Distributed photovoltaic grid-connected power distribution network electric energy quality monitoring system and method

Abstract

The invention relates to the technical field of photovoltaic power quality monitoring, and discloses a distributed photovoltaic grid-connected power distribution network power quality monitoring system and method, and the method comprises the steps: firstly, comprehensively collecting the power quality data and environment data of a distributed photovoltaic grid-connected point, and carrying out the data preprocessing to form a standardized data set; and then, constructing a power quality prediction model based on the data, and predicting power quality indexes of key nodes in the power distribution network in combination with environmental data and distributed photovoltaic power generation characteristics. Then, acquiring data in real time, inputting the data into the model, and identifying an electric energy quality problem; and comparing and analyzing the real-time abnormal data and the prediction index to determine the abnormal reason and distribution. And finally, the electric energy quality is evaluated according to a preset standard, an evaluation report is generated, and abnormal early warning information is sent in time, so that related personnel take measures to improve the electric energy quality. According to the method, the dynamic electric energy quality problem caused by distributed photovoltaic grid connection is effectively monitored and predicted.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic power quality monitoring, and in particular to a distributed photovoltaic grid-connected distribution network power quality monitoring system and method. Background Art

[0002] With the transformation of the global energy structure and the vigorous development of renewable energy, the penetration rate of distributed photovoltaic power generation systems in distribution networks is increasing due to their clean, efficient and flexible characteristics. However, the large-scale grid connection of distributed photovoltaics has also brought new challenges to the power quality of distribution networks. Since the power generation of distributed photovoltaics is significantly affected by environmental factors (such as solar irradiance, temperature, etc.), its output power is intermittent and uncertain, which may lead to power quality problems such as voltage fluctuations, current anomalies, frequency deviations, unqualified power factors, excessive harmonics, three-phase imbalance, and voltage flicker in the distribution network.

[0003] Current power quality monitoring methods often focus on monitoring and analyzing steady-state power quality indicators, but lack effective monitoring and prediction of dynamic power quality problems caused by distributed photovoltaic grid connection. At the same time, due to the dispersion and diversity of distributed photovoltaic grid connection points, traditional monitoring methods are difficult to achieve comprehensive coverage and real-time monitoring of all grid connection points. In addition, the impact of environmental factors on distributed photovoltaic power generation also increases the complexity and uncertainty of power quality issues, making it difficult for traditional monitoring methods to accurately predict and identify power quality problems caused by distributed photovoltaic grid connection. Summary of the invention

[0004] The object of the present invention is to provide a distributed photovoltaic grid-connected distribution network power quality monitoring system and method to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: a method for monitoring power quality of a distributed photovoltaic grid-connected distribution network, comprising: Step 1: Collect power quality data of the distribution network of distributed photovoltaic grid-connected points, including voltage, current, frequency, power factor, harmonic content, three-phase unbalance, voltage fluctuation and flicker, as well as the power generation power of distributed photovoltaics in the grid and the location information of the grid-connected points; also collect environmental data such as temperature, humidity, wind speed and solar irradiance; Step 2: Preprocess the data collected in step 1, including data cleaning, data alignment and time synchronization, to form a standardized data set; Step 3: Based on the normalized data set, a power quality prediction model is constructed. The model combines environmental data and the power generation characteristics of distributed photovoltaics to predict the power quality indicators of key nodes in the distribution network. Step 4: Obtain power quality data and environmental data of distributed photovoltaic grid-connected points in real time; Step 5: Input the data obtained in step 4 into the power quality prediction model, and receive the predicted power quality index output by the model; Step 6: Perform anomaly detection on the power quality data acquired in real time to identify power quality problems, including voltage over-limit, current abnormality, frequency deviation, unqualified power factor, excessive harmonics, three-phase imbalance, voltage fluctuation and flicker; Step 7: Compare and analyze the power quality anomaly data identified in step 6 with the predicted power quality index output by the model in step 5 to determine the cause and distribution of the anomaly; Step 8: Evaluate the power quality of the distribution network according to the preset power quality standards and thresholds, and generate a power quality evaluation report; Step 9: Send the power quality assessment report and abnormal warning information to relevant personnel so that timely measures can be taken to improve the power quality.

[0006] Preferably, in step 2, the data preprocessing method includes: For missing values, the average value of adjacent data points or linear interpolation is used to fill them; For outliers, statistical algorithms are used to identify them and replace or delete them with reasonable values; Perform time synchronization on the data to make the time labels of each data point consistent.

[0007] Preferably, in step 3, the step of constructing a power quality prediction model includes: Step 3.1: Use the normalized dataset as input and divide the dataset into training and test sets. Step 3.2: Select appropriate machine learning algorithm and build power quality prediction model; Step 3.3: Use the training set to train the model and optimize the model parameters through cross-validation and grid search methods; Step 3.4: Use the test set to evaluate the trained model and verify the model's predictive performance; Step 3.5: Adjust and optimize the model according to the evaluation results until the prediction accuracy of the model reaches a satisfactory level; Step 3.6: Save and deploy the trained power quality prediction model.

[0008] Preferably, in step 3.2, the selected machine learning algorithm is a long short-term memory network LSTM, and the steps of constructing the LSTM model include: Set the number of layers, number of hidden layer units and activation function of the LSTM network; Configure the input layer, hidden layer, and output layer structure of the LSTM network; The LSTM model is trained using the training set, and the model parameters are iteratively optimized through the back propagation algorithm and the gradient descent algorithm. The calculation formula of the gradient descent algorithm is: in, Represents the model parameters, including all weights and biases that the model needs to learn; Represents the learning rate, which is used to control the step size of parameter update; Represents the loss function About parameters The gradient of is a vector pointing to the direction where the loss function grows fastest; Indicates an assignment operation, that is, updating parameters The value of .

[0009] Preferably, the input layer of the LSTM model receives a normalized data set as input, including historical power quality data, environmental data, distributed photovoltaic power generation and location information of the grid connection point; the hidden layer contains multiple LSTM units for capturing long-term dependencies in time series data; the output layer outputs predicted power quality indicators, including voltage, current, frequency, power factor, harmonic content, three-phase imbalance, and voltage fluctuation and flicker.

[0010] Preferably, in step 6, the method of abnormality detection includes: Set the threshold range of power quality indicators, including upper and lower limits of voltage, allowable range of current, and deviation limit of frequency; Check the power quality data obtained in real time point by point to determine whether each indicator exceeds the set threshold range; Mark indicators that exceed the threshold, and record the time, location, and specific indicator value of the anomaly.

[0011] Preferably, in step 7, the comparative analysis step includes: Time-align the abnormal data marked in step 6 with the predicted power quality indicators output by the model in step 5; Calculate the deviation between abnormal data and predicted data, and analyze the causes and trends of the deviation; Based on the deviation amount and cause analysis, determine whether the abnormal power quality is caused by distributed photovoltaic grid connection or by other factors.

[0012] Preferably, in step 8, the step of generating the power quality assessment report includes: Conduct statistics and analysis on the power quality anomalies identified in step 7, and summarize the types, quantity, distribution and severity of the anomalies; According to the preset power quality standards and thresholds, the power quality of the distribution network is evaluated as a whole, and the evaluation results and grades are given; Propose measures to improve power quality, including adjusting the grid connection mode of distributed photovoltaics, optimizing the grid structure and strengthening power quality monitoring; Prepare power quality assessment report, including assessment results, abnormality analysis and recommended measures.

[0013] Preferably, in step 9, the step of sending the power quality assessment report and abnormal warning information includes: Send the power quality assessment report to relevant personnel by email, file transfer or system push; Provide real-time warning of abnormal power quality and notify relevant personnel via SMS, phone, email or system interface; Determine the level and frequency of warnings based on the urgency and severity of the abnormal situation.

[0014] Preferably, a distributed photovoltaic grid-connected distribution network power quality monitoring system comprises: The data acquisition module is used to collect the power quality data of the distribution network of the distributed photovoltaic grid-connected points, including voltage, current, frequency, power factor, harmonic content, three-phase imbalance, voltage fluctuation and flicker, as well as the power generation power of the distributed photovoltaic in the grid and the location information of the grid-connected points; it also collects environmental data, including temperature, humidity, wind speed and solar irradiance; The data preprocessing module is connected to the data acquisition module and is used to preprocess the collected data, including data cleaning, data alignment and time synchronization, to form a standardized data set; A power quality prediction model building module is connected to the data preprocessing module and is used to build a power quality prediction model based on a normalized data set, combined with environmental data and distributed photovoltaic power generation characteristics; Real-time data acquisition module, used to obtain power quality data and environmental data of distributed photovoltaic grid-connected points in real time; A prediction index output module is connected to the power quality prediction model building module and the real-time data acquisition module, and is used to input real-time data into the power quality prediction model and receive the predicted power quality index output by the model; An anomaly detection module, connected to the real-time data acquisition module, is used to perform anomaly detection on the power quality data acquired in real time and identify power quality problems; A comparison and analysis module, connected to the anomaly detection module and the prediction index output module, is used to compare and analyze the identified power quality anomaly data with the predicted power quality index to determine the cause and distribution of the anomaly; The power quality assessment module is connected to the comparison and analysis module and is used to assess the power quality of the distribution network according to preset power quality standards and thresholds and generate a power quality assessment report; The report sending module is connected to the power quality assessment module and is used to send the power quality assessment report and abnormal warning information to relevant personnel.

[0015] Compared with the prior art, the present invention has the following beneficial effects: By comprehensively collecting the power quality data of distributed photovoltaic grid-connected points and related environmental data, the present invention can realize comprehensive monitoring of the power quality indicators of key nodes in the distribution network. At the same time, the power quality prediction model constructed by combining environmental data and distributed photovoltaic power generation characteristics can accurately predict the future trend of power quality changes, making it possible to take preventive measures in advance.

[0016] Current power quality monitoring methods often focus on monitoring steady-state indicators, while the present invention pays special attention to dynamic power quality problems caused by distributed photovoltaic grid connection. By acquiring data in real time and inputting it into the prediction model, combined with the abnormal detection mechanism, it can quickly identify power quality problems such as voltage fluctuations, current anomalies, frequency deviations, unqualified power factors, excessive harmonics, three-phase imbalance, and voltage flicker, and issue early warnings in a timely manner, which helps to quickly respond to and deal with potential risks.

[0017] By comparing and analyzing the abnormal power quality data obtained in real time with the indicators output by the prediction model, the present invention can accurately locate the location of power quality problems and deeply analyze the causes and distribution patterns of the problems. This provides a scientific basis for formulating targeted improvement measures and helps to improve the efficiency and accuracy of power quality management.

[0018] By timely identifying, warning and handling power quality issues, the present invention helps reduce distribution network failures and power outages caused by distributed photovoltaic grid connection, thereby improving the stability and safety of the distribution network. At the same time, regular power quality assessment and report generation also provide important references for the long-term planning and management of the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A working step diagram of the method for monitoring power quality of a distributed photovoltaic grid-connected distribution network according to the present invention; Figure 2 Flowchart for building power quality prediction model; Figure 3 Flowchart for comparative analysis of abnormal data and predicted data. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] See also Figure 1-3 The present invention provides a technical solution: a method for monitoring the power quality of a distributed photovoltaic grid-connected distribution network, the method comprising: Step 1: Use power quality monitoring devices and environmental monitoring devices deployed at distributed photovoltaic grid-connected points to collect power quality data and environmental data of the distribution network in real time. Power quality data includes but is not limited to voltage, current, frequency, power factor, harmonic content, three-phase imbalance, voltage fluctuation and flicker, as well as the power generation of distributed photovoltaics in the power grid and the location information of the grid-connected points. Environmental data includes factors that may affect the power generation of distributed photovoltaics, such as temperature, humidity, wind speed and solar irradiance.

[0022] Step 2: Preprocess the collected data. This includes data cleaning to remove invalid or erroneous data; data alignment to ensure the temporal consistency of data from different sources; and time synchronization to ensure that all data is based on a unified time reference. After preprocessing, a standardized data set is formed to provide a basis for subsequent analysis.

[0023] Step 3: Based on the normalized data set, a power quality prediction model is constructed. The model uses machine learning or deep learning algorithms, combined with environmental data and distributed photovoltaic power generation characteristics, such as the relationship between light intensity and power generation, the impact of temperature on photovoltaic cell efficiency, etc., to train the model to predict the power quality indicators of key nodes in the distribution network.

[0024] Step 4: Continue to obtain real-time power quality data and environmental data of distributed photovoltaic grid-connected points to ensure that the input data of the model is always up to date.

[0025] Step 5: Input the real-time acquired data into the power quality prediction model, and the model outputs the predicted power quality indicators based on the input data. These indicators reflect the possible power quality status of key nodes in the distribution network in the future.

[0026] Steps: Perform anomaly detection on the power quality data obtained in real time. By setting reasonable thresholds and judgment logic, identify power quality problems such as voltage over-limit, current anomaly, frequency deviation, unqualified power factor, excessive harmonics, three-phase imbalance, voltage fluctuation and flicker.

[0027] Step 7: Compare and analyze the identified power quality anomaly data with the predicted power quality indicators output by the model. Through comparative analysis, the cause and distribution of the anomaly can be determined, such as whether it is due to fluctuations in distributed photovoltaic power generation caused by changes in environmental factors or due to problems in the power grid itself.

[0028] Steps: Evaluate the power quality of the distribution network according to the preset power quality standards and thresholds. The evaluation results are presented in the form of a power quality evaluation report, which lists in detail the evaluation results and existing problems of various power quality indicators.

[0029] Step 9: Send the power quality assessment report and abnormal warning information to relevant personnel via email, SMS or other communication methods. Relevant personnel can take timely measures to improve power quality based on the report and warning information, such as adjusting the distributed photovoltaic power generation plan, optimizing the grid operation mode or performing equipment maintenance.

[0030] Through the above steps, the present invention realizes comprehensive monitoring and prediction of the power quality of the distributed photovoltaic grid-connected distribution network, providing a strong guarantee for the safe and stable operation of the distribution network.

[0031] The present invention will be further described below in conjunction with Examples 1 to 3: Embodiment 1: In the steps of building a power quality prediction model, the detailed implementation is as follows: Step 3.1: Take the normalized dataset obtained by preprocessing as input and divide it into training set and test set reasonably. The training set is used for model training and optimization, while the test set is used to evaluate the prediction performance of the model. The division ratio can be adjusted according to the actual situation. It is usually recommended that the training set accounts for a larger proportion to ensure that the model can be fully trained.

[0032] Step 3.2: Select the long short-term memory network (LSTM) as the machine learning algorithm to build the power quality prediction model. LSTM is a recurrent neural network (RNN) variant that is particularly suitable for processing time series data. It can capture long-term dependencies in the data and is very suitable for tasks such as power quality prediction that have time series characteristics.

[0033] When building an LSTM model, first set the number of layers, number of hidden layer units, and activation function of the LSTM network. The number of layers can be determined based on the complexity of the task and the amount of data. Generally, you can start with one layer and gradually increase the number of layers until the model performance is no longer significantly improved. The number of hidden layer units determines the representation ability of the model, and it is also necessary to find the optimal value through experiments. The activation function usually chooses ReLU (Rectified Linear Unit) or its variants to introduce nonlinear characteristics.

[0034] Then, configure the input layer, hidden layer, and output layer structure of the LSTM network. The input layer receives a normalized data set as input, which includes historical power quality data (such as voltage, current, etc.), environmental data (such as temperature, humidity, etc.), distributed photovoltaic power generation, and the location information of the grid connection point. These data are organized into time series and input into the LSTM model. The hidden layer contains multiple LSTM units, which control the flow of information through internal gating mechanisms (forget gate, input gate, and output gate) to capture long-term dependencies in time series data. The output layer outputs predicted power quality indicators, which also include voltage, current, frequency, power factor, harmonic content, three-phase unbalance, voltage fluctuation and flicker, etc.

[0035] Step 3.3: Use the training set to train the LSTM model. Cross-validation and grid search methods are used to optimize model parameters during training. Cross-validation evaluates the performance of the model and prevents overfitting by dividing the training set into multiple subsets, using one of the subsets as the validation set and the remaining subsets as the training set in turn. Grid search traverses the parameter space to find the parameter combination that optimizes the model performance. Back-propagation algorithm and gradient descent algorithm are used during training to iteratively optimize model parameters until the preset stop condition (such as the number of iterations, loss function convergence, etc.) is reached. The calculation formula of the gradient descent algorithm is: in, Represents the model parameters, including all weights and biases that the model needs to learn; Represents the learning rate, which is used to control the step size of parameter update; Represents the loss function About parameters The gradient of is a vector pointing to the direction where the loss function grows fastest; Indicates an assignment operation, that is, updating parameters The value of .

[0036] Step 3.4: Use the test set to evaluate the trained LSTM model. Evaluation indicators can be selected such as mean square error (MSE), root mean square error (RMSE), mean absolute error (MAE), etc. to quantify the prediction performance of the model. If the prediction performance of the model is not ideal, you need to return to step 3.3 to continue optimizing the model parameters or adjusting the model structure.

[0037] Step 3.5: Adjust and optimize the model based on the evaluation results. This may include increasing or decreasing the number of hidden layer units, adjusting the learning rate, introducing regularization terms, etc. Through continuous iterative training and optimization, until the prediction accuracy of the model reaches a satisfactory level.

[0038] Step 3.6: Save and deploy the trained power quality prediction model. To save the model, you can use the model saving function provided by deep learning frameworks such as TensorFlow and PyTorch to save the model parameters and structure as a file. To deploy the model, you need to integrate the model into the power quality monitoring system to achieve real-time prediction and monitoring functions. In this way, when new power quality data and environmental data arrive, the model can quickly give prediction results, providing strong support for the safe and stable operation of the distribution network.

[0039] Embodiment 2: In the steps of anomaly detection and comparative analysis, the detailed implementation is as follows: For anomaly detection in step 6: First, set the threshold range of power quality indicators. These thresholds are determined based on national power quality standards, grid operation experience, and the characteristics of distributed photovoltaic grid connection. For example, the upper and lower limits of voltage are set to ±10% of the rated voltage, the allowable range of current is ±20% of the rated current, and the frequency deviation limit is ±0.5Hz. These thresholds serve as a benchmark for judging whether the power quality is abnormal.

[0040] Check the power quality data acquired in real time point by point. By traversing each data point in the data set, determine whether each indicator exceeds the set threshold range. For example, for voltage data, check whether each data point is lower than the voltage lower limit or higher than the voltage upper limit; for current data, check whether it is lower than the lower limit of the allowable range or higher than the upper limit; for frequency data, check whether it deviates from the set deviation limit.

[0041] When an indicator is found to exceed the set threshold range, the indicator will be marked. The marking content includes the time and location of the abnormality (i.e. the location information of the grid connection point) and the specific indicator value. This information will be used for subsequent comparative analysis and problem location.

[0042] Comparative analysis in step 7: First, the abnormal data marked in step 6 is time-aligned with the predicted power quality indicators output by the model in step 5. Since the abnormal data is acquired in real time and the predicted data is generated based on historical data, the two need to be aligned through timestamps to ensure that the data compared are at the same time point.

[0043] Calculate the deviation between abnormal data and predicted data. For each abnormal indicator, calculate the difference between the measured value and the predicted value to obtain the deviation. The size of the deviation reflects the degree of deviation between the actual power quality and the predicted value.

[0044] Analyze the causes and trends of deviations. Combine environmental factors (such as temperature, humidity, wind speed, solar irradiance, etc.) and the power generation characteristics of distributed photovoltaics to analyze whether the deviation is caused by fluctuations in distributed photovoltaic power generation caused by changes in environmental factors or by problems in the power grid itself (such as equipment failure, line aging, etc.). At the same time, observe the trend of the deviation to determine whether the anomaly is sporadic or continuous.

[0045] Finally, based on the results of the deviation and cause analysis, determine whether the power quality abnormality is caused by distributed photovoltaic grid connection or other factors. If the deviation is large and closely related to changes in environmental factors, and there is also a significant fluctuation in the distributed photovoltaic power generation power, then it can be determined that the abnormality is caused by distributed photovoltaic grid connection. If the deviation is small or has no obvious correlation with environmental factors, it may be caused by problems in the power grid itself. This judgment result will provide an important basis for subsequent troubleshooting and power quality improvement.

[0046] Embodiment 3: In the steps of generating and sending the power quality assessment report and abnormal warning, the detailed implementation method is as follows: For power quality assessment report generation in step 8: Conduct detailed statistics and analysis on the power quality anomalies identified in step 7. This includes the type of anomaly (such as voltage fluctuation, frequency deviation, excessive harmonic content, etc.), quantity (the total number of abnormal events), distribution (geographical distribution of anomalies in the distribution network), and severity (the degree of impact of anomalies on grid operation and equipment safety). Through these analyses, a comprehensive understanding of the power quality anomaly can be obtained.

[0047] According to the preset power quality standards and thresholds, the power quality of the distribution network is evaluated as a whole. This includes comparing various power quality indicators with national and industry standards, judging whether they meet the standards, and giving corresponding evaluation results and grades (such as excellent, good, qualified, unqualified, etc.). The evaluation results should objectively reflect the overall status of the power quality of the distribution network.

[0048] Based on the results of abnormal analysis and overall assessment, proposed measures to improve power quality are put forward. These measures may include adjusting the grid connection mode of distributed photovoltaic (such as changing the location of the grid connection point, adjusting the grid connection capacity, etc.), optimizing the grid structure (such as strengthening the transmission capacity of the grid, improving the power supply reliability of the grid, etc.), and strengthening power quality monitoring (such as increasing monitoring points, improving the accuracy and real-time performance of monitoring equipment, etc.). The proposed measures should be targeted at practical problems and be operational and effective.

[0049] Write a power quality assessment report. The report should include three parts: assessment results, abnormal analysis, and recommended measures. The assessment results part should list the assessment results and grades of various power quality indicators in detail; the abnormal analysis part should describe the abnormal situation, cause and impact; the recommended measures part should propose specific improvement measures and implementation plans. The report should have a clear structure, complete content, and accurate language to facilitate understanding and use by relevant personnel.

[0050] For the power quality assessment report sending and abnormal warning in step 9: Send the power quality assessment report to relevant personnel in an appropriate manner. This can be achieved through email, file transfer (such as FTP, SFTP, etc.) or system push (such as corporate internal communication system, power quality management system, etc.). Ensure that the report can be delivered to relevant personnel in a timely and accurate manner so that they can understand the power quality situation and take appropriate measures.

[0051] Provide real-time warning for power quality anomalies. When an anomaly is detected, relevant personnel should be notified immediately via SMS, phone, email or system interface. The warning information should include key information such as the type, location, severity and time of occurrence of the anomaly, so that relevant personnel can respond quickly and handle the anomaly.

[0052] Determine the level and frequency of warnings based on the urgency and severity of the abnormal situation. For urgent and serious abnormalities, high-level warning information should be sent, and the frequency of warnings should be increased to ensure that relevant personnel can be informed and handled in a timely manner. For non-urgent or minor abnormalities, low-level warning information can be sent, and the frequency of warnings can be appropriately reduced to avoid excessive interference. By setting reasonable warning levels and frequencies, the effectiveness and pertinence of warning information can be ensured.

[0053] The present invention also includes a distributed photovoltaic grid-connected distribution network power quality monitoring system, the system comprising: The data acquisition module is used to collect the power quality data of the distribution network of the distributed photovoltaic grid-connected points, including voltage, current, frequency, power factor, harmonic content, three-phase imbalance, voltage fluctuation and flicker, as well as the power generation power of the distributed photovoltaic in the grid and the location information of the grid-connected points; it also collects environmental data, including temperature, humidity, wind speed and solar irradiance; The data preprocessing module is connected to the data acquisition module and is used to preprocess the collected data, including data cleaning, data alignment and time synchronization, to form a standardized data set; A power quality prediction model building module is connected to the data preprocessing module and is used to build a power quality prediction model based on a normalized data set, combined with environmental data and distributed photovoltaic power generation characteristics; Real-time data acquisition module, used to obtain power quality data and environmental data of distributed photovoltaic grid-connected points in real time; A prediction index output module is connected to the power quality prediction model building module and the real-time data acquisition module, and is used to input real-time data into the power quality prediction model and receive the predicted power quality index output by the model; An anomaly detection module, connected to the real-time data acquisition module, is used to perform anomaly detection on the power quality data acquired in real time and identify power quality problems; A comparison and analysis module, connected to the anomaly detection module and the prediction index output module, is used to compare and analyze the identified power quality anomaly data with the predicted power quality index to determine the cause and distribution of the anomaly; The power quality assessment module is connected to the comparison and analysis module and is used to assess the power quality of the distribution network according to preset power quality standards and thresholds and generate a power quality assessment report; The report sending module is connected to the power quality assessment module and is used to send the power quality assessment report and abnormal warning information to relevant personnel.

[0054] The implementation of the system refers to the above embodiment and will not be described in detail in the specification.

[0055] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0056] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for monitoring power quality of a distributed photovoltaic grid-connected distribution network, characterized in that: The method comprises: Step 1: Collect power quality data of the distribution network of distributed photovoltaic grid-connected points, including voltage, current, frequency, power factor, harmonic content, three-phase unbalance, voltage fluctuation and flicker, as well as the power generation power of distributed photovoltaics in the grid and the location information of the grid-connected points; also collect environmental data such as temperature, humidity, wind speed and solar irradiance; Step 2: Preprocess the data collected in step 1, including data cleaning, data alignment and time synchronization, to form a standardized data set; Step 3: Based on the normalized data set, a power quality prediction model is constructed. The model combines environmental data and the power generation characteristics of distributed photovoltaics to predict the power quality indicators of key nodes in the distribution network. Step 4: Obtain power quality data and environmental data of distributed photovoltaic grid-connected points in real time; Step 5: Input the data obtained in step 4 into the power quality prediction model, and receive the predicted power quality index output by the model; Step 6: Perform anomaly detection on the power quality data acquired in real time to identify power quality problems, including voltage over-limit, current abnormality, frequency deviation, unqualified power factor, excessive harmonics, three-phase imbalance, voltage fluctuation and flicker; Step 7: Compare and analyze the power quality anomaly data identified in step 6 with the predicted power quality index output by the model in step 5 to determine the cause and distribution of the anomaly; Step 8: Evaluate the power quality of the distribution network according to the preset power quality standards and thresholds, and generate a power quality evaluation report; Step 9: Send the power quality assessment report and abnormal warning information to relevant personnel so that timely measures can be taken to improve the power quality.

2. The method for monitoring power quality of a distributed photovoltaic grid-connected distribution network according to claim 1, characterized in that: In step 2, the data preprocessing method includes: For missing values, the average value of adjacent data points or linear interpolation is used to fill them; For outliers, statistical algorithms are used to identify them and replace or delete them with reasonable values; Perform time synchronization on the data to make the time labels of each data point consistent.

3. The method for monitoring power quality of a distributed photovoltaic grid-connected distribution network according to claim 1, characterized in that: In step 3, the steps of constructing a power quality prediction model include: Step 3.1: Use the normalized dataset as input and divide the dataset into training and test sets. Step 3.2: Select appropriate machine learning algorithm and build power quality prediction model; Step 3.3: Use the training set to train the model and optimize the model parameters through cross-validation and grid search methods; Step 3.4: Use the test set to evaluate the trained model and verify the model's predictive performance; Step 3.5: Adjust and optimize the model according to the evaluation results until the prediction accuracy of the model reaches a satisfactory level; Step 3.6: Save and deploy the trained power quality prediction model.

4. The method for monitoring power quality of a distributed photovoltaic grid-connected distribution network according to claim 3, characterized in that: In step 3.2, the selected machine learning algorithm is a long short-term memory network LSTM, and the steps of building an LSTM model include: Set the number of layers, number of hidden layer units and activation function of the LSTM network; Configure the input layer, hidden layer, and output layer structure of the LSTM network; The LSTM model is trained using the training set, and the model parameters are iteratively optimized through the back propagation algorithm and the gradient descent algorithm. The calculation formula of the gradient descent algorithm is: in, Represents the model parameters, including all weights and biases that the model needs to learn; Represents the learning rate, which is used to control the step size of parameter update; Represents the loss function About parameters The gradient of is a vector pointing to the direction where the loss function grows fastest; Indicates an assignment operation, that is, updating parameters The value of .

5. The method for monitoring power quality of a distributed photovoltaic grid-connected distribution network according to claim 4, characterized in that: The input layer of the LSTM model receives a normalized data set as input, including historical power quality data, environmental data, distributed photovoltaic power generation and location information of the grid connection point; the hidden layer contains multiple LSTM units for capturing long-term dependencies in time series data; the output layer outputs predicted power quality indicators, including voltage, current, frequency, power factor, harmonic content, three-phase unbalance, and voltage fluctuation and flicker.

6. The method for monitoring power quality of a distributed photovoltaic grid-connected distribution network according to claim 1, characterized in that: In step 6, the method for abnormality detection includes: Set the threshold range of power quality indicators, including upper and lower limits of voltage, allowable range of current, and deviation limit of frequency; Check the power quality data obtained in real time point by point to determine whether each indicator exceeds the set threshold range; Mark indicators that exceed the threshold, and record the time, location, and specific indicator value of the anomaly.

7. The method for monitoring power quality of a distributed photovoltaic grid-connected distribution network according to claim 1, characterized in that: In step 7, the comparative analysis step includes: Time-align the abnormal data marked in step 6 with the predicted power quality indicators output by the model in step 5; Calculate the deviation between abnormal data and predicted data, and analyze the causes and trends of the deviation; Based on the deviation amount and cause analysis, determine whether the abnormal power quality is caused by distributed photovoltaic grid connection or by other factors.

8. The method for monitoring power quality of a distributed photovoltaic grid-connected distribution network according to claim 1, characterized in that: In step 8, the step of generating the power quality assessment report includes: Conduct statistics and analysis on the power quality anomalies identified in step 7, and summarize the types, quantity, distribution and severity of the anomalies; According to the preset power quality standards and thresholds, the power quality of the distribution network is evaluated as a whole, and the evaluation results and grades are given; Propose measures to improve power quality, including adjusting the grid connection mode of distributed photovoltaics, optimizing the grid structure and strengthening power quality monitoring; Prepare power quality assessment report, including assessment results, abnormality analysis and recommended measures.

9. The method for monitoring power quality of a distributed photovoltaic grid-connected distribution network according to claim 1, characterized in that: In step 9, the step of sending the power quality assessment report and abnormal warning information includes: Send the power quality assessment report to relevant personnel by email, file transfer or system push; Provide real-time warning of abnormal power quality and notify relevant personnel via SMS, phone, email or system interface; Determine the level and frequency of warnings based on the urgency and severity of the abnormal situation.

10. A distributed photovoltaic grid-connected distribution network power quality monitoring system, characterized in that: The system comprises: The data acquisition module is used to collect the power quality data of the distribution network of the distributed photovoltaic grid-connected points, including voltage, current, frequency, power factor, harmonic content, three-phase imbalance, voltage fluctuation and flicker, as well as the power generation power of the distributed photovoltaic in the grid and the location information of the grid-connected points; it also collects environmental data, including temperature, humidity, wind speed and solar irradiance; The data preprocessing module is connected to the data acquisition module and is used to preprocess the collected data, including data cleaning, data alignment and time synchronization, to form a standardized data set; A power quality prediction model building module is connected to the data preprocessing module and is used to build a power quality prediction model based on a normalized data set, combined with environmental data and distributed photovoltaic power generation characteristics; Real-time data acquisition module, used to obtain power quality data and environmental data of distributed photovoltaic grid-connected points in real time; A prediction index output module is connected to the power quality prediction model building module and the real-time data acquisition module, and is used to input real-time data into the power quality prediction model and receive the predicted power quality index output by the model; An anomaly detection module, connected to the real-time data acquisition module, is used to perform anomaly detection on the power quality data acquired in real time and identify power quality problems; A comparison and analysis module, connected to the anomaly detection module and the prediction index output module, is used to compare and analyze the identified power quality anomaly data with the predicted power quality index to determine the cause and distribution of the anomaly; The power quality assessment module is connected to the comparison and analysis module and is used to assess the power quality of the distribution network according to preset power quality standards and thresholds and generate a power quality assessment report; The report sending module is connected to the power quality assessment module and is used to send the power quality assessment report and abnormal warning information to relevant personnel.

Images