CN117949887A - Fault detection and prevention method for electric energy metering system of energy controller - Google Patents
Fault detection and prevention method for electric energy metering system of energy controller Download PDFInfo
- Publication number
- CN117949887A CN117949887A CN202410087542.0A CN202410087542A CN117949887A CN 117949887 A CN117949887 A CN 117949887A CN 202410087542 A CN202410087542 A CN 202410087542A CN 117949887 A CN117949887 A CN 117949887A
- Authority
- CN
- China
- Prior art keywords
- fault
- data
- metering system
- electric energy
- energy metering
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 39
- 238000000034 method Methods 0.000 title claims abstract description 30
- 230000002265 prevention Effects 0.000 title claims abstract description 10
- 238000012423 maintenance Methods 0.000 claims abstract description 36
- 238000012544 monitoring process Methods 0.000 claims abstract description 27
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 18
- 230000007246 mechanism Effects 0.000 claims abstract description 18
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 16
- 238000007405 data analysis Methods 0.000 claims abstract description 16
- 238000005516 engineering process Methods 0.000 claims abstract description 14
- 230000008439 repair process Effects 0.000 claims abstract description 14
- 230000004044 response Effects 0.000 claims abstract description 5
- 238000004458 analytical method Methods 0.000 claims description 30
- 238000011176 pooling Methods 0.000 claims description 25
- 238000013528 artificial neural network Methods 0.000 claims description 23
- 230000015654 memory Effects 0.000 claims description 17
- 230000006870 function Effects 0.000 claims description 16
- 238000012706 support-vector machine Methods 0.000 claims description 15
- 238000012545 processing Methods 0.000 claims description 13
- 210000004027 cell Anatomy 0.000 claims description 12
- 239000011159 matrix material Substances 0.000 claims description 12
- 238000007781 pre-processing Methods 0.000 claims description 9
- 238000012360 testing method Methods 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 9
- 238000011161 development Methods 0.000 claims description 8
- 230000002159 abnormal effect Effects 0.000 claims description 7
- 238000003066 decision tree Methods 0.000 claims description 7
- 230000004913 activation Effects 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 6
- 238000000513 principal component analysis Methods 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 6
- 238000007689 inspection Methods 0.000 claims description 5
- 238000005065 mining Methods 0.000 claims description 4
- 238000013021 overheating Methods 0.000 claims description 4
- 238000003909 pattern recognition Methods 0.000 claims description 4
- 230000000737 periodic effect Effects 0.000 claims description 4
- 238000012351 Integrated analysis Methods 0.000 claims description 3
- 238000011263 association rule mining technique Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000008859 change Effects 0.000 claims description 3
- 238000007635 classification algorithm Methods 0.000 claims description 3
- 238000010586 diagram Methods 0.000 claims description 3
- 230000002708 enhancing effect Effects 0.000 claims description 3
- 238000011156 evaluation Methods 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims description 3
- 210000002569 neuron Anatomy 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 230000018109 developmental process Effects 0.000 description 6
- 230000007423 decrease Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000001052 transient effect Effects 0.000 description 3
- 238000005299 abrasion Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 230000003449 preventive effect Effects 0.000 description 2
- 230000032683 aging Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000006378 damage Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000009413 insulation Methods 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 230000009528 severe injury Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
Landscapes
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention relates to the technical field of power system monitoring, in particular to a fault detection and prevention method of an electric energy metering system of an energy controller, which comprises the following steps: s1: monitoring the running state of the electric energy metering system in real time; s2: identifying potential failure modes and trends; s3: applying convolutional neural networks to precisely locate the specific component or region where the fault occurred; s4: automatically classifying the detected faults and analyzing the reasons of the faults; s5: establishing a fault early warning mechanism; s6: formulating and implementing a targeted maintenance and repair strategy; s7: and periodically collecting fault data, and continuously optimizing a fault detection algorithm and an early warning mechanism based on the collected fault data. According to the invention, by implementing advanced data analysis and fault prediction technology, the fault detection accuracy and response speed of the electric energy metering system are obviously improved, so that the maintenance strategy is optimized, the operation and maintenance cost is reduced, and the safety and stability of the whole energy controller are finally enhanced.
Description
Technical Field
The invention relates to the technical field of power system monitoring, in particular to a fault detection and prevention method for an electric energy metering system of an energy controller.
Background
The energy controller (Energy control and monitoring terminal unit, ECU) is arranged in a public transformer or private transformer area, can realize flexible access of metering and sensing equipment at a client side and a power distribution side, and has the functions of data acquisition, intelligent fee control, clock synchronization, accurate metering, ordered charging, energy management, loop state inspection, user transformer relation identification, power failure event reporting and the like.
The energy controller adopts a modularized design, and in different application scenes, the requirements on various input/output interfaces are different, so that various types of functional modules are generated, and the energy controller redefines the terminal form through the cooperation of the different types of functional modules. The stability and the reliability of the electric energy metering system are critical to the safe operation of the energy controller, and the system is not only responsible for accurately measuring the electric energy consumption, but also relates to the load management and fault detection of the energy controller.
The current electric energy metering system has a plurality of challenges in fault detection and prevention, firstly, the traditional fault detection methods often rely on manual inspection or a simple automatic monitoring system, the methods have limited effects in early fault identification and accurate positioning, and in addition, due to the lack of an efficient data analysis and fault prediction mechanism, system maintenance is often passive and lagged, and faults cannot be effectively prevented, so that the operation and maintenance cost is increased, and the stable operation and the power supply reliability of an energy controller are also influenced.
Therefore, there is a great need for an efficient method for monitoring the operation state of an electric energy metering system in real time, accurately predicting and timely responding to faults.
Disclosure of Invention
Based on the above purpose, the invention provides a fault detection and prevention method for an electric energy metering system of an energy controller.
A fault detection and prevention method for an electric energy metering system of an energy controller comprises the following steps:
s1: presetting a fault detection device with a plurality of sensors, wherein the fault detection device is used for monitoring the running state of an electric energy metering system in real time;
S2: based on a data analysis and fault prediction algorithm of a neural network and a support vector machine, processing data from the S1, and identifying potential fault modes and trends;
S3: performing real-time data monitoring, and accurately positioning specific components or areas where faults occur by using a convolutional neural network;
s4: automatically classifying the detected faults, analyzing the reasons of the faults, and predicting the fault development trend by utilizing historical data and pattern recognition technology;
S5: based on the analysis results of the S2 and the S4, a fault early warning mechanism is established and used for timely notifying a maintenance team of an impending fault;
S6: according to the fault detection and analysis results, a targeted maintenance and repair strategy is formulated and implemented;
s7: and periodically collecting fault data, and continuously optimizing a fault detection algorithm and an early warning mechanism based on the collected fault data.
Further, the fault detection device in S1 includes a current sensor, a voltage sensor, a temperature sensor, and a data preprocessing unit, which is configured to comprehensively monitor an operation parameter of the electric energy metering system; wherein,
A current sensor: the current sensor is arranged on a main power supply line of the electric energy metering system and used for monitoring the current quantity passing through the electric wire in real time, and can detect the magnitude and fluctuation of the current so as to identify the fault condition of overload or open circuit;
a voltage sensor: the voltage sensor is arranged at the input end and the output end of the system and is used for measuring the voltage level of the electric energy metering system, the voltage sensor can monitor the stability and deviation of the voltage, and abnormal conditions of unstable or over-high and over-low voltage can be found;
Temperature sensor: the key components fixed on the electric energy metering system comprise a transformer and a cable joint part, are used for monitoring the operation temperature of the corresponding components, and can timely find out the overheating problem caused by overload or short circuit fault by monitoring the change of the temperature;
A data preprocessing unit: the data preprocessing unit is used for receiving data from all sensors and performing preliminary processing on the received data, and particularly adopts real-time data analysis technology including signal filtering, normalization processing and preliminary trend analysis to perform preliminary evaluation on parameters of current, voltage and temperature so as to ensure the quality of the provided data.
Further, the step S2 specifically includes:
S21: firstly, receiving data of current I, voltage V and temperature T processed in the step S1, and presetting an input layer of a neural network to receive n characteristics (F 1,F2, …, fn), wherein each characteristic is extracted from original data, and a specific characteristic extraction formula is as follows: f i＝fi (I, V, T), where F i is a feature extraction function for extracting a feature F i from current I, voltage V, and temperature T;
s22: the neural network further processes the characteristics through a hidden layer H, and the hidden layer is provided with m nodes, each node applies an activation function a to process input, and a specific hidden layer processing formula is as follows:
Where w ij is the weight, b j is the bias term, H j is the output of the j-th hidden layer node;
S23: the output layer O gives out initial judgment of fault prediction according to the output of the hidden layer, and the output layer judgment formula is as follows:
Wherein w j 'and b' are the weight and bias terms of the output layer, respectively;
s24: the support vector machine is used for classifying the characteristics extracted by the neural network to judge whether the system has faults, and particularly, the support vector machine is used for constructing a classification hyperplane for separating data in a normal state and data in an abnormal state, and the classification formula of the support vector machine is as follows:
Where α k is the Lagrangian multiplier, y k is the label of the training sample, K is the kernel function, x k and x are the eigenvectors of the training sample and the new sample, respectively, b "is the bias term, and y is the classification result.
Further, the step S3 specifically includes:
s31: firstly, monitoring current, voltage and temperature data of an electric energy metering system in real time, wherein the data are used as input of a convolutional neural network and used for fault location analysis;
s32: constructing a convolutional neural network model, wherein the convolutional neural network model consists of a plurality of convolutional layers, a pooling layer and a full-connection layer, and the specific function of each layer is as follows:
The convolution layer is used for extracting the characteristics of input data, specifically, if the convolution kernel of the convolution layer is K and the size is m multiplied by n, the convolution operation is expressed as:
Wherein, C ij is an element in the output feature diagram of the convolution layer, K ab is an element with a position (a, b) in the convolution kernel, D i+a,j+b is an element multiplied by a position corresponding to the convolution kernel in the input data matrix, m×n is a size of the convolution kernel, a, b is an index in the convolution kernel;
The pooling layer is used for reducing feature dimension and enhancing robustness of the features, and the pooling operation maximum pooling is expressed as: p ij＝max(Dkl), for k, l e window (i, j), where P ij is an element in the pooling layer output feature map, D kl is an element in the convolved feature map, max (D kl) is a pooling operation that selects the maximum value within a particular window, window (i, j) is a particular window region considered in the pooling operation, centered at (i, j);
The full-connection layer is used for connecting the output of the pooling layer to the full-connection layer for final classification after flattening;
s33: specific failure modes will be identified by training convolutional neural network models, each mode corresponding to a failure of a specific component or region, and each neuron of the output layer corresponding to a potential failure location in the system.
Further, the step S4 specifically includes:
S41: performing fault classification, automatically classifying the detected fault conditions into different categories including current faults, voltage faults or temperature faults according to the data collected from the steps S1 and S2 by using a decision tree classification algorithm, wherein each internal node represents a test of an attribute, each branch represents a result of the test, and each leaf node represents a fault category, and the decision tree rule is as follows:
If (A 1≤v1)∧(A2>v2)∧…∧(An＝vn) THENCLASSC, wherein A i is an attribute, v i is a test value of the attribute, and C is a fault class;
s42: analyzing potential reasons of various faults by adopting an association rule mining analysis technology, wherein the association rule mining can reveal the reasons behind different fault categories, and the association rule formula is as follows: rule: y, withsupportsandconfidencec, wherein X and Y are a collection of data items, the support s representing the frequency of simultaneous occurrence of X and Y, the confidence c representing the conditional probability of Y occurring in the case of X occurring;
S43: analyzing the historical fault data by using a long-short-period memory network to predict the fault development trend, wherein the specific calculation formula of the long-short-period memory network is as follows:
ft＝σ(Wf·[ht-1,xt]+bf)，
ot＝σ(Wo·[ht-1,xt]+bo)，
h t＝ot*tanh(Ct), where f t is a forget gate, controlling how much previous memory is reserved, i t is an input gate, controlling the impact of the current input, Is a candidate memory cell, C t is an updated memory cell, o t is an output gate, h t is the output at the current time, W and b are weights and bias parameters, σ is sigmoidi activation function, which represents the inter-element multiplication.
Further, the step S5 specifically includes:
S51: firstly, integrating a fault prediction result obtained in the step S2 by using a neural network and a support vector machine with a fault classification, a cause analysis and a trend prediction result obtained in the step S4;
S52: based on the integrated analysis result, setting a threshold value of the fault early warning, wherein the threshold value is determined according to historical data and system operation standards and is used for judging when to generate the fault early warning;
S53: when the analysis result shows that a certain parameter is about to exceed the normal operation range or a specific type of fault is predicted to possibly occur, an early warning signal is automatically generated;
S55: an automatic notification mechanism is established, and once an early warning signal is generated, early warning information is immediately sent to an operation and maintenance team in the mode of electronic mail, short messages or internal notification, wherein the early warning information comprises the prediction type, possible reasons, predicted occurrence time and suggested response measures of the fault.
Further, the step S6 specifically includes:
S61: firstly, comprehensively considering a fault prediction result of S2 and a fault classification and reason analysis result of S4 to determine the type, reason and severity of the fault;
s62: according to the result of fault analysis, a specific maintenance and repair strategy is formulated, and the specific steps are as follows:
for predicted minor faults, periodic checks and minor adjustments are scheduled;
For detected moderate failures, a more detailed inspection and partial component replacement is arranged;
For severe fault early warning, immediately taking emergency maintenance measures, including shutdown maintenance and quick replacement of key components;
S63: and according to the formulated strategy, arranging resources and personnel to carry out maintenance and repair work, and ensuring that all operations accord with the safety standard.
Further, the step S7 specifically includes:
s71: collecting monthly fault data from the power metering system including current, voltage, temperature and detailed fault records detailing time, duration, type of fault and affected system scope of the fault;
S72: analyzing the data collected monthly by applying a data analysis technology of principal component analysis to identify potential failure modes and trends, wherein the principal component analysis formula is as follows: z=xw, where X is an original data matrix, W is a principal component weight matrix extracted from data, and Z is a converted data matrix;
S73: according to the analysis result, the fault detection algorithm is adjusted and optimized, and the weight and bias of the neural network are specifically adjusted, wherein the formula for adjusting the weight of the neural network is as follows:
w new＝Wold +Δw, where W new is the adjusted weight, W old is the original weight, and Δw is the weight adjustment amount obtained based on fault data analysis;
S74: and updating parameters and threshold values of the early warning mechanism by using a newly optimized fault detection algorithm.
The invention has the beneficial effects that:
The invention obviously improves the accuracy and timeliness of fault detection in the electric energy metering system by introducing an advanced data analysis technology, and the method can monitor and analyze the running state of the system in real time and timely identify potential fault modes and trends.
According to the method, the faults can be accurately detected, the targeted maintenance and repair strategies can be formulated according to the types and the severity of the faults, and specific guidance can be provided for operation and maintenance teams through deep analysis of fault causes and comprehensive utilization of historical data, so that the operation and maintenance teams can be helped to plan maintenance work more effectively and respond to the faults rapidly, and the downtime and the maintenance cost of the system are greatly reduced.
The invention can obviously improve the safety and stability of the whole energy controller by improving the fault detection capability and response efficiency of the electric energy metering system, thereby being beneficial to guaranteeing the continuity and reliability of power supply, promoting the sustainable development of the power industry and providing solid basic support for social and economic activities.
Drawings
In order to more clearly illustrate the invention or the technical solutions of the prior art, the drawings which are used in the description of the embodiments or the prior art will be briefly described, it being obvious that the drawings in the description below are only of the invention and that other drawings can be obtained from them without inventive effort for a person skilled in the art.
Fig. 1 is a schematic flow chart of a fault detection and prevention method of an electric energy metering system of an energy controller according to an embodiment of the invention.
Detailed Description
The present invention will be further described in detail with reference to specific embodiments in order to make the objects, technical solutions and advantages of the present invention more apparent.
It is to be noted that unless otherwise defined, technical or scientific terms used herein should be taken in a general sense as understood by one of ordinary skill in the art to which the present invention belongs. The terms "first," "second," and the like, as used herein, do not denote any order, quantity, or importance, but rather are used to distinguish one element from another. The word "comprising" or "comprises", and the like, means that elements or items preceding the word are included in the element or item listed after the word and equivalents thereof, but does not exclude other elements or items. The terms "connected" or "connected," and the like, are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "upper", "lower", "left", "right", etc. are used merely to indicate relative positional relationships, which may also be changed when the absolute position of the object to be described is changed.
As shown in fig. 1, a fault detection and prevention method for an electric energy metering system of an energy controller includes the following steps:
s1: presetting a fault detection device with a plurality of sensors, wherein the fault detection device is used for monitoring the running state of an electric energy metering system in real time;
S2: based on a data analysis and fault prediction algorithm of a neural network and a support vector machine, processing data from the S1, and identifying potential fault modes and trends;
S3: performing real-time data monitoring, and accurately positioning specific components or areas where faults occur by using a convolutional neural network;
s4: automatically classifying the detected faults, analyzing the reasons of the faults, and predicting the fault development trend by utilizing historical data and pattern recognition technology;
s5: based on the analysis results of the S2 and the S4, a fault early warning mechanism is established and used for timely notifying a maintenance team of an impending fault so as to take preventive measures;
S6: according to the fault detection and analysis results, a targeted maintenance and repair strategy is formulated and implemented, so that the electric energy metering system is ensured to resume normal operation in the shortest time;
S7: and periodically collecting fault data, and continuously optimizing a fault detection algorithm and an early warning mechanism based on the collected fault data so as to improve the stability and reliability of the system.
The fault detection device in the S1 comprises a current sensor, a voltage sensor, a temperature sensor and a data preprocessing unit, and is used for comprehensively monitoring the operation parameters of the electric energy metering system; wherein,
A current sensor: the current sensor is arranged on a main power supply line of the electric energy metering system and used for monitoring the current quantity passing through the electric wire in real time, and can detect the magnitude and fluctuation of the current so as to identify the fault condition of overload or open circuit;
a voltage sensor: the voltage sensor is arranged at the input end and the output end of the system and is used for measuring the voltage level of the electric energy metering system, the voltage sensor can monitor the stability and deviation of the voltage, and abnormal conditions of unstable or over-high and over-low voltage can be found;
Temperature sensor: the key components fixed on the electric energy metering system comprise a transformer and a cable joint part, are used for monitoring the operation temperature of the corresponding components, and can timely find out the overheating problem caused by overload or short circuit fault by monitoring the change of the temperature;
A data preprocessing unit: the data preprocessing unit is used for receiving data from all sensors and carrying out preliminary processing on the received data, and particularly adopts real-time data analysis technology comprising signal filtering, normalization processing and preliminary trend analysis, and carries out preliminary evaluation on parameters of current, voltage and temperature so as to ensure the quality of the provided data and provide reliable and clear basic data for subsequent more complex data analysis and fault prediction.
S2 specifically comprises:
S21: firstly, receiving data of current I, voltage V and temperature T processed in the step S1, and presetting an input layer of a neural network to receive n characteristics (F 1,F2, …, fn), wherein each characteristic is extracted from original data, and a specific characteristic extraction formula is as follows: f i＝fi (I, V, T), where F i is a feature extraction function for extracting a feature F i from current I, voltage V, and temperature T;
s22: the neural network further processes the characteristics through a hidden layer H, and the hidden layer is provided with m nodes, each node applies an activation function a to process input, and a specific hidden layer processing formula is as follows:
Where w ij is the weight, b j is the bias term, H j is the output of the j-th hidden layer node;
S23: the output layer O gives out initial judgment of fault prediction according to the output of the hidden layer, and the output layer judgment formula is as follows:
Wherein w j 'and b' are the weight and bias terms of the output layer, respectively;
s24: the support vector machine is used for classifying the characteristics extracted by the neural network to judge whether the system has faults, and particularly, the support vector machine is used for constructing a classification hyperplane for separating data in a normal state and data in an abnormal state, and the classification formula of the support vector machine is as follows:
Wherein α k is the Lagrangian multiplier, y k is the label of the training sample, K is the kernel function, x k and x are the eigenvectors of the training sample and the new sample, respectively, b "is the bias term, and y is the classification result;
By combining the pattern recognition capability of the neural network and the classification accuracy of the SVM, the potential faults of the electric energy metering system can be effectively recognized and predicted from a large amount of data, the neural network is responsible for extracting key features from complex data, and the SVM is responsible for accurately classifying the features into normal or abnormal states, so that the potential faults can be effectively recognized.
S3 specifically comprises:
s31: firstly, monitoring current, voltage and temperature data of an electric energy metering system in real time, wherein the data are used as input of a convolutional neural network and used for fault location analysis;
s32: constructing a convolutional neural network model, wherein the convolutional neural network model consists of a plurality of convolutional layers, a pooling layer and a full-connection layer, and the specific function of each layer is as follows:
The convolution layer is used for extracting the characteristics of input data, specifically, if the convolution kernel of the convolution layer is K and the size is m multiplied by n, the convolution operation is expressed as:
Wherein, C ij is an element in the output feature diagram of the convolution layer, K ab is an element with a position (a, b) in the convolution kernel, D i+a,j+b is an element multiplied by a position corresponding to the convolution kernel in the input data matrix, m×n is a size of the convolution kernel, a, b is an index in the convolution kernel;
the pooling layer is used for reducing feature dimension and enhancing robustness of the features, and the pooling operation maximum pooling is expressed as: p ij＝max(Dkl), for k, l e window (i, j), where P ij is an element in the pooling layer output feature map, D kl is an element in the convolved feature map, max (D kl) is a pooling operation (max pooling) that selects the maximum value within a specific window, window (i, j) is a specific window area considered in the pooling operation, and (i, j) is centered;
The full-connection layer is used for connecting the output of the pooling layer to the full-connection layer for final classification after flattening;
S33: specific fault modes are identified by training convolutional neural network models, each mode corresponds to the fault of a specific component or region, each neuron of the output layer corresponds to a potential fault position in the system, and the models predict the most probable position of the fault according to input data;
By the method, the step S3 can accurately position the fault position in the electric energy metering system by utilizing the real-time monitoring data and the strong analysis capability of the convolutional neural network, and the deep learning capability of the CNN is particularly suitable for processing and analyzing complex sensor data, so that the accurate prediction of the fault position is realized.
S4 specifically comprises the following steps:
S41: performing fault classification, automatically classifying the detected fault conditions into different categories including current faults, voltage faults or temperature faults by using a decision tree classification algorithm according to the data collected from the steps S1 and S2, wherein each internal node represents a test of an attribute, each branch represents a result of the test, and each leaf node represents a fault category, and the decision tree rule is as follows:
If (A 1≤v1)∧(A2>v2)∧…∧(An＝vn) THENCLASSC, wherein A i is an attribute, v i is a test value of the attribute, and C is a fault class;
s42: and the potential reasons of various faults are analyzed by adopting an association rule mining analysis technology, the association rule mining can reveal the reasons behind different fault categories, such as equipment aging, overload operation or external interference, and the like, and an association rule formula is as follows: rule: withsupportsandconfidencec, where X and Y are a collection of data items, the support s represents the frequency of simultaneous occurrence of X and Y, and the confidence c represents the conditional probability of Y occurring in the case where X occurs;
S43: the long-period memory network is used for analyzing the historical fault data so as to predict the fault development trend, and the specific calculation formula of the long-period memory network is as follows:
ft＝σ(Wf·[ht-1,xt]+bf)，
ot＝σ(Wo·[ht-1,xt]+bo)，
h t＝ot*tanh(Ct), where f t is a forget gate, controlling how much previous memory is reserved, i t is an input gate, controlling the impact of the current input, Is a candidate memory cell, C t is an updated memory cell, o t is an output gate, h t is the output at the current time, W and b are weights and bias parameters, σ is sigmoidi activation function, representing inter-element multiplication;
In a long and short term memory network, each unit includes a forget gate (f t), an input gate (i t), an output gate (o t) and a memory unit (C t), which work together to enable LSTM to retain information over long time intervals while avoiding long term dependency problems; forget door (f t): deciding which information should be discarded from the state of the cell; for example, if the history data shows that some type of fault has not occurred, the LSTM may "forget" this information; input gate (i t) and candidate memory cell Deciding which new information is stored in the state of the cell; for example, a new fault type or a changed fault pattern will be added here; output gate (o t) and cell state (h t): determining what the output of the next step is; the output is based on the current state of the cell, and information learned from the input and the previous cell state, and the LSTM output can be interpreted as a prediction of future failure trends; for example: if LSTM predicts an increase in the frequency of a particular type of failure, this may mean that a certain component in the system begins to age or a new problem arises; if the LSTM predicts an increase in fault severity, this may indicate an increasingly serious potential problem in the system, requiring immediate attention;
Through the method, the step S4 provides a comprehensive fault analysis and prediction framework, the decision tree algorithm effectively classifies faults, association rule mining reveals potential causes of the faults, LSTM predicts the development trend of the faults by using historical data, and powerful technical support is provided for maintenance and optimization of an electric energy metering system.
S5 specifically comprises the following steps:
S51: firstly, integrating a fault prediction result obtained in the step S2 by using a neural network and a support vector machine with a fault classification, a cause analysis and a trend prediction result obtained in the step S4;
S52: based on the integrated analysis result, setting a threshold value of the fault early warning, wherein the threshold value is determined according to historical data and system operation standards and is used for judging when to generate the fault early warning;
s53: when the analysis result shows that a certain parameter (such as current, voltage or temperature) is about to exceed the normal operation range or a specific type of fault is predicted to possibly occur, an early warning signal is automatically generated;
s55: an automatic notification mechanism is established, and once an early warning signal is generated, early warning information is immediately sent to an operation and maintenance team in the mode of electronic mail, short messages or internal notification, wherein the early warning information comprises the prediction type, possible reasons, predicted occurrence time and suggested response measures of faults;
by the method, the step S5 provides a timely and effective fault early warning mechanism, so that an operation and maintenance team can take necessary preventive measures or prepare maintenance work before a fault occurs, the system downtime is reduced, and the reliability and the safety of the electric energy metering system are improved.
S6 specifically comprises the following steps:
S61: firstly, comprehensively considering a fault prediction result of S2 and a fault classification and reason analysis result of S4 to determine the type, reason and severity of the fault;
s62: according to the result of fault analysis, a specific maintenance and repair strategy is formulated, and the specific steps are as follows:
for predicted minor faults, periodic checks and minor adjustments are scheduled;
For detected moderate failures, a more detailed inspection and partial component replacement is arranged;
For severe fault early warning, immediately taking emergency maintenance measures, including shutdown maintenance and quick replacement of key components;
the cause of the slight failure is as follows:
Slight deviations in parameters, currents or voltages slightly exceeding or falling below the normal operating range, but not reaching dangerous levels;
transient fluctuations, transient voltage fluctuations or current surges in the system, may be caused by external grid fluctuations or transient load changes;
the initial abrasion phenomenon, slight abrasion of equipment components such as a switch, a contactor and the like is generated, and the overall performance of the system is not influenced temporarily;
the cause of the moderate failure is as follows:
Continuous performance decreases, the efficiency of the device gradually decreases, such as a decrease in transformer efficiency, or an increase in cable resistance;
Frequent minor faults, such as frequent breaker trips, indicate potential circuit problems;
environmental factors such as temperature, humidity and the like are not in an ideal range for a long time, and cause continuous negative influence on equipment;
the reason for serious fault early warning is as follows:
critical component faults, such as main transformer faults, main control system faults, which directly affect the operation of the whole electric energy metering system;
Potential safety hazards, such as severe overheating of cables, insulation damage, etc., may lead to electrical fires or other safety accidents;
external severe disturbances, such as severe damage caused by external factors such as extreme weather, accidents, etc., like lightning strikes, flood attacks, etc.
S63: according to the formulated strategy, resources and personnel are arranged to carry out maintenance and repair work, all operations are ensured to meet the safety standard, after maintenance and repair are implemented, the repair effect is evaluated, if the fault is solved, details of maintenance and repair are recorded for future reference, if the problem is not completely solved, the strategy is adjusted according to the actual situation, and the implementation is repeated;
By the method, the step S6 ensures that the electric energy metering system can be timely and effectively maintained and repaired when the potential or actual fault is detected, so that the high-efficiency and stable operation of the system is ensured.
S7 specifically comprises the following steps:
s71: collecting monthly fault data from the power metering system including current, voltage, temperature and detailed fault records detailing time, duration, type of fault and affected system scope of the fault;
s72: and analyzing the data collected monthly by applying a data analysis technology of principal component analysis to identify potential failure modes and trends, wherein the principal component analysis formula is as follows: z=xw, where X is an original data matrix, W is a principal component weight matrix extracted from data, and Z is a converted data matrix;
s73: according to the analysis result, the fault detection algorithm is adjusted and optimized, and the weight and bias of the neural network are specifically adjusted, wherein the formula for adjusting the weight of the neural network is as follows:
w new＝Wold +Δw, where W new is the adjusted weight, W old is the original weight, and Δw is the weight adjustment amount obtained based on fault data analysis;
S74: updating parameters and threshold values of an early warning mechanism by using a newly optimized fault detection algorithm so as to ensure the accuracy and timeliness of early warning;
the step S7 not only realizes the effective and periodic collection of fault data of the electric energy metering system, but also ensures the continuous update and optimization of a detection algorithm and an early warning mechanism, thereby improving the performance and stability of the system, and in summary, the step S7 provides an effective way for continuously improving the performance of the electric energy metering system by periodically collecting the fault data each month and continuously optimizing the fault detection algorithm and the early warning mechanism based on the data.
The present invention is intended to embrace all such alternatives, modifications and variances which fall within the broad scope of the appended claims. Therefore, any omission, modification, equivalent replacement, improvement, etc. of the present invention should be included in the scope of the present invention.
Claims (8)
1. The fault detection and prevention method for the electric energy metering system of the energy controller is characterized by comprising the following steps of:
s1: presetting a fault detection device with a plurality of sensors, wherein the fault detection device is used for monitoring the running state of an electric energy metering system in real time;
S2: based on a data analysis and fault prediction algorithm of a neural network and a support vector machine, processing data from the S1, and identifying potential fault modes and trends;
S3: performing real-time data monitoring, and accurately positioning specific components or areas where faults occur by using a convolutional neural network;
s4: automatically classifying the detected faults, analyzing the reasons of the faults, and predicting the fault development trend by utilizing historical data and pattern recognition technology;
S5: based on the analysis results of the S2 and the S4, a fault early warning mechanism is established and used for timely notifying a maintenance team of an impending fault;
S6: according to the fault detection and analysis results, a targeted maintenance and repair strategy is formulated and implemented;
s7: and periodically collecting fault data, and continuously optimizing a fault detection algorithm and an early warning mechanism based on the collected fault data.
2. The method for detecting and preventing faults of an electric energy metering system of an energy controller according to claim 1, wherein the fault detection device in S1 comprises a current sensor, a voltage sensor, a temperature sensor and a data preprocessing unit, and is used for comprehensively monitoring operation parameters of the electric energy metering system; wherein,
A current sensor: the current sensor is arranged on a main power supply line of the electric energy metering system and used for monitoring the current quantity passing through the electric wire in real time, and can detect the magnitude and fluctuation of the current so as to identify the fault condition of overload or open circuit;
a voltage sensor: the voltage sensor is arranged at the input end and the output end of the system and is used for measuring the voltage level of the electric energy metering system, the voltage sensor can monitor the stability and deviation of the voltage, and abnormal conditions of unstable or over-high and over-low voltage can be found;
Temperature sensor: the key components fixed on the electric energy metering system comprise a transformer and a cable joint part, are used for monitoring the operation temperature of the corresponding components, and can timely find out the overheating problem caused by overload or short circuit fault by monitoring the change of the temperature;
A data preprocessing unit: the data preprocessing unit is used for receiving data from all sensors and performing preliminary processing on the received data, and particularly adopts real-time data analysis technology including signal filtering, normalization processing and preliminary trend analysis to perform preliminary evaluation on parameters of current, voltage and temperature so as to ensure the quality of the provided data.
3. The method for detecting and preventing a fault in an electric energy metering system of an energy controller according to claim 2, wherein S2 specifically comprises:
S21: firstly, receiving data of current I, voltage V and temperature T processed in the step S1, and presetting an input layer of a neural network to receive n characteristics (F 1,F2, …, fn), wherein each characteristic is extracted from original data, and a specific characteristic extraction formula is as follows: f i＝fi (I, V, T), where F i is a feature extraction function for extracting a feature F i from current I, voltage V, and temperature T;
s22: the neural network further processes the characteristics through a hidden layer H, and the hidden layer is provided with m nodes, each node applies an activation function a to process input, and a specific hidden layer processing formula is as follows:
Where w ij is the weight, b j is the bias term, H j is the output of the j-th hidden layer node;
S23: the output layer O gives out initial judgment of fault prediction according to the output of the hidden layer, and the output layer judgment formula is as follows:
Wherein w j 'and b' are the weight and bias terms of the output layer, respectively;
s24: the support vector machine is used for classifying the characteristics extracted by the neural network to judge whether the system has faults, and particularly, the support vector machine is used for constructing a classification hyperplane for separating data in a normal state and data in an abnormal state, and the classification formula of the support vector machine is as follows:
Where α k is the Lagrangian multiplier, y k is the label of the training sample, K is the kernel function, x k and x are the eigenvectors of the training sample and the new sample, respectively, b "is the bias term, and y is the classification result.
4. The method for detecting and preventing a fault in an electric energy metering system of an energy controller according to claim 3, wherein S3 specifically comprises:
s31: firstly, monitoring current, voltage and temperature data of an electric energy metering system in real time, wherein the data are used as input of a convolutional neural network and used for fault location analysis;
s32: constructing a convolutional neural network model, wherein the convolutional neural network model consists of a plurality of convolutional layers, a pooling layer and a full-connection layer, and the specific function of each layer is as follows:
The convolution layer is used for extracting the characteristics of input data, specifically, if the convolution kernel of the convolution layer is K and the size is m multiplied by n, the convolution operation is expressed as:
Wherein, C ij is an element in the output feature diagram of the convolution layer, K ab is an element with a position (a, b) in the convolution kernel, D i+a,j+b is an element multiplied by a position corresponding to the convolution kernel in the input data matrix, m×n is a size of the convolution kernel, a, b is an index in the convolution kernel;
The pooling layer is used for reducing feature dimension and enhancing robustness of the features, and the pooling operation maximum pooling is expressed as: p ij＝max(Dkl), for k, l e window (i, j), where P ij is an element in the pooling layer output feature map, D kl is an element in the convolved feature map, max (D kl) is a pooling operation that selects the maximum value within a particular window, window (i, j) is a particular window region considered in the pooling operation, centered at (i, j);
The full-connection layer is used for connecting the output of the pooling layer to the full-connection layer for final classification after flattening;
s33: specific failure modes will be identified by training convolutional neural network models, each mode corresponding to a failure of a specific component or region, and each neuron of the output layer corresponding to a potential failure location in the system.
5. The method for detecting and preventing a fault in an electric energy metering system of an energy controller according to claim 4, wherein S4 specifically comprises:
S41: performing fault classification, automatically classifying the detected fault conditions into different categories including current faults, voltage faults or temperature faults according to the data collected from the steps S1 and S2 by using a decision tree classification algorithm, wherein each internal node represents a test of an attribute, each branch represents a result of the test, and each leaf node represents a fault category, and the decision tree rule is as follows:
If (A 1≤v1)∧(A2>v2)∧…∧(An＝vn) THENCLASSC, wherein A i is an attribute, v i is a test value of the attribute, and C is a fault class;
s42: analyzing potential reasons of various faults by adopting an association rule mining analysis technology, wherein the association rule mining can reveal the reasons behind different fault categories, and the association rule formula is as follows:
Rule: withsupportsandconfidencec, where X and Y are a collection of data items, the support s represents the frequency of simultaneous occurrence of X and Y, and the confidence c represents the conditional probability of Y occurring in the case where X occurs;
S43: analyzing the historical fault data by using a long-short-period memory network to predict the fault development trend, wherein the specific calculation formula of the long-short-period memory network is as follows:
ft＝σ(Wf·[ht-1,xt]+bf)，
ot＝σ(Wo·[ht-1,xt]+bo)，
h t＝ot*tanh(Ct), where f t is a forget gate, controlling how much previous memory is reserved, i t is an input gate, controlling the impact of the current input, Is a candidate memory cell, C t is an updated memory cell, o t is an output gate, h t is the output at the current time, W and b are weights and bias parameters, σ is sigmoidi activation function, which represents the inter-element multiplication.
6. The method for detecting and preventing a fault in an electric energy metering system of an energy controller according to claim 5, wherein S5 specifically comprises:
S51: firstly, integrating a fault prediction result obtained in the step S2 by using a neural network and a support vector machine with a fault classification, a cause analysis and a trend prediction result obtained in the step S4;
S52: based on the integrated analysis result, setting a threshold value of the fault early warning, wherein the threshold value is determined according to historical data and system operation standards and is used for judging when to generate the fault early warning;
S53: when the analysis result shows that a certain parameter is about to exceed the normal operation range or a specific type of fault is predicted to possibly occur, an early warning signal is automatically generated;
S55: an automatic notification mechanism is established, and once an early warning signal is generated, early warning information is immediately sent to an operation and maintenance team in the mode of electronic mail, short messages or internal notification, wherein the early warning information comprises the prediction type, possible reasons, predicted occurrence time and suggested response measures of the fault.
7. The method for detecting and preventing a fault in an electric energy metering system of an energy controller according to claim 6, wherein S6 specifically comprises:
S61: firstly, comprehensively considering a fault prediction result of S2 and a fault classification and reason analysis result of S4 to determine the type, reason and severity of the fault;
s62: according to the result of fault analysis, a specific maintenance and repair strategy is formulated, and the specific steps are as follows:
for predicted minor faults, periodic checks and minor adjustments are scheduled;
For detected moderate failures, a more detailed inspection and partial component replacement is arranged;
For severe fault early warning, immediately taking emergency maintenance measures, including shutdown maintenance and quick replacement of key components;
S63: and according to the formulated strategy, arranging resources and personnel to carry out maintenance and repair work, and ensuring that all operations accord with the safety standard.
8. The method for detecting and preventing a fault in an electric energy metering system of an energy controller according to claim 7, wherein S7 specifically comprises:
s71: collecting monthly fault data from the power metering system including current, voltage, temperature and detailed fault records detailing time, duration, type of fault and affected system scope of the fault;
S72: analyzing the data collected monthly by applying a data analysis technology of principal component analysis to identify potential failure modes and trends, wherein the principal component analysis formula is as follows: z=xw, where X is an original data matrix, W is a principal component weight matrix extracted from data, and Z is a converted data matrix;
S73: according to the analysis result, the fault detection algorithm is adjusted and optimized, and the weight and bias of the neural network are specifically adjusted, wherein the formula for adjusting the weight of the neural network is as follows:
w new＝Wold +Δw, where W new is the adjusted weight, W old is the original weight, and Δw is the weight adjustment amount obtained based on fault data analysis;
S74: and updating parameters and threshold values of the early warning mechanism by using a newly optimized fault detection algorithm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410087542.0A CN117949887A (en) | 2024-01-22 | 2024-01-22 | Fault detection and prevention method for electric energy metering system of energy controller |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410087542.0A CN117949887A (en) | 2024-01-22 | 2024-01-22 | Fault detection and prevention method for electric energy metering system of energy controller |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117949887A true CN117949887A (en) | 2024-04-30 |
Family
ID=90791737
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410087542.0A Pending CN117949887A (en) | 2024-01-22 | 2024-01-22 | Fault detection and prevention method for electric energy metering system of energy controller |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117949887A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118152172A (en) * | 2024-05-13 | 2024-06-07 | 南京新联电子股份有限公司 | Method and system for self-checking and self-recovering energy controller software faults |
-
2024
- 2024-01-22 CN CN202410087542.0A patent/CN117949887A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118152172A (en) * | 2024-05-13 | 2024-06-07 | 南京新联电子股份有限公司 | Method and system for self-checking and self-recovering energy controller software faults |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Appiah et al. | Long short-term memory networks based automatic feature extraction for photovoltaic array fault diagnosis | |
| CN117949887A (en) | Fault detection and prevention method for electric energy metering system of energy controller | |
| CN117277592B (en) | Protection switching method for monitoring high-voltage circuit signals | |
| CN117674140B (en) | Power distribution network measurement and control system and method | |
| CN113763667A (en) | Fire early warning and state monitoring device and method based on 5G edge calculation | |
| CN113902946A (en) | Power system fault direction judging method and device, terminal equipment and storage medium | |
| CN111044845B (en) | Power distribution network accident identification method and system based on Apriori algorithm | |
| CN117612345A (en) | Power equipment state monitoring and alarming system and method | |
| CN105205732A (en) | Risk assessment and maintenance method based on equipment risk characteristic model | |
| CN114487705A (en) | Power grid equipment fault positioning detection method | |
| CN117350548B (en) | Power distribution equipment potential safety hazard investigation method | |
| Sun et al. | A multi-model-integration-based prediction methodology for the spatiotemporal distribution of vulnerabilities in integrated energy systems under the multi-type, imbalanced, and dependent input data scenarios | |
| CN117494009A (en) | Electrical equipment state evaluation method based on insulating material pyrolysis analysis and cloud platform | |
| CN113758652A (en) | Converter transformer oil leakage detection method and device, computer equipment and storage medium | |
| CN117372954A (en) | Charging station safety monitoring method, device, system and computer equipment | |
| Lan et al. | Unsupervised power system event detection and classification using unlabeled pmu data | |
| CN114358165A (en) | Detection method for preventing ground fault of photovoltaic module based on multi-source data fusion | |
| CN114077809A (en) | Method and monitoring system for monitoring performance of decision logic of controller | |
| Adhikari et al. | Real-Time Short-Term Voltage Stability Assessment using Temporal Convolutional Neural Network | |
| CN118261577A (en) | Operation state grading method for power equipment of substation in front-end system | |
| CN117407824B (en) | Health detection method, equipment and medium of power time synchronization device | |
| CN117805607B (en) | DC level difference matching test method for power plant DC system | |
| Rajesh et al. | ENHANCING POWER SYSTEM STABILITY WITH AI-BASED RELAYING ALGORITHMS–A REVIEW. | |
| CN118035814A (en) | State fault early warning method and system based on digital twin field data analysis equipment | |
| CN117609861A (en) | BP neural network algorithm and temperature measurement system capable of being used for temperature prediction of electrical equipment of transformer substation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |