Building power data processing method and system, control device and storage medium

Abstract

The invention provides a building power data processing method and system, a control device and a storage medium, and the method comprises the steps: determining the actual electrical data, sent by a plurality of collection devices, of a building power system, inputting the actual electrical data into a fault detection model obtained through pre-training, obtaining a fault control strategy of the actual electrical data, and managing the building power system according to the fault control strategy. The building power system is detected through the fault detection model according to the electrical data, the method does not depend on experience of maintenance personnel, fault analysis is more accurate, a reliable fault early warning mechanism is established, closed-loop control formed from early warning to automatic execution is achieved, and the fault early warning efficiency is improved. An intelligent, automatic and extensible building power system fault prevention solution is formed, and the operation reliability and the maintenance efficiency of the building power system are remarkably improved.

Description

Technical Field

[0001] The present application relates to the technical field of fault detection, and in particular to a building power data processing method, system, control device and storage medium. Background Art

[0002] With the development of power systems, the reliable operation and maintenance efficiency of power systems can be ensured through fault monitoring and prevention of building power systems.

[0003] In the prior art, fault monitoring and prevention of building power systems are achieved in two ways. The first is regular manual inspections, specifically, maintenance personnel regularly use handheld test equipment to check electrical equipment. The second is an automated detection method, such as deploying a Supervisory Control And Data Acquisition (SCADA) system to monitor electrical parameters, or using an infrared thermal imaging system to monitor equipment temperature.

[0004] However, the first method in the prior art has a low detection frequency, discontinuous data, and cannot detect defects such as sudden failures in a timely manner. Manual inspection depends on the experience level of maintenance personnel, and it is difficult to establish a reliable fault warning mechanism. The second method has a single monitoring dimension and cannot fully reflect the health status of the equipment. Summary of the invention

[0005] The purpose of this application is to provide a building power data processing method, system, control device and storage medium to address the deficiencies in the above-mentioned prior art, so as to solve the problems of low reliability and single monitoring dimension in the prior art.

[0006] To achieve the above objectives, the technical solutions adopted in this application are as follows:

[0007] In a first aspect, the present application provides a building power data processing method, the method comprising:

[0008] Determine actual electrical data of the building power system sent by multiple acquisition devices, the actual electrical data including: current electrical data and historical electrical data within a preset time period before the current moment, the current electrical data and the historical electrical data both including: basic electrical data, temperature data, insulation monitoring data and discharge data;

[0009] Inputting the actual electrical data into a pre-trained fault detection model to obtain a fault control strategy for the actual electrical data, wherein the fault control strategy includes: a non-intervention control strategy, a first intervention control strategy, or a second intervention control strategy, wherein the first intervention control strategy is used to instruct to increase the detection frequency and fault inspection, and the second intervention control strategy is used to instruct to cut off the power supply;

[0010] The building power system is managed according to the fault control strategy.

[0011] Optionally, inputting the actual electrical data into a pre-trained fault detection model and outputting a fault control strategy for the actual electrical data includes:

[0012] Inputting the actual electrical data into a pre-trained fault detection model;

[0013] The fault detection model preprocesses the actual electrical data to obtain processed electrical data, wherein the preprocessing includes data cleaning and data normalization;

[0014] The fault detection model performs feature extraction and ensemble learning on the processed electrical data to determine the operation score of the actual electrical data;

[0015] The fault detection model determines a fault control strategy based on the operation score.

[0016] Optionally, the fault detection model performs feature extraction and ensemble learning on the processed electrical data to determine the running score of the actual electrical data, including:

[0017] Extracting features from the processed electrical data to obtain a plurality of electrical features, wherein the electrical features include: time domain features, frequency domain features, and statistical features;

[0018] Perform feature fusion on multiple electrical features to obtain multiple fused features;

[0019] In the model layer of the fault detection model, an anomaly detection model in the fault detection model performs an anomaly detection analysis on the multiple fused features to obtain an anomaly detection analysis result, a trend detection model in the fault detection model performs a trend prediction analysis on the multiple fused features to obtain a trend prediction analysis result, and a classification model in the fault detection model performs a classification analysis on the multiple fused features to obtain a classification analysis result;

[0020] An operation score of actual electrical data is predicted based on the anomaly detection analysis result, the trend prediction analysis result, and the classification analysis result.

[0021] Optionally, determining a fault control strategy according to the operation score includes:

[0022] Determining whether the operation score of the actual electrical data is greater than a normal operation score threshold;

[0023] If yes, determining that the fault control strategy is a non-intervention control strategy;

[0024] If not, the fault control strategy is determined and an alarm signal is output according to the operation score of the actual electrical data and the fault operation score threshold.

[0025] Optionally, determining the fault control strategy and outputting an alarm signal according to the operation score of the actual electrical data and the fault operation score threshold includes:

[0026] determining whether the operation score is greater than the fault operation score threshold, and if so, determining that the fault control strategy is the first intervention control strategy, and outputting a first alarm signal, where the first alarm signal is used to indicate a minor fault state;

[0027] If not, it is determined that the fault control strategy is the second intervention control strategy, and a second alarm signal is output, where the second alarm signal is used to indicate a serious fault state.

[0028] Optionally, the method further includes:

[0029] According to a preset time period, the fault detection model is optimized and adjusted according to the acquired historical electrical data within the period to obtain an adjusted fault detection model.

[0030] Optionally, after controlling the building power data according to the fault control strategy, the method further includes:

[0031] Obtain regular assessment results corresponding to actual electrical data;

[0032] The model parameters and the operation score threshold group in the fault detection model are dynamically adjusted and optimized according to the periodic evaluation results to obtain an optimized fault detection model.

[0033] In a second aspect, the present application provides a building power data processing device, the method comprising:

[0034] A determination module, used to determine actual electrical data of a building power system sent by a plurality of acquisition devices, wherein the actual electrical data includes: current electrical data and historical electrical data within a preset period before the current moment, wherein the current electrical data and the historical electrical data both include: basic electrical data, temperature data, insulation monitoring data and discharge data;

[0035] An input module, used to input the actual electrical data into a pre-trained fault detection model to obtain a fault control strategy for the actual electrical data, wherein the fault control strategy includes: a non-intervention control strategy, a first intervention control strategy, or a second intervention control strategy, wherein the first intervention control strategy is used to instruct to increase the detection frequency and fault inspection, and the second intervention control strategy is used to instruct to cut off the power supply;

[0036] A management module is used to manage the building power system according to the fault control strategy.

[0037] Optionally, the input module is specifically used for:

[0038] Inputting the actual electrical data into a pre-trained fault detection model;

[0039] The fault detection model preprocesses the actual electrical data to obtain processed electrical data, wherein the preprocessing includes data cleaning and data normalization;

[0040] The fault detection model performs feature extraction and ensemble learning on the processed electrical data to determine the operation score of the actual electrical data;

[0041] The fault detection model determines a fault control strategy based on the operation score.

[0042] Optionally, the input module is specifically used for:

[0043] Extracting features from the processed electrical data to obtain a plurality of electrical features, wherein the electrical features include: time domain features, frequency domain features, and statistical features;

[0044] Perform feature fusion on multiple electrical features to obtain multiple fused features;

[0045] In the model layer of the fault detection model, an anomaly detection model in the fault detection model performs an anomaly detection analysis on the multiple fused features to obtain an anomaly detection analysis result, a trend detection model in the fault detection model performs a trend prediction analysis on the multiple fused features to obtain a trend prediction analysis result, and a classification model in the fault detection model performs a classification analysis on the multiple fused features to obtain a classification analysis result;

[0046] An operation score of actual electrical data is predicted based on the anomaly detection analysis result, the trend prediction analysis result, and the classification analysis result.

[0047] Optionally, the input module is specifically used for:

[0048] Determining whether the operation score of the actual electrical data is greater than a normal operation score threshold;

[0049] If yes, determining that the fault control strategy is a non-intervention control strategy;

[0050] If not, the fault control strategy is determined and an alarm signal is output according to the operation score of the actual electrical data and the fault operation score threshold.

[0051] Optionally, the input module is specifically used for:

[0052] determining whether the operation score is greater than the fault operation score threshold, and if so, determining that the fault control strategy is the first intervention control strategy, and outputting a first alarm signal, where the first alarm signal is used to indicate a minor fault state;

[0053] If not, it is determined that the fault control strategy is the second intervention control strategy, and a second alarm signal is output, where the second alarm signal is used to indicate a serious fault state.

[0054] Optionally, the management module is further configured to:

[0055] According to a preset time period, the fault detection model is optimized and adjusted according to the acquired historical electrical data within the period to obtain an adjusted fault detection model.

[0056] Optionally, the management module is further configured to:

[0057] Obtain regular assessment results corresponding to actual electrical data;

[0058] The model parameters and the operation score threshold group in the fault detection model are dynamically adjusted and optimized according to the periodic evaluation results to obtain an optimized fault detection model.

[0059] In a third aspect, the present application provides a building power data processing system, the system comprising a plurality of insulation monitoring devices, a plurality of infrared thermal imaging devices, a plurality of partial discharge monitoring devices, a basic acquisition device, a monitoring and data acquisition device, a control device and a building power system, the building power system comprising a distribution cabinet, a transformer and a plurality of distribution boxes;

[0060] The multiple insulation monitoring devices, multiple infrared thermal imaging devices, multiple partial discharge monitoring devices, and basic acquisition devices are used to detect the current electrical data of the distribution cabinets, transformers, and multiple distribution boxes in the building power system in real time, and send the actual electrical data to the monitoring and data acquisition device;

[0061] The monitoring and data acquisition device sends the current electrical data to the control device, and the control device executes the building power data processing method as described in the first aspect above.

[0062] In a fourth aspect, the present application provides a control device, comprising: a processor, a storage medium and a bus, wherein the storage medium stores machine-readable instructions executable by the processor. When the control device is running, the processor communicates with the storage medium through the bus, and the processor executes the machine-readable instructions to perform the steps of the method described in the first aspect above.

[0063] In a fifth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the method described in the first aspect are executed.

[0064] The beneficial effect of the present application is that actual electrical data including current electrical data and historical electrical data within a preset time period before the current moment is input into a fault detection model, so that the fault detection model can comprehensively analyze the electrical data and the changing trend of the electrical data based on the current data and the historical data, thereby outputting a fault control strategy that takes into account the changing trend of the electrical data, and then issuing an alarm before a fault occurs, thereby improving preventive measures. The fault detection model detects the building power system based on each electrical data, does not rely on the experience of maintenance personnel, and makes the fault analysis more accurate, thereby establishing a reliable fault warning mechanism. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0066] Figure 1 It is a structural schematic diagram of a building power data processing system provided in an embodiment of the present application;

[0067] Figure 2 It is a flowchart of a building power data processing method provided in an embodiment of the present application;

[0068] Figure 3 It is a flow chart of a fault control strategy for outputting actual electrical data provided by an embodiment of the present application;

[0069] Figure 4 is a schematic diagram of a flow chart of determining an operating score of actual electrical data provided by an embodiment of the present application;

[0070] Figure 5 It is a flowchart of a fault control strategy improved by an embodiment of the present application;

[0071] Figure 6 It is a schematic diagram of another flow chart for determining a fault control strategy improved by an embodiment of the present application;

[0072] Figure 7 It is a structural schematic diagram of a building power data processing device provided in an embodiment of the present application;

[0073] Figure 8It is a structural schematic diagram of a control device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0074] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the drawings in the present application only serve the purpose of explanation and description and are not used to limit the scope of protection of the present application. In addition, it should be understood that the schematic drawings are not drawn in real proportion. The flowchart used in this application shows the operations implemented according to some embodiments of the present application. It should be understood that the operations of the flowchart can be implemented out of sequence, and the steps without logical context can be reversed in order or implemented simultaneously. In addition, those skilled in the art can add one or more other operations to the flowchart under the guidance of the content of the present application, or remove one or more operations from the flowchart.

[0075] In addition, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0076] It should be noted that the term "comprising" will be used in the embodiments of the present application to indicate the existence of the features declared thereafter, but does not exclude the addition of other features.

[0077] In the prior art, fault detection and prevention of building power systems are mainly achieved through two technical solutions. The first is the traditional electrical manual inspection solution, in which maintenance personnel regularly use handheld test equipment to check each electrical equipment, including using a thermometer to detect the equipment temperature of the electrical equipment, and using a megohmmeter to test the insulation resistance of each electrical equipment. Although this solution is simple to operate, it has problems such as low detection frequency, discontinuous data, and inability to detect sudden faults in time. In addition, the repeatability of manual detection is poor, and the detection quality is heavily dependent on the experience level of maintenance personnel, making it difficult to establish a reliable fault warning mechanism. The second is a single automated monitoring solution, such as deploying SCADA alone for electrical parameter monitoring, or using data acquisition equipment alone to collect relevant data of electrical equipment. This method has a single monitoring dimension and is difficult to fully reflect the health status of the equipment. For example, the insulation degradation problem cannot be found by analyzing the temperature detection data, and only analyzing the real-time electrical data may miss the early signs of equipment failure, and thus fail to issue an alarm before the failure occurs, which is insufficient in prevention.

[0078] Based on this, the present application proposes a building power data processing method, which inputs the current electrical data and the historical electrical data within a preset time period before the current moment into a fault detection model, and determines the fault control strategy through the preprocessing, feature extraction, integrated learning and operation score analysis of the actual electrical data by the fault detection model, thereby improving the detection accuracy and continuity while comprehensively reflecting the health status of the equipment through multi-dimensional electrical data, and improving preventive measures through integrated learning.

[0079] Next, refer to Figure 1 The structure of building power data processing system is introduced. Figure 1 Schematic diagram of a building power data processing system provided by an embodiment of the present application. Figure 1 As shown, the building power data processing system includes multiple insulation monitoring devices, multiple infrared thermal imaging devices, multiple partial discharge monitoring devices, monitoring and data acquisition devices, control devices, basic acquisition devices and a building power system. The building power system includes a distribution cabinet, a transformer and multiple distribution boxes. Figure 1 Two insulation monitoring devices, two infrared thermal imaging devices and two partial discharge detection devices are used as examples for illustration.

[0080] In the building power system, the distribution cabinet is connected to the transformer and multiple distribution boxes. The distribution cabinet and transformer can be deployed in the distribution room, and multiple distribution boxes can be deployed in the power shaft on each floor. There can be multiple distribution cabinets and transformers, and one distribution cabinet can be connected to multiple distribution boxes. This embodiment limits the number and connection relationship of distribution cabinets, transformers, and distribution boxes in the building power system. Figure 1Take a power distribution cabinet and a transformer as an example for explanation.

[0081] The insulation monitoring device in the building power data processing system can be an insulation monitoring sensor. The insulation monitoring device is deployed on high-voltage equipment such as transformers and important lines to continuously monitor the insulation resistance and leakage current of electrical equipment. The infrared thermal imaging device can be a fixed infrared thermal imager, which is deployed around equipment such as transformers and distribution cabinets to monitor the temperature changes of the equipment according to the preset time interval of the infrared thermal imaging device. The preset time interval of the infrared thermal imaging device of the partial discharge monitoring device can be, for example, 5 minutes. The infrared thermal imaging device can be deployed on equipment such as transformers to collect discharge signals through various sensors such as sound, light and electricity. The basic acquisition device is used to collect basic electrical data such as voltage, current and power. The basic acquisition device is deployed in the distribution room.

[0082] After each insulation monitoring device, each infrared thermal imaging device, multiple partial discharge monitoring devices and basic acquisition device detect the current electrical data of the main control cabinet, distribution cabinet, transformer and multiple distribution boxes in the building power system, the current electrical data is sent to the monitoring and data acquisition device. The monitoring and data acquisition device is used to monitor and collect data according to the preset time interval of the monitoring and data acquisition device, wherein the preset time interval of the monitoring and data acquisition device can be, for example, 1 second.

[0083] As an optional implementation, the monitoring and data acquisition device sends the current electrical data to the control device via industrial Ethernet, and the control device executes the building power data processing method based on the received current electrical data, thereby realizing fault detection and prevention of the building power system.

[0084] As an optional implementation, Figure 1 As shown, the control device may include a central control console, a data storage server, and an analysis and processing server. The central control console receives the real-time electrical data sent by the monitoring and data acquisition device, and stores the real-time electrical data in the data storage server. The data storage server may include a relational database and a time series database, and the real-time electrical data may be stored in each database according to the type of the real-time electrical data. The data storage server may regularly perform quality checks on the stored data, remove abnormal values, and regularly back up and archive the data to ensure the integrity and availability of the data.

[0085] When processing building power data, actual electrical data is extracted from the data storage server and sent to the analysis and processing server, so that the analysis and processing server generates a fault control strategy based on the actual electrical data, and sends the fault control strategy to the central control console. The central control console sends control instructions to the monitoring and data acquisition device based on the fault control strategy, so that the monitoring and data acquisition device controls the equipment in the building power system to perform specific actions in the fault control strategy based on the control instructions.

[0086] Next, refer to Figure 2 The specific implementation steps of the building power data processing method are introduced. Figure 2 It is a flow chart of a building power data processing method provided in an embodiment of the present application.

[0087] S201. Determine actual electrical data of a building power system sent by multiple acquisition devices. The actual electrical data includes: current electrical data and historical electrical data within a preset time period before the current moment. Both the current electrical data and the historical electrical data include: basic electrical data, temperature data, insulation monitoring data and discharge data.

[0088] The acquisition equipment may include: multiple insulation monitoring devices, multiple infrared thermal imaging devices, multiple partial discharge monitoring devices, and basic acquisition devices. Basic electrical data is acquired by the basic acquisition device, temperature data is acquired by multiple infrared thermal imaging devices, insulation monitoring data is acquired by multiple insulation monitoring devices, and discharge data is acquired by multiple partial discharge monitoring devices.

[0089] Optionally, in the actual electrical data, the current electrical data is the electrical data acquired at the current moment, and the preset period before the current moment may be, for example, the electrical data within 1 hour before the current moment. For example, if the current moment is 10:00 a.m. on January 1, the actual electrical data may be the latest electrical data acquired at 10:00 a.m. on January 1 and all electrical data acquired between 9:00 a.m. and 10:00 a.m. on January 1.

[0090] S202. Input the actual electrical data into the pre-trained fault detection model to obtain a fault control strategy for the actual electrical data. The fault control strategy includes: a non-intervention control strategy, a first intervention control strategy or a second intervention control strategy. The first intervention control strategy is used to instruct to increase the detection frequency and fault inspection, and the second intervention control strategy is used to instruct to cut off the power supply.

[0091] Optionally, after receiving multi-source heterogeneous actual electrical data, the fault detection model can perform data preprocessing, feature extraction, ensemble learning, and operation score analysis to determine the fault control strategy. Specifically, the fault detection model can use a sliding window method to comprehensively consider short-term and long-term change trends, thereby calculating the operation score of various factors, and determining the fault control strategy based on the operation score.

[0092] Optionally, in the fault control strategy, a non-intervention control strategy is applied to deal with the non-fault state of the building electrical system, a first intervention control strategy is applied to deal with the minor fault state of the building electrical system, and a second intervention control strategy is applied to deal with the serious fault state of the building electrical system. Among them, the first intervention control strategy can indicate the fault type and fault location, and indicate to increase the detection frequency and fault inspection. Increasing the detection frequency means reducing the preset time intervals corresponding to multiple insulation monitoring devices, multiple infrared thermal imaging devices, multiple partial discharge monitoring devices and basic acquisition devices. Fault inspection is to perform specific fault inspections on the fault location. The second intervention control strategy can indicate the fault type and fault location, and indicate to cut off the power supply. As another optional implementation, the second intervention control strategy can also indicate load transfer. Cutting off the power supply means cutting off the power supply at the fault location, and transferring the load means transferring the load of the line at the fault location to other lines.

[0093] It is worth mentioning that when the output fault control strategy is the first intervention control strategy or the second intervention control strategy, the fault-related characteristic values ​​and timestamps can also be recorded and output, thereby providing the staff with a more detailed response plan.

[0094] S203. Manage the building power system according to the fault control strategy.

[0095] Optionally, after the fault detection module outputs the fault control strategy, it will generate control instructions according to the fault control strategy and send the control instructions to the monitoring and data acquisition device, so that the monitoring and data acquisition device controls the equipment in the building power system based on the control instructions to perform specific actions in the fault control strategy.

[0096] As an optional implementation, the control device can generate multiple types of operation reports, including daily monitoring reports, abnormal event reports, and periodic analysis reports. Among them, the daily monitoring report is used to record the monitoring status of the building power system within a preset time, including normal conditions and abnormal conditions. The abnormal event report is used to record all abnormal events and the corresponding fault control strategies. The periodic analysis report can record the fault cause analysis and maintenance suggestions, etc.

[0097] As an optional implementation, the control device periodically sends each report to the corresponding processor according to a preset distribution principle, and creates a report index in the database to facilitate retrieval and analysis.

[0098] In this embodiment, actual electrical data including current electrical data and historical electrical data within a preset time period before the current moment is input into the fault detection model, so that the fault detection model can comprehensively analyze the electrical data and the changing trend of the electrical data based on the current data and the historical data, thereby outputting a fault control strategy that takes the changing trend of the electrical data into consideration, and then issuing an alarm before a fault occurs, thereby improving preventive measures. The fault detection model detects the building power system based on each electrical data, does not rely on the experience of maintenance personnel, and makes the fault analysis more accurate, thereby establishing a reliable fault warning mechanism.

[0099] Next, refer to Figure 3 The specific steps of inputting the actual electrical data into the pre-trained fault detection model in the above step S202 and outputting the fault control strategy of the actual electrical data are introduced. Figure 3 It is a flow chart of a fault control strategy for outputting actual electrical data provided in an embodiment of the present application.

[0100] Optionally, the fault detection model may include an input layer, a feature layer, a model layer, and an output layer.

[0101] S301 , inputting actual electrical data into a pre-trained fault detection model.

[0102] Specifically, actual electrical data is input into the input layer of the fault detection model.

[0103] S302, the fault detection model pre-processes the actual electrical data to obtain processed electrical data, and the pre-processing includes data cleaning and data normalization.

[0104] Specifically, the input layer of the fault detection model can preprocess the actual electrical data, including data cleaning and data normalization of the actual electrical data, to obtain processed electrical data.

[0105] Among them, data cleaning is used to remove abnormal electrical data. As an optional implementation, multiple abnormal electrical data thresholds can be pre-set, and abnormal electrical data can be removed according to the abnormal electrical data thresholds. As another optional implementation, the actual electrical data can be sorted by time, and it can be determined whether there are mutation values ​​in the electrical data arranged by time. If so, the mutation values ​​are removed, thereby removing noise and abnormal values ​​in the actual electrical data.

[0106] S303, the fault detection model performs feature extraction and ensemble learning on the processed electrical data to determine the operating score of the actual electrical data.

[0107] Optionally, the input layer of the fault detection model inputs the processed electrical data into the feature layer for feature extraction, and inputs the extracted features into the model layer for ensemble learning, and inputs the results of the ensemble learning into the output layer of the fault detection model, thereby outputting the running score of the actual electrical data.

[0108] S304: The fault detection model determines a fault control strategy according to the running score.

[0109] Specifically, the output layer of the fault detection model can determine the fault control strategy according to the operation score and the operation score threshold. The operation score threshold includes a normal operation score threshold and an abnormal operation score threshold. The normal operation score threshold is used to distinguish between the normal state and the abnormal state of the building power system, and the abnormal operation score threshold is used to distinguish between the minor fault state and the serious fault state of the building power system.

[0110] In this embodiment, the fault control strategy is determined by preprocessing, feature extraction, integrated learning and running score analysis of actual electrical data, thereby determining a comprehensive and accurate fault analysis result, thereby establishing a reliable fault warning mechanism.

[0111] Further, refer to Figure 4 The specific steps of extracting features and performing integrated learning on the processed electrical data by the fault detection model in step S303 to determine the running score of the actual electrical data are introduced. Figure 4 It is a flowchart of determining the operating score of actual electrical data provided by an embodiment of the present application.

[0112] S401 , extracting features from the processed electrical data to obtain a plurality of electrical features, where the electrical features include: time domain features, frequency domain features, and statistical features.

[0113] Optionally, time domain features may be extracted through time domain analysis, frequency domain features may be extracted through frequency domain analysis, and statistical features may be extracted through statistical analysis.

[0114] S402: Fusing multiple electrical features to obtain multiple fused features.

[0115] Optionally, feature fusion is performed on the time domain features, frequency domain features and statistical features to obtain multiple fused features. Feature fusion is used to merge features from different sources into a unified feature representation, thereby enhancing the performance and generalization ability of the fault detection model through multiple features and improving the comprehensiveness of the fault control strategy output by the fault detection model.

[0116] S403. At the model layer of the fault detection model, the anomaly detection model in the fault detection model performs anomaly detection analysis on the multiple fused features to obtain anomaly detection analysis results, the trend detection model in the fault detection model performs trend prediction analysis on the multiple fused features to obtain trend prediction analysis results, and the classification model in the fault detection model performs classification analysis on the multiple fused features to obtain classification analysis results.

[0117] Optionally, the fault detection model may integrate an anomaly detection model, a trend detection model, and a classification model.

[0118] Among them, the anomaly detection model is used to perform anomaly detection analysis based on the fused features, so as to obtain anomaly detection analysis results, which can be used to characterize whether the data represented by the features is abnormal. In the training process of the fault detection model, a machine learning algorithm can be used to train the anomaly detection model. The machine learning algorithm can be a vector machine and a random forest. The model performance of the anomaly detection model can be evaluated through cross-validation, and the model parameters can be optimized according to the verification results to ensure that the model has good generalization ability.

[0119] The trend detection model is used to analyze data change trends based on historical electrical data and current electrical data, and determine whether a failure is about to occur based on the data change trend. For example, the aging of the line is predicted based on data involving voltage and power consumption.

[0120] The classification model is used to classify features into normal data features, abnormal data features, and fault data features, thereby achieving classification.

[0121] S404: Predict and obtain an operating score of actual electrical data according to the anomaly detection analysis result, the trend prediction analysis result, and the classification analysis result.

[0122] Specifically, the anomaly detection analysis results, trend prediction analysis results and classification analysis results are respectively for different types of fault features. Based on the anomaly detection analysis results, trend prediction analysis results and classification analysis results, integrated learning is performed, and the integrated information method can be bagging, boosting and stacking.

[0123] Among them, the operation score is generated by integrating the results of anomaly detection analysis, trend prediction analysis and classification analysis, and can comprehensively reflect the status of the building power system.

[0124] In this embodiment, feature extraction and feature fusion are performed on the processed electrical data, and then the anomaly detection model, trend prediction model and classification model in the fault detection model are used to obtain anomaly detection analysis results, trend prediction analysis results and classification analysis results respectively, so that multiple sub-models are used to analyze different types of fault characteristics respectively, thereby improving the generalization ability of the fault detection model.

[0125] Next, refer to Figure 5 The step of determining the fault control strategy according to the running score in the above step S304 is introduced. Figure 5 It is a flowchart of a fault control strategy determined in an improved embodiment of the present application.

[0126] S501 , determining whether the operation score of the actual electrical data is greater than a normal operation score threshold.

[0127] Optionally, the higher the operation score of the actual electrical data, the better the operation status of the building power system, and the lower the operation score, the higher the fault degree of the building power system.

[0128] S502: If yes, determine that the fault control strategy is a non-intervention control strategy.

[0129] As an optional implementation, when the fault control strategy is a non-intervention control strategy, no intervention is performed on the building power system.

[0130] S503: If not, determine a fault control strategy and output an alarm signal according to the operation score of the actual electrical data and the fault operation score threshold.

[0131] Optionally, the faulty operation score threshold is less than the normal operation score threshold.

[0132] Optionally, if the operation score of the actual electrical data is less than the normal operation score threshold, the fault control strategy may be the first intervention control strategy or the second intervention control strategy.

[0133] Optionally, the alarm signal is used to notify the staff via SMS and email, so that the staff can respond to the fault condition according to the alarm signal. The alarm signal may include the fault location and the fault type.

[0134] It is worth mentioning that in the process of determining the fault control strategy, all strategies, control instructions and execution actions are recorded in real time to generate a record table. The record table includes the fault location, fault time, fault type, fault cause, fault control strategy and execution action. Among them, the execution action includes the action performed on each device in the building power system.

[0135] In this embodiment, the fault control strategy is determined according to whether the operation score of the actual electrical data is greater than the normal operation score threshold, thereby realizing a hierarchical early warning mechanism.

[0136] Furthermore, the process of how to determine the fault control strategy and output the alarm signal in the above step S503 is introduced.

[0137] Optionally, it is determined whether the operation score is greater than a fault operation score threshold. If so, the fault control strategy is determined to be a first intervention control strategy, and a first alarm signal is output, where the first alarm signal is used to indicate a minor fault state.

[0138] Optionally, on the basis that the operation score is less than the normal operation score threshold, if the operation score is greater than the fault operation score threshold, the fault control strategy is determined to be the first intervention control strategy, and a first alarm signal is output. The first intervention control strategy is used to indicate an increase in detection frequency and fault inspection. The first alarm signal can be output by SMS, phone call, and email.

[0139] Optionally, if not, the fault control strategy is determined to be a second intervention control strategy, and a second alarm signal is output, where the second alarm signal is used to indicate a serious fault state.

[0140] Optionally, if the operation score is less than the fault operation score threshold, it means that the current building power system fault is serious. The output second intervention control strategy includes cutting off the power supply, and the second alarm signal can be output by SMS, phone, email, and issuing an on-site buzzer alarm, so as to promptly remind the staff to deal with the fault problem.

[0141] As another optional implementation, the first intervention control strategy may be recording and continuous monitoring. Specifically, continuous monitoring of the fault location. The second intervention control strategy may include switching to a backup system and power failure protection.

[0142] In this embodiment, a hierarchical early warning mechanism is implemented by judging whether the operation score is greater than the fault operation score threshold.

[0143] As an optional implementation, refer to Figure 6Confirm the overall steps for determining the fault control strategy. Collect the current power data in the building power system, and then input the real-time power data and the historical electrical data in the preset period before the current moment into the input layer of the fault detection model, perform data cleaning and data normalization, and obtain processed electrical data. Input the processed electrical data into the feature layer, extract the features of the processed electrical data, obtain time domain features, frequency domain features and statistical features, and then fuse the features to obtain multiple fused features. Input the multiple fused features into the anomaly detection model, trend prediction model and classification model in the model layer of the fault detection model for analysis, and obtain the anomaly detection analysis results, trend prediction analysis results and classification analysis results respectively. The anomaly detection analysis results, trend prediction analysis results and classification analysis results are integrated and learned to obtain the operation score. Determine whether the operation score is greater than the normal operation score threshold. If so, determine that the fault control strategy is a non-intervention control strategy. If not, determine whether the operation score is greater than the abnormal operation score threshold. If so, determine that the fault control strategy is a first intervention control strategy. If not, determine that the fault control strategy is a second intervention control strategy.

[0144] As an optional implementation, the building power data processing method also includes a training process of a fault detection model.

[0145] Specifically, when initially training the fault model, the feature space of the normal operation state in the initial fault detection model is first constructed according to the training samples, so as to perform iterative training and generate the fault detection model. The training samples include a plurality of electrical data.

[0146] Optionally, during actual use of the fault detection model, the fault detection model is optimized and adjusted according to the historical electrical data acquired within the period according to a preset time period to obtain an adjusted fault detection model.

[0147] Specifically, after the initial fault detection model is trained to obtain the fault detection model, the fault detection model is optimized and adjusted according to a preset time period, so that the fault detection model outputs a more accurate fault control strategy.

[0148] For example, after obtaining the fault detection model by training the initial fault detection module on January 1, the fault detection model is optimized and adjusted once a month to obtain an adjusted fault detection model. The historical electrical data within the period may be electrical data within one month. During the use of the model, the latest adjusted fault detection model is used.

[0149] Optionally, the fault detection model is optimized and adjusted according to the acquired electrical data at a preset time period to obtain an adjusted fault detection model, thereby improving the accuracy of the model output.

[0150] As an optional implementation, after the above step S203, i.e., controlling the building power data according to the fault control strategy, the following steps may also be included:

[0151] Optionally, periodic evaluation results corresponding to actual electrical data are obtained.

[0152] The regular evaluation results may be the results obtained by the staff regularly evaluating the fault conditions, the corresponding fault control strategies and the control effects.

[0153] Optionally, model parameters and an operation score threshold group in the fault detection model are dynamically adjusted and optimized according to the periodic evaluation results to obtain an optimized fault detection model.

[0154] As an optional implementation, the normal operation score threshold and the abnormal operation score threshold, as well as the specific contents of the first intervention control strategy and the second intervention control strategy may also be updated according to the usage of each device in the building power system.

[0155] As another optional implementation, model performance indicators can be automatically collected to evaluate the operating efficiency of the anomaly detection model, trend detection model and classification model in the fault detection model, and optimization suggestions can be generated to optimize the fault detection model, thereby ensuring the continuous and effective operation of the system.

[0156] In this embodiment, the fault detection model is dynamically adjusted and optimized according to the periodic evaluation results, thereby improving the effectiveness and accuracy of the model output.

[0157] Based on the same inventive concept, a building power data processing device corresponding to the building power data processing method is also provided in the embodiment of the present application. Since the principle of solving the problem by the device in the embodiment of the present application is similar to the above-mentioned building power data processing method in the embodiment of the present application, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be repeated.

[0158] Reference Figure 7 , which is a schematic diagram of the structure of a building power data processing device provided in an embodiment of the present application, the device includes: a determination module 701, an input module 702 and a management module 703; wherein:

[0159] The determination module 701 is used to determine the actual electrical data of the building power system sent by multiple acquisition devices, wherein the actual electrical data includes: current electrical data and historical electrical data within a preset period before the current moment, wherein the current electrical data and the historical electrical data both include: basic electrical data, temperature data, insulation monitoring data and discharge data;

[0160] An input module 702 is used to input the actual electrical data into a pre-trained fault detection model to obtain a fault control strategy for the actual electrical data, wherein the fault control strategy includes: a non-intervention control strategy, a first intervention control strategy, or a second intervention control strategy, wherein the first intervention control strategy is used to instruct to increase the detection frequency and fault inspection, and the second intervention control strategy is used to instruct to cut off the power supply;

[0161] The management module 703 is used to manage the building power system according to the fault control strategy.

[0162] Optionally, the input module 702 is specifically used for:

[0163] Inputting the actual electrical data into a pre-trained fault detection model;

[0164] The fault detection model preprocesses the actual electrical data to obtain processed electrical data, wherein the preprocessing includes data cleaning and data normalization;

[0165] The fault detection model performs feature extraction and ensemble learning on the processed electrical data to determine the operation score of the actual electrical data;

[0166] The fault detection model determines a fault control strategy based on the operation score.

[0167] Optionally, the input module 702 is specifically used for:

[0168] Extracting features from the processed electrical data to obtain a plurality of electrical features, wherein the electrical features include: time domain features, frequency domain features, and statistical features;

[0169] Perform feature fusion on multiple electrical features to obtain multiple fused features;

[0170] In the model layer of the fault detection model, an anomaly detection model in the fault detection model performs an anomaly detection analysis on the multiple fused features to obtain an anomaly detection analysis result, a trend detection model in the fault detection model performs a trend prediction analysis on the multiple fused features to obtain a trend prediction analysis result, and a classification model in the fault detection model performs a classification analysis on the multiple fused features to obtain a classification analysis result;

[0171] An operation score of actual electrical data is predicted based on the anomaly detection analysis result, the trend prediction analysis result, and the classification analysis result.

[0172] Optionally, the input module 702 is specifically used for:

[0173] Determining whether the operation score of the actual electrical data is greater than a normal operation score threshold;

[0174] If yes, determining that the fault control strategy is a non-intervention control strategy;

[0175] If not, the fault control strategy is determined and an alarm signal is output according to the operation score of the actual electrical data and the fault operation score threshold.

[0176] Optionally, the input module 702 is specifically used for:

[0177] determining whether the operation score is greater than the fault operation score threshold, and if so, determining that the fault control strategy is the first intervention control strategy, and outputting a first alarm signal, where the first alarm signal is used to indicate a minor fault state;

[0178] If not, it is determined that the fault control strategy is the second intervention control strategy, and a second alarm signal is output, where the second alarm signal is used to indicate a serious fault state.

[0179] Optionally, the management module 703 is further used to:

[0180] According to a preset time period, the fault detection model is optimized and adjusted according to the acquired historical electrical data within the period to obtain an adjusted fault detection model.

[0181] Optionally, the management module 703 is further used to:

[0182] Obtain regular assessment results corresponding to actual electrical data;

[0183] The model parameters and the operation score threshold group in the fault detection model are dynamically adjusted and optimized according to the periodic evaluation results to obtain an optimized fault detection model.

[0184] For descriptions of the processing flow of each module in the device and the interaction flow between each module, reference may be made to the relevant descriptions in the above method embodiment, which will not be described in detail here.

[0185] The present application also provides a control device, such as Figure 8 801, a control device according to an embodiment of the present application, including: a processor 801, a memory 802 and a bus. The memory 802 stores machine-readable instructions executable by the processor 801 (for example, Figure 7 In the device, the execution instructions corresponding to the determination module 701, the input module 702 and the management module 703 are determined, etc. When the computer device is running, the processor 801 communicates with the memory 802 through a bus, and when the machine-readable instructions are executed by the processor 801, the processing of the above-mentioned building power data processing method is performed.

[0186] An embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned building power data processing method are executed.

[0187] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, the specific working process of the system and device described above can refer to the corresponding process in the method embodiment, and will not be repeated in this application. In the several embodiments provided in this application, it should be understood that the disclosed system, device and method can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the modules is only a logical function division. There may be other division methods in actual implementation. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0188] In addition, each functional unit in each embodiment of the present application can be integrated into a processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or part of the technical solution that contributes to the prior art or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, RandomAccess Memory), disk or optical disk and other media that can store program code.

[0189] The above are only specific implementation methods of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be covered by the protection scope of the present application.

Claims

1. A building power data processing method, characterized in that: The method comprises: Determine actual electrical data of the building power system sent by multiple acquisition devices, the actual electrical data including: current electrical data and historical electrical data within a preset time period before the current moment, the current electrical data and the historical electrical data both including: basic electrical data, temperature data, insulation monitoring data and discharge data; Inputting the actual electrical data into a pre-trained fault detection model to obtain a fault control strategy for the actual electrical data, wherein the fault control strategy includes: a non-intervention control strategy, a first intervention control strategy, or a second intervention control strategy, wherein the first intervention control strategy is used to instruct to increase the detection frequency and fault inspection, and the second intervention control strategy is used to instruct to cut off the power supply; The building power system is managed according to the fault control strategy.

2. The building power data processing method according to claim 1, characterized in that: The step of inputting the actual electrical data into a pre-trained fault detection model and outputting a fault control strategy for the actual electrical data comprises: Inputting the actual electrical data into a pre-trained fault detection model; The fault detection model preprocesses the actual electrical data to obtain processed electrical data, wherein the preprocessing includes data cleaning and data normalization; The fault detection model performs feature extraction and ensemble learning on the processed electrical data to determine the operation score of the actual electrical data; The fault detection model determines a fault control strategy based on the operation score.

3. The building power data processing method according to claim 2, characterized in that: The fault detection model performs feature extraction and ensemble learning on the processed electrical data to determine the operating score of the actual electrical data, including: Extracting features from the processed electrical data to obtain a plurality of electrical features, wherein the electrical features include: time domain features, frequency domain features, and statistical features; Perform feature fusion on multiple electrical features to obtain multiple fused features; In the model layer of the fault detection model, an anomaly detection model in the fault detection model performs an anomaly detection analysis on the multiple fused features to obtain an anomaly detection analysis result, a trend detection model in the fault detection model performs a trend prediction analysis on the multiple fused features to obtain a trend prediction analysis result, and a classification model in the fault detection model performs a classification analysis on the multiple fused features to obtain a classification analysis result; An operation score of actual electrical data is predicted based on the anomaly detection analysis result, the trend prediction analysis result, and the classification analysis result.

4. The building power data processing method according to claim 2, characterized in that: Determining a fault control strategy according to the operation score includes: Determining whether the operation score of the actual electrical data is greater than a normal operation score threshold; If yes, determining that the fault control strategy is a non-intervention control strategy; If not, the fault control strategy is determined and an alarm signal is output according to the operation score of the actual electrical data and the fault operation score threshold.

5. The building power data processing method according to claim 4, characterized in that: The determining the fault control strategy and outputting an alarm signal according to the operation score of the actual electrical data and the fault operation score threshold comprises: determining whether the operation score is greater than the fault operation score threshold, and if so, determining that the fault control strategy is the first intervention control strategy, and outputting a first alarm signal, where the first alarm signal is used to indicate a minor fault state; If not, it is determined that the fault control strategy is the second intervention control strategy, and a second alarm signal is output, where the second alarm signal is used to indicate a serious fault state.

6. The building power data processing method according to claim 1, characterized in that: The method further comprises: According to a preset time period, the fault detection model is optimized and adjusted according to the acquired historical electrical data within the period to obtain an adjusted fault detection model.

7. The building power data processing method according to claim 1, characterized in that: After controlling the building power data according to the fault control strategy, the method further includes: Obtain regular assessment results corresponding to actual electrical data; The model parameters and the operation score threshold group in the fault detection model are dynamically adjusted and optimized according to the periodic evaluation results to obtain an optimized fault detection model.

8. A building power data processing system, characterized in that: The system includes multiple insulation monitoring devices, multiple infrared thermal imaging devices, multiple partial discharge monitoring devices, basic acquisition devices, monitoring and data acquisition devices, control devices and building power systems, and the building power systems include power distribution cabinets, transformers and multiple distribution boxes; The multiple insulation monitoring devices, multiple infrared thermal imaging devices, multiple partial discharge monitoring devices, and basic acquisition devices are used to detect the current electrical data of the distribution cabinets, transformers, and multiple distribution boxes in the building power system in real time, and send the actual electrical data to the monitoring and data acquisition device; The monitoring and data acquisition device sends the current electrical data to the control device, and the control device executes the building power data processing method as described in any one of claims 1 to 7 above.

9. A control device, characterized in that: include: A processor and a memory, wherein the memory stores machine-readable instructions executable by the processor, and when the control device is running, the processor executes the machine-readable instructions to perform the steps of the building power data processing method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the building power data processing method as claimed in any one of claims 1 to 7 are executed.

Images