Fault analysis method and device of weak current system, computer equipment and storage medium

Abstract

The invention discloses a fault analysis method and device for a weak current system, computer equipment and a storage medium, and the method comprises the steps: obtaining real-time data and historical data of the weak current system; inputting the real-time data and the historical data into a trained fault analysis model to obtain a fault analysis result; performing troubleshooting on the weak current system based on the fault analysis result; wherein the fault analysis model is established based on a deep learning algorithm and / or a support vector machine. Therefore, the accuracy and timeliness of fault diagnosis of the weak current system can be improved.

Description

Technical Field

[0001] The present application belongs to the field of computer technology, and in particular, relates to a fault analysis method and device for a weak current system, a computer device, and a storage medium. Background Art

[0002] In current intelligent buildings, the weak current system is an indispensable part, which includes multiple subsystems such as communication, monitoring, fire protection, and security.

[0003] Fault diagnosis is an important part of the current weak current system management. Existing weak current system fault diagnosis technologies mostly rely on manual experience or simple threshold judgments, which often fail to effectively identify early potential faults in the system. As the complexity of weak current systems increases, traditional fault diagnosis methods face challenges in accuracy and timeliness.

[0004] Therefore, a new fault analysis strategy for weak current systems is urgently needed to solve the above problems. Summary of the invention

[0005] In order to overcome the above-mentioned defects, the present application proposes a new fault diagnosis strategy, which can improve the accuracy and timeliness of weak current system fault diagnosis.

[0006] To solve the above problems, in a first aspect, the present application provides a fault analysis method for a weak current system, the method comprising: obtaining real-time data and historical data of the weak current system; inputting the real-time data and the historical data into a trained fault analysis model to obtain a fault analysis result; and troubleshooting the weak current system based on the fault analysis result; wherein the fault analysis model is constructed based on a deep learning algorithm and / or a support vector machine.

[0007] In an optional embodiment, the real-time data and the historical data are input into a trained fault analysis model to obtain a fault analysis result, including: extracting target features from the real-time data and the historical data through a trained deep learning model; inputting the target features into a trained fault analysis model to obtain a fault analysis result, wherein the fault analysis model is constructed based on a support vector machine.

[0008] In an optional embodiment, the deep learning model is trained based on a back propagation algorithm and a gradient descent method.

[0009] In an optional embodiment, the training steps of the deep learning model include: obtaining training data and preprocessing the training data to obtain intermediate data; dividing the intermediate data into a training set, a validation set and a test set; obtaining an initial model and randomly initializing the weights and biases of the initial model; training the initial model based on the training set, updating the weights and biases of the initial model, and obtaining the deep learning model; evaluating the performance of the deep learning model through the validation set; and evaluating the generalization ability of the deep learning model through the test set.

[0010] In an optional embodiment, the optimization problem of the support vector machine is solved by a sequential minimal optimization (SMO) algorithm.

[0011] In an optional embodiment, the nonlinear problem of the support vector machine is processed by a kernel function; wherein the formula of the kernel function is as follows:

[0012]

[0013] Among them, K(x i ,x) is the kernel function; α i is the Lagrange multiplier; y i is the category label; b is the bias term, i takes the value of 1, 2, …n; n is a natural number greater than or equal to 1.

[0014] In an optional embodiment, before acquiring the real-time data, the method further includes: acquiring initial data of the weak current system; and preprocessing the initial data to obtain the real-time data.

[0015] In a second aspect, the present application also provides a fault analysis device for a weak current system, the device comprising: a data acquisition module for acquiring real-time data and historical data of the weak current system; a model analysis module for inputting the real-time data and the historical data into a trained fault analysis model to obtain a fault analysis result; a fault troubleshooting module for troubleshooting the weak current system based on the fault analysis result; wherein the fault analysis model is constructed based on a deep learning algorithm and / or a support vector machine.

[0016] In a third aspect, the present application further provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the program can implement the steps of any one of the above methods when executed by the processor.

[0017] In a fourth aspect, the present application also provides a storage medium for storing a computer program, wherein the program implements the steps of any of the methods described above when executed by a computer or a processor.

[0018] Compared with the prior art, the technical solution of the embodiment of the present application has the following beneficial effects:

[0019] The fault analysis method of the weak current system of the embodiment of the present application can improve the accuracy and timeliness of fault diagnosis of the weak current system.

[0020] Specifically, it is possible to obtain the real-time data collected by the current sensor and the historical data stored in the weak current system over a period of time, and analyze the current fault status of the weak current system based on these data. During the analysis process, the target features that reflect the fault status can be mined and extracted from a large amount of data, and then the target features can be analyzed based on the deep learning algorithm or SVM to obtain the fault analysis results. The deep learning algorithm has the ability to automatically extract features and is suitable for the recognition of complex patterns. It can include a large number of feature dimensions contained in the target features, while SVM has excellent generalization capabilities and performs outstandingly in the field of fault diagnosis. Therefore, using them separately or in combination can obtain a highly accurate and timely fault diagnosis model. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A schematic flow chart of a first method for analyzing a fault in a weak current system according to an embodiment of the present application;

[0022] Figure 2 A schematic flow chart of a second method for analyzing a fault in a weak current system according to an embodiment of the present application;

[0023] Figure 3 A schematic diagram of the structure of a data management device for a weak current system according to an embodiment of the present application;

[0024] Figure 4 A structural schematic diagram of a fault analysis method for a weak current system according to an embodiment of the present application. DETAILED DESCRIPTION

[0025] As mentioned in the background art, existing weak current systems face challenges in terms of accuracy and timeliness of fault diagnosis.

[0026] Specifically, although big data analysis and machine learning technologies have been widely used in many fields, their application in the field of weak current system fault diagnosis is still insufficient. The main problems existing in the existing technology include incomplete fault data collection, difficulty in fault feature extraction, low algorithm training efficiency, and low fault prediction accuracy. For example, traditional threshold judgment methods may not be able to identify systemic problems behind small voltage fluctuations, which may indicate an impending serious fault. In addition, since the amount of data in weak current systems is usually large, how to extract useful information from massive data has become a technical problem.

[0027] In order to solve the above problems, the purpose of this application is to provide a fault analysis method and device, computer equipment, and storage medium for a weak current system, which can comprehensively collect the operating data of the weak current system and use efficient data mining and machine learning algorithms to perform early detection and accurate diagnosis of potential system faults. This technology aims to improve the operating stability of the weak current system, reduce the occurrence of failures, reduce maintenance costs, and improve the overall management level of the system. Through this technology, early warning of weak current system failures can be achieved, thereby avoiding equipment damage and production interruptions caused by failures, which is of great significance for improving the continuity and reliability of industrial production.

[0028] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a description of the weak current system proposed by the present invention in combination with the accompanying drawings and preferred embodiments.

[0029] The specific implementation methods and functions of the fault analysis method, device, equipment, and storage medium are described in detail below.

[0030] In a first aspect, the present application provides a fault analysis method for a weak current system, which can be executed by the weak current system, and specifically by a controller arranged in the weak current system. In an optional embodiment, the weak current system can be arranged in an intelligent building to manage and control various devices in the intelligent building.

[0031] See also Figure 1 , Figure 1 This is a flow chart of a fault analysis method for a weak current system of the present application. The method may mainly include the following steps S101 to S103, and each step is described in detail as follows.

[0032] Step S101, acquiring real-time data of the weak current system, and reading historical data corresponding to the real-time data from a storage node.

[0033] Among them, real-time data is the data of the weak current system currently collected, which can better represent the current fault status of the weak current system. Generally speaking, the data format of real-time data may be relatively messy and the information therein may be relatively scattered.

[0034] In an alternative embodiment, Figure 1 Before obtaining the real-time data in step S101, the method further includes: obtaining initial data of the weak current system; and preprocessing the initial data to obtain the real-time data.

[0035] The initial data is data collected by the sensor, and the initial data may be preprocessed before being used. The preprocessing includes at least one of the following: removing noise data; removing erroneous data; and standardizing the initial data.

[0036] Historical data refers to the data collected by the weak current system in the past. It is generally processed and stored in one or more storage nodes. Compared with real-time data, the format of historical data is generally more orderly, and historical data usually excludes some data of little use, so the information is more concentrated.

[0037] In the weak current system, multiple sensors of different types can be set up, each sensor collects data according to the set sampling period and sends the collected data to the data processing center. The data processing center processes and analyzes the collected data and then stores it in one or more storage nodes.

[0038] When performing fault analysis on weak current systems, it is necessary to obtain both the real-time data collected by the current sensors and the historical data collected over a period of time. Figure 1 The device of the method can obtain real-time data collected by sensors and can also call historical data stored in storage nodes.

[0039] In an optional embodiment, after acquiring real-time data and historical data, these data can also be preprocessed, which includes cleaning and formatting the data. During the data cleaning process, noise and outliers can be removed by designing a filtering algorithm, which may mislead the fault diagnosis results. Data preprocessing also includes filling in missing values ​​and normalization processing to ensure the quality and consistency of the data. In this way, data preparation can be performed for accurately extracting target features.

[0040] Step S102: extracting target features from the real-time data and the historical data.

[0041] The target features are features related to the fault status of the weak current system, such as voltage, current, rate of change of voltage, etc.

[0042] In an optional embodiment, the target features may be extracted by a feature extraction algorithm, wherein the feature extraction algorithm may include one or more of principal component analysis (PCA), linear discriminant analysis (LDA), or an autoencoder.

[0043] In another optional embodiment, the target features may also be extracted through a big data model or a neural network model, for example, through a deep learning model.

[0044] Step S103: input the target feature into a fault analysis model to obtain a fault analysis result.

[0045] Step S104: troubleshoot the weak current system based on the fault analysis result.

[0046] Wherein, the fault analysis model is constructed based on a deep learning algorithm and / or a support vector machine (Support Vector Machine, referred to as SVM).

[0047] pass Figure 1 The fault analysis method can obtain the real-time data collected by the current sensor and the historical data stored in the weak current system for a period of time in the past, and analyze the current fault status of the weak current system based on these data. In the analysis process, the target features that reflect the fault status can be mined and extracted from a large amount of data, and then the target features can be analyzed based on the deep learning algorithm or SVM to obtain the fault analysis results. The deep learning algorithm has the ability to automatically extract features and is suitable for the recognition of complex patterns. It can contain a large number of feature dimensions of the target features, while SVM has excellent generalization ability and performs outstandingly in the field of fault diagnosis. Therefore, using them separately or in combination can obtain a highly accurate and timely fault diagnosis model.

[0048] In a specific embodiment of the first fault analysis model, Figure 1 In step S103, the fault analysis model is constructed based on a deep learning algorithm.

[0049] Optionally, the training steps of the deep learning model include: obtaining training data, and preprocessing the training data to obtain intermediate data; dividing the intermediate data into a training set, a validation set, and a test set; obtaining an initial model, and randomly initializing the weights and biases of the initial model; training the initial model based on the training set, updating the weights and biases of the initial model, and obtaining the deep learning model; evaluating the performance of the deep learning model through the validation set; and evaluating the generalization ability of the deep learning model through the test set. Among them, the training set is used to train the model, the validation set is used to adjust the model parameters, and the test set is used to evaluate the generalization ability of the model.

[0050] Optionally, the deep learning model is trained based on a back propagation algorithm and a gradient descent method.

[0051] Among them, in deep learning model training, back-propagation algorithm (Back-propagation) and gradient descent method (Gradient Descent) are usually used in combination. The following are the specific training steps combining these two algorithms:

[0052] Step 1: Initialize model parameters. Before training the initial model, randomly initialize the model's weights and biases.

[0053] Step 2: Data preprocessing: The training data can be divided into a training set, a validation set, and a test set, and the data in the training set, validation set, and test set can be subjected to preprocessing steps such as standardization or normalization; alternatively, the training data can be preprocessed to obtain intermediate data, and then the intermediate data can be divided into a training set, a validation set, and a test set.

[0054] Step 3, forward propagation. Input a training sample into the initial model. The initial model includes multiple layers, the first layer is the input layer, the last layer is the output layer, and the others are hidden layers. The initial model calculates the output of each layer through the linear transformation and nonlinear activation function of each layer. Until the last layer, the predicted value is obtained. Preferably, each hidden layer can use ReLU as the activation function to increase the nonlinearity of the model.

[0055] Step 4: Loss Computation: Use a loss function (such as mean square error, cross entropy, etc.) to calculate the difference between the predicted value and the true value to get the loss value.

[0056] Step 5, backpropagation. Starting from the output layer, calculate the gradient of each layer in reverse. Using the chain rule, pass the gradient of the previous layer to the next layer and calculate the gradient of the weight and bias of each layer.

[0057] The specific steps are as follows: calculate the gradient of the loss function with respect to the output of each layer; calculate the gradient of the activation function of each layer; and calculate the gradient of the weights of each layer according to the chain rule.

[0058] Step 6, Gradient Descent. Use the calculated gradient to update the weights and biases of the model. The learning rate learning_rate determines the step size of the updated parameters. For example, updating the weights and biases of each layer can be expressed as weight1=weight0-learning_rate×gradient. Weight1 is the updated weight, weight0 is the weight before the update, and gradient is the gradient.

[0059] Step 7: Repeat steps 3 to 6 above. Use all samples in the training set for multiple iterations (epochs), and perform forward propagation, loss calculation, backpropagation, and gradient descent in each iteration.

[0060] Step 8, validation and testing. After each iteration or after a certain number of iterations, use the validation set to evaluate the model performance and adjust the hyperparameters (such as learning rate, network structure, etc.). After training is completed, use the test set to evaluate the generalization ability of the model.

[0061] Step 9: Convergence judgment: Determine whether the model has converged based on the performance of the validation set, or set an early termination condition (such as the validation set performance does not improve for n consecutive iterations).

[0062] Step 10: Save the model. After the training is completed, save the trained model parameters for subsequent use or deployment, thereby obtaining a trained deep learning model.

[0063] Through the above steps, combined with the back-propagation algorithm and gradient descent method, the deep learning model can be effectively trained.

[0064] Furthermore, in order to speed up the convergence of the deep learning model and improve the training effect, the momentum term described by the following mathematical formula can be used:

[0065]

[0066] θ t =θ t―1 +v t

[0067] Among them, v t is the momentum term at time step t, μ is the momentum coefficient, η is the learning rate, J(θ t―1 ) is θ t―1 The corresponding loss function is, is the gradient of the loss function with respect to the parameters θ, where θ is the weight and bias.

[0068] In the first embodiment of the fault analysis model, a suitable deep learning model architecture is first selected according to the data characteristics of the target feature, and the parameters of the model are initialized. The model is then trained using the training set, while the performance on the validation set is monitored to adjust the learning rate and model structure. During the training process, a momentum term can be applied to accelerate the gradient descent, and the performance of the model on the test set can be regularly checked to evaluate its generalization ability. In this way, the fault analysis model can be trained quickly and accurately.

[0069] In a specific embodiment of the second fault analysis model, Figure 1 In step S103, the fault analysis model is constructed based on SVM.

[0070] Optionally, the optimization problem of the support vector machine (SVM) is solved by a sequential minimal optimization (SMO) algorithm.

[0071] Optionally, the nonlinear problem of the support vector machine is processed by a kernel function; wherein the formula of the kernel function is as follows:

[0072]

[0073] Among them, K(x i ,x) is the kernel function; α i is the Lagrange multiplier; y i is the category label; b is the bias term, i takes the value of 1, 2, …n; n is a natural number greater than or equal to 1.

[0074] The kernel function maps the input data to a high-dimensional feature space, making the originally linearly inseparable data linearly separable in the new space. This mapping is implicit and does not require explicit definition of the transformed feature space. This can handle nonlinear problems in fault analysis well.

[0075] In the second embodiment of the fault analysis model, for the SVM model, the optimal Lagrange multiplier can be found through the SMO algorithm, and the type and parameters of the kernel function can be determined, thereby obtaining an optimized fault analysis model that can accurately predict faults in weak current systems.

[0076] In a specific embodiment of the third fault analysis model, Figure 1 The fault analysis model in step S103 is constructed based on the deep learning algorithm and SVM.

[0077] For example, two fault analysis models can be constructed, the first fault analysis model is constructed based on a deep learning algorithm, and the second fault analysis model is constructed based on an SVM. The target features are input into the first fault analysis model and the second fault analysis model respectively, and two fault analysis results are obtained respectively. The two fault analysis results are comprehensively analyzed to obtain the final fault analysis result.

[0078] For another example, deep learning algorithms and SVMs can be integrated to build a fault analysis model.

[0079] Since the target features of weak-current circuits usually do not follow linear relationships and may have complex nonlinear mapping relationships, in practical applications, in the real-time and historical data of weak-current systems, target features related to faults are often difficult to obtain. In particular, target features usually only account for a small part of the overall data, and the data in normal states is usually far more than the data in fault states, resulting in an unbalanced data set. Due to the influence of measurement equipment or environmental factors, the data may contain noise and outliers. Based on these differences in the characteristics of weak-current circuits, we chose deep learning and support vector machines (SVM) as the main training algorithms.

[0080] In this embodiment, the deep learning algorithm is suitable for the recognition of complex patterns due to its ability to automatically extract features. It can better extract features from a large number of feature dimensions contained in the target features. These features may come from various sensors such as voltage, current, temperature, vibration, etc. At the same time, SVM performs outstandingly in the field of fault diagnosis due to its excellent generalization ability. Therefore, using them in combination can obtain a highly accurate fault diagnosis model.

[0081] In an optional embodiment, before inputting the target features into the above three fault analysis models, Figure 1 Step S102 extracts target features from the real-time data and the historical data, which may include: identifying and extracting target features through a feature extraction algorithm, wherein the feature extraction algorithm may include one or more of PCA, LDA or autoencoder.

[0082] Feature extraction algorithms can extract the most representative features from high-dimensional data. The following provides a specific process of feature extraction: perform statistical analysis on real-time data and historical data to determine the distribution characteristics of the data; apply feature extraction algorithms, select appropriate parameters based on the distribution characteristics of the data, and extract key features.

[0083] Feature extraction algorithms can perform in-depth analysis of data and identify key patterns in the data, which are closely related to the fault status of the weak current system. Specifically, feature extraction algorithms can be used to calculate the main components of the data and select the components with the highest contribution rate as target features. These target features can effectively represent the fault information in real-time data and historical data. The following is a mathematical description of feature extraction:

[0084] Let X be the original data matrix, X T is the transposed matrix of X, k is the number of samples in the matrix X, and the covariance matrix Σ can be expressed as follows:

[0085]

[0086] Perform eigenvalue decomposition on Σ to obtain eigenvalues ​​and eigenvectors, and select several eigenvectors with the largest eigenvalues ​​to form the matrix of the target features.

[0087] In an alternative embodiment, Figure 1 Step S102 in the method includes: extracting target features from the real-time data and the historical data through a deep learning algorithm; and the fault analysis model in step S103 is constructed based on a support vector machine (SVM).

[0088] For details, see Figure 2 , Figure 2A flow chart of a second method for analyzing a fault in a weak current system of the present application is provided. The method may include:

[0089] Step S101, acquiring real-time data of the weak current system, and reading historical data corresponding to the real-time data from a storage node;

[0090] Step S202, extracting target features from the real-time data and the historical data by using a deep learning algorithm;

[0091] Step S203, inputting the target feature into a fault analysis model to obtain a fault analysis result, wherein the fault analysis model is constructed based on a support vector machine;

[0092] Step S104: troubleshoot the weak current system based on the fault analysis result.

[0093] In this embodiment, since the real-time data and historical data of the weak current system contain a large number of data types and these features contain a large number of feature dimensions, the model training of the deep learning algorithm can be used to adaptively extract features from these data to improve the accuracy of feature extraction. The training steps of the deep learning model are similar to the training method of the first fault analysis model, which will not be repeated here. After obtaining the target features, they are input into the SVM, which can accurately and quickly perform fault diagnosis through the good generalization ability of the SVM. The specific scheme of the SVM can refer to the relevant description of the second fault analysis model, which will not be repeated here.

[0094] See also Figure 3 , Figure 3 The data management device 30 of the weak current system (hereinafter referred to as the device 30 ) may include a sensor 301 , a data acquisition device 302 , a data processing center 303 , a first storage node 304 and a second storage node 305 .

[0095] Initial data is collected by sensors 301 set at multiple nodes of the weak current system, and the initial data is transmitted to the data acquisition device 302; the data acquisition device 302 performs a first preprocessing on the initial data to obtain target data, and the target data is sent to the data processing center 303. The corresponding copy strategy is executed by each first storage node 304, and the data fragment is copied to the second storage node 305, so that the second storage node 305 stores the data fragment.

[0096] Among them, the sensor 301 may include at least one of the following: a temperature sensor, a humidity sensor, a light sensor, a personnel detection sensor, a voltage sensor, a current sensor, a power sensor, etc. Multiple sensors 301 are carefully deployed at each key node of the weak current system to ensure that the sensor can comprehensively and uninterruptedly collect the status information in the weak current system. The initial data is the unprocessed data collected by each sensor 301, such as the voltage signal collected by the voltage sensor, the current signal collected by the current sensor, etc.

[0097] In an optional embodiment, the sensor 301 and the data acquisition device 302 convert the acquired parameters, such as current, voltage, power, etc., into digital signals. Furthermore, the conversion of digital signals can be achieved through an analog-to-digital converter (ADC).

[0098] In an optional embodiment, the data acquisition device 302 performs a first preprocessing on the initial data (including preliminary cleaning and standardization) to ensure the quality and consistency of the data, thereby removing noise and erroneous data in the initial data and reducing the amount of data for subsequent processing. The first preprocessing unifies the initial data into the same data format, which can also reduce the computational pressure of subsequent data processing steps, and can also ensure the real-time and accuracy of the collected data, laying a solid foundation for subsequent data analysis and visualization.

[0099] Optionally, the initial data is standardized according to the following formula:

[0100]

[0101] Among them, x is the original data, x normalized is the data after X is standardized, μ is the mean of the data set, and σ is the standard deviation of the data set.

[0102] In an optional embodiment, each sensor 301 is equipped with a data acquisition device 302, and the data acquisition device 302 can pre-process the initial data collected by the sensor 301 in real time, thereby effectively improving the efficiency of the pre-processing.

[0103] In another optional embodiment, multiple sensors 301 in each area are equipped with a data acquisition device 302, and the data acquisition device 302 is responsible for preprocessing the initial data collected by the multiple sensors 301 in the area. This can reduce the layout cost of the data acquisition device. In addition, multiple sensors 301 and data acquisition devices 302 in the same area can communicate based on the intranet of the area (which can be a wired network or a wireless network) to ensure the stability of communication.

[0104] In an optional embodiment, the data processing center 303 includes a buffer zone, which can be used to store the target data and perform a second preprocessing on the target data through the buffer zone.

[0105] Specifically, after the data processing center 303 obtains the target data sent by the multiple data acquisition devices 302, it first sends the data to the data buffer, which is responsible for temporarily storing the data and performing the second preprocessing. The target data after the second preprocessing is allocated to different data partitions and transmitted to each first storage node 304 through the network.

[0106] The second preprocessing includes but is not limited to: unifying the format of the target data; fusing and summarizing the target data to form a unified data set.

[0107] In an optional embodiment, after the data processing center 303 integrates and summarizes the acquired target data, it can perform data partitioning, that is, the data processing center 303 can divide the large-scale target data into multiple small blocks (that is, data fragments), and these data fragments are distributedly stored on multiple first storage nodes 304 to improve data reading and writing efficiency.

[0108] In an optional embodiment, after each first storage node 304 receives the data fragments sent by the data processing center 303, it executes a corresponding replication mechanism, that is, each first storage node 304 creates one or more replicas for each data fragment, and stores these replicas in a distributed manner on one or more second storage nodes 305 to enhance the fault tolerance of the data, thereby completing the persistent storage of the data.

[0109] In the method described in the present application, stored data can be retrieved from the first storage node 304 and / or the second storage node 305 to obtain historical data.

[0110] Through the data management device 30 of the above-mentioned weak current system, the weak current system collects data through multiple internal sensors 301, and the data collection device 302 at the sensor 301 end performs preprocessing to reduce the calculation pressure of the subsequent data processing steps. Multiple data collection devices 302 send the preprocessed data to the data processing center 303, and the data processing center 303 centrally processes these data, and then performs data partitioning on the processed data. The data partitioning can divide the large-scale target data into small data fragments, and these data fragments can be distributedly processed to improve the read and write efficiency of the data. Then, these data fragments are scattered and stored on multiple first storage nodes 304, even in the case of failure of some storage nodes, the reliability and availability of other data can be guaranteed. Finally, the first storage node 304 also stores the copies of each data fragment in one or more second storage nodes 305 according to the set copy mechanism, which can further enhance the stability and fault tolerance of the data, thereby realizing the persistent storage of data.

[0111] It should be noted that the difference between weak current circuits and other circuits is that the voltage of weak current circuits is below 36V, the current is small, and the main function is to transmit signals, such as digital signals, analog signals, etc. These signals are sensitive to noise and require good signal integrity. Therefore, the big data models required for the weak current system circuits and other circuits, such as the civil 220V and factory 380V strong current circuits, which are mainly used for power transmission, are quite different. Compared with the high voltage and high current transmission mode of other strong current circuits, the weak current circuits are more susceptible to external electromagnetic interference (EMI) and radio frequency interference (RFI), so more attention will be paid to the interference part in the optimization of the model. The accuracy of the sensors used is also more refined.

[0112] See attached Figure 4 , Figure 4 : is a structural diagram of a fault analysis device 40 for a weak current system according to an embodiment of the present application. The fault analysis device 40 for a weak current system provided by this embodiment may mainly include:

[0113] The data acquisition module 401 is used to acquire the real-time data of the weak current system and read the historical data corresponding to the real-time data from the storage node;

[0114] A feature extraction module 402 is used to extract target features from the real-time data and the historical data;

[0115] The model analysis module 403 is used to obtain target features, input the target features into the fault analysis model, and obtain fault analysis results;

[0116] A fault troubleshooting module 403, used for troubleshooting the weak current system based on the fault analysis result;

[0117] Wherein, the fault analysis model is constructed based on a deep learning algorithm and / or a support vector machine.

[0118] In an optional embodiment, target features are extracted from the real-time data and the historical data by a deep learning algorithm, and the fault analysis model is constructed based on a support vector machine.

[0119] In an optional embodiment, the deep learning model is trained based on a back propagation algorithm and a gradient descent method.

[0120] In an optional embodiment, the fault analysis device 40 of the weak current system may further include a training module of a deep learning model, the module including:

[0121] A training data acquisition unit, used to acquire training data and preprocess the training data to obtain intermediate data;

[0122] A data set division unit, used for dividing the intermediate data into a training set, a validation set and a test set;

[0123] A model initialization unit, used to obtain an initial model and randomly initialize the weights and biases of the initial model;

[0124] A model training unit, used to train the initial model based on the training set, update the weights and biases of the initial model, and obtain the deep learning model;

[0125] A verification unit, used to evaluate the performance of the deep learning model through the verification set;

[0126] A testing unit is used to evaluate the generalization ability of the deep learning model through the test set.

[0127] In an optional embodiment, the optimization problem of the support vector machine is solved by a sequential minimal optimization (SMO) algorithm.

[0128] In an optional embodiment, the nonlinear problem of the support vector machine is processed by a kernel function;

[0129] The formula of the kernel function is as follows:

[0130]

[0131] Among them, K(x i ,x) is the kernel function; α i is the Lagrange multiplier; y i is the category label; b is the bias term, i takes the value of 1, 2, …n; n is a natural number greater than or equal to 1.

[0132] In an optional embodiment, before the data acquisition module, the fault analysis device 40 of the weak current system further includes: an initial data acquisition module for acquiring initial data of the weak current system; and a preprocessing module for preprocessing the initial data to obtain the real-time data.

[0133] For more details about the working principle and working mode of the fault analysis device 40 for the weak current system, please refer to the relevant description about the fault analysis method for the weak current system, which will not be repeated here.

[0134] The embodiment of the present application also provides a storage medium, which can be a computer-readable storage medium for storing a computer program, wherein the program is executed by a computer or a processor in any of the steps of the above-mentioned weak current system fault analysis method. The computer-readable storage medium may include a non-volatile memory or a non-transitory memory, and may also include an optical disk, a mechanical hard disk, a solid-state hard disk, etc.

[0135] An embodiment of the present application also provides a computer device, which may include a computer device including a memory and a processor, wherein the memory stores a computer program, and when the program is executed by the processor, it can implement the steps of any one of the above-mentioned weak current system fault analysis methods.

[0136] In the embodiments of the present application, the processor may be a general-purpose processor, a digital signal processor, an application-specific integrated circuit, a field programmable gate array or other programmable logic device, a discrete gate or transistor logic device, or a discrete hardware component, and may implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of the present application. The general-purpose processor may be a microprocessor or any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application may be directly embodied as being executed by a hardware processor, or may be executed by a combination of hardware and software modules in the processor.

[0137] In the embodiment of the present application, the memory may be a non-volatile memory, such as a hard disk drive (HDD) or a solid-state drive (SSD), etc., or a volatile memory (volatile memory), such as a random-access memory (RAM). The memory is any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory in the embodiment of the present application may also be a circuit or any other device that can realize a storage function, for storing computer programs and / or data.

[0138] In the fault analysis method of the weak current system provided in the embodiment of the present application, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program is loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, a network device, a user device or other programmable device. The computer program may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instruction may be transmitted from a website site, a computer, a server or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server or a data center that includes one or more available media integrated. The available medium may be a magnetic medium (eg, a floppy disk, a hard disk, a magnetic tape), an optical medium (eg, a digital video disc (DVD)), or a semiconductor medium (eg, an SSD).

[0139] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

[0140] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0141] In the above embodiments, the description of each embodiment has its own emphasis, and any multiple embodiments can be used in combination. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0142] In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in a processor or an instruction in the form of software. The steps of the method disclosed in conjunction with the embodiment of the present application can be directly embodied as a hardware processor for execution, or a combination of hardware and software units in a processor for execution. The software unit can be located in a storage medium mature in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in a memory, and the processor executes the instructions in the memory, and completes the steps of the above method in conjunction with its hardware. To avoid repetition, it is not described in detail here.

[0143] In the embodiment of the present application, the processor of the above-mentioned device may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0144] The present application also provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to execute some or all of the steps of any method described in the above method embodiment. The computer program product may be a software installation package.

[0145] In the several embodiments provided in the present application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of the above-mentioned units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0146] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiments of the present application.

[0147] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0148] If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or all or part of the technical solution, can be embodied in the form of a software product, which is stored in a memory and includes several instructions for a computer device (which can be a personal computer, server or TRP, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned memory includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk and other media that can store program codes.

[0149] It should be understood that the term "and / or" in this article is only a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article indicates that the associated objects before and after are in an "or" relationship.

[0150] The "plurality" appearing in the embodiments of the present application refers to two or more.

[0151] The first, second, etc. descriptions appearing in the embodiments of the present application are only used for illustration and distinction of the description objects. There is no order, nor do they indicate any special limitation on the number of devices in the embodiments of the present application, and cannot constitute any limitation on the embodiments of the present application.

[0152] The "connection" that appears in the embodiments of the present application refers to various connection methods such as direct connection or indirect connection to achieve communication between devices, and the embodiments of the present application do not impose any limitations on this.

[0153] The embodiments of the present application are introduced in detail above. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method and core idea of ​​the present application. At the same time, for general technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A fault analysis method for a weak current system, characterized in that: The method comprises: Acquire the real-time data of the weak current system, and read the historical data corresponding to the real-time data from the storage node; Extracting target features from the real-time data and the historical data; Inputting the target feature into a fault analysis model to obtain a fault analysis result; Performing fault troubleshooting on the weak current system based on the fault analysis result; Wherein, the fault analysis model is constructed based on a deep learning algorithm and / or a support vector machine.

2. The method according to claim 1, characterized in that Target features are extracted from the real-time data and the historical data by a deep learning algorithm, and the fault analysis model is constructed based on a support vector machine.

3. The method according to claim 1 or 2, characterized in that: The deep learning model is trained based on the back propagation algorithm and the gradient descent method.

4. The method according to claim 1 or 2, characterized in that: The training steps of the deep learning model include: Acquiring training data, and preprocessing the training data to obtain intermediate data; Dividing the intermediate data into a training set, a validation set and a test set; Obtain an initial model, and randomly initialize the weights and biases of the initial model; Training the initial model based on the training set, updating the weights and biases of the initial model, and obtaining the deep learning model; Evaluating the performance of the deep learning model using the validation set; The generalization ability of the deep learning model is evaluated by the test set.

5. The method according to claim 1 or 2, characterized in that: The optimization problem of the support vector machine is solved by a sequential minimal optimization (SMO) algorithm.

6. The method according to claim 1 or 2, characterized in that: Processing the nonlinear problem of the support vector machine by using a kernel function; The formula of the kernel function is as follows: Among them, K(x i ,x) is the kernel function; α i is the Lagrange multiplier; y i is the category label; b is the bias term, i takes the value of 1, 2, …n; n is a natural number greater than or equal to 1.

7. The method according to claim 1, characterized in that Before acquiring the real-time data, the method further includes: acquiring initial data of the weak current system; The initial data is preprocessed to obtain the real-time data.

8. A fault analysis device for a weak current system, characterized in that: The device comprises: A data acquisition module, used to acquire real-time data of the weak current system and read historical data corresponding to the real-time data from a storage node; A feature extraction module, used to extract target features from the real-time data and the historical data; A model analysis module, used for inputting the target features into a fault analysis model to obtain a fault analysis result; A fault troubleshooting module is used to perform fault troubleshooting on the weak current system based on the fault analysis results; wherein the fault analysis model is constructed based on a deep learning algorithm and / or a support vector machine.

9. A computer device comprising a processor and a storage device, wherein the storage device is suitable for storing a plurality of program codes, characterized in that: The program code is suitable for being loaded and executed by the processor to perform the method of any one of claims 1 to 7.

10. A storage medium storing a plurality of program codes, characterized in that: The program code is suitable for being loaded and executed by a processor to execute the method according to any one of claims 1 to 7.

Images