the high-voltage and low-voltage sides of the photovoltaic substation in real time, perform pre-processing, and fuse them using the dynamic weighted fusion method to generate fused dynamic data; The fault prediction module is used to input the fused dynamic data into the fault prediction model and output the fault prediction results of the high-voltage side and the low-voltage side in real time; The voltage control module is used to build a protection strategy library, dynamically match protection strategies based on real-time fault prediction results, dynamically balance and control the high and low voltage sides of the photovoltaic substation according to protection instructions, and output the optimal predicted voltage control benchmark solution; The fault monitoring module is used to monitor the actual high and low voltage side voltage data in real time, and compare it with the optimal predicted voltage control benchmark solution to generate a fault detection report.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the high and low voltage control and monitoring method for a photovoltaic substation according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the high and low voltage control and monitoring method for a photovoltaic substation according to any one of claims 1 to 7 are implemented.

Images