Research on measurement technology of optical fiber angle sensor based on MEA-BP

High-voltage disconnectors operate under strong electromagnetic radiation for long periods, making it prone to incomplete opening and closing. Current electronic sensors used to detect their status are susceptible to electromagnetic interference, leading to inaccurate measurement data. Therefore, based on the rotation characteristics of the moving contact during opening and closing, a method for measuring rotation angles in a straight line is proposed. A fiber optic angle sensor is designed according to the characteristics of reflective fiber optics to measure the rotation angle of disconnectors. The Mean Evolution Algorithm (MEA) is used to optimize the BP neural network model (MEA-BP) to compensate for the nonlinearity of the designed fiber optic angle sensor. Through experimental simulations, the nonlinearity compensation effect of the MEA-BP model is compared with that of the traditional BP neural network. The mean square error of nonlinearity compensation decreases from 0.0916 to 0.0211, and the fitting accuracy improves from 83.57% to 95.06%. The linear error decreases from 6.85% to 0.81%. An experimental platform using SBE27.5 kV/2000A disconnectors as the research object is established to test the accuracy of the fiber optic angle sensor. The experimental results show that the accuracy of the fiber optic angle sensor improves from 8.043% to 0.781% after nonlinearity compensation, indicating that this solution has higher accuracy and stronger feasibility.