Partial discharge data processing method and system based on big data model

Abstract

The invention discloses a partial discharge data processing scheme and system based on a big data model, and relates to the field of partial discharge data processing, and the method comprises the steps: obtaining a partial discharge signal of power equipment, carrying out the time-frequency analysis of the partial discharge signal, obtaining the current discharge characteristic data, inputting the current discharge characteristic data into a discharge classification model, and obtaining a discharge classification model; obtaining different types of current discharge events, forming a current discharge time sequence characteristic based on a change rule of discharge frequency, intensity and time interval, and carrying out dynamic time warping calculation on the current discharge time sequence characteristic and a standard discharge mode to obtain a current discharge development stage; determining a fault risk level according to the current discharge development stage and the current discharge intensity, and determining a maintenance time window according to the fault risk level; and calculating the maintenance cost in the maintenance time window, and sending maintenance information containing the maintenance cost to the target client. By implementing the method, the accuracy of determining the maintenance opportunity of the power equipment can be improved.

Description

Technical Field

[0001] The present application relates to the field of partial discharge data processing, and in particular to a partial discharge data processing solution and system based on a big data model. Background Art

[0002] With the continuous expansion of the scale of power systems and the continuous increase of power equipment, the operation status monitoring and fault diagnosis of power equipment are becoming more and more important. Partial discharge is an important indicator for evaluating the insulation status of power equipment. Its detection and analysis are of great significance to ensure the safe and stable operation of power systems.

[0003] In related technologies, partial discharge detection of power equipment mainly adopts the method of regular detection, collects partial discharge signals through portable detection equipment, and performs manual analysis and judgment based on characteristic parameters such as amplitude and frequency of the discharge signal. Based on experience, the detection personnel analyze the collected discharge data, determine the discharge type and equipment status, and then formulate corresponding maintenance plans based on the equipment status.

[0004] However, due to the dynamic changes in the equipment operating environment and load conditions, the partial discharge characteristics present a complex time-series change pattern, and the existing fixed-cycle maintenance method is difficult to accurately grasp the optimal maintenance time. Summary of the invention

[0005] The present application provides a partial discharge data processing solution and system based on a big data model, which is used to improve the accuracy of determining the maintenance timing of power equipment.

[0006] In the first aspect, the present application provides a local discharge data processing method based on a big data model, which is applied to a local discharge data processing system, the method comprising: obtaining a local discharge signal of an electrical device, and performing time-frequency analysis on the local discharge signal to obtain current discharge characteristic data, wherein the current discharge characteristic data includes amplitude data, frequency data and energy data; inputting the current discharge characteristic data into a discharge classification model to obtain different types of current discharge events; performing a time series analysis on the current discharge event to obtain the changing law of the discharge frequency, intensity and time interval, and forming the current discharge time series characteristics based on the changing law of the discharge frequency, the intensity and the time interval; performing dynamic time regularization calculation on the current discharge time series characteristics and the standard discharge mode to obtain the current discharge development stage; determining the fault risk level according to the current discharge development stage and the current discharge intensity, and determining the maintenance time window according to the fault risk level; calculating the maintenance cost within the maintenance time window, and sending the maintenance information containing the maintenance cost to the target client.

[0007] By adopting the above technical solution, the partial discharge signal of power equipment is obtained and time-frequency analysis is performed to obtain discharge characteristic data including amplitude, frequency and energy data, providing a comprehensive data basis for subsequent analysis. It is input into the discharge classification model to identify the type of discharge event, and then the discharge timing characteristics are formed through time series analysis. Then, the discharge development stage is determined by comparing with the standard discharge mode, and the fault risk level and maintenance time window are determined by comprehensively considering the discharge development stage and intensity. Finally, the maintenance cost is calculated and information is sent, realizing accurate monitoring and evaluation of the partial discharge status of power equipment, which can effectively grasp the equipment status and improve the accuracy of determining the maintenance timing of power equipment.

[0008] In combination with some embodiments of the first aspect, in some embodiments, the step of performing a timing analysis on the current discharge event to obtain the changing law of the discharge frequency, intensity and time interval, and forming the current discharge timing characteristics based on the changing law of the discharge frequency, the intensity and the time interval, specifically includes: sorting the current discharge events in chronological order, and obtaining the time interval sequence between adjacent current discharge events; calculating the number of current discharge events in each preset time window to obtain the discharge frequency sequence, and obtaining the amplitude corresponding to each current discharge event as the discharge intensity sequence; decomposing the time interval sequence, the discharge frequency sequence and the discharge intensity sequence to obtain the trend component and the fluctuation component of the current discharge event; calculating the discharge development trend characteristics based on the rate of change of the trend component, and calculating the discharge fluctuation characteristics based on the variance of the fluctuation component; combining the discharge development trend characteristics and the discharge fluctuation characteristics to form the current discharge timing characteristics.

[0009] By adopting the above technical solution, when performing time series analysis on the current discharge event, the time interval sequence is first sorted, the number of discharge events in the preset time window is calculated to obtain the discharge frequency sequence, and the discharge amplitude is obtained to obtain the discharge intensity sequence, and then the three sequences are decomposed to obtain the trend and fluctuation components. The discharge development trend characteristics and the discharge fluctuation characteristics are calculated based on the trend component change rate and the fluctuation component variance, respectively, and finally combined to form the current discharge time series characteristics. The characteristics are interrelated, the time interval sequence reflects the discharge time law, the frequency and intensity sequences reflect the degree of discharge activity, and the trend and fluctuation characteristics characterize the discharge state from different angles, which can accurately capture the time series characteristics of the discharge activity and provide a key basis for accurately judging the discharge development stage.

[0010] In combination with some embodiments of the first aspect, in some embodiments, the step of performing dynamic time warping calculation on the current discharge timing characteristics and the standard discharge pattern to obtain the current discharge development stage specifically includes: selecting a standard discharge pattern matching the current discharge type from a standard discharge pattern library, the standard discharge pattern including the typical timing characteristics of the discharge development stage, the discharge development stage including the initial stage, the development stage and the acceleration stage; calculating the first feature similarity, the second similarity and the third similarity of the current discharge timing characteristics and the typical timing characteristics of all discharge development stages; if the first feature similarity is greater than a preset feature threshold, the current discharge development stage is the initial stage, if the second feature similarity is greater than the preset feature threshold, the current discharge development stage is the development stage, and if the third feature similarity is greater than the preset feature threshold, the current discharge development stage is the acceleration stage.

[0011] By adopting the above technical scheme, a pattern matching the current discharge type is selected from the standard discharge pattern library, which contains typical timing characteristics of different development stages, and the similarity between the current discharge timing characteristics and the typical characteristics of each stage is calculated. By setting a preset feature threshold, the current discharge development stage is determined based on the comparison of the similarity of the first, second and third characteristics with the threshold. The typical characteristics of the standard discharge pattern are fully utilized as a reference. Through precise calculation and reasonable threshold judgment, the development stage of the current discharge can be accurately identified, such as the initial stage, development stage or acceleration stage, so as to provide accurate basic information for subsequent fault risk level determination and maintenance decision-making.

[0012] In combination with some embodiments of the first aspect, in some embodiments, the step of determining the fault risk level according to the current discharge development stage and the current discharge intensity, and determining the maintenance time window according to the fault risk level, specifically includes: inputting the current discharge development stage and the current discharge intensity into a preset risk level determination table to obtain the fault risk level; determining the maintenance time window according to the fault risk level and a preset maintenance time threshold, wherein the maintenance time threshold includes a first maintenance time threshold and a second maintenance time threshold; when the fault risk level is greater than the first risk threshold, determining that the maintenance time window is less than the first maintenance time threshold; when the fault risk level is between the first risk threshold and the second risk threshold, determining that the maintenance time window is between the first maintenance time threshold and the second maintenance time threshold.

[0013] By adopting the above technical solution, the fault risk level is determined according to the current discharge development stage and discharge intensity with the help of a preset risk level determination table. The determination table is derived from the statistics of different discharge conditions and failure rates in historical maintenance records. The maintenance time window is then determined based on the relationship between the risk level and the maintenance time threshold. The historical data is used to make risk judgments based on evidence, and reasonable threshold settings make the maintenance timing selection accurate, which improves the rationality of maintenance arrangements.

[0014] In combination with some embodiments of the first aspect, in some embodiments, the calculation of the maintenance cost within the maintenance time window specifically includes: obtaining the historical maintenance records of the power equipment, and establishing a maintenance cost prediction model based on the historical maintenance records, wherein the historical maintenance records include historical maintenance costs and historical fault types; inputting the current discharge development stage into the maintenance cost prediction model to obtain predicted maintenance costs; calculating a time cost coefficient based on the maintenance time window, and determining the maintenance cost within the maintenance time window based on the predicted maintenance cost and the time cost coefficient.

[0015] By adopting the above technical solution, the model is established using the historical maintenance records of power equipment, and the maintenance cost is predicted with the current discharge development stage as the key input. The final cost is determined and notified to the client in combination with the maintenance time window calculation coefficient. The historical records provide the data foundation for the model, the discharge development stage dominates the prediction trend, and the time factor adjusts the final cost. This method can achieve accurate cost estimation, make maintenance decisions more economical, and help to plan resources rationally.

[0016] In combination with some embodiments of the first aspect, in some embodiments, after the step of calculating the maintenance cost within the maintenance time window and sending the maintenance information including the maintenance cost to the target client, the method also includes: comparing the maintenance cost with the equipment replacement cost; when the maintenance cost exceeds a preset proportion of the equipment replacement cost, generating an equipment replacement suggestion and sending it to the target client.

[0017] By adopting the above technical solution, after completing the cost calculation and information sending, the maintenance and replacement costs are compared, and a decision is made based on the preset ratio whether to generate a replacement suggestion and send it to the client. Accurate maintenance cost calculation and reasonable ratio setting are the core. This comparison mechanism can detect high-cost maintenance in a timely manner, adding a decision-making perspective to equipment management.

[0018] In combination with some embodiments of the first aspect, in some embodiments, before the step of inputting the current discharge development stage and the current discharge intensity into a preset risk level determination table to obtain the fault risk level, the method also includes: based on historical maintenance records, statistically analyzing equipment failure rates under different combinations of discharge development stages and discharge intensities; and establishing a risk level determination table based on the equipment failure rate.

[0019] By adopting the above technical solution, historical maintenance records are used to count the equipment failure rates under different discharge stages and intensity combinations. The failure rate data is the key to constructing a risk level determination table. By establishing an accurate determination table, a reliable standard is provided for subsequent accurate judgment of the fault risk level.

[0020] In a second aspect, an embodiment of the present application provides a partial discharge data processing system, which includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the partial discharge data processing system to perform the method described in the first aspect and any possible implementation method of the first aspect.

[0021] In a third aspect, an embodiment of the present application provides a computer program product comprising instructions. When the above-mentioned computer program product runs on a partial discharge data processing system, the above-mentioned partial discharge data processing system executes the method described in the first aspect and any possible implementation method of the first aspect.

[0022] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, comprising instructions. When the instructions are executed on a partial discharge data processing system, the partial discharge data processing system executes the method described in the first aspect and any possible implementation method of the first aspect.

[0023] It can be understood that the partial discharge data processing system provided in the second aspect, the computer program product provided in the third aspect, and the computer storage medium provided in the fourth aspect are all used to execute the method provided in the embodiment of the present application. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding method, which will not be repeated here.

[0024] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. This application obtains the partial discharge signal of power equipment and performs time-frequency analysis to obtain discharge characteristic data including amplitude, frequency and energy data, providing a comprehensive data basis for subsequent analysis. It is input into the discharge classification model to identify the type of discharge event, and then the discharge timing characteristics are formed through time series analysis. Then, the discharge development stage is determined by comparing with the standard discharge mode. The fault risk level and maintenance time window are determined by comprehensively considering the discharge development stage and intensity. Finally, the maintenance cost is calculated and information is sent, realizing accurate monitoring and evaluation of the partial discharge status of power equipment, effectively grasping the equipment status, and improving the accuracy of determining the maintenance timing of power equipment.

[0025] 2. In the present application, when performing a time series analysis on the current discharge event, the application first sorts the time interval sequence to obtain the time interval sequence, calculates the number of discharge events in the preset time window to obtain the discharge frequency sequence, and obtains the discharge amplitude to obtain the discharge intensity sequence. The three sequences are then decomposed to obtain the trend and fluctuation components. The discharge development trend characteristics and the discharge fluctuation characteristics are respectively calculated based on the trend component change rate and the fluctuation component variance. Finally, the current discharge time series characteristics are combined. The characteristics are interrelated. The time interval sequence reflects the law of discharge time, the frequency and intensity sequences reflect the degree of discharge activity, and the trend and fluctuation characteristics characterize the discharge state from different angles. The discharge activity can be accurately captured in the time series characteristics, providing a key basis for accurately judging the discharge development stage.

[0026] 3. The present application selects a pattern that matches the current discharge type from a standard discharge pattern library, which includes typical timing characteristics of different development stages, calculates the similarity between the current discharge timing characteristics and the typical characteristics of each stage, sets a preset feature threshold, and determines the current discharge development stage based on the comparison of the similarity between the first, second and third characteristics and the threshold. It makes full use of the typical characteristics of the standard discharge pattern as a reference, and through precise calculation and reasonable threshold judgment, it can accurately identify the development stage of the current discharge, such as the initial stage, development stage or acceleration stage, and provide accurate basic information for subsequent fault risk level determination and maintenance decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a flow chart of a partial discharge data processing method based on a big data model in an embodiment of the present application; Figure 2 is another flow chart of the partial discharge data processing method based on the big data model in the embodiment of the present application; Figure 3 It is another flow chart of the partial discharge data processing method based on the big data model in the embodiment of the present application; Figure 4 It is a schematic diagram of the structure of a physical device of the local discharge data processing system in the embodiment of the present application. DETAILED DESCRIPTION

[0028] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to be used as limitations to the present application. As used in the specification of the present application, the singular expressions "one", "a kind of", "above", "the" and "this" are intended to also include plural expressions, unless there is a clear indication to the contrary in the context. It should also be understood that the term "and / or" used in the present application refers to any or all possible combinations comprising one or more of the listed items.

[0029] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as suggesting or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, unless otherwise specified, "plurality" means two or more.

[0030] For ease of understanding, the following method is described in detail. Figure 1 , which is a flow chart of a local discharge data processing method based on a big data model in an embodiment of the present application.

[0031] S101. Obtain a partial discharge signal of an electric power device, and perform time-frequency analysis on the partial discharge signal to obtain current discharge characteristic data, wherein the current discharge characteristic data includes amplitude data, frequency data, and energy data.

[0032] Among them, power equipment refers to high-voltage equipment such as transformers, switchgear, cables, etc. used in power systems. Partial discharge signals refer to electromagnetic signals generated by tiny discharge phenomena occurring in the insulating medium inside the equipment. Time-frequency analysis refers to a method of analyzing signals in both the time domain and the frequency domain, which is used to obtain the time-varying and frequency characteristics of signals. Discharge characteristic data refers to the discharge signal characteristics extracted after signal processing, including amplitude data, frequency data, and energy data. Among them, amplitude data represents the intensity of the discharge signal, frequency data represents the frequency distribution characteristics of the discharge signal, and energy data represents the energy carried by the discharge signal.

[0033] This step is performed after the system starts the signal acquisition module. Specifically, the system first obtains the partial discharge signal through the ultra-high frequency sensor or ultrasonic sensor installed on the power equipment. The acquired original signal is pre-amplified and filtered to remove power frequency interference and high-frequency noise. Then the pre-processed signal is subjected to wavelet transformation to decompose the signal into different frequency bands to obtain the time-frequency distribution diagram of the signal. Based on the time-frequency distribution diagram, the system extracts the amplitude, main frequency and energy characteristics of each discharge pulse to form discharge characteristic data.

[0034] In some embodiments, the acquisition and feature extraction of partial discharge signals can be achieved in a variety of ways: optionally, using a UHF sensor to collect signals, extracting discharge pulses through bandpass filtering and envelope detection, and then using short-time Fourier transform to calculate time-frequency features; optionally, using an ultrasonic sensor to collect signals, using wavelet denoising and empirical mode decomposition to extract discharge features, and then using Hilbert transform to obtain instantaneous features. It is understandable that other sensing methods and signal processing methods can also be used to achieve the acquisition and feature extraction of discharge signals, which are not limited here.

[0035] S102: Input the current discharge feature data into a discharge classification model to obtain different types of current discharge events.

[0036] The discharge classification model refers to a machine learning model used to identify different types of discharges. The current discharge event represents a single discharge phenomenon detected. Discharge types include different forms of discharge modes such as suspended discharge, surface discharge, and corona discharge. The discharge feature data is the input of the model and contains a multi-dimensional feature vector that characterizes the discharge characteristics.

[0037] This step is performed after obtaining the discharge feature data. Specifically, the system inputs the extracted feature data into a pre-trained discharge classification model, which classifies the current discharge event into different discharge types by analyzing the distribution pattern of the feature data. The classification result includes the discharge type identification and the corresponding confidence level, which is used for subsequent status evaluation.

[0038] In some embodiments, the classification of discharge events can be achieved in a variety of ways: optionally, a support vector machine is used to build a classification model, and features are mapped to a high-dimensional space through a kernel function for classification to achieve recognition of discharge types; optionally, a deep neural network is used to build a classification model, and abstract representations of features are extracted through multi-layer nonlinear transformations to complete the discrimination of discharge types. It is understandable that other machine learning methods can also be used to achieve the classification of discharge events, which are not limited here.

[0039] S103, performing a time series analysis on the current discharge event to obtain a variation pattern of the discharge frequency, intensity and time interval, and forming a current discharge time series feature based on the variation pattern of the discharge frequency, the intensity and the time interval.

[0040] Among them, time series analysis refers to the method of analyzing data with time sequence. Discharge frequency refers to the number of times discharge occurs per unit time, which is used to measure the frequency of discharge. Discharge intensity refers to the amplitude of the discharge signal, reflecting the intensity of the discharge. The time interval refers to the time difference between two adjacent discharge events. The change law refers to the trend and characteristics of these parameters changing over time. The current discharge time series feature refers to the feature set that comprehensively reflects the time characteristics of the discharge activity.

[0041] This step is performed after obtaining the discharge event classification results. Specifically, the system arranges the detected discharge events in chronological order, counts the number of discharges in each preset time window to form a frequency sequence, extracts the amplitude of each discharge to form an intensity sequence, and calculates the time difference between adjacent discharge events to form an interval sequence. Then, trend analysis and fluctuation analysis are performed on these three sequences to obtain the development trend and stability characteristics of the discharge activity, and finally these features are combined to form a feature vector that describes the discharge timing characteristics.

[0042] In some embodiments, the extraction of discharge timing features can be achieved in a variety of ways: optionally, first divide the discharge event into time windows, calculate the statistical features within the window, then use the sliding average method to extract trend features, and finally obtain fluctuation features through variance analysis, and normalize all features and combine them; optionally, first perform empirical mode decomposition on the discharge sequence, extract trend components and fluctuation components, then calculate the rate of change and fluctuation intensity of each component, and finally construct a timing feature vector through a feature fusion method. It is understandable that other data analysis methods can also be used to extract discharge timing features, which are not limited here.

[0043] S104, performing dynamic time regularization calculation on the current discharge timing characteristics and the standard discharge mode to obtain the current discharge development stage.

[0044] Among them, the standard discharge mode refers to the typical characteristic mode exhibited by different types of discharges during their development. Dynamic time warping calculation is an algorithm for calculating the similarity of two time series, which can handle sequences of different lengths and different sampling rates. The current discharge development stage represents the time stage of the discharge degradation degree, which is usually divided into the initial stage, development stage and acceleration stage. The discharge timing feature is a feature vector that describes the time characteristics of the discharge activity.

[0045] This step is performed after obtaining the discharge timing characteristics. Specifically, the system selects a standard mode that matches the current discharge type from a pre-established standard discharge mode library, which contains typical timing characteristics of different development stages. Then, the dynamic time warping algorithm is used to calculate the similarity between the current discharge timing characteristics and the typical characteristics of each stage, and the development stage of the current discharge is determined based on the size of the similarity.

[0046] In some embodiments, the determination of the discharge development stage can be achieved in a variety of ways: optionally, first construct a standard discharge pattern library, including feature templates of various types of discharges at different stages, then calculate the dynamic time warping distance between the current feature and the template, and finally determine the development stage by setting a threshold; optionally, first establish a state assessment model based on pattern matching, then input the current feature into the model for similarity calculation, and finally determine the development stage through fuzzy reasoning. It is understandable that other pattern recognition methods can also be used to determine the discharge development stage, which is not limited here.

[0047] S105: Determine a fault risk level according to the current discharge development stage and the current discharge intensity, and determine a maintenance time window according to the fault risk level.

[0048] Among them, the fault risk level refers to the possibility of equipment failure, which is usually divided into three levels: low, medium, and high. The current discharge development stage indicates the time progression of the degree of discharge degradation, including the initial stage, development stage, and acceleration stage. The current discharge intensity refers to the amplitude of the discharge signal, reflecting the severity of the discharge. The maintenance time window refers to the time range in which maintenance work is recommended, which is inversely proportional to the risk level. The preset maintenance time threshold is used to define the boundaries of the maintenance time window, including the emergency maintenance time threshold and the regular maintenance time threshold.

[0049] This step is performed after the discharge development stage is determined. Specifically, the system uses the current discharge development stage and discharge intensity as input parameters, queries the preset risk level determination table, and obtains the fault risk level under the current state. Then, the corresponding maintenance time threshold is selected according to the risk level. The emergency maintenance time threshold is used for high-risk states, and the regular maintenance time threshold is used for medium-risk states. Finally, the time range for performing maintenance work is determined.

[0050] In some embodiments, fault risk assessment and maintenance time determination can be achieved in a variety of ways: optionally, first establish a risk assessment model based on historical fault data, then input current state parameters to calculate risk probability, and finally divide risk levels according to probability and determine maintenance time range; optionally, first build a rule-based expert system, then perform rule reasoning based on discharge development stage and intensity, and finally determine risk level and maintenance time window based on reasoning results. It is understandable that other risk assessment methods can also be used to achieve maintenance decision-making, which is not limited here.

[0051] S106: Calculate the maintenance cost within the maintenance time window, and send maintenance information including the maintenance cost to the target client.

[0052] Among them, maintenance cost refers to the expenses required to perform maintenance work, including labor cost, material cost and equipment cost. The maintenance time window indicates the time range in which maintenance work is allowed to be performed. Maintenance information refers to maintenance work related information including cost budget, recommended implementation time, etc. The target client refers to the terminal device or system that receives maintenance information. The cost prediction model is a maintenance cost estimation model trained based on historical data.

[0053] This step is performed after the maintenance time window is determined. Specifically, the system first establishes a maintenance cost prediction model based on historical maintenance records, and inputs the current state parameters into the model to obtain the basic maintenance cost. Then, considering the impact of the maintenance time window, the time cost coefficient is calculated, and the basic maintenance cost is multiplied by the time cost coefficient to obtain the final maintenance cost. Finally, the maintenance cost, recommended maintenance time and other information are packaged and sent to the specified client device through the network interface.

[0054] In some embodiments, maintenance cost prediction and information transmission can be achieved in a variety of ways: optionally, firstly, data cleaning and feature extraction of historical maintenance records are performed, then a regression model is trained to predict basic costs, and finally, actual costs are calculated and maintenance recommendations are generated in combination with time factors; optionally, a cost estimation system based on case reasoning is first established, then similar historical cases are retrieved to estimate costs, and finally, the final cost is determined considering the impact of time and maintenance information is sent. It is understandable that other cost prediction methods can also be used to achieve maintenance cost estimation and information release, which are not limited here.

[0055] The following is a more detailed description of the process of the method provided by this implementation. Figure 2 , is another flow chart of the local discharge data processing method based on the big data model in an embodiment of the present application.

[0056] S201. Obtain a partial discharge signal of an electric power device, and perform time-frequency analysis on the partial discharge signal to obtain current discharge characteristic data, wherein the current discharge characteristic data includes amplitude data, frequency data, and energy data.

[0057] Partial discharge refers to the phenomenon of tiny discharges occurring in local areas inside power equipment. Partial discharge signals contain electrical signals that reflect discharge characteristics, usually in the form of voltage or current pulses. Time-frequency analysis is a method that simultaneously analyzes the characteristics of a signal in the time domain and frequency domain, and can obtain the frequency characteristics of a signal that changes over time. The amplitude data in the discharge characteristic data reflects the discharge intensity, the frequency data reflects the discharge spectrum characteristics, and the energy data reflects the amount of energy released by a single discharge.

[0058] The system collects partial discharge signals through sensors installed on power equipment. The collected raw signals are first pre-processed to remove external interference. Then the system uses time-frequency analysis methods such as wavelet transform to decompose the signal into different frequency bands and extract the amplitude, frequency and energy characteristics of each frequency band. In specific implementation, the system can use continuous wavelet transform, select Mexican hat wavelet as the mother wavelet, set the analysis scale to 1-8, obtain the wavelet coefficients of the signal at different scales, and then calculate the characteristic data.

[0059] S202: Input the current discharge feature data into a discharge classification model to obtain different types of current discharge events.

[0060] The discharge classification model is a machine learning model used to identify discharge types. It achieves automatic classification by learning the characteristics of known types of discharges. A discharge event refers to a single complete discharge process. Different types of discharge events have different characteristic patterns. The current discharge event refers to the specific discharge phenomenon currently detected.

[0061] The system inputs the feature data obtained in S201 into a pre-trained discharge classification model. The model can use a support vector machine (SVM) or a deep neural network structure, with the input layer receiving feature data and the output layer giving a probability distribution of the discharge type. The model can classify discharge events into specific types such as suspended discharge, surface discharge, and corona discharge. The classification result includes a discharge type label and a corresponding confidence level.

[0062] It should be noted that the discharge classification model adopts a deep learning architecture, and establishes a mapping relationship from input features to discharge types through training with a large amount of labeled data. The training data sources include laboratory simulation data and historical data collected on site. Each training sample contains amplitude data, frequency data, and energy data as input features, as well as discharge type labels annotated by experts. In the data preprocessing stage, the system uses the wavelet threshold denoising method to eliminate background noise, normalizes features of different dimensions to the [-1, 1] interval, and performs data enhancement by adding random noise, time scale transformation, and other methods, and balances the number of various types of discharge samples. The model adopts a deep convolutional neural network structure, including an input layer, 3 convolutional layers (convolution kernel size 3×3), a pooling layer, 3 fully connected layers (the number of neurons is 512, 256, and 128, respectively), and a softmax output layer. The training process uses the cross entropy loss function and Adam optimizer, the initial learning rate is set to 0.001 and a dynamic adjustment strategy is adopted, the batch size is 64, the number of training rounds is 100, of which 20% of the training data is used for validation, and the early stopping mechanism is triggered when the performance of the validation set does not improve for 5 consecutive rounds. The model evaluation uses indicators such as accuracy, recall, F1 score and confusion matrix, and continuously optimizes the model performance through feature selection, parameter tuning, ensemble learning and online update.

[0063] S203: Sort the current discharge events in chronological order, and obtain a time interval sequence between adjacent current discharge events.

[0064] Time sequence refers to the order in which discharge events occur. Time interval sequence records the time difference between two adjacent discharge events. Adjacent discharge events refer to two discharges that occur consecutively in time.

[0065] The system first sorts all events in ascending order according to the timestamps of the discharge events to form time series data. Then the timestamp differences of adjacent events are calculated to obtain the interval sequence. For example, if three consecutive discharges occur at t1=1s, t2=1.5s, and t3=2.3s, the corresponding time interval sequence is [0.5s, 0.8s]. This sequence reflects the time pattern of discharge occurrence and can be used for subsequent trend analysis.

[0066] S204: Calculate the number of current discharge events in each preset time window to obtain a discharge frequency sequence, and obtain the amplitude corresponding to each current discharge event as a discharge intensity sequence.

[0067] The preset time window refers to a fixed-length time period used to count the characteristics of discharge events, usually set to the minute or hour level. The discharge frequency series records the number of discharge events that occurred in each time window, reflecting the frequency of discharge activities. The discharge intensity series consists of the amplitude of each discharge event. The amplitude indicates the intensity of the discharge signal in millivolts or microvolts.

[0068] The system first divides the time axis evenly into preset time windows (such as 10 minutes), counts the discharge events in each window, and forms a discharge frequency sequence. At the same time, the amplitude data of each discharge event is extracted and arranged in chronological order to form a discharge intensity sequence. For example, if 5, 8, and 3 discharges are detected in three consecutive 10-minute windows, and the amplitudes of each discharge are [2.1, 1.8, 2.3, 2.5, 1.9], [2.4, 2.8, 2.2, 2.6, 2.3, 2.1, 2.4, 2.7], and [2.9, 2.5, 2.4] millivolts, respectively, then the frequency sequence [5, 8, 3] and the corresponding intensity sequence are obtained.

[0069] S205 , decomposing the time interval sequence, the discharge frequency sequence and the discharge intensity sequence to obtain a trend component and a fluctuation component of the current discharge event.

[0070] The trend component reflects the long-term change trend of the signal and reflects the overall direction of the discharge development. The fluctuation component represents the short-term fluctuation of the signal around the trend and reflects the instability of the discharge activity. Series decomposition is a mathematical process that separates time series data into different characteristic components.

[0071] The system uses the empirical mode decomposition (EMD) method to decompose the three sequences. EMD decomposes the sequence into several intrinsic mode functions (IMFs) and a residual term through iterative screening. The trend component is represented by the residual term, and the volatility component is obtained by the superposition of IMFs. In the specific implementation, the extreme points are first determined and the envelope is constructed by cubic spline interpolation. The mean envelope is calculated as the local mean, and the local mean is subtracted from the original data to obtain the detail signal. This process is repeated until the IMF conditions are met, and finally the complete decomposition result is obtained.

[0072] S206: Calculate the discharge development trend characteristics based on the change rate of the trend component, and calculate the discharge fluctuation characteristics based on the variance of the fluctuation component.

[0073] The rate of change refers to the change amplitude of the trend component in a unit time, reflecting the speed of discharge development. The discharge development trend characteristics are used to characterize the evolution direction and speed of discharge activities. The variance measures the degree to which the fluctuation component deviates from the mean, and the discharge fluctuation characteristics reflect the stability of the discharge activity.

[0074] The system calculates the first-order difference of the trend component to obtain the change rate sequence, and obtains the discharge development trend characteristics through weighted average, with the weight decaying over time to highlight recent changes. The sliding variance is calculated for the fluctuation component, and the window size is the same as the preset time window, to obtain a characteristic sequence reflecting the local fluctuation intensity. These two characteristics reflect the development trend and stability of the discharge activity, respectively, and provide important basis for subsequent status assessment. For example, a larger positive trend feature indicates that the discharge activity is intensifying, and a larger fluctuation feature indicates that the discharge activity is unstable, which are early warning signs of potential faults.

[0075] S207: Combining the discharge development trend feature and the discharge fluctuation feature to form a current discharge timing feature.

[0076] The discharge development trend feature is a numerical indicator that describes the long-term change trend of the discharge activity, including the change direction and rate information. The discharge fluctuation feature is a numerical indicator that describes the short-term fluctuation degree of the discharge activity, reflecting the discharge stability. The current discharge timing feature is a feature vector that comprehensively characterizes the timing characteristics of the discharge activity, including information in two dimensions: trend and fluctuation.

[0077] The system uses a feature fusion method to construct the discharge timing features. First, the discharge development trend features and discharge fluctuation features are normalized so that the numerical ranges of the two types of features are unified to the interval [0, 1]. Then the normalized features are combined into a feature vector according to a predefined structure, and the vector dimension is the sum of the trend feature dimension and the fluctuation feature dimension. For example, if the trend feature contains two components: the average rate of change and the cumulative change, and the fluctuation feature contains two components: the variance and the peak-to-valley ratio, then the combined timing feature is a 4-dimensional vector [trend feature 1, trend feature 2, fluctuation feature 1, fluctuation feature 2].

[0078] S208. Select a standard discharge mode that matches the current discharge type from a standard discharge mode library, where the standard discharge mode includes typical timing characteristics of a discharge development stage, and the discharge development stage includes an initial stage, a development stage, and an acceleration stage.

[0079] The standard discharge pattern library is a database that stores typical discharge development patterns, including standard features of different types of discharges at various development stages. The discharge development stage is the different periods of the discharge degradation process. The initial stage is characterized by sporadic low-intensity discharges, the development stage is characterized by steadily increasing discharge activities, and the acceleration stage is characterized by rapidly deteriorating discharge states. Typical timing features are representative timing feature templates for each development stage.

[0080] Based on the discharge type identified in the previous step, the system retrieves the standard pattern of the corresponding type from the standard discharge pattern library. The standard pattern contains typical time series feature templates of three stages, each of which has the same dimension and structure as the current discharge time series feature. These templates are obtained through statistical analysis of a large amount of historical data and reflect the general law of discharge development. For example, for the surface discharge type, the initial template shows low frequency and low intensity characteristics, the development template shows medium frequency and slow growth characteristics, and the acceleration template shows high frequency and rapid growth characteristics.

[0081] S209, calculating a first feature similarity, a second similarity, and a third similarity between the current discharge timing feature and the typical timing features of all discharge development stages.

[0082] Feature similarity is a numerical indicator that measures the similarity between two feature vectors. The first feature similarity measures the matching degree of the current feature with the typical feature of the initial stage, the second similarity measures the matching degree with the typical feature of the development stage, and the third similarity measures the matching degree with the typical feature of the acceleration stage.

[0083] The system uses the cosine similarity method to calculate the similarity between feature vectors. First, the current discharge timing characteristics and the typical timing characteristics of the three stages are respectively vector inner product operations, and then divided by the product of the vector modulus length to obtain three standardized similarity values, with a value range of [-1, 1]. The calculation formula is: cos(θ)=(A·B) / (|A|·|B|), where A is the current feature vector and B is the typical feature vector. For example, if the current feature vector is [0.3, 0.4, 0.2, 0.3], the similarity calculation result with the initial typical feature [0.2, 0.3, 0.2, 0.2] is 0.98, indicating that the current discharge characteristics are highly similar to the initial characteristics. These three similarity values ​​together reflect the development stage of the current discharge activity.

[0084] S210. If the first feature similarity is greater than the preset feature threshold, the current discharge development stage is the initial stage; if the second feature similarity is greater than the preset feature threshold, the current discharge development stage is the development stage; if the third feature similarity is greater than the preset feature threshold, the current discharge development stage is the acceleration stage.

[0085] The preset feature threshold is a standard value for judging feature similarity, usually set between 0.8-0.95. The determination of the discharge development stage is the process of classifying the current discharge state into the initial stage, development stage, or acceleration stage. The initial stage features low-frequency, low-intensity sporadic discharges; the development stage features steady growth in discharge frequency and intensity; and the acceleration stage features rapid deterioration of discharge activity.

[0086] The system uses a threshold judgment method to determine the discharge development stage. First, a unified preset feature threshold (such as 0.9) is set, and the three feature similarity values ​​are compared with the threshold. When the first feature similarity exceeds the threshold, the system determines that it is currently in the initial stage; when the second feature similarity exceeds the threshold, it is determined to be in the development stage; when the third feature similarity exceeds the threshold, it is determined to be in the acceleration stage. For example, if the three similarity values ​​are [0.95, 0.75, 0.60], and the preset threshold is 0.9, then because the first feature similarity 0.95 is greater than the threshold, the system determines that it is currently in the initial stage.

[0087] After combining the above scenarios, the following is a more detailed description of the process of the method provided by this implementation. Figure 3 , is another flow chart of the local discharge data processing method based on the big data model in an embodiment of the present application.

[0088] S301. Based on historical maintenance records, calculate equipment failure rates at different discharge development stages and discharge intensity combinations.

[0089] Historical maintenance records are operation and maintenance data sets for power equipment, including equipment status, fault information, and maintenance measures. Equipment failure rate is the statistical value of the probability of failure under specific conditions. Discharge intensity combination refers to the combination of discharge activities of different intensity levels, which are usually divided into three levels: low, medium, and high.

[0090] The system performs data mining and statistical analysis on historical maintenance records. First, the discharge intensity is divided into multiple levels according to the amplitude, such as low intensity (<2mV), medium intensity (2-5mV), and high intensity (>5mV). Then the frequency of equipment failures at different discharge development stages (initial, development, acceleration) and discharge intensity combinations is counted. The calculation formula is: Failure rate = number of failures under specific conditions / total number of observations. For example, the failure rate is 5% in the initial stage and when the discharge intensity is low, while the failure rate reaches 85% in the acceleration stage and when the discharge intensity is high.

[0091] S302: Establish a risk level determination table based on the equipment failure rate.

[0092] The risk level determination table is a multi-dimensional matrix used to map the discharge development stage and discharge intensity to the corresponding risk level. Risk levels are usually divided into multiple levels, such as low risk, medium risk and high risk, and each level corresponds to different disposal requirements. Equipment failure rate is the basic data for establishing risk level determination.

[0093] The system constructs a risk level determination table based on the statistically obtained failure rate data. First, determine the risk level classification criteria, such as failure rate <20% is low risk, 20%-60% is medium risk, and >60% is high risk. Then establish a three-dimensional matrix with discharge development stage, discharge intensity, and risk level as the three dimensions. Each element in the matrix corresponds to the risk level under a specific discharge development stage and discharge intensity combination. For example, for low-intensity discharge in the initial stage, the failure rate is 5%, which is determined to be low risk; for high-intensity discharge in the acceleration period, the failure rate is 85%, which is determined to be high risk. This determination table provides a basis for subsequent maintenance decisions.

[0094] S303: Input the current discharge development stage and the current discharge intensity into a preset risk level determination table to obtain the fault risk level.

[0095] The preset risk level determination table is a multi-dimensional mapping matrix that stores the risk levels corresponding to different discharge development stages and discharge intensity combinations. The current discharge development stage is a time-dimensional classification of the degree of discharge degradation, including the initial stage, development stage, and acceleration stage. The current discharge intensity is a quantitative level of the discharge signal amplitude, which is usually divided into three levels: low, medium, and high. The fault risk level is a quantitative assessment of the possibility of equipment failure, which is divided into multiple levels based on the statistical results of the failure rate.

[0096] The system uses a table lookup method to determine the fault risk level. First, the current discharge development stage (such as development stage) and discharge intensity (such as medium intensity) are used as input parameters to locate the corresponding table entry in the preset risk level determination table. The determination table is implemented using a three-dimensional array structure. The first dimension of the array represents the discharge development stage (0-initial stage, 1-development stage, 2-acceleration stage), the second dimension represents the discharge intensity level (0-low intensity, 1-medium intensity, 2-high intensity), and the third dimension stores the corresponding risk level value. For example, risk_level=risk_table[1][1] means querying the risk level under the conditions of development stage and medium intensity.

[0097] S304: Determine a maintenance time window according to the fault risk level and a preset maintenance time threshold, wherein the maintenance time threshold includes a first maintenance time threshold and a second maintenance time threshold.

[0098] The maintenance time window is the recommended execution time range for equipment maintenance operations. The preset maintenance time threshold is the boundary parameter of the maintenance time window, where the first maintenance time threshold corresponds to the time limit for emergency maintenance (such as 24 hours), and the second maintenance time threshold corresponds to the time limit for routine maintenance (such as 72 hours). There is a corresponding relationship between the risk level and the maintenance time window. The higher the risk level, the shorter the maintenance time window.

[0099] The system selects the corresponding maintenance time threshold based on the value of the fault risk level. The maintenance time threshold is set in stages, with the first maintenance time threshold for high-risk situations and the second maintenance time threshold for medium-risk situations. In specific implementation, the system compares the risk level value with the preset risk threshold and selects the corresponding time threshold range based on the comparison result. For example, for a risk level of 0.8, because it exceeds the first risk threshold of 0.7, the system limits the maintenance time window to the first maintenance time threshold (24 hours).

[0100] S305: When the fault risk level is greater than a first risk threshold, determine that the maintenance time window is less than the first maintenance time threshold.

[0101] The first risk threshold is the critical value of the risk level to determine whether emergency maintenance is required, which is usually set to a higher value (such as 0.7 or 0.8). When the fault risk level exceeds this threshold, it indicates that the equipment has a greater risk of failure and needs to be maintained in a short time. The first maintenance time threshold is the maximum allowable delay time in emergency maintenance.

[0102] The system performs risk threshold judgment and maintenance time setting. When the calculated fault risk level value is greater than the first risk threshold, the system will start the emergency maintenance process. The maintenance time window is limited to the first maintenance time threshold to ensure that maintenance work is carried out in a timely manner. For example, if the fault risk level is 0.85, which is greater than the first risk threshold of 0.7, and the first maintenance time threshold is set to 24 hours, the system will set the maintenance time window to the range of 0-24 hours, requiring maintenance work to be completed within 24 hours.

[0103] S306: When the fault risk level is between the first risk threshold and the second risk threshold, determine that the maintenance time window is between the first maintenance time threshold and the second maintenance time threshold.

[0104] The second risk threshold is the critical value of the risk level for judging the need for routine maintenance, which is usually set to a medium value (such as 0.4). The interval between the first risk threshold and the second risk threshold corresponds to a medium risk state, and maintenance needs to be arranged within a reasonable time. The first maintenance time threshold and the second maintenance time threshold are the upper and lower boundaries of the maintenance time window, such as 24 hours and 72 hours. The maintenance time window is the specific time range for performing maintenance work.

[0105] The system determines the maintenance time window by comparing the thresholds. When the value of the fault risk level is between the first risk threshold (such as 0.7) and the second risk threshold (such as 0.4), the system sets the maintenance time window between the first maintenance time threshold (such as 24 hours) and the second maintenance time threshold (such as 72 hours). In specific implementation, the system first determines whether the risk level meets the interval condition: risk_level<first_threshold&&risk_level> second_threshold, when the conditions are met, the maintenance time window is set to [first_time_threshold, second_time_threshold]. For example, when the risk level is 0.6, the system sets the maintenance time window to 24-72 hours.

[0106] S307: Obtain historical maintenance records of the power equipment, and establish a maintenance cost prediction model based on the historical maintenance records, where the historical maintenance records include historical maintenance costs and historical fault types.

[0107] Historical maintenance records are maintenance work record datasets for power equipment, including maintenance date, maintenance content, maintenance cost, and fault type. Historical maintenance costs record the specific costs of each maintenance work, including labor costs, material costs, and equipment costs. Historical fault types describe the specific forms of equipment failures, such as insulation aging and mechanical damage. The maintenance cost prediction model is a machine learning model trained based on historical data and is used to predict the cost of future maintenance work.

[0108] The system uses machine learning methods to build a maintenance cost prediction model. First, the historical maintenance records are preprocessed, including data cleaning, feature extraction and standardization. Then, an ensemble learning algorithm (such as random forest or gradient boosting tree) is used to train the prediction model. The input features include fault type, equipment operation time, environmental factors, etc., and the output is the predicted maintenance cost. The model training uses a cross-validation method to evaluate performance and select the model parameters with the smallest prediction error. For example, by analyzing historical records, it was found that the average maintenance cost of a certain type of fault at different stages was 5,000 yuan in the initial stage, 15,000 yuan in the development stage, and 30,000 yuan in the acceleration stage.

[0109] It should be noted that the maintenance cost prediction model is established based on historical maintenance record data, and a decision tree structure is used to achieve the mapping of equipment status to maintenance cost. The training data comes from the maintenance records of power equipment, which contains the following key information: basic equipment information (type, operating time), fault characteristics (discharge development stage, discharge type), and maintenance information (maintenance measure type, maintenance cost). The maintenance records are cleaned in the data preprocessing stage, including removing obviously erroneous records, unifying cost measurement units, and supplementing missing category labels. Feature engineering mainly considers the three main influencing factors of discharge development stage, equipment operating time, and historical maintenance times, and converts categorical features into numerical codes. The model adopts a basic decision tree structure, takes the discharge development stage as the primary split feature, and establishes a segmented mapping relationship from status to cost. The training process uses 80% of the historical data as the training set and 20% as the validation set, and the prediction effect is evaluated by the mean square error. The model input is the current discharge development stage and related state characteristics, and the output is the predicted maintenance cost range under the corresponding working conditions. This prediction method based on historical data statistics can provide a preliminary cost reference for maintenance decisions.

[0110] S308: Input the current discharge development stage into the maintenance cost prediction model to obtain predicted maintenance cost.

[0111] The current discharge development stage is an indicator of the equipment status, including the initial stage, development stage, and acceleration stage. The maintenance cost prediction model is a decision tree structure established based on historical data, which maps the discharge development stage to the maintenance cost. The predicted maintenance cost is the estimated value of the maintenance cost output by the model, reflecting the capital investment required to complete the maintenance work under the current status.

[0112] The system calls the maintenance cost prediction model to perform cost prediction. First, the current discharge development stage is used as the main input feature, and auxiliary features such as equipment operation time and historical maintenance times are input into the pre-trained decision tree model. The model outputs the predicted maintenance cost value based on the cost data under similar working conditions in the historical maintenance records. For example, for equipment in the development stage, if the historical records show that the maintenance cost of this stage is mainly distributed in the range of 8,000-12,000 yuan, the predicted cost output by the model is 10,000 yuan.

[0113] S309: Calculate a time cost coefficient based on the maintenance time window, determine the maintenance cost within the maintenance time window according to the predicted maintenance cost and the time cost coefficient, and send maintenance information including the maintenance cost to the target client.

[0114] The time cost coefficient is a regulatory factor that reflects the impact of maintenance delay on costs and is related to the selection of the maintenance time window. The maintenance time window is the time range for performing maintenance work. The maintenance cost is the final cost after comprehensively considering the predicted cost and time impact. The maintenance information includes the cost budget, the recommended implementation time, etc. The target client is the terminal device that receives the maintenance information.

[0115] The system calculates the time cost coefficient based on the selected maintenance time window. The time cost coefficient increases as the maintenance delay increases, reflecting the cost increase caused by delayed maintenance. The system multiplies the predicted maintenance cost by the time cost coefficient to obtain the final maintenance cost after considering the time factor. The maintenance information includes the maintenance cost value, maintenance time suggestion, maintenance measure suggestion, etc., which is sent to the specified client device through the network interface. For example, if the predicted cost is 10,000 yuan and the time cost coefficient is 1.2, the final maintenance cost is 12,000 yuan.

[0116] S310: Compare the maintenance cost with the equipment replacement cost.

[0117] S311. When the maintenance cost exceeds a preset proportion of the equipment replacement cost, generate an equipment replacement suggestion and send it to the target client.

[0118] Maintenance cost is the total expenditure for performing maintenance work. Equipment replacement cost is the total cost of purchasing and installing new equipment. The preset ratio is the critical value for judging whether to recommend equipment replacement, which is usually set at 0.6-0.8. Equipment replacement recommendations are disposal opinions given based on cost comparison.

[0119] The system performs cost comparison and recommendation generation. First, the standard replacement cost of the equipment is obtained, and the calculated maintenance cost is calculated with the replacement cost. When the ratio exceeds the preset critical ratio, it indicates that the economic efficiency of continued maintenance is poor, and the system generates a device replacement recommendation. The replacement recommendation includes cost comparison analysis, replacement reasons, and replacement time recommendations, and is sent to the target client through the network interface. For example, if the maintenance cost is 80,000 yuan and the equipment replacement cost is 100,000 yuan, the ratio of 0.8 exceeds the preset ratio of 0.7, and the system will generate a replacement recommendation.

[0120] In some embodiments, steps S301 to S312 may be performed after step S210 or after step S106, which is not limited here.

[0121] The following describes the partial discharge data processing system in the embodiment of the present invention from the perspective of hardware processing. Figure 4 , which is a schematic diagram of a physical device structure of a local discharge data processing system in an embodiment of the present application.

[0122] It should be noted that Figure 4 The structure of the partial discharge data processing system shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0123] like Figure 4 As shown, the partial discharge data processing system includes a central processing unit (CPU) 401, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 402 or the program loaded from the storage part 408 to the random access memory (RAM) 403, such as executing the method described in the above embodiment. Various programs and data required for system operation are also stored in RAM 403. CPU 401, ROM 402 and RAM 403 are connected to each other through bus 404. Input / output (I / O) interface 405 is also connected to bus 404.

[0124] The following components are connected to the I / O interface 405: an input section 406 including an audio input device, a button switch, etc.; an output section 407 including a liquid crystal display (LCD) and an audio output device, an indicator light, etc.; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to the I / O interface 405 as needed. A removable medium 411, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 410 as needed so that a computer program read therefrom is installed into the storage section 408 as needed.

[0125] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 409, and / or installed from a removable medium 411. When the computer program is executed by the central processing unit (CPU) 401, various functions defined in the present invention are performed.

[0126] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, apparatus, or device.

[0127] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. Each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from that marked in the accompanying drawings.

[0128] Specifically, the partial discharge data processing system of this embodiment includes a processor and a memory. The memory stores a computer program. When the computer program is executed by the processor, the partial discharge data processing method based on the big data model provided in the above embodiment is implemented.

[0129] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the partial discharge data processing system described in the above embodiment; or may exist independently without being assembled into the partial discharge data processing system. The above storage medium carries one or more computer programs, and when the above one or more computer programs are executed by a processor of the partial discharge data processing system, the partial discharge data processing system implements the partial discharge data processing method based on the big data model provided in the above embodiment.

[0130] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

[0131] As used in the above embodiments, the term "when..." may be interpreted to mean "if..." or "after..." or "in response to determining..." or "in response to detecting...", depending on the context. Similarly, the phrases "upon determining..." or "if (the stated condition or event) is detected" may be interpreted to mean "if determining..." or "in response to determining..." or "upon detecting (the stated condition or event)" or "in response to detecting (the stated condition or event)", depending on the context.

[0132] Those skilled in the art can understand that to implement all or part of the processes in the above-mentioned embodiments, the processes can be completed by computer programs to instruct related hardware, and the programs can be stored in computer-readable storage media. When the programs are executed, they can include the processes of the above-mentioned method embodiments. The aforementioned storage media include: ROM or random access memory RAM, magnetic disk or optical disk and other media that can store program codes.

Claims

1. A partial discharge data processing method based on a big data model, characterized in that: Applied to a partial discharge data processing system, the method comprises: Acquire a partial discharge signal of the electric power equipment, and perform time-frequency analysis on the partial discharge signal to obtain current discharge characteristic data, wherein the current discharge characteristic data includes amplitude data, frequency data and energy data; Inputting the current discharge characteristic data into a discharge classification model to obtain different types of current discharge events; Performing a time series analysis on the current discharge event to obtain a variation rule of the discharge frequency, intensity and time interval, and forming a current discharge time series feature based on the variation rule of the discharge frequency, intensity and time interval; Perform dynamic time warping calculation on the current discharge timing characteristics and the standard discharge mode to obtain the current discharge development stage; Determine a fault risk level according to the current discharge development stage and the current discharge intensity, and determine a maintenance time window according to the fault risk level; The maintenance cost within the maintenance time window is calculated, and maintenance information including the maintenance cost is sent to the target client.

2. The method according to claim 1, characterized in that The step of performing a time series analysis on the current discharge event to obtain a change rule of the discharge frequency, intensity and time interval, and forming a current discharge time series feature based on the change rule of the discharge frequency, intensity and time interval, specifically includes: Sorting the current discharge events in chronological order and obtaining a time interval sequence between adjacent current discharge events; Calculating the number of current discharge events in each preset time window to obtain a discharge frequency sequence, and obtaining the amplitude corresponding to each current discharge event as a discharge intensity sequence; Decomposing the time interval sequence, the discharge frequency sequence and the discharge intensity sequence to obtain a trend component and a fluctuation component of the current discharge event; Calculating the discharge development trend characteristics based on the rate of change of the trend component, and calculating the discharge fluctuation characteristics based on the variance of the fluctuation component; The discharge development trend feature and the discharge fluctuation feature are combined to form a current discharge timing feature.

3. The method according to claim 1, characterized in that The step of performing dynamic time warping calculation on the current discharge timing characteristics and the standard discharge mode to obtain the current discharge development stage specifically includes: Selecting a standard discharge mode that matches the current discharge type from a standard discharge mode library, wherein the standard discharge mode includes typical timing characteristics of a discharge development stage, wherein the discharge development stage includes an initial stage, a development stage, and an acceleration stage; Calculating a first feature similarity, a second similarity, and a third similarity between the current discharge timing feature and the typical timing features of all discharge development stages; If the first feature similarity is greater than the preset feature threshold, the current discharge development stage is the initial stage; if the second feature similarity is greater than the preset feature threshold, the current discharge development stage is the development stage; if the third feature similarity is greater than the preset feature threshold, the current discharge development stage is the acceleration stage.

4. The method according to claim 1, characterized in that The step of determining the fault risk level according to the current discharge development stage and the current discharge intensity, and determining the maintenance time window according to the fault risk level specifically includes: Inputting the current discharge development stage and the current discharge intensity into a preset risk level determination table to obtain the fault risk level; Determine a maintenance time window according to the fault risk level and a preset maintenance time threshold, wherein the maintenance time threshold includes a first maintenance time threshold and a second maintenance time threshold; When the fault risk level is greater than a first risk threshold, determining that the maintenance time window is less than the first maintenance time threshold; When the failure risk level is between a first risk threshold and a second risk threshold, it is determined that the maintenance time window is between the first maintenance time threshold and the second maintenance time threshold.

5. The method according to claim 1, characterized in that The step of calculating the maintenance cost within the maintenance time window specifically includes: Acquire historical maintenance records of the power equipment, and establish a maintenance cost prediction model based on the historical maintenance records, wherein the historical maintenance records include historical maintenance costs and historical fault types; Inputting the current discharge development stage into the maintenance cost prediction model to obtain a predicted maintenance cost; A time cost coefficient is calculated based on the maintenance time window, and a maintenance cost within the maintenance time window is determined according to the predicted maintenance cost and the time cost coefficient.

6. The method according to claim 1, characterized in that After the step of calculating the maintenance cost within the maintenance time window and sending the maintenance information including the maintenance cost to the target client, the method further includes: Compare said maintenance costs to equipment replacement costs; When the maintenance cost exceeds a preset proportion of the equipment replacement cost, an equipment replacement suggestion is generated and sent to the target client.

7. The method according to claim 4, characterized in that Before the step of inputting the current discharge development stage and the current discharge intensity into a preset risk level determination table to obtain the fault risk level, the method further includes: Based on historical maintenance records, the equipment failure rate at different discharge development stages and discharge intensity combinations is calculated; A risk level determination table is established based on the equipment failure rate.

8. A partial discharge data processing system, characterized in that: The partial discharge data processing system includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the partial discharge data processing system to perform the method as described in any one of claims 1 to 7.

9. A computer-readable storage medium comprising instructions, characterized in that: When the instructions are executed on a partial discharge data processing system, the partial discharge data processing system is caused to execute the method according to any one of claims 1 to 7.

10. A computer program product, characterized in that When the computer program product runs on a partial discharge data processing system, the partial discharge data processing system is enabled to execute the method according to any one of claims 1 to 7.

Images