The Research On Robot Trajectory Planing Of Aero Engine Blade Automatic Ultrasonic Testing System

In recent years,with the development of the aerospace and astronautics by leaps and bounds in our country,the safety and reliability of the aircraft have been closely attended by more and more people.As the main power source of the aircraft,it has greatly important significance to maintain its performance.So it has great significance for improving the safety and reliability of the spacecraft to implement maintenance daily and accurately test evaluation in a timely manner to the engine blades.Traditional detection means low degree of automation,low efficiency and poor accuracy.But the automation ultrasonic nondestructive testing technology is a good way to solve these questions. And it can realize the real-time,intelligent,accurate and efficient nondestructive testing.In the process of the aircraft engine blade automatic ultrasonic nondestructive testing,it needs the end of the industrial robot to scan blade with ultrasound. In process of the whole detection,the end of robot ultrasound probe has to meet two conditions:firstly,the end of probe into the distance on the surface of the object to be tested is always equal;secondly,probe of ultrasonic wave incident direction is perpendicular to the surface of object to be tested. Therefore,in order to achieve the perfect automation process of aircraft engine blade ultrasonic nondestructive testing and defect detection the ideal result,it needs to establish a three-dimensional model of the object to be tested,and based on this model to complete the scanning of the trajectory planning of robot,which is the precondition of completing automatic ultrasonic nondestructive testing.According to these questions,these are some researches in aspects:At first,scan the kinematics model of the 6 degree of freedom robot is established,the posture and coordinate transformation of robot are described. Based on the analysis of the inverse solution of 6 degree of freedom robot,combined with this scanning robot which can satisfy the demands of vertical and isometric special trajectory in this topic,under the analysis of the Cartesian space coordinates of scanning robot,and provide a theoretical basis for scanning the trajectory planning of robot.Secondly,in order to get ideal 3d model of the object to be tested,this paper is based on the thoughts of machine learning as the introduction,using gradient descent algorithm completed a new fitting method,using the Python language in Tensorflowdistributed deep learning system to the implementation of the algorithm,and through the Matlab simulation software for the algorithm of fitting accuracy and the fitting speed is verified.Finally,a new fitting method for 3d model as the foundation and to scan the robot’s path planning. Aiming at the problem of complexity of numerical method in traditional trajectory planning algorithm and the long time of operation time,this paper puts forward the improve numerical algorithm based on incremental refinement,under the premise of satisfying the demands of isometric and vertical trajectory,it approached that simplify the operation process,shorten the operation time.Eventually the Matlab Robotics Toolbox verified the trajectory planning of the numerical algprithm can satisfy the requirements of the trajectory planning of scanning robot in this topic.