Risk prediction method and system for switch cabinet
The GBDT algorithm establishes a switch cabinet temperature change model, which solves the problems of low temperature prediction accuracy and insufficient generalization ability of switch cabinets in the existing technology, achieves more accurate risk monitoring and prevention, and improves the safety management level of switch cabinets.
Patent Information
- Application Number
- CN202411004041.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-25
- Publication Date
- 2024-11-19
AI Technical Summary
The existing technology has problems such as low model accuracy, high feature engineering requirements and insufficient generalization capabilities in terms of switching cabinet temperature prediction and risk identification, which is difficult to meet the needs of potential data law mining of complex transmission systems.
A risk prediction method is proposed, by obtaining the temperature data, environmental factors and load conditions of the switch cabinet, preprocessing and inputting the pretrained switch cabinet temperature change model for temperature trend prediction, using the GBDT algorithm to establish a relational model, monitor and predict in real time The temperature changes, and an alarm signal is generated.
It improves the accuracy of risk monitoring and prediction of switch cabinets, reduces the requirements of feature engineering, enhances the degree of automation and generalization of the model, and improves the level of safety management and risk prevention capabilities.
Smart Images

Figure CN118982233A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of risk prediction for switch cabinets, and in particular to a risk prediction method and system for switch cabinets. Background Art
[0002] Temperature prediction and risk identification of switchgear in power transmission systems are important issues in power system operation and management. Excessive temperature of switchgear may lead to safety problems such as equipment overload, damage or even fire, so accurate prediction and risk identification of switchgear temperature are of great significance.
[0003] At present, the methods for switchgear temperature prediction and risk identification mainly include thermodynamic calculation methods based on physical models and machine learning methods based on statistical learning. The method based on physical models needs to consider a large number of equipment parameters and environmental factors, and the calculation is complex and the accuracy is limited; while the machine learning method based on statistical learning can better capture the complex relationship between switchgear temperature and environmental factors, load conditions, etc. in a data-driven way, and has better flexibility and applicability. The method based on physical models requires a large number of equipment parameters and environmental factors, and the calculation is complex and the accuracy is limited, which has certain limitations for practical engineering applications. Although machine learning methods based on statistical learning, such as support vector machines and neural networks, have improved the prediction accuracy and applicability to a certain extent, they still have problems such as insufficient model generalization ability and high requirements for feature engineering. The existing technology still has some problems in switchgear temperature prediction and risk identification, including low model accuracy, high requirements for feature engineering, and insufficient generalization ability. In addition, for complex transmission systems, the existing technology is often difficult to fully explore the potential laws in the data and cannot meet the needs of practical engineering applications. The existing technology has improved the accuracy and applicability of switchgear temperature prediction and risk identification to a certain extent, but there are still some limitations, such as low model accuracy and insufficient generalization ability. Summary of the invention
[0004] The purpose of the present invention is to propose a method and system for predicting the risks of a switch cabinet, so as to solve the technical problem of how to predict the risks of a switch cabinet more efficiently and accurately.
[0005] On the one hand, a risk prediction method for a switch cabinet is provided, comprising: Obtain temperature data, environmental factors and load conditions of the switch cabinet, and pre-process the acquired temperature data, environmental factor data and load condition data; The pre-processed temperature data, environmental factor data and load condition data are used as input items, and the pre-trained switch cabinet temperature change model is input to predict the temperature trend to obtain the temperature trend prediction result; wherein the switch cabinet temperature change model is trained by historical temperature data, historical environmental factor data and historical load condition data, and is used to monitor the switch cabinet temperature change in real time and predict the temperature trend; The potential safety risks of the switch cabinet are identified based on the temperature trend prediction results, and a corresponding alarm signal is generated when there is a safety risk.
[0006] Preferably, the switch cabinet temperature change model specifically includes:
[0007] in, represents the switch cabinet temperature prediction model, M represents the upper limit of the number of iterations, A vector representing the historical temperature data, environmental factors, and load condition values of the switch cabinet. represents the leaf node region of the regression tree, Represents the number of leaf nodes, represents the ordinal number of the leaf node, represents the residual calculated using the loss function, Represents the calculated value of the regression tree for the leaf node.
[0008] Preferably, the switch cabinet temperature variation model is pre-trained by the following process: Collect historical temperature data, historical environmental factor data, and historical load condition data, and divide the collected data into a training set and a test set according to a preset ratio value; wherein the training set at least includes a vector composed of data values under the dimensions of historical temperature data, historical environmental factor data, and historical load condition data, and the switch cabinet temperature corresponding to the vector; the test set at least includes a vector composed of data values under the dimensions of historical temperature data, historical environmental factor data, and historical load condition data, and the switch cabinet temperature corresponding to the vector; Input the training set into the preset switch cabinet temperature change model, initialize the switch cabinet temperature change model, and set the corresponding number of iterations; Calculate the corresponding residual for each sample, and fit a new regression tree based on the calculated residual. Calculate the leaf nodes of the regression tree to obtain the corresponding switch cabinet temperature prediction model. The test set is input into the obtained switch cabinet temperature prediction model to obtain the corresponding test result. If the test result meets the preset accuracy threshold, the final switch cabinet temperature prediction model is output.
[0009] Preferably, the preprocessing of the acquired temperature data, environmental factor data and load condition data specifically includes: Check whether there are missing values in the temperature data, environmental factor data and load condition data according to the preset detection rules. If there are missing values, select the corresponding filling mean to fill the corresponding missing value to obtain complete temperature data, environmental factor data and load condition data; The detection rule identifies the temperature data, environmental factor data and load condition data through a preset algorithm and determines whether there are missing values.
[0010] Preferably, the step of determining and identifying the potential safety risk of the switch cabinet according to the temperature trend prediction result specifically includes: When the temperature trend prediction result exceeds the preset safety threshold at a certain time point, the switch cabinet is judged to have a low safety risk; When the temperature trend prediction results exceed the preset safety threshold at several consecutive time nodes, it is determined that the switch cabinet has a high safety risk.
[0011] On the other hand, a risk prediction system for a switch cabinet is also provided to implement the risk prediction method for a switch cabinet, comprising: A data acquisition module is used to acquire temperature data, environmental factors and load conditions of the switch cabinet, and pre-process the acquired temperature data, environmental factor data and load condition data; A prediction module, which is used to input the pre-processed temperature data, environmental factor data and load condition data as input items, and input the pre-trained switch cabinet temperature change model to predict the temperature trend, and obtain the temperature trend prediction result; wherein the switch cabinet temperature change model is trained by historical temperature data, historical environmental factor data and historical load condition data, and is used to monitor the switch cabinet temperature change in real time and predict the temperature trend; The risk alarm module is used to identify the potential safety risks of the switch cabinet based on the temperature trend prediction results, and generate a corresponding alarm signal when there is a safety risk.
[0012] Preferably, it also includes: A pre-training module, used to collect historical temperature data, historical environmental factor data, and historical load condition data, and divide the collected data into a training set and a test set according to a preset ratio value; wherein the training set at least includes a vector composed of data values under the dimensions of historical temperature data, historical environmental factor data, and historical load condition data, and the switch cabinet temperature corresponding to the vector; the test set at least includes a vector composed of data values under the dimensions of historical temperature data, historical environmental factor data, and historical load condition data, and the switch cabinet temperature corresponding to the vector; Input the training set into the preset switch cabinet temperature change model, initialize the switch cabinet temperature change model, and set the corresponding number of iterations; Calculate the corresponding residual for each sample, and fit a new regression tree based on the calculated residual. Calculate the leaf nodes of the regression tree to obtain the corresponding switch cabinet temperature prediction model. The test set is input into the obtained switch cabinet temperature prediction model to obtain the corresponding test result. If the test result meets the preset accuracy threshold, the final switch cabinet temperature prediction model is output.
[0013] Preferably, the data acquisition module is specifically used to check whether there are missing values in the temperature data, environmental factor data and load condition data according to a preset detection rule, and if there are missing values, select a corresponding filling mean to fill the corresponding missing value to obtain complete temperature data, environmental factor data and load condition data; The detection rule identifies the temperature data, environmental factor data and load condition data through a preset algorithm and determines whether there are missing values.
[0014] Preferably, the risk alarm module is specifically used to determine that the switch cabinet has a low safety risk when the temperature trend prediction result exceeds a preset safety threshold at a certain time node; When the temperature trend prediction results exceed the preset safety threshold at several consecutive time nodes, it is determined that the switch cabinet has a high safety risk.
[0015] In summary, the implementation of the embodiments of the present invention has the following beneficial effects: The risk prediction method and system for switch cabinets provided by the present invention construct a relationship model between the historical temperature of the switch cabinet, environmental factors, load conditions and the temperature of the switch cabinet. This model is crucial for improving the safety management level and risk prevention capabilities of the switch cabinet and comprehensively evaluating the risk status of the safety cabinet. The risk monitoring and prediction of the switch cabinet can be carried out more accurately, thereby more effectively improving the safety level of the switch cabinet and achieving sustainable development. The factors affecting the temperature of the switch cabinet can be understood more comprehensively, and the factors affecting the safety level of the switch cabinet can be improved and prevented in a targeted manner. The requirements for feature engineering can be reduced, manual intervention can be reduced, and the degree of automation of the model can be improved, making the temperature prediction model more intelligent and efficient. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, without paying creative labor, other drawings obtained based on these drawings still belong to the scope of the present invention.
[0017] Figure 1 It is a schematic diagram of the main process of a risk prediction method for a switch cabinet in an embodiment of the present invention.
[0018] Figure 2 Schematic diagram of a risk prediction system for a switch cabinet in an embodiment of the present invention. DETAILED DESCRIPTION
[0019] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings.
[0020] like Figure 1 FIG. 1 is a schematic diagram of an embodiment of a risk prediction method for a switch cabinet provided by the present invention. In this embodiment, the method includes the following steps: Step S1, obtain the temperature data, environmental factors and load conditions of the switch cabinet, and pre-process the obtained temperature data, environmental factor data and load condition data; understandably, collect multi-dimensional data such as temperature data, environmental factors, load conditions, and perform pre-processing and feature engineering, and then use the GBDT algorithm for model training and optimization. Check whether there are missing values in the data. For missing values, you can choose to fill the mean to ensure data integrity.
[0021] In one embodiment, the preprocessing of the acquired temperature data, environmental factor data and load condition data specifically includes checking whether there are missing values in the temperature data, environmental factor data and load condition data according to a preset detection rule, and if there are missing values, selecting a corresponding filling mean to fill the corresponding missing value to obtain complete temperature data, environmental factor data and load condition data; wherein the detection rule identifies the temperature data, environmental factor data and load condition data through a preset algorithm and determines whether there are missing values. The above detection rules are commonly used algorithms in this field and are not described in detail here.
[0022] Step S2, taking the pre-processed temperature data, environmental factor data and load condition data as input items, inputting the pre-trained switch cabinet temperature change model to predict the temperature trend, and obtaining the temperature trend prediction result; wherein, the switch cabinet temperature change model is trained by historical temperature data, historical environmental factor data, and historical load condition data, and is used to monitor the switch cabinet temperature change in real time and predict the temperature trend; it can be understood that based on the GBDT algorithm, by learning and analyzing multi-dimensional data such as historical temperature data, environmental factors, and load conditions, a switch cabinet temperature prediction and risk identification model is established. The model can automatically select features, process nonlinear relationships, and has strong generalization ability, and can more accurately predict the switch cabinet temperature change trend and identify potential risks. In practical applications, the model can monitor the switch cabinet temperature change in real time, predict the temperature trend, and identify potential safety risks, thereby providing timely warning and decision support. GBDT (Gradient Boosting Decision Tree) is a powerful machine learning algorithm with good generalization ability and automatic feature selection ability, which can effectively capture the complex relationship between the switch cabinet temperature and environmental factors, load conditions, etc. Therefore, the use of GBDT algorithm to realize switch cabinet temperature prediction and risk identification has important theoretical and practical significance. Through the research and application of this method, accurate prediction of switch cabinet temperature change trend and intelligent identification of potential risks can be achieved, which helps to take measures in advance to avoid safety problems such as equipment overload, damage and even fire.
[0023] In one embodiment, the switch cabinet temperature change model specifically includes:
[0024] in, represents the switch cabinet temperature prediction model, M represents the upper limit of the number of iterations, Represents a vector of data values in the dimensions of switch cabinet historical temperature data, environmental factors, and load conditions. represents the leaf node region of the regression tree, Represents the number of leaf nodes, represents the ordinal number of the leaf node, represents the residual calculated using the loss function, Represents the calculated value of the regression tree for the leaf node.
[0025] In one embodiment, the model of switch cabinet temperature change is pre-trained through the following process, historical temperature data, historical environmental factor data, and historical load condition data are collected, and the collected data is divided into a training set and a test set according to a preset ratio value; wherein the training set at least includes a vector composed of data values under the dimensions of historical temperature data, historical environmental factor data, and historical load condition data, and the switch cabinet temperature corresponding to the vector; the test set at least includes a vector composed of data values under the dimensions of historical temperature data, historical environmental factor data, and historical load condition data, and the switch cabinet temperature corresponding to the vector; it is understandable that the data samples are divided according to the ratio of 80% training set and 20% test set. Usually, machine learning needs to perform operations such as data standardization before model training, but the gradient boosting decision tree (GBDT) does not need to standardize the data. The main reason is that GBDT is a nonlinear model based on a tree structure, and its construction process mainly depends on the relative order of features rather than the specific numerical value. Since the decision tree focuses on the relative order of features when dividing nodes, and is not affected by the absolute scale, data standardization is not necessary for the construction and prediction of the GBDT model. This feature enables GBDT to show good robustness when processing features of different scales and ranges, without the need for linear transformation of features, and can more flexibly and effectively capture complex nonlinear relationships.
[0026] Input the training set into the preset switch cabinet temperature change model, initialize the switch cabinet temperature change model, and set the corresponding number of iterations; calculate the corresponding residual for each sample, and fit a new regression tree based on the calculated residual, calculate the leaf nodes of the regression tree, and obtain the corresponding switch cabinet temperature prediction model; The test set is input into the obtained switch cabinet temperature prediction model to obtain the corresponding test result. If the test result meets the preset accuracy threshold, the final switch cabinet temperature prediction model is output.
[0027] In one embodiment, the training set in A vector composed of the switch cabinet historical temperature data, environmental factors, load conditions and other dimensions , for The corresponding switch cabinet temperature is is the loss function.
[0028] Initialize the switchgear temperature prediction model ,
[0029] right , where M is the number of iterations; for each sample , calculate the residual (The residual here is approximated by the negative gradient of the loss function):
[0030] Based on the previous step Fit a new regression tree, whose leaf node area is recorded as , where J represents the number of leaf nodes in the mth tree. ,calculate:
[0031] Update the current learner and get:
[0032] The final switch cabinet temperature prediction model is obtained: .
[0033] In this embodiment, the main parameters involved in the training process of the GBDT model are as follows: (1) n_estimators n_estimators represents the number of decision trees, that is, the number of model iterations. Generally speaking, the larger the n_estimators value, the stronger the model fit, but too large a n_estimators value may cause the model to overfit. Therefore, when selecting n_estimators, it is necessary to consider the model's fit and efficiency. The default value of n_estimators is 100.
[0034] (2) learning_rate learning_rate represents the learning rate of the model. The learning rate represents the contribution weight of each base learner in the entire iteration. The default value is generally 0.1, which is used to control the step size of each tree model update, also known as the step size factor. Generally speaking, the smaller the learning rate, the slower the convergence speed of the model, and more iterations are required to achieve the desired effect. Therefore, when adjusting the learning rate, it is necessary to comprehensively consider the number of iterations and the size of the learning rate to determine the appropriate parameter combination.
[0035] (3) Subsample Subsample is used to control the subsample ratio used by each base learner, and to prevent overfitting by sampling samples. The value range of subsample is (0,1], and it is generally recommended to be between [0.5,0.8]. The default value is 1. When the value is 1, it means that during the training process of the model, the training set used each time is the same as the original sample set, that is, no sample is sampled.
[0036] (4) max_depth max_depth indicates the depth limit of each decision tree, which is used to control the complexity of the model. When max_depth is small, the model's fitting ability is weak and it is easy to underfit; when max_depth is too large, the model is easy to overfit. The default value of max_depth is 3. When the depth is 1, the generated trees are all binary trees with only two leaf nodes, so there is no interaction between the variables. When the depth is 2, the generated tree model has interaction between two variables. By analogy, deeper base trees can explore very complex variable interactions, but they are also more likely to cause overfitting.
[0037] In the process of model parameter adjustment, the K-fold cross-validation method is usually used to find the optimal parameters. K-fold cross-validation is a commonly used model evaluation technique that divides the data set into K non-overlapping equal subsets and uses K training and testing to evaluate the performance of the model. In each cross-validation, one subset is used as the test set, and the remaining K-1 subsets are used as training sets to train the model. The final result is the average of all cross-validation results. The 10-fold cross-validation and grid search methods are used for parameter optimization. First, the range of parameter combinations is determined through grid search, and then the model is evaluated in combination with 10-fold cross-validation to find the optimal parameter combination.
[0038] In one embodiment, model evaluation is a very important part of machine learning. A reasonable model evaluation process can assist in judging the pros and cons of the model and improve the model's ability in prediction. Therefore, it is necessary to select appropriate evaluation indicators to comprehensively evaluate the performance of the model, find out the problems and determine the direction of optimization. According to different model tasks, the evaluation indicators can be divided into two categories: classification indicators and regression indicators. Classification indicators are usually used to measure the performance of classification algorithms, such as accuracy, recall rate, ROC (Receiver Operating Characteristic) curve, AUC (Area Under Curve), etc.; while regression indicators are used to evaluate the performance of regression algorithms, including MSE (Mean Squared Error), MAE (Mean Absolute Error), MAPE (Mean Absolute Percentage Error), R² (Coefficient of Determination), etc. Different models and application scenarios require the selection of appropriate evaluation indicators to measure the performance and applicability of the model. The present invention selects MAE and R² as evaluation indicators of the model to measure the accuracy of model prediction.
[0039] Step S3, determine and identify the potential safety risk of the switch cabinet according to the temperature trend prediction result, and generate a corresponding alarm signal when there is a safety risk. It can be understood that the temperature change of the switch cabinet is monitored in real time, the temperature trend is predicted, and the potential safety risk is identified, so as to provide timely warning and decision support.
[0040] In one embodiment, the identifying of potential safety risks of the switch cabinet based on the temperature trend prediction results specifically includes: when the temperature trend prediction results exceed a preset safety threshold at a certain time node, determining that the switch cabinet has a low safety risk; when the temperature trend prediction results exceed the preset safety threshold at several consecutive time nodes, determining that the switch cabinet has a high safety risk.
[0041] like Figure 2 As shown, an embodiment of the present invention further provides a risk prediction system for a switch cabinet, which is used to implement the risk prediction method for a switch cabinet, including: A data acquisition module is used to acquire temperature data, environmental factors and load conditions of the switch cabinet, and pre-process the acquired temperature data, environmental factor data and load condition data; A prediction module, which is used to input the pre-processed temperature data, environmental factor data and load condition data as input items, and input the pre-trained switch cabinet temperature change model to predict the temperature trend, and obtain the temperature trend prediction result; wherein the switch cabinet temperature change model is trained by historical temperature data, historical environmental factor data and historical load condition data, and is used to monitor the switch cabinet temperature change in real time and predict the temperature trend; The risk alarm module is used to identify the potential safety risks of the switch cabinet based on the temperature trend prediction results, and generate a corresponding alarm signal when there is a safety risk.
[0042] A pre-training module, used to collect historical temperature data, historical environmental factor data, and historical load condition data, and divide the collected data into a training set and a test set according to a preset ratio value; wherein the training set at least includes a vector composed of data values under the dimensions of historical temperature data, historical environmental factor data, and historical load condition data, and the switch cabinet temperature corresponding to the vector; the test set at least includes a vector composed of data values under the dimensions of historical temperature data, historical environmental factor data, and historical load condition data, and the switch cabinet temperature corresponding to the vector; Input the training set into the preset switch cabinet temperature change model, initialize the switch cabinet temperature change model, and set the corresponding number of iterations; Calculate the corresponding residual for each sample, and fit a new regression tree based on the calculated residual. Calculate the leaf nodes of the regression tree to obtain the corresponding switch cabinet temperature prediction model. The test set is input into the obtained switch cabinet temperature prediction model to obtain the corresponding test result. If the test result meets the preset accuracy threshold, the final switch cabinet temperature prediction model is output.
[0043] In one embodiment, the data acquisition module is specifically used to check whether there are missing values in the temperature data, environmental factor data and load condition data according to a preset detection rule, and if there are missing values, select a corresponding filling mean to fill the corresponding missing value to obtain complete temperature data, environmental factor data and load condition data; The detection rule identifies the temperature data, environmental factor data and load condition data through a preset algorithm and determines whether there are missing values.
[0044] In one embodiment, the risk alarm module is specifically used to determine that the switch cabinet has a low safety risk when the temperature trend prediction result exceeds a preset safety threshold at a certain time node; When the temperature trend prediction results exceed the preset safety threshold at several consecutive time nodes, it is determined that the switch cabinet has a high safety risk.
[0045] It should be noted that the system described in the above embodiment corresponds to the method described in the above embodiment. Therefore, the undetailed parts of the system described in the above embodiment can be obtained by referring to the contents of the method described in the above embodiment, and will not be repeated here.
[0046] In summary, the implementation of the embodiments of the present invention has the following beneficial effects: The risk prediction method and system for switch cabinets provided by the present invention construct a relationship model between the historical temperature of the switch cabinet, environmental factors, load conditions and the temperature of the switch cabinet. This model is crucial for improving the safety management level and risk prevention capabilities of the switch cabinet and comprehensively evaluating the risk status of the safety cabinet. The risk monitoring and prediction of the switch cabinet can be carried out more accurately, thereby more effectively improving the safety level of the switch cabinet and achieving sustainable development. The factors affecting the temperature of the switch cabinet can be understood more comprehensively, and the factors affecting the safety level of the switch cabinet can be improved and prevented in a targeted manner. The requirements for feature engineering can be reduced, manual intervention can be reduced, and the degree of automation of the model can be improved, making the temperature prediction model more intelligent and efficient.
[0047] The above disclosure is only the preferred embodiment of the present invention, which certainly cannot be used to limit the scope of the present invention. Therefore, equivalent changes made according to the claims of the present invention are still within the scope of the present invention.
Claims
1. A risk prediction method for a switch cabinet, characterized in that: include: Obtain temperature data, environmental factors and load conditions of the switch cabinet, and pre-process the acquired temperature data, environmental factor data and load condition data; The pre-processed temperature data, environmental factor data and load condition data are used as input items, and the pre-trained switch cabinet temperature change model is input to predict the temperature trend to obtain the temperature trend prediction result; wherein the switch cabinet temperature change model is trained by historical temperature data, historical environmental factor data and historical load condition data, and is used to monitor the switch cabinet temperature change in real time and predict the temperature trend; The potential safety risks of the switch cabinet are identified based on the temperature trend prediction results, and a corresponding alarm signal is generated when there is a safety risk.
2. The method according to claim 1, characterized in that The switch cabinet temperature change model specifically includes: in, represents the switch cabinet temperature prediction model, M represents the upper limit of the number of iterations, x represents the vector composed of the switch cabinet historical temperature data, environmental factors, and load condition data values, R mj represents the leaf node area of the regression tree, J represents the number of leaf nodes, j represents the ordinal number of the leaf node, I represents the residual calculated by the loss function, and c m j Represents the calculated value of the regression tree for the leaf node.
3. The method according to claim 2, characterized in that The model of switch cabinet temperature variation is pre-trained through the following process, Collect historical temperature data, historical environmental factor data, and historical load condition data, and divide the collected data into a training set and a test set according to a preset ratio value; wherein the training set at least includes a vector composed of data values under the dimensions of historical temperature data, historical environmental factor data, and historical load condition data, and the switch cabinet temperature corresponding to the vector; the test set at least includes a vector composed of data values under the dimensions of historical temperature data, historical environmental factor data, and historical load condition data, and the switch cabinet temperature corresponding to the vector; Input the training set into the preset switch cabinet temperature change model, initialize the switch cabinet temperature change model, and set the corresponding number of iterations; Calculate the corresponding residual for each sample, and fit a new regression tree based on the calculated residual. Calculate the leaf nodes of the regression tree to obtain the corresponding switch cabinet temperature prediction model. The test set is input into the obtained switch cabinet temperature prediction model to obtain the corresponding test result. If the test result meets the preset accuracy threshold, the final switch cabinet temperature prediction model is output.
4. The method according to claim 3, characterized in that The preprocessing of the acquired temperature data, environmental factor data and load condition data specifically includes: Check whether there are missing values in the temperature data, environmental factor data and load condition data according to the preset detection rules. If there are missing values, select the corresponding filling mean to fill the corresponding missing value to obtain complete temperature data, environmental factor data and load condition data; The detection rule identifies the temperature data, environmental factor data and load condition data through a preset algorithm and determines whether there are missing values.
5. The method according to claim 4, characterized in that The identification of potential safety risks of the switch cabinet according to the temperature trend prediction result specifically includes: When the temperature trend prediction result exceeds the preset safety threshold at a certain time point, the switch cabinet is judged to have a low safety risk; When the temperature trend prediction results exceed the preset safety threshold at several consecutive time nodes, it is determined that the switch cabinet has a high safety risk.
6. A risk prediction system for a switch cabinet, used to implement the method according to any one of claims 1 to 5, characterized in that: include, A data acquisition module is used to acquire temperature data, environmental factors and load conditions of the switch cabinet, and pre-process the acquired temperature data, environmental factor data and load condition data; A prediction module, which is used to input the pre-processed temperature data, environmental factor data and load condition data as input items, and input the pre-trained switch cabinet temperature change model to predict the temperature trend, and obtain the temperature trend prediction result; wherein the switch cabinet temperature change model is trained by historical temperature data, historical environmental factor data and historical load condition data, and is used to monitor the switch cabinet temperature change in real time and predict the temperature trend; The risk alarm module is used to identify the potential safety risks of the switch cabinet based on the temperature trend prediction results, and generate a corresponding alarm signal when there is a safety risk.
7. The system according to claim 6, characterized in that Also includes, A pre-training module, used to collect historical temperature data, historical environmental factor data, and historical load condition data, and divide the collected data into a training set and a test set according to a preset ratio value; wherein the training set at least includes a vector composed of data values under the dimensions of historical temperature data, historical environmental factor data, and historical load condition data, and the switch cabinet temperature corresponding to the vector; the test set at least includes a vector composed of data values under the dimensions of historical temperature data, historical environmental factor data, and historical load condition data, and the switch cabinet temperature corresponding to the vector; Input the training set into the preset switch cabinet temperature change model, initialize the switch cabinet temperature change model, and set the corresponding number of iterations; Calculate the corresponding residual for each sample, and fit a new regression tree based on the calculated residual. Calculate the leaf nodes of the regression tree to obtain the corresponding switch cabinet temperature prediction model. The test set is input into the obtained switch cabinet temperature prediction model to obtain the corresponding test result. If the test result meets the preset accuracy threshold, the final switch cabinet temperature prediction model is output.
8. The system according to claim 7, characterized in that The data acquisition module is specifically used to check whether there are missing values in the temperature data, environmental factor data and load condition data according to the preset detection rules. If there are missing values, the corresponding filling mean is selected to fill the corresponding missing value to obtain complete temperature data, environmental factor data and load condition data; The detection rule identifies the temperature data, environmental factor data and load condition data through a preset algorithm and determines whether there are missing values.
9. The system according to claim 8, characterized in that The risk alarm module is specifically used to determine that the switch cabinet has a low safety risk when the temperature trend prediction result exceeds a preset safety threshold at a certain time node; When the temperature trend prediction results exceed the preset safety threshold at several consecutive time nodes, it is determined that the switch cabinet has a high safety risk.
Citation Information
Cited By
Heat dissipation control method and system for network control cabinet
CN120825913A
