Research On Electrical Fault Diagnosis Algorithm Based On Support Vector Machines And Compressed Sensing

In recent years,building electrical systems have played an increasingly important role in modern buildings.While bringing convenience to people’s lives,building electrical systems will inevitably fail.However,in the field of building electrical systems at this stage,most of the faults of building electrical systems rely on manual detection for monitoring and diagnosis.The research on intelligent fault diagnosis algorithms in this field at home and abroad is still in its infancy.On the one hand,it is because the building electrical is a low-voltage power distribution system,which is at the end of the entire power grid and is easily ignored by people;on the other hand,the research on fault diagnosis in the electrical field is more concentrated on large electrical equipment such as generators and transformers.,and the building electrical system has not been paid much attention by researchers because of its relatively simple structure.In view of the above problems,this paper takes the fault of the building electrical system as the research object,relies on the building electrical experiment platform MA2067 to simulate common building electrical faults,and uses the low-voltage electrical comprehensive tester Eurotest61557 to collect fault data.A fault diagnosis model of building electrical system combined with algorithms;considering the problem of low diagnostic efficiency of building electrical system,this paper further proposes a fault diagnosis model based on the combination of compressed sensing and K-nearest neighbor algorithm,aiming to improve the fault diagnosis of building electrical system.diagnostic efficiency.The main research contents of this paper are as follows:（1）For the fault diagnosis of building electrical system,considering the difficulty of building electrical system fault data acquisition,this paper proposes a fault diagnosis model based on the combination of wavelet analysis and support vector machine algorithm.Firstly,wavelet analysis is used to denoise the collected original fault signal,and then the denoised signal is used as the input of the support vector machine algorithm to classify the faults of the building electrical system.The proposed fault diagnosis model can effectively classify the faults of building electrical systems.This part also studies the superiority of support vector machine algorithm in fault classification under small sample data.（2）Aiming at the low efficiency of existing algorithms for diagnosing building electrical system faults,this paper further proposes a building electrical system fault diagnosis model based on compressive sensing and K-nearest neighbor algorithm.The model combines compressive sensing algorithm.After the signal is compressed and sampled,the original signal can still be reconstructed.First,the original fault data of the building electrical system is dimensionally reduced,and then the dimensionality-reduced data is used as the input of the K-nearest neighbor algorithm to classify the fault.Finally,it is verified by experiments.The results show that the proposed fault diagnosis model can not only effectively classify the faults of the building electrical system,but also greatly improve the diagnosis efficiency,which proves the effectiveness of the model.（3）This paper conducts in-depth research on the building electrical experimental platform,conducts detailed circuit analysis on its composition structure and corresponding principle,and uses the experimental platform to simulate 22 kinds of common faults of the building electrical system,and then uses the low-voltage electrical comprehensive tester to collect relevant data,which provides data basis for the experiment of this paper.