CN114881162A - Method, apparatus, device and medium for predicting failure of metering automation master station - Google Patents
Method, apparatus, device and medium for predicting failure of metering automation master station Download PDFInfo
- Publication number
- CN114881162A CN114881162A CN202210548132.2A CN202210548132A CN114881162A CN 114881162 A CN114881162 A CN 114881162A CN 202210548132 A CN202210548132 A CN 202210548132A CN 114881162 A CN114881162 A CN 114881162A
- Authority
- CN
- China
- Prior art keywords
- target
- master station
- metering automation
- data
- automation master
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/047—Probabilistic or stochastic networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Abstract
The invention discloses a method, a device, equipment and a medium for predicting faults of a metering automation master station, the operation data of the metering automation master station is subjected to data cleaning to obtain target operation data, and by utilizing a convolutional neural network in a target fault prediction model, extracting the characteristics of the target operation data to obtain a characteristic sequence of the target operation data, utilizing a long-term and short-term memory network in the target fault prediction model, and according to the characteristic sequence, carrying out fault prediction on the metering automation master station to obtain a fault probability value of the metering automation master station at a future target moment, comparing the fault probability value with a preset fault threshold value to determine the state information of the metering automation master station at a future target moment, the state information is a failure state or a non-failure state, thereby improving the prediction accuracy using the advantages of the CNN and LSTM.
Description
Technical Field
The invention relates to the technical field of metering automation master stations, in particular to a fault prediction method, a fault prediction device, fault prediction equipment and a fault prediction medium of a metering automation master station.
Background
The metering automation master station is a computer system connected with various metering automation terminals, collects the information of the whole metering automation system in a control center, collects and controls the information of the metering automation terminals through a remote communication channel or a lower unit system interface, and analyzes, comprehensively processes and releases electric energy data. In order to ensure the normal operation of the metering automation system, the fault early warning of the metering automation master station is very important.
At present, a metering automation master station usually performs fault prediction based on a traditional data mining technology, for example, a fault prediction model of a master station system is established based on a nearest neighbor algorithm, a decision tree classifier or a clustering method, and fault prediction of the master station is performed through the fault prediction model. However, due to the large scale and the large number of data sets of the metering automation master station, the traditional data mining technology has the problems of low prediction efficiency and low prediction accuracy. For example, the nearest neighbor algorithm has a large calculation amount and low operation efficiency; the convergence speed of the clustering method on large-scale data is low, and the condition that the data is converged to a local minimum value may occur; the decision tree classifier has difficulty in extracting correlation between data and is prone to overfitting phenomena.
Disclosure of Invention
The invention provides a method, a device, equipment and a medium for predicting faults of a metering automation master station, and aims to solve the technical problem that the accuracy of predicting faults of the current metering automation master station is low.
In order to solve the above technical problem, in a first aspect, the present invention provides a method for predicting a failure of a metering automation master station, including:
performing data cleaning on the operation data of the metering automation master station to obtain target operation data;
performing feature extraction on the target operation data by using a convolutional neural network in a target fault prediction model to obtain a feature sequence of the target operation data;
carrying out fault prediction on the metering automation master station by utilizing a long-short term memory network in the target fault prediction model according to the characteristic sequence to obtain a fault probability value of the metering automation master station at a future target moment;
and comparing the fault probability value with a preset fault threshold value to determine the state information of the metering automation master station at a future target moment, wherein the state information is a fault state or a non-fault state.
Preferably, before the feature extraction is performed on the target operation data by using a convolutional neural network in a target fault prediction model to obtain the feature data of the target operation data, the method further includes:
performing data cleaning on historical operating data of the metering automation master station to obtain an operating data sample;
oversampling is carried out on the operation data sample to obtain a target operation data sample;
sampling the target operation data sample based on a preset two-stage time window to obtain a target sample set comprising the two-stage time window, wherein a first-stage time window of the target sample set is used for recording the target operation data sample, a second-stage time window is used for recording a fault state label of the metering automation master station at a future moment, and the time lengths of the first-stage time window and the second-stage time window are the same;
and training a convolutional neural network and a long-term and short-term memory network of a preset fault prediction model by using the target sample set until the preset fault prediction model is converged to obtain the target fault prediction model.
Preferably, the data cleaning of the historical operation data of the metering automation master station to obtain the operation data sample includes:
acquiring historical operating data acquired by a metering automation master station;
and performing outlier processing on the historical operating data to remove redundant data of the historical operating data to obtain the operating data sample.
Preferably, the oversampling the operation data sample to obtain a target operation data sample includes:
classifying the running data samples to obtain a positive sample set and a negative sample set;
generating a random number and randomly taking a positive sample from the set of positive samples and a negative sample from the set of negative samples based on the random number;
and generating a new negative sample according to the positive sample and the negative sample, and adding the new negative sample into the negative sample set until the number of samples in the positive sample set is the same as that in the negative sample set.
Preferably, the performing feature extraction on the target operation data by using a convolutional neural network in a target fault prediction model to obtain a feature sequence of the target operation data includes:
converting the target operation data into a vector matrix by using the convolutional neural network;
performing convolution operation on the vector matrix based on the multi-stage convolution layers of the convolutional neural network to obtain a feature matrix output by each stage of convolution layer;
and performing feature aggregation on the feature matrix output by each stage of convolutional layer based on the multistage pooling layers of the convolutional neural network to obtain a feature sequence of the target operation data.
Preferably, the expression of the convolution operation is:
S＝f(WZ+b)；
wherein S is a feature matrix obtained after convolution operation, W is a weight matrix of the convolution layer, Z is a vector matrix, b is an offset vector, and f represents convolution operation.
Preferably, the performing fault prediction on the metering automation master station according to the feature sequence by using a long-short term memory network in the target fault prediction model to obtain a fault probability value of the metering automation master station at a future target time includes:
performing feature screening on the feature sequence by using the long-term and short-term memory network to obtain a vector matrix;
and fully connecting and activating the vector matrix to obtain the fault probability value of the metering automation master station at the future target moment.
In a second aspect, the present invention provides a failure prediction device for a metering automation master station, including:
the cleaning module is used for cleaning the operation data of the metering automation master station to obtain target operation data;
the extraction module is used for extracting the characteristics of the target operation data by using a convolutional neural network in a target fault prediction model to obtain a characteristic sequence of the target operation data;
the prediction module is used for predicting the faults of the metering automation master station according to the characteristic sequence by utilizing a long-term and short-term memory network in the target fault prediction model to obtain the fault probability value of the metering automation master station at the future target moment;
and the comparison module is used for comparing the fault probability value with a preset fault threshold value so as to determine the state information of the metering automation master station at a future target moment, wherein the state information is a fault state or a non-fault state.
In a third aspect, the invention provides a computer device comprising a processor and a memory for storing a computer program which, when executed by the processor, implements the method of fault prediction for a metering automation master station according to the first aspect.
In a fourth aspect, the present invention provides a computer-readable storage medium storing a computer program which, when executed by a processor, implements the method for failure prediction of a metering automation master according to the first aspect.
Compared with the prior art, the invention has the following beneficial effects:
according to the invention, the target operation data is obtained by cleaning the operation data of the metering automation master station, so that the data calculation amount is reduced, and the data processing efficiency is improved; performing feature extraction on the target operation data by using a convolutional neural network in a target fault prediction model to obtain a feature sequence of the target operation data, so that the advantage of the convolutional neural network is utilized to improve the feature extraction depth of the target operation data, thereby extracting more important features in the target operation data and ensuring the relevance between time sequence features; then, a long-short term memory network in the target fault prediction model is utilized to carry out fault prediction on the metering automation master station according to the characteristic sequence to obtain a fault probability value of the metering automation master station at a future target moment, and target operation data with time sequence characteristics are predicted by using an LSTM (local state transformation) so as to improve the precision of the fault probability value at the future moment; and finally, comparing the fault probability value with a preset fault threshold value to determine the state information of the metering automation master station at a future target moment, so that intelligent fault early warning is realized based on the CNN and the LSTM, and the prediction accuracy is improved by utilizing the advantages of the CNN and the LSTM.
Drawings
FIG. 1 is a schematic flow diagram illustrating a method for predicting a failure of a metering automation master, according to an embodiment of the present invention;
FIG. 2 illustrates an embodiment of the present invention;
FIG. 3 illustrates an embodiment of the present invention;
FIG. 4 is a schematic structural diagram of a failure prediction apparatus of a metering automation master station according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of a computer device according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, fig. 1 is a flowchart illustrating a failure prediction method of a metering automation master station according to an embodiment of the present invention. The method for predicting the fault of the metering automation master station can be applied to computer equipment, wherein the computer equipment comprises but is not limited to equipment such as a smart phone, a notebook computer, a tablet computer, a desktop computer, a physical server and a cloud server. As shown in fig. 1, the method for predicting the failure of the metering automation master station of the present embodiment includes steps S101 to S104, which are detailed as follows:
and step S101, performing data cleaning on the operation data of the metering automation master station to obtain target operation data.
In this step, the operation data includes, but is not limited to, data traffic, server CPU occupancy, TCP connection number, and the like. The related operation data generated by the metering automation master station has large data volume and much redundant data, so the operation data is subjected to data cleaning. Based on the data points where the running data at the abnormal position of the network can present outliers, the data preprocessing method for eliminating the redundant non-outliers is adopted to reduce the data volume, and the outliers are defined to be points which are far away from most of the points and exceed a certain specific numerical value.
Optionally, the normal points are aggregated first and sorted according to data density. Similar points of a small set formed in the data set are removed, one point is selected, and the distance between the point and other points is calculated. If the distance from the center of the other set is less than the set radius, the point is considered to belong to the set; otherwise, the point is considered as a new set center. And keeping the sets with the point number smaller than k in the sets, and deleting all the other sets. Based on the operation, partial redundant non-outliers can be removed, the data volume needing to be processed by the fault prediction model is reduced, and the prediction efficiency is improved.
And S102, extracting the characteristics of the target operation data by using a convolutional neural network in a target fault prediction model to obtain a characteristic sequence of the target operation data.
In this step, as shown in the schematic diagram of the target failure prediction model shown in fig. 2, the Convolutional Neural Network (CNN) of this embodiment includes a convolutional layer and a pooling layer, which is used for feature extraction of the operation data, wherein the pooling layer may be a K-max pooling layer. Optionally, the convolution layer and the pooling layer of the CNN perform feature extraction from the running data through convolution operation and pooling operation, respectively, to obtain the most important feature information in the running data.
In an embodiment, the step S102 includes:
converting the target operation data into a vector matrix by using the convolutional neural network;
performing convolution operation on the vector matrix based on the multi-stage convolution layers of the convolutional neural network to obtain a feature matrix output by each stage of convolution layer;
and performing feature aggregation on the feature matrix output by each stage of convolutional layer based on the multistage pooling layers of the convolutional neural network to obtain a feature sequence of the target operation data.
In this embodiment, the CNN adopts a local connection and weight sharing manner, which can reduce the complexity of the neural network model and reduce the number of weights. The original data are alternately processed by the convolution layer and the pooling layer, so that the local features of the data can be automatically extracted, and corresponding feature vectors are established. The feature extraction engineering comprises the following three steps:
(1) the embedding layer (input layer). The operational data needs to be input from the embedding layer and vectorized into a matrix for input into the network. The data length of the operational data is limited to m, and each data is converted into an n-dimensional data vector. The output of the embedding layer is an n × m two-dimensional matrix Z ═ W 1 ,...,W i ,...,W m ]Wherein W is i ＝[X i1 ,...,X ij ,...,X in ]Is data W i The data vector of (2).
(2) And (4) rolling up the layers. Convolutional layers are used to extract features, with each convolutional kernel corresponding to a particular portion of the extracted feature. The convolution kernel performs the following convolution operation on the output matrix Z of the embedding layer:
S＝f(WZ+b)；
wherein S is a feature matrix obtained after convolution operation, W is a weight matrix of the convolution layer, Z is a vector matrix, b is an offset vector, and f represents convolution operation. W and b are parameters obtained by web learning. In the convolution operation, the weight of the filter is kept unchanged, and the number of model parameters to be learned is reduced, so that the neural network structure is simplified. After passing through the convolutional layer, the activation function is added to the output of the convolutional layer, playing a nonlinear role in the output of the neural network convolutional layer. Optionally, the activation function adopted by this embodiment is:
ReLU(x)＝max(0,x)；
the ReLU function has the characteristics of high convergence rate and simple gradient solution.
(3) And (4) a pooling layer. After the convolution operation, the extracted feature matrix is fed into the K-max pooling layer. The pooling layer further aggregates features, selecting the K most prominent features to simplify feature expression.
And step S103, carrying out fault prediction on the metering automation master station by utilizing a long-short term memory network in the target fault prediction model according to the characteristic sequence to obtain the fault probability value of the metering automation master station at a future target moment.
In this step, as shown in fig. 2, the CNN signature sequence is used as input to LSTM for predicting faults. The long-short term memory (LSTM) network is an improved method based on a recurrent neural network, and has certain advantages in the aspect of learning long-term dependence. The LSTM of this embodiment predicts the probability of failure at a future time from the operation data within a certain time range. LSTM can solve the problem of disappearance of gradients present in long sequence data processing from long term memory deficits.
In an embodiment, the step S103 includes:
performing feature screening on the feature sequence by using the long-term and short-term memory network to obtain a vector matrix;
and fully connecting and activating the vector matrix to obtain the fault probability value of the metering automation master station at the future target moment.
In this embodiment, fig. 3 shows a schematic network structure of LSTM. Optionally, to prevent the data from having the over-fitting problem, the vector matrix output by the LSTM model is input to the discarding layer for feature screening, where the feature screening process is as follows: the basic unit of the LSTM network includes a forgetting gate, an input gate, and an output gate. As shown in the figure, f t To forget the door, i t To the input gate o t Is an output gate. Sigma is sigmoid activation function. The input of the forgetting gate is x t Intermediate input h t-1 And a state storage unit C t-1 The three parts together determine which information needs to be forgotten by the state storage unit. On the other hand, x of the input gate t The output of the sigma and tanh functions is used to determine which information should be retained in the state memory cells. Based on the structure of the input gate and the forgetting gate, the LSTM network can more effectively decide which information should be retained and which information should be forgotten. Subsequently, C is updated t And an output o t And based on the two updates, an intermediate output h is obtained t And completing a complete updating process. Illustratively, metersThe calculation formula is shown as the formula:
f t ＝σ(W t ·[h t-1 ,x t ]+b f )；
i t ＝σ(W i ·[h t-1 ,x t ]+b i )；
o t ＝σ(W o ·[h t-1 ,x t ]+b o )；
h t ＝o t ·tanh(C t )；
wherein, W t ，W i ，W o ，W C Respectively corresponding weight values of the forgetting gate, the input gate, the output gate and the state storage unit, b f ，b i ，b o ，b C Respectively, the offset of the corresponding status gate. σ (x) and tanh (x) are activation functions.
Optionally, the vector matrix is input into the full-connection layer to be subjected to dimension reduction processing, the vector dimension is reduced, the output of the full-connection layer is processed by using an activation function, and the probability p of the prediction result is obtained, wherein p belongs to [0,1 ].
And step S104, comparing the fault probability value with a preset fault threshold value to determine the state information of the metering automation master station at a future target moment, wherein the state information is a fault state or a non-fault state.
In the step, when the prediction probability is larger than the threshold value, the metering automation master station is considered to be in failure; otherwise, no failure is considered to occur.
In an embodiment, based on the embodiment shown in fig. 1, the method further includes:
performing data cleaning on historical operating data of the metering automation master station to obtain an operating data sample;
oversampling the operation data sample to obtain a target operation data sample;
sampling the target operation data sample based on a preset two-stage time window to obtain a target sample set comprising the two-stage time window, wherein a first-stage time window of the target sample set is used for recording the target operation data sample, a second-stage time window is used for recording a fault state label of the metering automation master station at a future moment, and the time lengths of the first-stage time window and the second-stage time window are the same;
and training a convolutional neural network and a long-term and short-term memory network of a preset fault prediction model by using the target sample set until the preset fault prediction model is converged to obtain the target fault prediction model.
In the embodiment, in the process of fault early warning of the metering automation master station, faults such as network congestion and interface damage are typical fault types. Therefore, data traffic, server CPU occupancy, TCP connection count, and other related data need to be collected first. Because the related network data generated by the metering automation master station system has the characteristics of large data volume, more redundant data, less abnormal data and more defective data, the data collected by the metering automation master station needs to be preprocessed based on a scheme of eliminating redundant non-abnormal data points and oversampling. Subsequently, sample information is extracted using two levels of time windows. The first-level window is used for recording the current operation data of the metering automation master station system and taking the current operation data as an input sequence of the prediction model. The second-level window is used for recording the future communication network security state to be predicted and using the future communication network security state as a label. The length of the two-stage window should be kept uniform. Inputting the data after the data preprocessing process into a preset fault prediction model constructed based on CNN-LSTM for training until the loss function of the preset fault prediction model is smaller than a preset threshold value or the iteration number of the training process reaches a preset iteration upper limit number, and then converging the preset fault prediction model to obtain a target fault prediction model.
Optionally, the data cleaning is performed on the historical operation data of the metering automation master station to obtain an operation data sample, and the method includes:
acquiring historical operating data acquired by a metering automation master station;
and performing outlier processing on the historical operating data to remove redundant data of the historical operating data to obtain the operating data sample.
In the present embodiment, the data cleansing is the same as the data cleansing process of step S101, and reference may be made to the related description of step S101.
Optionally, the oversampling the operation data sample to obtain a target operation data sample includes:
classifying the running data samples to obtain a positive sample set and a negative sample set;
generating a random number and randomly taking a positive sample from the set of positive samples and a negative sample from the set of negative samples based on the random number;
and generating a new negative sample according to the positive sample and the negative sample, and adding the new negative sample into the negative sample set until the number of samples in the positive sample set is the same as that in the negative sample set.
In this embodiment, since the metering automation master station is in a normal operation state at most of the time, the acquired abnormal data information is very limited, which is not favorable for the training of the fault prediction model. Based on this, the number of normal samples and abnormal samples is balanced using an oversampling technique. The problem that the computer cannot learn the characteristics of abnormal information due to data imbalance can be effectively solved through oversampling.
Optionally. The collected data is divided into a positive (normal) sample set and a negative (abnormal) sample set. At this time, the amount of data in the negative sample set is much smaller than the amount of data in the positive sample set. Then, a random number r is generated, r belongs to [0.1,0.4], a sample in a negative sample is randomly extracted and is recorded as x, a positive sample is randomly extracted and is recorded as y, and a new negative sample z is generated, namely r.y + (1-r). x. This step is repeated until the sum of the number of elements of the negative sample set and the newly generated negative sample set is equal to the number of elements of the positive sample set, and the oversampling is ended.
It should be noted that, compared with the prior art, the invention has the following advantages:
(1) compared with the method which only depends on manual daily routing inspection, the method for predicting the faults of the metering automation master station based on the convolutional neural network and the long-term and short-term memory network is safer and faster and has predictability.
(2) By removing part of redundant non-abnormal information and oversampling, the problems of large data collection amount, more redundant data, less abnormal data and more defective data of the metering automation master station can be solved.
(3) Compared with a CNN training model or a random forest prediction model and the like which are used independently, the CNN-LSTM-based fault prediction model has higher prediction accuracy.
In order to execute the fault prediction method of the metering automation master station corresponding to the method embodiment, corresponding functions and technical effects are realized. Referring to fig. 4, fig. 4 is a block diagram illustrating a failure prediction apparatus of a metering automation master station according to an embodiment of the present invention. For convenience of explanation, only the parts related to the present embodiment are shown, and the failure prediction apparatus for a metering automation master station according to the present embodiment includes:
a cleaning module 401, configured to perform data cleaning on the operation data of the metering automation master station to obtain target operation data;
an extraction module 402, configured to perform feature extraction on the target operation data by using a convolutional neural network in a target fault prediction model to obtain a feature sequence of the target operation data;
the prediction module 403 is configured to perform fault prediction on the metering automation master station according to the feature sequence by using a long-term and short-term memory network in the target fault prediction model, so as to obtain a fault probability value of the metering automation master station at a future target time;
a comparison module 404, configured to compare the fault probability value with a preset fault threshold value, so as to determine state information of the metering automation master station at a future target time, where the state information is a fault state or a non-fault state.
In an embodiment, the failure prediction apparatus further includes:
the cleaning module is used for carrying out data cleaning on historical operating data of the metering automation master station to obtain an operating data sample;
the oversampling module is used for oversampling the operation data sample to obtain a target operation data sample;
the sampling module is used for sampling the target operation data sample based on a preset two-stage time window to obtain a target sample set containing the two-stage time window, wherein a first-stage time window of the target sample set is used for recording the target operation data sample, a second-stage time window of the target sample set is used for recording a fault state label of the metering automation master station at a future moment, and the time lengths of the first-stage time window and the second-stage time window are the same;
and the training module is used for training a convolutional neural network and a long-term and short-term memory network of a preset fault prediction model by using the target sample set until the preset fault prediction model is converged to obtain the target fault prediction model.
In an embodiment, the cleaning module is specifically configured to:
acquiring historical operating data acquired by a metering automation master station;
and performing outlier processing on the historical operating data to remove redundant data of the historical operating data to obtain the operating data sample.
In an embodiment, the oversampling module is specifically configured to:
classifying the running data samples to obtain a positive sample set and a negative sample set;
generating a random number, and randomly taking a positive sample from the set of positive samples and a negative sample from the set of negative samples based on the random number;
and generating a new negative sample according to the positive sample and the negative sample, and adding the new negative sample into the negative sample set until the number of samples in the positive sample set is the same as that in the negative sample set.
In an embodiment, the extracting module 402 is specifically configured to:
converting the target operation data into a vector matrix by using the convolutional neural network;
performing convolution operation on the vector matrix based on the multi-stage convolution layers of the convolutional neural network to obtain a feature matrix output by each stage of convolution layer;
and performing feature aggregation on the feature matrix output by each stage of convolutional layer based on the multistage pooling layers of the convolutional neural network to obtain a feature sequence of the target operation data.
In one embodiment, the convolution operation is expressed as:
S＝f(WZ+b)；
wherein S is a feature matrix obtained after convolution operation, W is a weight matrix of the convolution layer, Z is a vector matrix, b is an offset vector, and f represents convolution operation.
In an embodiment, the prediction module 403 is specifically configured to:
performing feature screening on the feature sequence by using the long-term and short-term memory network to obtain a vector matrix;
and fully connecting and activating the vector matrix to obtain the fault probability value of the metering automation master station at the future target moment.
The failure prediction device of the metering automation master station can implement the failure prediction method of the metering automation master station of the method embodiment. The alternatives in the above-described method embodiments are also applicable to this embodiment and will not be described in detail here. The rest of the embodiments of the present invention may refer to the contents of the above method embodiments, and in this embodiment, details are not repeated.
Fig. 5 is a schematic structural diagram of a computer device according to an embodiment of the present invention. As shown in fig. 5, the computer device 5 of this embodiment includes: at least one processor 50 (only one shown in fig. 5), a memory 51, and a computer program 52 stored in the memory 51 and executable on the at least one processor 50, the processor 50 implementing the steps of any of the above-described method embodiments when executing the computer program 52.
The computer device 5 may be a computing device such as a smart phone, a tablet computer, a desktop computer, and a cloud server. The computer device may include, but is not limited to, a processor 50, a memory 51. Those skilled in the art will appreciate that fig. 5 is merely an example of the computer device 5 and does not constitute a limitation of the computer device 5, and may include more or less components than those shown, or combine some of the components, or different components, such as input output devices, network access devices, etc.
The Processor 50 may be a Central Processing Unit (CPU), and the Processor 50 may be other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic device, discrete hardware component, etc. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The memory 51 may in some embodiments be an internal storage unit of the computer device 5, such as a hard disk or a memory of the computer device 5. The memory 51 may also be an external storage device of the computer device 5 in other embodiments, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like, which are provided on the computer device 5. Further, the memory 51 may also include both an internal storage unit and an external storage device of the computer device 5. The memory 51 is used for storing an operating system, an application program, a BootLoader (BootLoader), data, and other programs, such as program codes of the computer program. The memory 51 may also be used to temporarily store data that has been output or is to be output.
In addition, an embodiment of the present invention further provides a computer-readable storage medium, where a computer program is stored, and when the computer program is executed by a processor, the computer program implements the steps in any of the method embodiments described above.
Embodiments of the present invention provide a computer program product, which when running on a computer device, enables the computer device to implement the steps in the above method embodiments when executed.
In several embodiments provided by the present invention, it will be understood that each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved.
The functions, if implemented in the form of software functional modules and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
The above-mentioned embodiments are provided to further explain the objects, technical solutions and advantages of the present invention in detail, and it should be understood that the above-mentioned embodiments are only examples of the present invention and are not intended to limit the scope of the present invention. It should be understood that any modifications, equivalents, improvements and the like, which come within the spirit and principle of the invention, may occur to those skilled in the art and are intended to be included within the scope of the invention.
Claims (10)
1. A method of predicting a failure of a metering automation master, comprising:
performing data cleaning on the operation data of the metering automation master station to obtain target operation data;
performing feature extraction on the target operation data by using a convolutional neural network in a target fault prediction model to obtain a feature sequence of the target operation data;
carrying out fault prediction on the metering automation master station by utilizing a long-short term memory network in the target fault prediction model according to the characteristic sequence to obtain a fault probability value of the metering automation master station at a future target moment;
and comparing the fault probability value with a preset fault threshold value to determine the state information of the metering automation master station at a future target moment, wherein the state information is a fault state or a non-fault state.
2. The method of claim 1, wherein before the extracting the characteristic of the target operation data by using a convolutional neural network in a target failure prediction model to obtain the characteristic data of the target operation data, the method further comprises:
performing data cleaning on historical operating data of the metering automation master station to obtain an operating data sample;
oversampling the operation data sample to obtain a target operation data sample;
sampling the target operation data sample based on a preset two-stage time window to obtain a target sample set comprising the two-stage time window, wherein a first-stage time window of the target sample set is used for recording the target operation data sample, a second-stage time window is used for recording a fault state label of the metering automation master station at a future moment, and the time lengths of the first-stage time window and the second-stage time window are the same;
and training a convolutional neural network and a long-term and short-term memory network of a preset fault prediction model by using the target sample set until the preset fault prediction model is converged to obtain the target fault prediction model.
3. The method of predicting failure of a metering automation master station of claim 2, wherein the data cleansing of historical operational data of the metering automation master station to obtain operational data samples comprises:
acquiring historical operating data acquired by a metering automation master station;
and performing outlier processing on the historical operating data to remove redundant data of the historical operating data to obtain the operating data sample.
4. The method of predicting a failure of a metering automation master station of claim 2, the oversampling the operational data samples to obtain target operational data samples comprising:
classifying the running data samples to obtain a positive sample set and a negative sample set;
generating a random number and randomly taking a positive sample from the set of positive samples and a negative sample from the set of negative samples based on the random number;
and generating a new negative sample according to the positive sample and the negative sample, and adding the new negative sample into the negative sample set until the number of samples in the positive sample set is the same as that in the negative sample set.
5. The method of claim 1, wherein the performing feature extraction on the target operation data by using a convolutional neural network in a target failure prediction model to obtain a feature sequence of the target operation data comprises:
converting the target operation data into a vector matrix by using the convolutional neural network;
performing convolution operation on the vector matrix based on the multi-stage convolution layers of the convolutional neural network to obtain a feature matrix output by each stage of convolution layer;
and performing feature aggregation on the feature matrix output by each stage of convolutional layer based on the multistage pooling layers of the convolutional neural network to obtain a feature sequence of the target operation data.
6. The method of predicting failure of a metering automation master station of claim 5, wherein the convolution operation is expressed by:
S＝f(WZ+b)；
wherein S is a feature matrix obtained after convolution operation, W is a weight matrix of the convolution layer, Z is a vector matrix, b is an offset vector, and f represents convolution operation.
7. The method for predicting the failure of the metering automation master station as claimed in claim 1, wherein the step of predicting the failure of the metering automation master station according to the characteristic sequence by using the long-short term memory network in the target failure prediction model to obtain the failure probability value of the metering automation master station at the future target time comprises:
performing feature screening on the feature sequence by using the long-term and short-term memory network to obtain a vector matrix;
and fully connecting and activating the vector matrix to obtain the fault probability value of the metering automation master station at the future target moment.
8. A failure prediction apparatus for a metering automation master station, comprising:
the cleaning module is used for cleaning the operation data of the metering automation master station to obtain target operation data;
the extraction module is used for extracting the characteristics of the target operation data by using a convolutional neural network in a target fault prediction model to obtain a characteristic sequence of the target operation data;
the prediction module is used for predicting the faults of the metering automation master station according to the characteristic sequence by utilizing a long-term and short-term memory network in the target fault prediction model to obtain the fault probability value of the metering automation master station at the future target moment;
and the comparison module is used for comparing the fault probability value with a preset fault threshold value so as to determine the state information of the metering automation master station at a future target moment, wherein the state information is a fault state or a non-fault state.
9. A computer device comprising a processor and a memory for storing a computer program which, when executed by the processor, implements the method of fault prediction for a metering automation master as claimed in any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that it stores a computer program which, when executed by a processor, implements the method of fault prediction of a metering automation master as claimed in any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210548132.2A CN114881162A (en) | 2022-05-19 | 2022-05-19 | Method, apparatus, device and medium for predicting failure of metering automation master station |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210548132.2A CN114881162A (en) | 2022-05-19 | 2022-05-19 | Method, apparatus, device and medium for predicting failure of metering automation master station |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114881162A true CN114881162A (en) | 2022-08-09 |
Family
ID=82678537
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210548132.2A Pending CN114881162A (en) | 2022-05-19 | 2022-05-19 | Method, apparatus, device and medium for predicting failure of metering automation master station |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114881162A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117270482A (en) * | 2023-11-22 | 2023-12-22 | 博世汽车部件(苏州)有限公司 | Automobile factory control system based on digital twin |
-
2022
- 2022-05-19 CN CN202210548132.2A patent/CN114881162A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117270482A (en) * | 2023-11-22 | 2023-12-22 | 博世汽车部件(苏州)有限公司 | Automobile factory control system based on digital twin |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Prajwala | A comparative study on decision tree and random forest using R tool | |
| WO2021000556A1 (en) | Method and system for predicting remaining useful life of industrial equipment, and electronic device | |
| CN110874550A (en) | Data processing method, device, equipment and system | |
| US11551076B2 (en) | Event-driven temporal convolution for asynchronous pulse-modulated sampled signals | |
| CN111431819B (en) | Network traffic classification method and device based on serialized protocol flow characteristics | |
| CN111797122A (en) | Method and device for predicting change trend of high-dimensional reappearance concept drift stream data | |
| CN111368887B (en) | Training method of thunderstorm weather prediction model and thunderstorm weather prediction method | |
| CN111008224B (en) | Time sequence classification and retrieval method based on deep multitasking representation learning | |
| CN115801463B (en) | Industrial Internet platform intrusion detection method and device and electronic equipment | |
| CN114330541A (en) | Road traffic accident risk prediction deep learning algorithm | |
| CN115016965A (en) | Method, device, equipment and storage medium for detecting faults of metering automation master station | |
| CN114615010B (en) | Edge server-side intrusion prevention system design method based on deep learning | |
| CN115905959A (en) | Method and device for analyzing relevance fault of power circuit breaker based on defect factor | |
| CN114881162A (en) | Method, apparatus, device and medium for predicting failure of metering automation master station | |
| CN114021425B (en) | Power system operation data modeling and feature selection method and device, electronic equipment and storage medium | |
| CN113541834B (en) | Abnormal signal semi-supervised classification method and system and data processing terminal | |
| CN114530210A (en) | Drug molecule screening method and system | |
| Mobtahej et al. | An lstm-autoencoder architecture for anomaly detection applied on compressors audio data | |
| CN117596191A (en) | Power Internet of things abnormality detection method, device and storage medium | |
| CN116400168A (en) | Power grid fault diagnosis method and system based on depth feature clustering | |
| CN115035966B (en) | Superconductor screening method, device and equipment based on active learning and symbolic regression | |
| CN115175192A (en) | Vehicle networking intrusion detection method based on graph neural network | |
| CN114037051A (en) | Deep learning model compression method based on decision boundary | |
| CN114202746A (en) | Road surface state identification method and device, terminal equipment and storage medium | |
| CN112926269A (en) | Method and system for grouping and cleaning data of edge nodes of power plant |
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 |