| At this stage,the vast majority of substation rooms in China still adopt the traditional operation and maintenance mode based on manual inspection.This mode has the defects of high labor intensity,low operation and maintenance efficiency and high investment cost.It is gradually difficult to adapt to the changes of power grid under the new situation.It is in urgent need of an economic,safe,scientific and efficient intelligent operation and maintenance mode to replace it.With the innovation of intelligent manufacturing technology,the power inspection robot aiming to achieve automatic operation and maintenance and predictive maintenance is also constantly exploring and developing,becoming the key to the transition from traditional operation and maintenance mode to intelligent operation and maintenance mode.Based on the existing research and application of inspection robot technology at home and abroad,this topic has carried out the research on the key technologies of the track inspection robot system in the substation,explored the key technologies of the control system and image recognition,and verified the feasibility and effectiveness of the system through field tests.The specific research contents and results are as follows:1.Aiming at the problems of low hardware integration,poor expansion ability and incomplete functions of the inspection robot control system,a inspection robot control system based on STM32F103ZET6 microprocessor is proposed.Adhering to the design concept of modularity,easy expansion and strong compatibility,the balance and optimization design of the overall hardware structure,device selection and hardware circuit is studied,which ensures the coordination and linkage between various links and promotes the improvement of the flexibility of the system.2.Aiming at the problems of poor real-time,low maintenance and insufficient reliability of the control system software of inspection robot,a control system software of inspection robot based on Free RTOS is designed.Through multi task management,the resources of STM32 microprocessor are used more reasonably and effectively,and the real-time and reliability of the system are better guaranteed.In order to solve the problems of low positioning accuracy,poor speed regulation performance and low motion efficiency,the motion model of the track inspection robot is constructed,the quantitative relationship between the motion posture and the pulse signal is quantitatively analyzed and studied,and an incremental PID position and speed double closedloop motion control algorithm is proposed to achieve the effect of smooth speed regulation and accurate positioning.In order to solve the problem that patrol inspection takes too long,a route optimization algorithm based on bubble sorting method is proposed.The calling sequence of preset positions is optimized according to the spatial coordinates of preset positions,which greatly improves the patrol inspection efficiency of the substation.In order to solve the problems of unknown fault cause and poor reliability,the fault information table is defined,and a multi-sensor joint fault diagnosis method is proposed,which enhances the accuracy and reliability of fault diagnosis.3.Aiming at the low accuracy and robustness of instrument recognition,an instrument image recognition algorithm based on Python and Open CV is proposed.In order to simplify the image and filter out the noise,the original image is preprocessed by using the algorithms of geometric transformation,weighted average graying,Gaussian filtering and Otsu’s binarization,and the dial area is segmented by using the algorithms of morphology,Canny detection,contour screening and perspective transformation.In order to realize the automatic reading of the pointer instrument,the homomorphic filtering algorithm is used to improve the visual effect of the dial image;Secondly,the contour line of the scale line is fitted,and the center of the scale line fitting circle is found based on the Mean-Shift algorithm;Thirdly,Hough line transform algorithm is used to extract the pointer;Finally,the coordinate system of automatic meter reading is established,and the automatic meter reading is completed according to the relationship between range and angle.In order to realize the automatic reading of the digital display instrument,the morphological transformation and projection segmentation method are used to extract the single digital character in the digital display panel,and then the normalized correlation coefficient matching method is used to complete the automatic reading.4.Comprehensively and objectively test and analyze the track inspection robot system in the substation through the built field test environment.The test results of key performance indicators are as follows: the average absolute error of horizontal positioning is 1.3mm,and the average absolute error of vertical positioning is 0.6mm;The miss rate of image recognition of pointer instrument is 2.93%,and the overall accuracy is 98.20%;The Miss recognition rate of digital display instrument image recognition is 1.94%,and the overall accuracy is 93.18%.The test results meet the expected requirements.The track inspection robot system of the substation studied in this subject has indeed relieved the pressure of the traditional operation and maintenance mode,improved the operation and maintenance efficiency and automation level of the substation,promoted the unattended process of the substation,and has certain significance for the transition from the traditional operation and maintenance mode to the intelligent operation and maintenance mode at this stage. |