AC / DC power supply system monitoring device and method

Abstract

The present invention relates to the field of power electronics and energy management technology. An AC / DC power system monitoring device and method are provided, comprising the steps of: constructing a historical data set; preprocessing the historical data set to obtain a preprocessed data set; constructing a system monitoring model based on the preprocessed data set and a machine learning algorithm, and presetting a voltage anomaly determination threshold of the system monitoring model; obtaining real-time power data of the AC / DC power system, as well as the real-time operating status and load change data of the power grid, inputting the data into the system monitoring model, and identifying the pattern, trend and potential anomaly of the voltage fluctuation of the AC / DC power system; dynamically adjusting the preset voltage anomaly determination threshold, and providing the corresponding AC / DC power system adjustment strategy and early warning information. The present invention solves the problems of low efficiency, frequent false alarms and missed alarms, difficulty in adapting to the complex and changeable power grid environment, and lack of intelligent analysis and early warning functions in the AC / DC power system monitoring method in the prior art.

Description

Technical Field

[0001] The present invention relates to the technical field of power electronics and energy management, and in particular to an AC / DC power supply system monitoring device and method. Background Art

[0002] In the power system, the stable operation of AC and DC power systems is crucial to ensure the quality and reliability of power supply. Traditional power system monitoring methods often rely on manually set thresholds and regular on-site inspections. This method is not only inefficient, but also prone to false alarms and missed alarms when facing complex and changing power grid environments, making it difficult to detect and solve voltage anomalies in a timely and accurate manner.

[0003] With the expansion of the scale of power grids and the diversification of power loads, voltage fluctuations and abnormal conditions in power grids are becoming increasingly complex. Traditional monitoring methods can no longer meet the high requirements of modern power grids for voltage stability and reliability. Especially in AC / DC hybrid power systems, due to the introduction of DC transmission technology and the large-scale access of renewable energy, the dynamic characteristics and uncertainties of the power grid have further increased, making voltage monitoring more difficult and complex. Traditional voltage monitoring methods usually lack intelligent analysis and early warning functions, and cannot conduct in-depth mining and analysis of voltage data, making it difficult to identify the patterns and causes of voltage fluctuations in a timely and accurate manner. At the same time, traditional voltage regulation methods often rely on manual intervention, which not only has a slow response speed, but also has limited regulation effects, making it difficult to meet the needs of modern power grids for fast and accurate voltage regulation. Summary of the invention

[0004] The purpose of the present invention is to provide an AC / DC power supply system monitoring method, aiming to solve the problems of low efficiency, frequent false alarms and missed alarms, difficulty in adapting to complex and changeable power grid environments, and lack of intelligent analysis and early warning functions in the AC / DC power supply system monitoring method in the prior art.

[0005] The present invention is achieved through the following technical solutions:

[0006] A method for monitoring an AC / DC power supply system comprises the following steps:

[0007] Acquire historical grid data on grid operation status and load changes, and historical system power data of AC and DC power systems corresponding to the historical grid data, and construct a historical data set;

[0008] Preprocessing the historical data set, and performing factor analysis on the preprocessed historical data set to obtain a preprocessed data set;

[0009] Based on the preprocessed data set and machine learning algorithm, a system monitoring model is constructed, and the voltage anomaly determination threshold of the system monitoring model is preset;

[0010] Acquire real-time power data of the AC and DC power supply systems, as well as the real-time operation status and load change data of the power grid, and input them into the system monitoring model. The system monitoring model is used to identify the patterns, trends and potential anomalies of the voltage fluctuations of the AC and DC power supply systems and obtain identification results.

[0011] According to the identification results and the actual situation of the power grid, the preset voltage anomaly judgment threshold is dynamically adjusted, and the corresponding AC / DC power system adjustment strategy and early warning information are given.

[0012] Optionally, the specific process of acquiring the historical data of the power grid on the power grid operation status and load change, and the historical power data of the AC and DC power supply systems corresponding to the historical data of the power grid, and constructing the historical data set is as follows:

[0013] Extract the historical operation status parameters of the power grid and the corresponding historical load change data from the power grid monitoring system and the recording system of the AC and DC power supply system to form the power grid historical data;

[0014] Collecting system historical power data of the AC and DC power systems corresponding to timestamps of the historical data of the power grid;

[0015] The historical data of the power grid and the historical power data of the system are aligned in time and in a unified format to construct a historical data set.

[0016] Optionally, the specific process of preprocessing the historical data set and performing factor analysis on the preprocessed historical data set to obtain the preprocessed data set is:

[0017] Clean the data in the historical data set to remove missing values, outliers, and duplicate values;

[0018] Standardize the data in the historical data set and convert the data to the same scale;

[0019] Using factor analysis techniques, the key factors that affect the voltage fluctuation of AC and DC power systems are extracted from historical data sets;

[0020] According to the results of factor analysis, the key factors are explained and named to obtain the preprocessed data set.

[0021] Optionally, the specific process of building the system monitoring model based on the preprocessed data set and the machine learning algorithm is:

[0022] According to the characteristics of the preprocessed data set and monitoring requirements, an initial system monitoring model is built based on the machine learning algorithm;

[0023] The preprocessed data set is divided into a training set and a test set. The system monitoring initial model is trained using the training set. During the training process, the hyperparameters of the system monitoring initial model are iteratively adjusted, and the model structure is optimized.

[0024] Verify the trained system monitoring initial model through the test set to evaluate the accuracy and stability of the system monitoring initial model;

[0025] According to the verification results, the model is adjusted and optimized until it meets the preset monitoring performance requirements and the system monitoring model is obtained.

[0026] Optionally, the specific process of constructing the system monitoring initial model based on the machine learning algorithm according to the characteristics of the preprocessed data set and the monitoring requirements is:

[0027] The initial model of system monitoring is constructed by long short-term memory network. The preprocessed data set is ,in Indicates at time The input feature vector is represents the total length of the time series; the corresponding output label set is ,in Indicates at time The actual voltage fluctuation state; the expression of the system monitoring initial model is shown in the following formula (1):

[0028]

[0029] in, Indicates time The hidden state of Represents the hidden state at the previous moment; Function representing a long short-term memory network;

[0030] The output layer expression of the initial system monitoring model is shown in the following formula (2):

[0031]

[0032] in, Indicates time The predicted value of represents the activation function; and denote the weight and bias of the output layer respectively.

[0033] Optionally, during the training process, the model parameters are optimized by minimizing the loss function, and the expression of the loss function is shown in the following formula (3):

[0034]

[0035] in, Represents the prediction error of the model over the entire time series; Indicates time The actual voltage fluctuation state, the value is 0 or 1, 0 means normal, 1 means abnormal;

[0036] The system monitors the weights and biases of the initial model using the back-propagation algorithm and gradient descent until the model reaches the preset performance requirements on the validation set.

[0037] Optionally, the specific process of setting the voltage anomaly determination threshold of the preset system monitoring model is:

[0038] Calculate the statistical characteristics of voltage anomaly events based on historical data sets and known voltage anomaly events;

[0039] According to the statistical characteristics of voltage anomaly events, combined with the safety standards of power grid operation and the design requirements of AC and DC power supply systems, a voltage anomaly judgment threshold is preliminarily set, and the preliminarily set voltage anomaly judgment threshold is applied to the system monitoring model.

[0040] Optionally, the specific process of dynamically adjusting the preset voltage anomaly determination threshold according to the recognition result and the actual situation of the power grid is:

[0041] Analyze the identification results output by the system monitoring model, evaluate the actual impact range and potential risks of voltage fluctuations, and obtain the evaluation results;

[0042] Based on the evaluation results, the initially set voltage anomaly determination threshold is dynamically fine-tuned;

[0043] The adjusted voltage anomaly judgment threshold is applied in real time to the system monitoring model to respond and regulate the voltage fluctuations of the AC and DC power supply systems, and is continuously optimized and adjusted based on the actual application effects.

[0044] Optionally, the specific process of providing the corresponding AC / DC power system adjustment strategy and warning information is as follows:

[0045] When the system monitoring model identifies voltage fluctuations or potential anomalies, the category and urgency of the voltage anomaly are determined based on the voltage fluctuation pattern, trend, and degree of anomaly in the identification results;

[0046] Based on the type and urgency of the voltage anomaly, a corresponding regulation strategy is selected from a preset regulation strategy library;

[0047] Generate an adjustment instruction including an adjustment strategy and specific operation steps, and send the adjustment instruction to the corresponding device of the AC / DC power supply system through a remote control system or an automatic control system to perform the adjustment operation;

[0048] Generate early warning information based on voltage anomalies and regulation strategies;

[0049] Send warning information to relevant power grid operators and managers and take corresponding actions;

[0050] According to the actual effect of the regulation operation and the changes in the operating status of the power grid, the output of the system monitoring model is continuously monitored, and the regulation strategy and early warning information are adjusted as needed.

[0051] The present invention also provides an AC / DC power system monitoring device, which is used to implement the AC / DC power system monitoring method, comprising:

[0052] The data acquisition module is used to obtain the historical data of the power grid on the operation status and load change of the power grid, and the historical power data of the AC and DC power supply systems corresponding to the historical data of the power grid, to construct a historical data set, and to obtain the real-time power data of the AC and DC power supply systems, as well as the real-time operation status and load change data of the power grid in real time;

[0053] A data preprocessing module is used to preprocess the historical data set, and to perform factor analysis on the preprocessed historical data set, to extract key factors affecting the voltage fluctuation of the AC and DC power supply systems, and to obtain a preprocessed data set;

[0054] A model building module, used to build a system monitoring model based on a preprocessed data set and a machine learning algorithm, and to preset a voltage anomaly determination threshold of the system monitoring model;

[0055] A monitoring and identification module is used to input the real-time acquired data into the system monitoring model, identify the pattern, trend and potential abnormality of the voltage fluctuation of the AC and DC power supply system through the system monitoring model, and obtain the identification result;

[0056] The threshold adjustment and strategy generation module is used to dynamically adjust the preset voltage anomaly judgment threshold according to the recognition results and the actual situation of the power grid, and provide the corresponding AC and DC power system adjustment strategy and early warning information;

[0057] The communication module is used to send adjustment instructions and warning information to relevant power grid operators and managers, and to send adjustment instructions to corresponding equipment in the AC and DC power supply system to perform adjustment operations.

[0058] The technical solution of the present invention has at least the following advantages and beneficial effects:

[0059] Improve monitoring efficiency and accuracy: By constructing historical data sets and performing preprocessing and factor analysis, we can deeply explore the intrinsic relationship between the grid operation status, load changes and AC / DC power system power data. Therefore, the system monitoring model built based on the machine learning algorithm can more accurately identify the patterns, trends and potential anomalies of voltage fluctuations. Compared with traditional manual threshold setting and regular inspections, this method significantly improves the efficiency and accuracy of monitoring and reduces false alarms and missed alarms.

[0060] Enhanced intelligent analysis and early warning capabilities: Through rapid processing of real-time data, the system can automatically identify voltage anomalies, issue early warning information in a timely manner, provide decision-making support for operation and maintenance personnel, effectively shorten fault response time, and improve the overall stability and reliability of the power system.

[0061] Adapt to complex power grid environments: By dynamically adjusting the voltage anomaly judgment threshold, a flexible response to power grid changes is achieved, which not only improves the accuracy and efficiency of voltage regulation, but also enhances the system's adaptability to complex power grid environments.

[0062] Optimize voltage regulation strategy: The regulation strategy and early warning information automatically generated by the system monitoring model can guide fast and accurate voltage regulation operations, effectively avoid the impact of voltage fluctuations on the stable operation of the power grid, and improve the quality and reliability of power supply. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 A schematic diagram of a flow chart of an AC / DC power system monitoring method according to an embodiment of the present invention;

[0064] Figure 2 Schematic diagram of the structure of an AC / DC power supply system monitoring device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0065] The following is a specific implementation method in conjunction with the drawings.

[0066] Reference Figure 1 , a method for monitoring an AC / DC power supply system, comprising the steps of:

[0067] Step 1: Obtain historical grid data on grid operation status and load changes, as well as historical system power data of AC and DC power systems corresponding to the historical grid data, and construct a historical data set.

[0068] In some embodiments, the specific process of acquiring the historical data of the power grid on the power grid operation status and load changes, and the system historical power data of the AC and DC power supply systems corresponding to the historical data of the power grid, and constructing the historical data set is as follows:

[0069] The historical operating status parameters of the power grid are extracted from the recording systems of the power grid monitoring system and the AC / DC power supply system. The historical operating status parameters of the power grid are extracted from the power grid monitoring system according to the time range (such as the past one or two years), including voltage level, current size, frequency fluctuation, etc.; as well as the corresponding historical load change data, including total load, peak and valley load, load change trend, etc., to form the historical data of the power grid.

[0070] Collect the system historical power data of the AC and DC power supply system corresponding to the timestamp of the historical data of the power grid; extract the system historical power data of the corresponding time point from the AC and DC power supply system recording system according to the timestamp of the historical data of the power grid, and ensure that the extracted data completely corresponds to the historical data of the power grid in time, so as to facilitate subsequent data analysis and processing.

[0071] The historical data of the power grid and the historical power data of the system are time-aligned and formatted to construct a historical data set. Since there may be time synchronization issues between the power grid monitoring system and the AC / DC power system recording system, the extracted data needs to be time-aligned to ensure that all data points are on the same timeline; the processed historical data of the power grid and the historical power data of the system are integrated to form a complete historical data set; the integrated historical data set is stored in a suitable data storage medium, such as a database, data warehouse or cloud storage, to ensure data security and accessibility for subsequent data analysis and model building.

[0072] Step 2: Preprocess the historical data set, and perform factor analysis on the preprocessed historical data set to obtain the preprocessed data set.

[0073] In some embodiments, the historical data set is preprocessed, and factor analysis is performed on the preprocessed historical data set to obtain the preprocessed data set in the following specific process:

[0074] Clean the data in the historical data set to remove missing values, outliers, and duplicate values. For missing values ​​in the historical data set, select appropriate filling methods (such as mean filling, median filling, interpolation, etc.) to fill them according to the characteristics of the data and the distribution of the missing values, or directly delete the records containing missing values; use statistical methods (such as box plots, Z scores, etc.) to identify and remove outliers in the historical data set to ensure data accuracy and consistency; check and delete duplicate records in the historical data set to avoid interference with subsequent analysis.

[0075] Standardize the data in the historical data set and convert the data to the same scale; you can use methods such as Z-score standardization and Min-Max standardization to convert the data to a mean of 0, a standard deviation of 1, or a value within a specified range.

[0076] Using factor analysis technology, the key factors that affect the voltage fluctuation of the AC and DC power supply system are extracted from the historical data set; the extracted key factors can reflect the main influencing factors of the voltage fluctuation of the AC and DC power supply system, such as grid load changes, grid voltage fluctuations, and the operating status of the AC and DC power supply system.

[0077] According to the results of factor analysis, the key factors are explained and named to obtain the preprocessed data set.

[0078] In some embodiments, relevant variables are selected from the preprocessed historical data set, and these variables cover the operating status of the power grid, load changes, and power data of the AC and DC power systems; ensure that the data has been standardized to eliminate the dimensional differences between different variables; calculate the correlation coefficient matrix between the selected variables to evaluate the linear correlation between the variables, and the elements of the correlation coefficient matrix represent the correlation coefficients between the variables, and their values ​​are between -1 and 1; according to the correlation coefficient matrix, select appropriate statistical methods (such as Kaiser-Guttman criterion, scree plot, etc.) to determine the number of factors; use principal component analysis (PCA) or other factor extraction methods to extract the factor loading matrix from the correlation coefficient matrix, and the factor loading matrix represents the load of the original variable on the factor, that is, the contribution of the variable to the factor; rotate the extracted factors to simplify the factor structure and improve the interpretability of the factors, and orthogonal rotation (such as Varimax) and oblique rotation (such as Promax) can be used; based on the factor loading matrix and the factor rotation results, explain the meaning of each factor and name it.

[0079] Step 3: Based on the preprocessed data set and machine learning algorithm, a system monitoring model is constructed, and the voltage anomaly determination threshold of the system monitoring model is preset.

[0080] In some embodiments, based on the preprocessed data set and the machine learning algorithm, the specific process of building the system monitoring model is as follows:

[0081] According to the characteristics of the preprocessed data set and the monitoring requirements, the initial system monitoring model is constructed based on the machine learning algorithm; the preprocessed data set is divided into a training set and a test set, the training set is used for model training, and the test set is used for model verification; the ratio of the training set and the test set can be adjusted according to actual needs, such as 70% of the training set and 30% of the test set. In the preprocessed data set, features closely related to the voltage fluctuation of the AC and DC power supply system are selected, including grid voltage, current, load change rate, grid frequency, etc.; through feature selection, the training efficiency and prediction accuracy of the model can be improved. The initial system monitoring model is trained using the training set. During the training process, the hyperparameters of the initial system monitoring model are iteratively adjusted, and the model structure is optimized; the trained initial system monitoring model is verified through the test set to evaluate the accuracy and stability of the initial system monitoring model; according to the verification results, the model is adjusted and optimized until it meets the preset monitoring performance requirements, and the system monitoring model is obtained.

[0082] In some embodiments, according to the characteristics of the preprocessed data set and the monitoring requirements, the specific process of building the system monitoring initial model based on the machine learning algorithm is as follows:

[0083] The initial model of system monitoring is constructed by long short-term memory network. The preprocessed data set is ,in Indicates at time The input feature vector is represents the total length of the time series; the corresponding output label set is ,in Indicates at time The actual voltage fluctuation state; the expression of the system monitoring initial model is shown in the following formula (1):

[0084]

[0085] in, Indicates time The hidden state of Represents the hidden state at the previous moment; Function representing a long short-term memory network;

[0086] The output layer expression of the initial system monitoring model is shown in the following formula (2):

[0087]

[0088] in, Indicates time The predicted value of represents the activation function; and denote the weight and bias of the output layer respectively.

[0089] In some embodiments, during the training process, the model parameters are optimized by minimizing the loss function, and the expression of the loss function is shown in the following formula (3):

[0090]

[0091] in, Represents the prediction error of the model over the entire time series; Indicates time The actual voltage fluctuation state, the value is 0 or 1, 0 means normal, 1 means abnormal;

[0092] The system monitors the weights and biases of the initial model using the back-propagation algorithm and gradient descent until the model reaches the preset performance requirements on the validation set.

[0093] In some embodiments, the specific process of presetting the voltage anomaly determination threshold of the system monitoring model is:

[0094] Based on historical data sets and known voltage anomaly events, calculate the statistical characteristics of voltage anomaly events, such as mean, variance, maximum, minimum, etc.;

[0095] According to the statistical characteristics of abnormal voltage events, combined with the safety standards of power grid operation and the design requirements of AC and DC power supply systems, the voltage anomaly determination threshold is preliminarily set, and the preliminarily set voltage anomaly determination threshold is applied to the system monitoring model. During the model verification process, the preliminarily set threshold can be dynamically fine-tuned according to the model's prediction results and the actual situation of abnormal voltage events to make the threshold more in line with the actual situation.

[0096] Step 4: Acquire real-time power data of the AC / DC power supply system, as well as the real-time operating status and load change data of the power grid, and input them into the system monitoring model. The system monitoring model is used to identify the pattern, trend and potential anomaly of the voltage fluctuation of the AC / DC power supply system to obtain the identification result.

[0097] In some embodiments, the real-time operating state parameters (such as voltage, current, frequency, etc.) and load change data of the power grid are extracted from the real-time data acquisition system of the power grid monitoring system and the AC / DC power supply system; at the same time, the real-time power data of the AC / DC power supply system is collected, including key parameters such as the output voltage, current, and power of each power module. The real-time data is cleaned to remove noise and outliers. This can be achieved by setting a reasonable threshold or using data smoothing technology; the real-time data and the historical data set are formatted in a unified manner to ensure data consistency and comparability. According to the input requirements of the system monitoring model, the pre-processed real-time data is converted into a format that the model can recognize; the converted real-time data is input into the system monitoring model, which can be achieved through a programming interface (API) or direct data import. After receiving the real-time data, the system monitoring model starts running and analyzing. The system monitoring model will process the data to identify the mode, trend, and potential anomaly of voltage fluctuation. After the model runs, the recognition result is output, and the recognition result includes the mode (such as periodic fluctuation, sudden fluctuation, etc.), trend (such as rising trend, falling trend, etc.) of voltage fluctuation, and potential anomaly (such as voltage is too high, too low, fluctuation is too large, etc.). Based on the preset evaluation criteria and historical experience, the identification results can be evaluated to determine the actual impact range and potential risks of voltage fluctuations. Based on the evaluation results, the output of the system monitoring model is fed back and adjusted. If there is a deviation between the identification results and the actual situation, the model can be fine-tuned or retrained to improve accuracy. If voltage anomalies or potential risks are identified, the system should immediately issue early warning information. Relevant grid operators and managers can be notified through SMS, email, system messages, etc.; based on the early warning information, corresponding response measures are formulated, including adjusting the output voltage, current and other parameters of the AC and DC power supply system, or starting the backup power supply, etc. Execute the response measures, and continuously monitor the voltage fluctuations of the AC and DC power supply system through the system monitoring model. If the anomaly persists or worsens, the response measures should be adjusted in time and the relevant personnel should be notified.

[0098] Step 5: According to the identification results and the actual situation of the power grid, dynamically adjust the preset voltage anomaly judgment threshold, and provide corresponding AC / DC power system adjustment strategy and early warning information.

[0099] In some embodiments, according to the recognition result and the actual situation of the power grid, the specific process of dynamically adjusting the preset voltage anomaly determination threshold is as follows:

[0100] Analyze the identification results output by the system monitoring model, evaluate the actual impact range and potential risks of voltage fluctuations, and obtain the evaluation results;

[0101] Based on the evaluation results, the initially set voltage anomaly judgment threshold is dynamically fine-tuned; based on the identification results and risk assessment, the initially set voltage anomaly judgment threshold is dynamically fine-tuned, and the actual situation of voltage fluctuations, the safety standards of the power grid and the design requirements of the AC and DC power supply systems are considered during the adjustment to ensure that the adjusted threshold can more accurately reflect the actual situation of voltage fluctuations and improve the sensitivity and accuracy of the monitoring system.

[0102] The adjusted voltage anomaly judgment threshold is applied in real time to the system monitoring model to respond and regulate the voltage fluctuations of the AC and DC power supply systems, and is continuously optimized and adjusted based on the actual application effects.

[0103] In some embodiments, the specific process of providing the corresponding AC / DC power system adjustment strategy and warning information is as follows:

[0104] When the system monitoring model identifies voltage fluctuations or potential anomalies, the category and urgency of the voltage anomaly are determined based on the voltage fluctuation pattern, trend, and degree of anomaly in the identification results;

[0105] Based on the type and urgency of the voltage anomaly, the corresponding regulation strategy is selected from the preset regulation strategy library; the regulation strategy library contains a variety of strategies, such as adjusting the output voltage, current, power factor, etc., as well as starting the backup power supply, etc. In combination with the actual situation of the power grid and the design requirements of the AC / DC power supply system, a specific regulation strategy is formulated to ensure that the regulation strategy can effectively respond to voltage anomalies while ensuring the stable operation of the power grid and the safety of the AC / DC power supply system.

[0106] Generate an adjustment instruction including an adjustment strategy and specific operation steps, and send the adjustment instruction to the corresponding device of the AC / DC power supply system through a remote control system or an automatic control system to perform the adjustment operation;

[0107] Generate warning information based on the voltage anomaly situation and regulation strategy; the warning information includes a specific description of the voltage anomaly, possible impact range, urgency, and recommended response measures.

[0108] Send warning information to relevant power grid operators and managers and take corresponding actions;

[0109] According to the actual effect of the regulation operation and the changes in the grid operation status, the output of the system monitoring model is continuously monitored, and the regulation strategy and early warning information are adjusted as needed. According to the actual effect of the regulation operation and the changes in the grid operation status, the effectiveness of the regulation strategy is evaluated. If the regulation effect is not good, the strategy is adjusted in time and the regulation instruction is resent.

[0110] Based on the same inventive concept, corresponding to any of the above embodiments, refer to Figure 2The present invention also provides an AC / DC power system monitoring device, which is used to implement the above-mentioned AC / DC power system monitoring method, comprising:

[0111] The data acquisition module is used to obtain the historical data of the power grid on the operation status and load change of the power grid, and the historical power data of the AC and DC power supply systems corresponding to the historical data of the power grid, to construct a historical data set, and to obtain the real-time power data of the AC and DC power supply systems, as well as the real-time operation status and load change data of the power grid in real time;

[0112] A data preprocessing module is used to preprocess the historical data set, and to perform factor analysis on the preprocessed historical data set, to extract key factors affecting the voltage fluctuation of the AC and DC power supply systems, and to obtain a preprocessed data set;

[0113] A model building module, used to build a system monitoring model based on a preprocessed data set and a machine learning algorithm, and to preset a voltage anomaly determination threshold of the system monitoring model;

[0114] A monitoring and identification module is used to input the real-time acquired data into the system monitoring model, identify the pattern, trend and potential abnormality of the voltage fluctuation of the AC and DC power supply system through the system monitoring model, and obtain the identification result;

[0115] The threshold adjustment and strategy generation module is used to dynamically adjust the preset voltage anomaly judgment threshold according to the recognition results and the actual situation of the power grid, and provide the corresponding AC and DC power system adjustment strategy and early warning information;

[0116] The communication module is used to send adjustment instructions and warning information to relevant power grid operators and managers, and to send adjustment instructions to corresponding equipment in the AC and DC power supply system to perform adjustment operations.

Claims

1. A method for monitoring an AC / DC power supply system, characterized in that: Includes steps: Acquire historical grid data on grid operation status and load changes, and historical system power data of AC and DC power systems corresponding to the historical grid data, and construct a historical data set; Preprocessing the historical data set, and performing factor analysis on the preprocessed historical data set to obtain a preprocessed data set; The specific process of preprocessing the historical data set and performing factor analysis on the preprocessed historical data set to obtain the preprocessed data set is as follows: Clean the data in the historical data set to remove missing values, outliers, and duplicate values; Standardize the data in the historical data set and convert the data to the same scale; Using factor analysis techniques, the key factors that affect the voltage fluctuation of AC and DC power systems are extracted from historical data sets; Based on the results of factor analysis, key factors are explained and named to obtain the preprocessed data set; Based on the preprocessed data set and machine learning algorithm, a system monitoring model is constructed, and the voltage anomaly determination threshold of the system monitoring model is preset; The specific process of building a system monitoring model based on the preprocessed data set and machine learning algorithm is as follows: According to the characteristics of the preprocessed data set and monitoring requirements, an initial system monitoring model is built based on the machine learning algorithm; The specific process of building the initial system monitoring model based on the machine learning algorithm according to the characteristics of the preprocessed data set and the monitoring requirements is as follows: The initial model of system monitoring is constructed by long short-term memory network. The preprocessed data set is ,in Indicates at time The input feature vector is represents the total length of the time series; the corresponding output label set is ,in Indicates at time The actual voltage fluctuation state; the expression of the system monitoring initial model is shown in the following formula (1): in, Indicates time The hidden state of Represents the hidden state at the previous moment; Function representing a long short-term memory network; The output layer expression of the initial system monitoring model is shown in the following formula (2): in, Indicates time The predicted value of represents the activation function; and Represent the weight and bias of the output layer respectively; During the training process, the model parameters are optimized by minimizing the loss function. The expression of the loss function is shown in the following formula (3): in, Represents the prediction error of the model over the entire time series; Indicates time The actual voltage fluctuation state, the value is 0 or 1, 0 means normal, 1 means abnormal; Use the back-propagation algorithm and gradient descent to update the weights and biases of the system monitoring the initial model until the model reaches the preset performance requirements on the validation set; The preprocessed data set is divided into a training set and a test set. The system monitoring initial model is trained using the training set. During the training process, the hyperparameters of the system monitoring initial model are iteratively adjusted and the model structure is optimized. The trained system monitoring initial model is verified using the test set to evaluate the accuracy and stability of the system monitoring initial model. Based on the verification results, the model is adjusted and optimized until it meets the preset monitoring performance requirements and the system monitoring model is obtained. Acquire real-time power data of the AC and DC power supply systems, as well as the real-time operation status and load change data of the power grid, and input them into the system monitoring model. The system monitoring model is used to identify the patterns, trends and potential anomalies of the voltage fluctuations of the AC and DC power supply systems and obtain identification results. According to the identification results and the actual situation of the power grid, the preset voltage anomaly judgment threshold is dynamically adjusted, and the corresponding AC / DC power system adjustment strategy and early warning information are given.

2. The AC / DC power system monitoring method according to claim 1, characterized in that: The specific process of acquiring the historical data of the power grid on the power grid operation status and load changes, and the historical power data of the AC and DC power supply systems corresponding to the historical data of the power grid, and constructing the historical data set is as follows: Extract the historical operation status parameters of the power grid and the corresponding historical load change data from the power grid monitoring system and the recording system of the AC and DC power supply system to form the power grid historical data; Collecting system historical power data of the AC and DC power systems corresponding to timestamps of the historical data of the power grid; The historical data of the power grid and the historical power data of the system are aligned in time and in a unified format to construct a historical data set.

3. The AC / DC power system monitoring method according to claim 1, characterized in that: The specific process of presetting the voltage anomaly determination threshold of the system monitoring model is as follows: Calculate the statistical characteristics of voltage anomaly events based on historical data sets and known voltage anomaly events; According to the statistical characteristics of voltage anomaly events, combined with the safety standards of power grid operation and the design requirements of AC and DC power supply systems, a voltage anomaly judgment threshold is preliminarily set, and the preliminarily set voltage anomaly judgment threshold is applied to the system monitoring model.

4. The AC / DC power system monitoring method according to claim 3, characterized in that: The specific process of dynamically adjusting the preset voltage anomaly determination threshold according to the identification result and the actual situation of the power grid is as follows: Analyze the identification results output by the system monitoring model, evaluate the actual impact range and potential risks of voltage fluctuations, and obtain the evaluation results; Based on the evaluation results, the initially set voltage anomaly determination threshold is dynamically fine-tuned; The adjusted voltage anomaly judgment threshold is applied in real time to the system monitoring model to respond and regulate the voltage fluctuation of the AC and DC power supply systems, and is continuously optimized and adjusted according to the actual application effects.

5. The AC / DC power system monitoring method according to claim 1, characterized in that: The specific process of providing the corresponding AC / DC power system adjustment strategy and warning information is as follows: When the system monitoring model identifies voltage fluctuations or potential anomalies, the category and urgency of the voltage anomaly are determined based on the voltage fluctuation pattern, trend, and degree of anomaly in the identification results; Based on the type and urgency of the voltage anomaly, a corresponding regulation strategy is selected from a preset regulation strategy library; Generate an adjustment instruction including an adjustment strategy and specific operation steps, and send the adjustment instruction to the corresponding device of the AC / DC power supply system through a remote control system or an automatic control system to perform the adjustment operation; Generate early warning information based on voltage anomalies and regulation strategies; Send warning information to relevant power grid operators and managers and take corresponding actions; According to the actual effect of the regulation operation and the changes in the operating status of the power grid, the output of the system monitoring model is continuously monitored, and the regulation strategy and early warning information are adjusted as needed.

6. An AC / DC power system monitoring device, used to implement the AC / DC power system monitoring method according to any one of claims 1 to 5, characterized in that: include: The data acquisition module is used to obtain the historical data of the power grid on the operation status and load change of the power grid, and the historical power data of the AC and DC power supply systems corresponding to the historical data of the power grid, to construct a historical data set, and to obtain the real-time power data of the AC and DC power supply systems, as well as the real-time operation status and load change data of the power grid in real time; A data preprocessing module is used to preprocess the historical data set, and to perform factor analysis on the preprocessed historical data set, to extract key factors affecting the voltage fluctuation of the AC and DC power supply systems, and to obtain a preprocessed data set; A model building module, used to build a system monitoring model based on a preprocessed data set and a machine learning algorithm, and to preset a voltage anomaly determination threshold of the system monitoring model; A monitoring and identification module is used to input the real-time acquired data into the system monitoring model, identify the pattern, trend and potential abnormality of the voltage fluctuation of the AC and DC power supply system through the system monitoring model, and obtain the identification result; The threshold adjustment and strategy generation module is used to dynamically adjust the preset voltage anomaly judgment threshold according to the recognition results and the actual situation of the power grid, and provide the corresponding AC and DC power system adjustment strategy and early warning information; The communication module is used to send adjustment instructions and warning information to relevant power grid operators and managers, and to send adjustment instructions to corresponding equipment in the AC and DC power supply system to perform adjustment operations.

Images