Application And Research Of Intelligent Patrol Robot In 500kV Transformer Substation

With the promotion of the lean management of substation operation quality of the China Southern Power Grid Corporation,the Hi-tech robot to help operator carry out the inspection in substation,which had greatly improved thequality of inspection,But to be fully competent for complex inspection,intelligent inspection robot needs to solve many problems,such as the accuracy of navigation and positioning,the coverage rate of patrol,the accurate identification of the indicator of the meter,the analysis of infrared temperature measurement.At the end of 2017,the trackless wheeled intelligent inspection robot began to be applied in 500kV Rongjiang substation.It's different from the traditional way of usingthe magnetic track guidance and the RFID label auxiliary positioning method to realize the navigation and positioning.In this paper we adopt a laser navigation system based on map matching,a map creation and location module implementation method in map stitching software is designed,the navigation path of all station equipment points is constructed..Finally,through the field test of the variable electric field,it is verified that the navigation path constructed by the map matching laser navigation system can achieve 100%inspection coverage,and the repetition localization precision can also meet the weight of the instrument recognition system,recover the requirement of image recognition.In the recognition of meter identification image,theintelligent robot adopts the image preprocessing algorithm based on visual saliency analysis,and fast positioning by reducing the feature template traversing the region.For different types of instrument identification recognition,the recognition method based on Opencv is adopted to reduce the complexity of the algorithm.The accuracy and robustness of the recognition algorithm are higher.Through a large number of field experiments and the comparison and analysis of the image manual and system identification of the robot client instrument,it is proved that the modeling method based onopencv can effectively realize the high precision recognition of the different types of instrument identification of the substation.In the course of testing,the results of instrument identification were analyzed and optimized.Finally,we study the patrol and maintenance mode of the inspection robot and optimized its function expansion.Robot to replace the manual inspection is unavoidable,streamlining management mode is an irresistible trend.The operators should constantly improve their own post skills in the direction of integrated inspection and inspection,and focus on the deepen application of the intelligent inspection robot and realize the real intelligence.