Substation operation state monitoring method and system
By constructing a substation operating status monitoring model based on CNN, LSTM and SVM, the problem of insufficient reliability and accuracy of substation status monitoring in the prior art is solved, and more efficient and accurate status monitoring is achieved.
Patent Information
- Application Number
- CN202411936169.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-26
- Publication Date
- 2025-04-25
AI Technical Summary
The existing substation status monitoring solutions are not reliable and have poor accuracy, and rely on manual monitoring and simple logical judgment.
The monitoring model based on CNN network, LSTM network and SVM classifier is adopted to obtain and clean historical operating status data, build and train the substation operating status monitoring model to achieve automated status monitoring.
It improves the reliability and accuracy of the operating status monitoring of the substation, and can more accurately predict and monitor the operating status of the substation.
Smart Images
Figure CN119884899A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the field of electrical automation, and in particular relates to a method and system for monitoring the operating status of a substation. Background Art
[0002] With the development of economy and technology and the improvement of people's living standards, electric energy has become an indispensable secondary energy source in people's production and life, bringing endless convenience to people's production and life. Therefore, ensuring the stable and reliable supply of electric energy has become one of the most important tasks of the power system. As an important part of the power system, the operating status of the substation is directly related to the safe and stable operation of the power system. Therefore, it is particularly important to monitor the status of the substation.
[0003] At present, the commonly used status monitoring scheme for substations is still the scheme of manual monitoring + simple logic judgment: the substation monitoring system obtains various operating data of the substation in real time and displays them in real time; at the same time, the substation staff manually monitors the operating status of the substation based on the current operating data; in addition, the substation monitoring system also uses simple logic judgment as the main method based on the current operating data, and realizes the monitoring of the substation status based on the relationship between important data such as real-time voltage, real-time current, real-time operating temperature, etc. and the set standards. However, this type of manual monitoring and simple logic judgment monitoring schemes both have the problems of low reliability and poor accuracy. Summary of the invention
[0004] One of the purposes of the present invention is to provide a method for monitoring the operation status of a substation with high reliability and good accuracy.
[0005] A second object of the present invention is to provide a system for implementing the substation operation status monitoring method.
[0006] The substation operation status monitoring method provided by the present invention comprises the following steps:
[0007] S1. Obtain historical operating status data information of the substation;
[0008] S2. Clean the data information obtained in step S1 to construct a training data set;
[0009] S3. Based on CNN network, LSTM network and SVM classifier, the initial model of substation operation status monitoring is constructed;
[0010] S4. Using the training data set obtained in step S2, the initial substation operation status monitoring model constructed in step S3 is trained to obtain a substation operation status monitoring model;
[0011] S5. Using the substation operation status monitoring model obtained in step S4, perform operation status monitoring of the target substation.
[0012] The data cleaning described in step S2 specifically includes the following steps:
[0013] Abnormal data identification: K-Means clustering method is used to perform cluster analysis on the data information obtained in step S1 to identify abnormal data; during the clustering process, Davies-Bouldin index is used to evaluate the effectiveness of the cluster analysis results and determine the optimal number of clusters;
[0014] Abnormal data deletion: Delete the identified abnormal data;
[0015] Data interpolation: EM interpolation and regression interpolation methods are used to calculate interpolation weights based on MAPE values to interpolate various data. When using EM interpolation for data interpolation, θ is used to continuously improve the interpretation function of the data until it converges to a set of parameters:
[0016]
[0017] Where θ (new) is the parameter value updated after one iteration of the EM interpolation method; θ is the probability distribution parameter of the data; n is the number of samples in the data set; X is the observed data; Z is the missing data; p(X,Z|θ) is the joint probability distribution of X and Z; E Z|X, θ[logp(X,Z|θ)] is the expected value of the log-likelihood of Z given X and parameter θ;
[0018] Data normalization: Use the following formula to normalize the data:
[0019]
[0020] In the formula, XX' is the normalized data; XX is the data before normalization; β is the weight and coefficient; b is the bias term, which is a constant used to adjust the output of the model.
[0021] The initial model for monitoring the operation status of a substation is constructed based on the CNN network, the LSTM network and the SVM classifier in step S3, which specifically includes the following steps:
[0022] The constructed initial model for substation operation status monitoring includes CNN module, LSTM module and SVM classifier module;
[0023] The CNN module, LSTM module and SVM classifier module are connected in series;
[0024] The CNN module is used to extract the features of the input substation operation data, obtain the corresponding feature map and upload it to the LSTM module;
[0025] The LSTM module is used to generate several serialized vectors based on the received feature maps and upload them to the SVM classifier module;
[0026] The SVM classifier module is used to classify the operating status of the substation according to the received serialized vector to achieve the prediction of the substation operating status.
[0027] The CNN module specifically includes the following steps:
[0028] The CNN module consists of two convolutional blocks;
[0029] Each convolution block includes a convolution layer, an activation function layer, and a maximum pooling layer;
[0030] The filter size of the convolution layer is 128, and the convolution kernel size is 3×3; the convolution layer is used to extract the features of the input data; the activation function layer uses the ReLU activation function to obtain the mapping feature matrix; the maximum pooling layer is used to process the obtained mapping feature matrix and obtain several one-dimensional feature maps.
[0031] The LSTM module specifically includes the following steps:
[0032] The LSTM module includes a fully connected layer and an LSTM layer;
[0033] The fully connected layer consists of 256 neurons;
[0034] The number of hidden nodes in the LSTM layer is 256.
[0035] The SVM classifier module specifically comprises the following steps:
[0036] The following formula is used as the decision function f(x) of the SVM classifier module:
[0037]
[0038] Where sign() is the sign function; n is the number of training samples; α i is the Lagrange multiplier, which is used to indicate the contribution of the i-th training sample to the decision boundary; y i is the category label of the i-th training sample; x i is the feature vector of the i-th training sample; K(x i ,x) is the kernel function, which is used to calculate the i-th training sample x i The inner product of the new data point x in the feature space; b is the bias term, which is used to indicate the location of the decision boundary.
[0039] The present invention also provides a system for implementing the substation operation status monitoring method, comprising a data acquisition module, a data cleaning module, a model construction module, a model training module and a status monitoring module; the data acquisition module, the data cleaning module, the model construction module, the model training module and the status monitoring module are connected in series in sequence; the data acquisition module is used to acquire the historical operation status data information of the substation, and upload the data information to the data cleaning module; the data cleaning module is used to perform data cleaning on the acquired data information according to the received data information to construct a training data set, and upload the data information to the model construction module; the model construction module is used to construct an initial model for substation operation status monitoring based on a CNN network, an LSTM network and an SVM classifier according to the received data information, and upload the data information to the model training module; the model training module is used to train the constructed initial model for substation operation status monitoring according to the received data information, using the obtained training data set, to obtain a substation operation status monitoring model, and upload the data information to the status monitoring module; the status monitoring module is used to perform operation status monitoring of the target substation according to the received data information, using the obtained substation operation status monitoring model.
[0040] The substation operation status monitoring method and system provided by the present invention construct a status monitoring model based on a CNN network, an LSTM network and an SVM classifier, and implement the training of the monitoring model based on historical data; therefore, the scheme of the present invention can not only realize the monitoring of the substation operation status, but also has higher reliability and better accuracy. BRIEF DESCRIPTION OF THE DRAWINGS
[0041] Figure 1 The figure is a schematic diagram of the method flow of the present invention.
[0042] Figure 2 Schematic diagram of the functional modules of the system of the present invention. DETAILED DESCRIPTION
[0043] like Figure 1 The method flow chart of the method of the present invention is shown as follows: The substation operation status monitoring method disclosed in the present invention comprises the following steps:
[0044] S1. Obtain historical operating status data information of the substation;
[0045] S2. Clean the data information obtained in step S1 to construct a training data set; specifically, the steps include:
[0046] Abnormal data identification: abnormal data collection will occur in abnormal situations such as abnormal sensor power supply and power surge; therefore, the K-Means clustering method is used to perform cluster analysis on the data information obtained in step S1 to identify abnormal data; in the clustering process, the Davies-Bouldin index is used to evaluate the effectiveness of the cluster analysis results, determine the optimal number of clusters, and solve the problem of selecting the number of clusters K;
[0047] Abnormal data deletion: Delete the identified abnormal data;
[0048] Data interpolation: EM interpolation and regression interpolation methods are used to calculate interpolation weights based on MAPE values to interpolate various data items. The EM algorithm helps estimate missing data by iteratively using observed data. When using the EM interpolation method for data interpolation, θ is used to continuously improve the explanatory function of the data until it converges to a set of parameters that can best describe the observed data and the expectations of the missing data:
[0049]
[0050] Where θ (new) is the parameter value updated after one iteration of the EM interpolation method; θ is the probability distribution parameter of voltage, current, temperature and other data; n is the number of samples in the data set; X is the observed data; Z is the missing data; p(X, Z|θ) is the joint probability distribution of X and Z; E Z|X, θ[logp(X,Z|θ)] is the expected value of the log-likelihood of Z given X and parameter θ;
[0051] Data normalization: Use the following formula to normalize the data:
[0052]
[0053] Where XX' is the normalized data; XX is the data before normalization; β is the weight and coefficient; b is the bias term, which is a constant used to adjust the output of the model;
[0054] S3. Based on CNN network, LSTM network and SVM classifier, the initial model of substation operation status monitoring is constructed; the specific steps include:
[0055] The constructed initial model for substation operation status monitoring includes CNN module, LSTM module and SVM classifier module;
[0056] The CNN module, LSTM module and SVM classifier module are connected in series;
[0057] The CNN module is used to extract the features of the input substation operation data, obtain the corresponding feature map and upload it to the LSTM module;
[0058] The LSTM module is used to generate several serialized vectors based on the received feature maps and upload them to the SVM classifier module;
[0059] The SVM classifier module is used to classify the operating status of the substation according to the received serialized vector to achieve the prediction of the substation operating status;
[0060] When implementing:
[0061] The CNN module consists of two convolutional blocks;
[0062] Each convolution block includes a convolution layer, an activation function layer, and a maximum pooling layer;
[0063] The filter size of the convolution layer is 128, and the convolution kernel size is 3×3; the convolution layer is used to extract the features of the input data; the activation function layer uses the ReLU activation function to obtain the mapping feature matrix; the maximum pooling layer is used to process the obtained mapping feature matrix and obtain several one-dimensional feature maps;
[0064] The LSTM module includes a fully connected layer and an LSTM layer;
[0065] The fully connected layer consists of 256 neurons;
[0066] The number of hidden nodes in the LSTM layer is 256.
[0067] The SVM classifier module includes the following steps:
[0068] The SVM classifier learning process is performed using the Optuna framework, and the optimal SVM kernel function K(x i ,x), bias term b and other parameters; the input of SVM is the serialized feature vector, and after the learning process, the parameters are optimized to obtain the trained SVM classifier;
[0069] The following formula is used as the decision function f(x) of the SVM classifier module:
[0070]
[0071] Where sign() is the sign function; n is the number of training samples; α i is the Lagrange multiplier, which is used to indicate the contribution of the i-th training sample to the decision boundary; y i is the category label of the i-th training sample; x i is the feature vector of the i-th training sample; K(x i,x) is the kernel function, which is used to calculate the i-th training sample x i The inner product of the new data point x in the feature space; b is the bias term, which is used to indicate the location of the decision boundary;
[0072] S4. Using the training data set obtained in step S2, the initial substation operation status monitoring model constructed in step S3 is trained to obtain a substation operation status monitoring model;
[0073] S5. Using the substation operation status monitoring model obtained in step S4, perform operation status monitoring of the target substation.
[0074] like Figure 2 The functional module schematic diagram of the system of the present invention is shown as follows: the system for implementing the substation operation status monitoring method disclosed in the present invention comprises a data acquisition module, a data cleaning module, a model construction module, a model training module and a status monitoring module; the data acquisition module, the data cleaning module, the model construction module, the model training module and the status monitoring module are connected in series in sequence; the data acquisition module is used to acquire the historical operation status data information of the substation, and upload the data information to the data cleaning module; the data cleaning module is used to perform data cleaning on the acquired data information according to the received data information to construct a training data set, and upload the data information to the model construction module; the model construction module is used to construct an initial model for substation operation status monitoring based on a CNN network, an LSTM network and an SVM classifier according to the received data information, and upload the data information to the model training module; the model training module is used to train the constructed initial model for substation operation status monitoring according to the received data information, using the obtained training data set, to obtain a substation operation status monitoring model, and upload the data information to the status monitoring module; the status monitoring module is used to perform operation status monitoring of the target substation according to the received data information, using the obtained substation operation status monitoring model.
Claims
1. A method for monitoring the operation status of a substation, comprising the following steps: S1. Obtain historical operating status data information of the substation; S2. Clean the data information obtained in step S1 to construct a training data set; S3. Based on CNN network, LSTM network and SVM classifier, the initial model of substation operation status monitoring is constructed; S4. Using the training data set obtained in step S2, the initial substation operation status monitoring model constructed in step S3 is trained to obtain a substation operation status monitoring model; S5. Using the substation operation status monitoring model obtained in step S4, perform operation status monitoring of the target substation.
2. The substation operation status monitoring method according to claim 1 is characterized in that The data cleaning described in step S2 specifically includes the following steps: Abnormal data identification: K-Means clustering method is used to perform cluster analysis on the data information obtained in step S1 to identify abnormal data; during the clustering process, Davies-Bouldin index is used to evaluate the effectiveness of the cluster analysis results and determine the optimal number of clusters; Abnormal data deletion: Delete the identified abnormal data; Data interpolation: EM interpolation and regression interpolation methods are used to calculate interpolation weights based on MAPE values to interpolate various data. When using EM interpolation for data interpolation, θ is used to continuously improve the interpretation function of the data until it converges to a set of parameters: Where θ (new) is the parameter value updated after one iteration of the EM interpolation method; θ is the probability distribution parameter of the data; n is the number of samples in the data set; X is the observed data; Z is the missing data; p(X,Z|θ) is the joint probability distribution of X and Z; E Z|X, θ[logp(X,Z|θ)] is the expected value of the log-likelihood of Z given X and parameter θ; Data normalization: Use the following formula to normalize the data: Where XX' is the normalized data; XX is the data before normalization; β is the weight and coefficient; b is the bias term.
3. The substation operation status monitoring method according to claim 2 is characterized in that The initial model for monitoring the operation status of a substation is constructed based on the CNN network, the LSTM network and the SVM classifier in step S3, which specifically includes the following steps: The constructed initial model for substation operation status monitoring includes CNN module, LSTM module and SVM classifier module; The CNN module, LSTM module and SVM classifier module are connected in series; The CNN module is used to extract the features of the input substation operation data, obtain the corresponding feature map and upload it to the LSTM module; The LSTM module is used to generate several serialized vectors based on the received feature maps and upload them to the SVM classifier module; The SVM classifier module is used to classify the operating status of the substation according to the received serialized vector to achieve the prediction of the substation operating status.
4. The substation operation status monitoring method according to claim 3 is characterized in that The CNN module specifically includes the following steps: The CNN module consists of two convolutional blocks; Each convolution block includes a convolution layer, an activation function layer, and a maximum pooling layer; The filter size of the convolution layer is 128, and the convolution kernel size is 3×3; the convolution layer is used to extract the features of the input data; The activation function layer uses the ReLU activation function to obtain the mapping feature matrix; the maximum pooling layer is used to process the obtained mapping feature matrix and obtain several one-dimensional feature maps.
5. The substation operation status monitoring method according to claim 3 is characterized in that The LSTM module specifically includes the following steps: The LSTM module includes a fully connected layer and an LSTM layer; The fully connected layer consists of 256 neurons; The number of hidden nodes in the LSTM layer is 256.
6. The substation operation status monitoring method according to claim 3 is characterized in that The SVM classifier module specifically comprises the following steps: The following formula is used as the decision function f(x) of the SVM classifier module: Where sign() is the sign function; n is the number of training samples; α i is the Lagrange multiplier, which is used to indicate the contribution of the i-th training sample to the decision boundary; y i is the category label of the i-th training sample; x i is the feature vector of the i-th training sample; K(x i ,x) is the kernel function, which is used to calculate the i-th training sample x i The inner product of the new data point x in the feature space; b is the bias term, which is used to indicate the location of the decision boundary.
7. A system for implementing the substation operation status monitoring method according to any one of claims 1 to 6, characterized in that It includes a data acquisition module, a data cleaning module, a model building module, a model training module and a status monitoring module; the data acquisition module, the data cleaning module, the model building module, the model training module and the status monitoring module are connected in series in sequence; the data acquisition module is used to obtain the historical operating status data information of the substation and upload the data information to the data cleaning module; The data cleaning module is used to clean the acquired data information according to the received data information to construct a training data set, and upload the data information to the model construction module; The model building module is used to build the initial model for substation operation status monitoring based on the received data information, CNN network, LSTM network and SVM classifier, and upload the data information to the model training module; The model training module is used to train the constructed substation operation status monitoring initial model based on the received data information and the obtained training data set, obtain the substation operation status monitoring model, and upload the data information to the status monitoring module; The status monitoring module is used to monitor the operating status of the target substation based on the received data information and the obtained substation operating status monitoring model.