Secondary Cable Short-Circuit Fault Detection Based on Data Mining and Integrated Learning Fusion Algorithm

This work provides a fault detection method that combines data mining and integrated learning to address the issue of high misjudgment rates caused by insufficient correlation mining of transient information in detecting short-circuit faults in secondary cable lines. Firstly, the fault traveling wave transmission mechanism is analyzed, the fault key information extraction objective function is established, the fault transient information is obtained, and then the information association rules are constructed; secondly, based on the Frequent Pattern Growth Algorithm (FP-growth algorithm), a Frequent Pattern Tree (FP- tree), which realizes the compression of fault information data and the mining of association rules, and obtains the frequent term sets that lead to short-circuit faults; then, a secondary cable line fault alarm framework is constructed based on Bootstrap Aggregating (Bagging) integrated learning method, which realizes the measurement of potential fault risks of secondary cable lines, and proposes the combination of data mining and Bagging integrated learning method to combine the fusion algorithm; finally, a secondary cable line fault simulation platform is built, and relevant experimental research on the validation and accuracy of the correlation of the transient information of secondary cable line faults is carried out, and the experimental results show that the detection accuracy of this method in 18000 sets of samples reaches 90.99 %, which is improved compared with the traditional method, and solves the problem of the high rate of misjudgment.