Automatic switching control of transformer simulated fault resistance network based on high-voltage integrated signal enhancement algorithm
Strong noise in high-voltage transformer environments can easily overwhelm weak fault signals during the switching control of simulated fault resistance networks. This interference hinders the accurate identification of fault types and levels, ultimately compromising the stability and accuracy of the switching process. To address this, we propose an automatic switching control method for simulated fault-resistant transformer networks based on a high-voltage integrated signal enhancement algorithm. By acquiring and enhancing signals from the high-voltage side of the transformer, we extract fault features from three domains — time, frequency, and nonlinearity — to construct a feature vector. An improved extreme learning machine is employed to identify fault types and levels, dynamically calculate fault-resistance parameters, and design a closed-loop automatic switching strategy using an FPGA controller. Experimental results show that, under 12 data sets representing four fault categories, the absolute deviation between the actual and theoretical resistance values is controlled within 0.0005 Ω to 0.0020 Ω. The switching delay is significantly lower than that of the comparison method, and the switching jitter time is substantially reduced.