Research and Design of Electromagnetic Navigation System of Power Transmission Line Inspection Robot
WU Zhou, FANG Yan-jun(Department of Automation, Wuhan University, Wuhan 430072, China)
To raise the efficiency of the high-voltage power transmission line's fault detection, a new method of inspection using the high-voltage line inspection robot is studied and realized. On the platform constructed by the PC104 hardware system and the embedded software system, the inspection robot, which can realize autonomous barriers pass and intelligent navigation, analyzes and processes the transmission line's images data for its fault detection. Based on the electromagnetic intensity detection, the navigation system of the robot consists of electromagnetic sensors' arrays separated on different parts and horizontal-vertical dual directions. According to the principle that inductive electromotive force around the high-voltage power transmission line is inversely proportioned to the distance with the line, PC104 adopts the method of the relative value detection. It compares the series of analogy voltage values output from the electromagnetic sensors' arrays, and judges the relative posture to the power transmission line, as well as recognizes the obstacles on power transmission line. Electromagnetic sensor's output signal is weak, and is apt to be influenced by the high-voltage environment on power transmission line. A MATLAB digital filter design method is optimized in order to process the characteristic signals. IIR digital filter is designed for choosing the amplified signals on the power current frequency, then the other frequencies disturbance signals are filtered. IIR optimizing algorithm is implemented on the PC104 CPU platform, satisfying the real time performance of the inspection robot system. The results of MATLAB simulation prove that the IIR output is linear to the amplitude input signals. The field tests of inspection robot on power line show that, the design of robot's electromagnetic navigation system based on IIR filtering is of simple structure and obviously effective.