由于变压器绕组漏磁强度分布不均匀,检测结果易产生较大误差。为此,提出了基于矩阵分解的扼流适配变压器绕组漏磁检测方法。通过多层聚类算法组建最优编码矩阵,利用最优权重译码算法对分类器输出进行优化,得到扼流适配变压器绕组漏磁类型。采用矩阵分解方法,计算扼流适配变压器绕组漏磁后的磁感应强度数据,根据漏磁后的磁感应强度数据布置磁感应强度测量点,实现了扼流适配变压器绕组漏磁检测。仿真试验结果表明,所提方法能够精准检测漏磁点,降低了检测延时,具有广泛的应用性。

Because of the uneven distribution of magnetic flux leakage intensity of transformer windings, the test results are prone to large errors. Therefore, a magnetic leakage detection method of choke adaptive transformer winding based on matrix decomposition is proposed. The optimal coding matrix is constructed by multi-layer clustering algorithm, and the type of magnetic leakage of choke adaptive transformer winding is than obtained. Matrix decomposition method is used to calculate the magnetic induction intensity data of choke adapter transformer winding after magnetic leakage, and magnetic induction intensity measurement points are arranged according to the magnetic induction intensity data post-magnetic leakage, so as to realize the magnetic leakage detection of choke adapter transformer winding. The simulation results show that the proposed method can accurately detect magnetic leakage points, reduce the detection delay, and has a wide range of applications.