Research on filtering of electromagnetic measurement while drilling signal in underground coal mine

In view of problems of slow convergence speed and long computation time existed in electromagnetic wireless measurement while drilling (EM-MWD) signals extracted by adaptive filtering least mean square algorithm and recursive least squares algorithm, a stabilized fast transversal filters for recursive least squares (SFT-RLS) algorithm was proposed, which was used for adaptive filtering unsteady electrical interference and power frequency, as well as double frequency and triple frequency interference, and extracting weak EM-MWD signals in real time. Based on the RLS algorithm, the algorithm uses four parallel filter structures to reduce the operation time, and uses the weighted least squares error for feedback to improve the stability. The simulation results show that the SFT-RLS algorithm can achieve adaptive filtering and adaptive notch of the EM-MWD signals with a sampling rate of 1 kHz, the average running time of each iteration is less than 156.98 μs, so as to realize fast and stable convergence operation and real-time adaptive filtering of EM-MWD signals; the adaptive filtering of the SFT-RLS algorithm can suppress unsteady electrical interference with a signal to interference ratio of -115 dB; the adaptive notch of the SFT-RLS algorithm can filter power frequency and its interference of double frequency and triple frequency in real time, and effectively extract the EM-MWD time position pulse modulation signal with 6.25 Hz, and provide reliable data for correct decoding of EM-MWD signals.