Research On Key Technologies Of Simulation And Optimization Of Aero-engine Installation Process Based On Virtual Reality

This thesis combines computer simulation and virtual reality technology to demonstrate the installation process in Unity3 D,addressing issues such as path planning and human-computer interaction efficiency in the virtual installation of aero engines in complex dynamic environments.The path planning and gesture recognition research was carried out based on the simulation scenario using the improved RRT* method and the improved KNN method,and the main work is as follows.1)An in-depth analysis of the aero-engine site layout and installation process was carried out,completing the process examine of test pipeline installation,engine transfer installation,main engine installation and other process of each installation module.The application requirements for the subsequent thesis were clarified.2)According to the installation environment of the aero-engine site,the 3D modeling software is used to build the 3D models of lifting fixture,transfer vehicle and engine,etc.,and the mapping operation of the related models was carried out to complete the construction of the whole virtual installation environment.Then the virtual installation environment was optimized by combining Polygon Cruncher tool and LOD technology.Finally the aero-engine installation process simulation was completed in the virtual installation environment by using Animation key frame animation technology.3)An improved RRT* algorithm was designed for the optimization of the transfer path of an aero-engine in dynamic and complex scenarios.Firstly,the directions are randomly selected in the search space with equal probability to expand the search.After the random directions are successfully selected,a fixed step is extended from the current node,and iterative search for the node with minimum overhead is continuously performed around the successful expansion point.Finally all nodes are connected to ensure the optimal path in the current region.The aero-engine path planning experiment was carried out with the RRT*algorithm,and an optimal transfer path is sought for the aero-engine transfer process.4)A fuzzy KNN gesture classification method based on weighted cardinality distance is designed for the problem of inefficient interaction of installers in the virtual installation process.Firstly,the sample is fuzzified,and the class affiliation of the fuzzified sample is regarded as a probability,which entropy is used to measure the fuzziness of the sample,as the probability entropy increases,the weight becomes smaller.The class affiliation of the final sample to be classified is obtained by the class affiliation of k neighbors and the cardinality distance weights of the neighbors together.The experiment proves that the method has a higher recognition accuracy in gesture recognition.Combining the improved gesture recognition method with the somatosensory controller,the virtual installation of an aero-engine using realistic gestures to control virtual gestures in a virtual installation environment is realized,which enables the installers to participate in the virtual simulation process conveniently.Finally,the simulation test and virtual installation effect demonstration of the overall aero-engine installation process were carried out.By combining the installation process with virtual reality technology,this system helps to shorten the transit time,improves the quality and efficiency of the aero-engine field installation,and provides some reference value for the optimization of the subsequent aero-engine installation process.