Design And Test Of The Resvervior Culvert Pipe Detecting Robot

The reservoir water culvert is one of the important hydraulic structures and it s a transmission channel for downstream irrigation and water supply. In order to guarantee the safe and steady operation of the water culvert, it is necessary to inspect it periodically. Currently, to check if the culvert is damage, fracture or weathering, etc rely mainly on artificial external inspection and by an extension rods which equipped with a camera, but the disadvantage of this method is short distance, poor detection effect and inefficient, so a pipeline robot to replace the manual detection has important significance.The current pipeline robot is divided into unpowered and powered systems. The mechanical structure of these pipe robots include wheel type, crawler, supporting wheel, peristaltic, spiral wheel, etc. The research of these pipeline robots has focused on the straight pipeline walking ability, pass through the bend pipe and branch pipeline ability. Once, when they enters the pipe similar to the reservoir culvert, it can t go ahead by avoiding the internal weathering pits or solid objects, which limits the application of current existing culvert pipe robot in the reservoir. To solve this problem, this paper developed a telescopic six support wheel pipe inspection robot, the main contents include research and design of the mechanical body, visual inspection of navigation and control systems. They are as follows:1) According to the interior environment of the reservoir culvert, the mechanical body uses six telescopic support wheels and the drive wheels can be turned when the robot is walking. The pipeline robot avoids the obstacle in the pipe by adjusting its attitude angle. After calculate the drag force and steering resistance, the paper complete the structural design of the pipe robot.2) The defects of the pipe to be observed by a camera and at the same, the in-pipe navigation is done by machine vision analysis. The program based on Matlab Image Toolbox, the operation is as follows: after the image gray-scale transformation, de-noising, binary processing, then by the erosion and dilation operations to separation pipe corrosion and obstacle information, and eventually fill the hole image. Analyze each connected domain feature for calculating the pipe center and obstructions center, then calculate the angle of the obstacle for adjusting pipe robot posture to avoid the obstacles.3) Complete the lower computer controller design, it communicate with the upper computer via RS232 serial port. When the lower computer controller receives instructions from the upper computer, it parses instructions and controls the motor for pipe robot auto navigating in the pipe. The manual control mode is also reserved in the lower computer to provide the pipe robot for getting out of the risk by the button control if it is necessary.4) Finish the pipe robot prototype and built a test pipeline, the outline size of the pipeline robot varies φ 500 mm ! 10 mm and the drag force reaches 149 N.After drive test in the pipe and the automatic navigation test by machine vision, the results show that, the pipe robot with six telescopic support wheels can work in the pipe by adjust its posture to avoid the obstacles, satisfy the testing needs of reservoir culvert.