Relay protection fault prediction and early warning method, system and device based on intelligent oscillograph and storage medium
By collecting and analyzing the operating parameters of the relay protection device in an intelligent recorder, constructing a multidimensional feature vector and utilizing multiple fault development trend models, the problems of low fault diagnosis accuracy and insufficient early warning in the existing technology are solved, accurate fault identification and early warning are achieved, and the safety and reliability of the power system are improved.
Patent Information
- Application Number
- CN202510559304.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-30
- Publication Date
- 2025-09-23
AI Technical Summary
Existing relay protection fault diagnosis technology based on intelligent recorders relies on expert experience and rule bases, resulting in low accuracy and efficiency in complex fault diagnosis. Most of the faults are diagnosed after the fact and cannot be warned in advance. There is a lack of prediction of fault development trends, making it difficult to provide effective preventive measures.
By collecting the operating parameters of the relay protection device, a multidimensional feature vector is established, and the operating status characteristics are extracted using a pre-trained fault feature extraction model. A fault development trend model is constructed, including Weibull distribution, hidden Markov model and differential equation model, to monitor the fault development trend in real time and issue early warnings.
It realizes accurate fault identification and trend prediction of relay protection devices, can provide early warning of faults, reduce power outages and economic losses, and improve the reliability and safety of the power system.
Smart Images
Figure CN120691304A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of oscilloscopes, and in particular to a relay protection fault prediction and early warning method, system, device and storage medium based on an intelligent oscilloscope. Background Art
[0002] Power system relay protection devices are critical equipment for ensuring safe and stable operation. To ensure power system reliability, they require regular inspection and maintenance. Traditional maintenance of relay protection devices relies primarily on manual inspections and periodic testing, which are inefficient, costly, and lack real-time monitoring capabilities. In recent years, with the rapid development of smart grid technology, relay protection fault diagnosis technology based on intelligent oscilloscopes has gained widespread application. Intelligent oscilloscopes can collect and store large amounts of power system operating data in real time, providing a rich data foundation for relay protection fault diagnosis.
[0003] However, existing relay protection fault diagnosis technology based on intelligent oscilloscopes still has several shortcomings: First, existing fault diagnosis methods rely primarily on expert experience and rule bases, resulting in low diagnostic accuracy and efficiency for complex fault scenarios. Second, existing fault diagnosis methods mostly rely on post-facto analysis, failing to provide early warning of faults and effectively prevent them from occurring. Finally, existing fault diagnosis methods lack the ability to predict fault development trends, making it difficult to provide effective fault prevention measures. Summary of the Invention
[0004] In view of the above-mentioned problems, the present invention is proposed.
[0005] Therefore, the technical problem solved by the present invention is that the existing relay protection fault diagnosis technology based on intelligent recorders relies on expert experience and rule bases, resulting in low accuracy and efficiency in complex fault diagnosis, most of the faults are analyzed post-hoc and cannot be warned in advance, and there is a lack of prediction of fault development trends, making it difficult to provide effective preventive measures.
[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:
[0007] In a first aspect, an embodiment of the present invention provides a relay protection fault prediction and early warning method based on an intelligent oscilloscope, comprising:
[0008] Collecting the operating parameters of the relay protection device and establishing a multidimensional feature vector, inputting the multidimensional feature vector into a pre-trained fault feature extraction model to extract the operating status characteristics of the relay protection device;
[0009] Identify the fault type of the relay protection device based on the operating status characteristics and build a fault development trend model;
[0010] According to the fault development trend model, fault early warning is issued for the relay protection device.
[0011] As a preferred solution for the relay protection fault prediction and early warning method based on the intelligent oscilloscope, the following are the solutions:
[0012] The collecting of operating parameters of the relay protection device and establishing a multidimensional feature vector, inputting the multidimensional feature vector into a pre-trained fault feature extraction model, and extracting the operating status features of the relay protection device include:
[0013] Various operating parameters of the relay protection device are collected and preprocessed, and a multidimensional feature vector is established based on the preprocessed parameters. A parameter correlation matrix is constructed based on the multidimensional feature vector to quantify the correlation between different parameters, and the parameter correlation matrix is subjected to feature dimensionality reduction to obtain the reduced eigenvector.
[0014] As a preferred solution for the relay protection fault prediction and early warning method based on the intelligent oscilloscope, the following are the solutions:
[0015] The collecting of operating parameters of the relay protection device and establishing a multidimensional feature vector, inputting the multidimensional feature vector into a pre-trained fault feature extraction model, and extracting the operating status features of the relay protection device further includes:
[0016] The reduced-dimensional feature vector is input into a pre-trained fault feature extraction model, and the model outputs the operating status characteristics of the relay protection device, which include the protection device operation stability index, the protection device action reliability index, and the protection device performance degradation index.
[0017] The beneficial effects of this preferred technical solution are: the operating status characteristics of the relay protection device are output through a pre-trained fault feature extraction model, and it clearly includes the protection device operation stability index, the protection device action reliability index, and the protection device performance degradation index. It can comprehensively and meticulously reflect the operating status of the relay protection device, and provide a rich and targeted basis for the subsequent accurate identification of fault types and the construction of fault development trend models.
[0018] As a preferred solution for the relay protection fault prediction and early warning method based on the intelligent oscilloscope, the following are the solutions:
[0019] The identification of the fault type of the relay protection device according to the operating status characteristics and the construction of the fault development trend model include:
[0020] Fault types include malfunction faults, malfunction faults, and refusal to operate faults; the identified fault type is matched with a preset fault feature library to obtain a fault development trend model corresponding to the fault type.
[0021] As a preferred solution for the relay protection fault prediction and early warning method based on the intelligent oscilloscope, the following are the solutions:
[0022] The acquiring of the fault development trend model corresponding to the fault type includes:
[0023] A historical fault feature library is constructed, which contains fault type identification, feature vector set, fault development process record, and key time node mark. The Mahalanobis distance between the operating status characteristics and each fault case in the historical fault feature library is calculated, and several fault cases with the smallest Mahalanobis distance are selected as the candidate set. A temporal weight factor is introduced to perform time-weighted feature similarity on each fault case in the candidate set to obtain the fault type of the relay protection device.
[0024] The beneficial effects of this preferred technical solution include: constructing a historical fault feature library and using a method combining Mahalanobis distance and time-series weighting factors to determine the fault type of the relay protection device. This method fully considers the distribution characteristics of fault features and the impact of fault features at different times on the current fault type judgment. This method can more accurately filter out cases similar to the current fault from historical fault cases, improving the accuracy and reliability of fault type identification.
[0025] As a preferred solution for the relay protection fault prediction and early warning method based on the intelligent oscilloscope, the following are the solutions:
[0026] The identifying of the fault type of the relay protection device according to the operating state characteristics and constructing the fault development trend model further includes:
[0027] According to the failure type, the first fault development trend model is established using Weibull distribution, and the shape parameters and scale parameters of the first fault development trend model are determined by the maximum likelihood estimation method;
[0028] For the type of malfunction, a hidden Markov model including normal state, mild abnormal state, moderate abnormal state, and severe abnormal state is constructed as the second fault development trend model; the state transition probability matrix is estimated based on the eigenvector of real-time monitoring as the observation sequence;
[0029] For the refusal to operate fault type, the fault feature sequence is accumulated and generated, a differential equation model is established as the third fault development trend model, the prediction parameters of the differential equation model are solved, and the time response function is constructed.
[0030] The beneficial effect of this preferred technical solution is that it uses different models to construct fault development trend models for different fault types, namely, a Weibull distribution model for malfunction faults, a hidden Markov model for malfunction faults, and a differential equation model for refusal to operate faults. This personalized modeling approach fully considers the characteristics and development patterns of different fault types, more accurately describing the development trends of various types of faults and providing a more reliable basis for subsequent fault warnings.
[0031] As a preferred solution for the relay protection fault prediction and early warning method based on the intelligent oscilloscope, the following are the solutions:
[0032] The fault early warning of the relay protection device according to the fault development trend model includes:
[0033] Real-time monitoring of the fault development rate of the first fault development trend model, the state transition probability of the second fault development trend model, and the prediction parameters of the third fault development trend model; triggering the first level warning when the fault development rate exceeds the preset rate threshold; triggering the second level warning when the state transition probability exceeds the preset probability threshold; triggering the third level warning when the prediction parameter exceeds the preset parameter threshold; and issuing a linkage warning signal when two or more levels of warnings are triggered at the same time.
[0034] In a second aspect, an embodiment of the present invention provides a relay protection fault prediction and warning system based on an intelligent oscilloscope, comprising:
[0035] A state feature extraction module is used to collect the operating parameters of the relay protection device and establish a multi-dimensional feature vector, input the multi-dimensional feature vector into a pre-trained fault feature extraction model, and extract the operating state characteristics of the relay protection device;
[0036] Fault development trend modeling module, used to identify the fault type of the relay protection device according to the operating status characteristics and build a fault development trend model;
[0037] The fault warning module is used to provide fault warning for the relay protection device based on the fault development trend model.
[0038] In a third aspect, an embodiment of the present invention provides an electronic device, including:
[0039] memory and processor;
[0040] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the one or more programs are executed by the one or more processors, the one or more processors implement the relay protection fault prediction and early warning method based on the intelligent recorder as described in any embodiment of the present invention.
[0041] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the relay protection fault prediction and early warning method based on an intelligent recorder.
[0042] Beneficial effects of the present invention: By analyzing operating parameters and extracting fault characteristics, the present invention can identify potential fault risks and avoid failure, malfunction or refusal of protection devices, thereby improving the safety of power system operation; based on the fault development trend model, the fault evolution law can be predicted, and measures can be taken in advance according to the warning level to prevent the expansion of the fault and reduce power outage losses; the generated fault warning report contains detailed fault information and prevention suggestions, providing maintenance personnel with a basis for decision-making, which helps to improve the efficiency of fault investigation and the pertinence of maintenance work. BRIEF DESCRIPTION OF THE DRAWINGS
[0043] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.
[0044] Figure 1 It is an overall flow chart of the relay protection fault prediction and early warning method based on the intelligent recorder described in the present invention. DETAILED DESCRIPTION
[0045] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.
[0046] Example 1, with reference to Figure 1 , which is the first embodiment of the present invention, provides a relay protection fault prediction and early warning method based on an intelligent oscilloscope, comprising:
[0047] S1: Collect the operating parameters of the relay protection device and establish a multi-dimensional feature vector, input the multi-dimensional feature vector into the pre-trained fault feature extraction model, and extract the operating status characteristics of the relay protection device;
[0048] S2: Identify the fault type of the relay protection device based on the operating status characteristics and build a fault development trend model;
[0049] S3: Provide fault warning for relay protection devices based on the fault development trend model.
[0050] It should be noted that through steps S1-S3, the stable operation of the relay protection device can be effectively guaranteed and the risk of power system failure can be reduced. Through a series of steps such as collecting operating parameters, extracting state characteristics, identifying fault types, building trend models, and performing fault warnings, a complete fault prediction and warning system is formed. In practical applications, potential fault hazards of relay protection devices can be discovered in advance, and appropriate measures can be taken in time to deal with them, avoiding further development and expansion of the fault, thereby improving the reliability and safety of the power system and reducing power outages and economic losses caused by relay protection device failures. At the same time, this method uses pre-trained fault feature extraction models and multiple fault development trend models to provide a scientific basis for accurate fault prediction and warning, and has strong practicality and promotion value.
[0051] Example 2, reference Figure 1 , which is an embodiment of the present invention, provides a relay protection fault prediction and early warning method based on an intelligent oscilloscope based on the previous embodiment, including:
[0052] In this embodiment, in step S1, the operating parameters of the relay protection device are collected and a multidimensional feature vector is established. The multidimensional feature vector is input into a pre-trained fault feature extraction model to extract the operating status features of the relay protection device. The following steps are performed:
[0053] The operating parameters of the relay protection device include protection setting parameters, protection action time parameters, protection start times parameters, protection action times parameters, protection device temperature parameters, and protection device power supply voltage parameters. A multidimensional feature vector is established based on the operating parameters of the relay protection device, and the multidimensional feature vector is input into a pre-trained fault feature extraction model to extract the operating status characteristics of the relay protection device.
[0054] The operating status characteristics include the protection device operation stability index, the protection device action reliability index, and the protection device performance degradation index.
[0055] In another possible embodiment, in the process of collecting the operating parameters of the relay protection device, the operating parameters may include protection setting parameters, protection action time parameters, protection device temperature parameters, protection start number parameters, protection action number parameters, power supply voltage fluctuation rate parameters, internal temperature change trend parameters, device restart number parameters and operating time parameters.
[0056] For example, the following parameters of a relay protection device can be collected: the protection setting is 2.5, the action time is 0.1 second, the temperature is 30 degrees Celsius, the number of starts is 10 times, the number of actions is 5 times, the power supply voltage fluctuation rate is 2%, the internal temperature change trend is rising, the number of device restarts is 1 time, and the operating time is 1000 hours.
[0057] Furthermore, the power supply voltage fluctuation rate data is divided into five sub-intervals within the voltage standard value range of 0.85 to 1.15, and the voltage fluctuation stability of each sub-interval is calculated using the Gaussian membership function; the temperature stability of the internal temperature change trend data is calculated using the trapezoidal membership function within the temperature change rate range of negative ten degrees per hour to positive ten degrees per hour; the number of device restarts is counted within a twenty-four-hour sliding time window, and the restart frequency stability of the device restart number data is calculated using the step membership function.
[0058] In another possible implementation, in the process of establishing a multidimensional feature vector, the collected parameters can be preprocessed: the protection setting parameters are normalized by minimum-maximum standardization, and the parameter values are mapped to between 0 and 1. The protection action time parameters and the protection device temperature parameters of the relay protection device are extracted using a sliding time window method. The statistical features include mean features, standard deviation features, skewness features, and kurtosis features. The protection start number parameters and the protection action number parameters of the relay protection device are processed using an exponential smoothing method to process the time series variation characteristics.
[0059] For example, if the minimum value of a set of protection setting parameters is 2.0 and the maximum value is 3.0, then the setting of 2.5 is normalized to (2.5-2.0) / (3.0-2.0)=0.5. The sliding time window method is used to extract statistical features of the protection action time parameters and the protection device temperature parameters, including mean, standard deviation, skewness and kurtosis. For example, a sliding time window with a length of 10 is selected to calculate the mean, standard deviation, skewness and kurtosis of the past 10 action times. The exponential smoothing method is used to process the time series change characteristics of the protection start number parameters and the protection action number parameters, giving greater weight to recent data.
[0060] Next, a parameter correlation matrix is constructed. This is done using a mutual information-based calculation method to quantify the correlations between different parameters. Principal component analysis is used to perform feature dimensionality reduction on the parameter correlation matrix, resulting in a 16-dimensional feature vector. This 16-dimensional feature vector is then fed into a fault feature extraction model with a three-layer long short-term memory (LSTM) network structure. The model's three-layer LSTM network structure includes 128 neurons, 64 neurons, and 32 neurons, respectively. An attention mechanism is introduced into the model to enhance the ability to extract temporal features. The fault feature extraction model takes a 16-dimensional feature vector as input, and outputs the operating status characteristics of the relay protection device, including indicators of its operational stability, operational reliability, and performance degradation.
[0061] In another possible implementation, after obtaining the operating status characteristics of the relay protection device, an evaluation may be performed;
[0062] Specifically, a fuzzy comprehensive evaluation model with five membership functions is adopted to evaluate the operation stability index according to the power supply voltage fluctuation rate parameter, the internal temperature change trend parameter and the device restart number parameter.
[0063] For example, five membership functions are defined, representing "very stable," "stable," "average," "unstable," and "very unstable," respectively. A dynamic time warping algorithm is used to evaluate the temporal consistency of the protection action time series, and a Bayesian network model is used to evaluate the action reliability index. For example, the current action time series is compared with the historical normal action time series, and their similarity is calculated. Finally, a lifespan prediction model based on the Weibull distribution is combined with the runtime parameters and the number of actions parameters, and a support vector regression algorithm is used to predict performance degradation indicators. For example, based on the known runtime and number of actions, the degree of performance degradation over a period of time in the future is predicted.
[0064] Furthermore, a hierarchical analysis method was used to construct a judgment matrix. The weight values of the power supply voltage fluctuation rate data, the internal temperature change trend data, and the device restart frequency data were determined based on the judgment matrix. A weight vector was constructed based on the weight values. A fuzzy relationship matrix was constructed based on the voltage fluctuation stability, temperature stability, and restart frequency stability. The weight vector and the fuzzy relationship matrix were fuzzy-comprehensively calculated to obtain the operation stability index.
[0065] The standard action timing curve of the relay protection device is used as the reference sequence, and the actual action timing curve of the relay protection device is used as the sequence to be evaluated. The dynamic time warping algorithm is used to construct the cumulative distance matrix. The minimum cumulative distance between the reference sequence and the sequence to be evaluated is calculated based on the dynamic programming method. The timing consistency score is obtained by backtracking the optimal path.
[0066] A Bayesian network was constructed, which included start-action transition nodes, misaction frequency nodes, timing consistency nodes, and reliability assessment nodes. The network parameters of the Bayesian network were determined using the maximum likelihood estimation method, and the action reliability index was calculated through probabilistic reasoning.
[0067] A two-dimensional Weibull distribution model is established based on the operating time data and cumulative operation number data of the relay protection device, and the distribution parameters of the two-dimensional Weibull distribution model are determined by the maximum likelihood estimation method. The normalized operating time, normalized operation number, Weibull distribution probability value and key performance parameter sequence are used as input features to construct a support vector regression model with a radial basis kernel function. The penalty factor and kernel parameters of the support vector regression model are determined through cross-validation, and the performance degradation index of the relay protection device is output.
[0068] In this embodiment, in step S2, the fault type of the relay protection device is identified according to the operating status characteristics, and the fault development trend model is constructed, which includes:
[0069] Fault types include malfunction faults, malfunction faults, and refusal to operate faults; the identified fault type is matched with a preset fault feature library to obtain a fault development trend model corresponding to the fault type;
[0070] Specifically, a historical fault feature library is constructed, which contains fault type identification, feature vector set, fault development process record, and key time node markers. The Mahalanobis distance between the operating status characteristics and each fault case in the historical fault feature library is calculated, and several fault cases with the smallest Mahalanobis distance are selected as the candidate set. A temporal weight factor is introduced to perform time-weighted feature similarity on each fault case in the candidate set to obtain the fault type of the relay protection device.
[0071] According to the failure type, the first fault development trend model is established using Weibull distribution, and the shape parameters and scale parameters of the first fault development trend model are determined by the maximum likelihood estimation method;
[0072] Specifically, the Weibull distribution is used to establish the first fault development trend model, and the formula is:
[0073]
[0074] Where F(t) is the cumulative probability of failure, β is the shape parameter, and η is the scale parameter. These two parameters are determined from historical data using the maximum likelihood estimation method.
[0075] By taking the derivative of the cumulative distribution function F(t) of the Weibull distribution, we can obtain the fault development rate function:
[0076]
[0077] During the operation of the relay protection device, the current time t is continuously obtained and substituted into the fault development rate function f(t) to calculate the current fault development rate.
[0078] For malfunction fault types, a hidden Markov model including normal state, mild abnormal state, moderate abnormal state, and severe abnormal state is constructed as the second fault development trend model. The Baum-Welch algorithm is used to estimate the state transition probability matrix based on the eigenvector of real-time monitoring as the observation sequence.
[0079] Specifically, the state space should be clarified, including normal state (various indicators of the protection device fluctuate within the allowable range), slightly abnormal state (individual monitoring parameters exceed the limit slightly), moderate abnormal state (multiple parameters deviate significantly at the same time), and seriously abnormal state (the protection device is in a critical state of malfunction).
[0080] The protection starting current-voltage ratio, measuring element phase angle deviation, protection starting time offset, tripping delay change rate, and logic criterion integrity index of the relay protection device are obtained, and a feature vector is constructed. The feature vector is normalized using the feature mean and feature standard deviation to obtain the observation sequence.
[0081] An initial state transition probability matrix is constructed based on expert experience, which is used to characterize the transition probability of the relay protection device between different states.
[0082] Based on the eigenvector of real-time monitoring as the observation sequence, the Baum-Welch algorithm is used to calculate the forward probability and backward probability corresponding to the observation sequence. The state transition statistics are calculated based on the forward probability and backward probability. The state transition statistics are used to iteratively optimize the initial state transition probability matrix to obtain the optimal state transition probability matrix.
[0083] During the operation of the device, new observation sequences are continuously acquired, and the Viterbi algorithm is used to calculate the most likely state sequence based on the observation sequences acquired in real time, thereby obtaining the transition probabilities between different states.
[0084] For the refusal to operate fault type, the fault feature sequence is accumulated and generated, a differential equation model is established as the third fault development trend model, the prediction parameters of the differential equation model are solved, and the time response function is constructed;
[0085] Specifically, for the refusal to operate fault type, the fault feature sequence is accumulated and generated, and a differential equation model is established as the third fault development trend model.
[0086] The protection start sensitivity coefficient, measurement loop impedance value, protection execution delay, and logic criterion response time of the relay protection device are obtained, and these data are sampled at equal intervals to obtain the original time series.
[0087] The original time series is accumulated to generate a cumulative series. The value at any moment in the cumulative series is the cumulative sum of the values at that moment and all previous moments in the original time series.
[0088] A differential equation model is established based on the cumulative sequence. The model includes development coefficients and grey action. The least squares estimation equation is used to solve the development coefficients and grey action, and the time response function of the differential equation model is obtained.
[0089] During the operation of the device, new fault feature data is continuously acquired, the original time series is updated, the cumulative series is recalculated, and the current prediction parameters are obtained based on the differential equation model and time response function.
[0090] The first fault development trend model, the second fault development trend model and the third fault development trend model jointly reflect the fault evolution law of the relay protection device.
[0091] In this embodiment, the above-mentioned step S3 performs fault early warning on the relay protection device according to the fault development trend model, including:
[0092] Real-time monitoring of the fault development rate of the first fault development trend model, the state transition probability of the second fault development trend model, and the prediction parameters of the third fault development trend model; triggering the first level warning when the fault development rate exceeds the preset rate threshold; triggering the second level warning when the state transition probability exceeds the preset probability threshold; triggering the third level warning when the prediction parameter exceeds the preset parameter threshold; and issuing a linkage warning signal when two or more levels of warnings are triggered at the same time.
[0093] When the early warning is triggered, a recording trigger instruction is sent to the intelligent recorder, and the intelligent recorder collects and stores the operating waveform data of the relay protection device according to the recording trigger instruction;
[0094] The fault warning results of the relay protection device are verified based on the operating waveform data, and a fault warning report is generated. The fault warning report includes a description of the fault type, fault warning level, fault development trend analysis, and fault prevention suggestions.
[0095] Example 3. The above is a schematic scheme of the relay protection fault prediction and warning method based on an intelligent oscilloscope recorder of this embodiment. It should be noted that the technical scheme of the relay protection fault prediction and warning system based on an intelligent oscilloscope recorder and the technical scheme of the relay protection fault prediction and warning method based on an intelligent oscilloscope recorder are of the same concept. For details not described in detail in the technical scheme of the relay protection fault prediction and warning system based on an intelligent oscilloscope recorder in this embodiment, please refer to the description of the technical scheme of the relay protection fault prediction and warning method based on an intelligent oscilloscope recorder.
[0096] This embodiment also provides a relay protection fault prediction and warning system based on an intelligent oscilloscope, comprising:
[0097] A state feature extraction module is used to collect the operating parameters of the relay protection device and establish a multi-dimensional feature vector, input the multi-dimensional feature vector into a pre-trained fault feature extraction model, and extract the operating state characteristics of the relay protection device;
[0098] Fault development trend modeling module, used to identify the fault type of the relay protection device according to the operating status characteristics and build a fault development trend model;
[0099] The fault warning module is used to provide fault warning for the relay protection device based on the fault development trend model.
[0100] This embodiment further provides an electronic device applicable to a relay protection fault prediction and early warning method based on an intelligent oscilloscope recorder, including:
[0101] Memory and processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the relay protection fault prediction and early warning method based on the intelligent recorder as proposed in the above embodiment.
[0102] This embodiment further provides a storage medium storing a computer program, which, when executed by a processor, implements the relay protection fault prediction and early warning method based on an intelligent oscilloscope as proposed in the above embodiment.
[0103] The storage medium proposed in this embodiment and the relay protection fault prediction and early warning method based on the intelligent recorder proposed in the above embodiment belong to the same inventive concept. The technical details not described in detail in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.
[0104] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.
Claims
1. A relay protection fault prediction and early warning method based on an intelligent oscilloscope, characterized in that: include: Collecting the operating parameters of the relay protection device and establishing a multidimensional feature vector, inputting the multidimensional feature vector into a pre-trained fault feature extraction model to extract the operating status characteristics of the relay protection device; Identify the fault type of the relay protection device based on the operating status characteristics and build a fault development trend model; According to the fault development trend model, fault early warning is issued for the relay protection device.
2. The relay protection fault prediction and early warning method based on the intelligent oscilloscope according to claim 1, characterized in that: The collecting of operating parameters of the relay protection device and establishing a multidimensional feature vector, inputting the multidimensional feature vector into a pre-trained fault feature extraction model, and extracting the operating status features of the relay protection device include: Various operating parameters of the relay protection device are collected and preprocessed, and a multidimensional feature vector is established based on the preprocessed parameters. A parameter correlation matrix is constructed based on the multidimensional feature vector to quantify the correlation between different parameters, and the parameter correlation matrix is subjected to feature dimensionality reduction to obtain the reduced eigenvector.
3. The relay protection fault prediction and early warning method based on the intelligent oscilloscope according to claim 2, characterized in that: The collecting of operating parameters of the relay protection device and establishing a multidimensional feature vector, inputting the multidimensional feature vector into a pre-trained fault feature extraction model, and extracting the operating status features of the relay protection device further includes: The reduced-dimensional feature vector is input into a pre-trained fault feature extraction model, and the model outputs the operating status characteristics of the relay protection device, which include the protection device operation stability index, the protection device action reliability index, and the protection device performance degradation index.
4. The relay protection fault prediction and early warning method based on the intelligent oscilloscope according to claim 3 is characterized in that: The identification of the fault type of the relay protection device according to the operating status characteristics and the construction of the fault development trend model include: Fault types include malfunction faults, malfunction faults, and refusal to operate faults; the identified fault type is matched with a preset fault feature library to obtain a fault development trend model corresponding to the fault type.
5. The relay protection fault prediction and early warning method based on the intelligent oscilloscope according to claim 4 is characterized in that: The acquiring of the fault development trend model corresponding to the fault type includes: A historical fault feature library is constructed, which contains fault type identification, feature vector set, fault development process record, and key time node mark. The Mahalanobis distance between the operating status characteristics and each fault case in the historical fault feature library is calculated, and several fault cases with the smallest Mahalanobis distance are selected as the candidate set. A temporal weight factor is introduced to perform time-weighted feature similarity on each fault case in the candidate set to obtain the fault type of the relay protection device.
6. The relay protection fault prediction and early warning method based on the intelligent oscilloscope according to claim 5, characterized in that: The identifying of the fault type of the relay protection device according to the operating state characteristics and constructing the fault development trend model further includes: According to the failure type, the first fault development trend model is established using Weibull distribution, and the shape parameters and scale parameters of the first fault development trend model are determined by the maximum likelihood estimation method. For the type of malfunction, a hidden Markov model including normal state, mild abnormal state, moderate abnormal state, and severe abnormal state is constructed as the second fault development trend model; the state transition probability matrix is estimated based on the eigenvector of real-time monitoring as the observation sequence; For the refusal to operate fault type, the fault feature sequence is accumulated and generated, a differential equation model is established as the third fault development trend model, the prediction parameters of the differential equation model are solved, and the time response function is constructed.
7. The relay protection fault prediction and early warning method based on the intelligent oscilloscope according to claim 6, characterized in that: The fault early warning of the relay protection device according to the fault development trend model includes: Real-time monitoring of the fault development rate of the first fault development trend model, the state transition probability of the second fault development trend model, and the prediction parameters of the third fault development trend model; triggering the first level warning when the fault development rate exceeds the preset rate threshold; triggering the second level warning when the state transition probability exceeds the preset probability threshold; triggering the third level warning when the prediction parameter exceeds the preset parameter threshold; and issuing a linkage warning signal when two or more levels of warnings are triggered at the same time.
8. A relay protection fault prediction and warning system based on an intelligent oscilloscope, applying the method according to any one of claims 1 to 7, characterized in that: include: A state feature extraction module is used to collect the operating parameters of the relay protection device and establish a multi-dimensional feature vector, input the multi-dimensional feature vector into a pre-trained fault feature extraction model, and extract the operating state characteristics of the relay protection device; Fault development trend modeling module, used to identify the fault type of the relay protection device according to the operating status characteristics and build a fault development trend model; The fault warning module is used to provide fault warning for the relay protection device based on the fault development trend model.
9. An electronic device comprising: memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the method according to any one of claims 1 to 7 are implemented.
10. A computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions, when executed by a processor, implement the steps of the method according to any one of claims 1 to 7.