Design Of Aircraft Harness Automatic Testing System Optimal Layout

Large aircraft are large in size.The number of avionics equipment onboard is large and the harness network is dense.The conduction and insulation of the harness account for more than 30%of the aircraft test period.There are some problems in manual detection,such as missing detection,wrong detection,and failure to realize insulation between wires.The construction of an automatic detection system for harness will greatly shorten the production cycle of aircraft and improve the quality of products.It is very significant.In this paper,a testing system for aircraft harness is established,which consists of distributed test box,line replaceable unit,terminal module and switching cable,aiming at the testing requirements of 40,000 test points for the aircraft,and the layout design of the system is optimized.The main research contents are as follows:Firstly,in order to improve the utilization of the matrix board in the test scheme of the distributed test box and the adapter cable,the neighbor exchange method is adopted.On the basis of satisfying the use of less cross-over in adapter cables,the automatic classification of the plug on the adapter cable is realized.The utilization rate of the matrix card is improved based on the 0-1 backpack solution method,which is improved from the original 89.7%to 100%,achieving the goal of system lightweight and economical design;Secondly,for the layout of distributed test boxes,aiming at the shortest total length of adapter cables,a step-by-step genetic particle swarm optimization algorithm is used to solve the coordinates of 27 test boxes,which reduces the total length of adapter cables by nearly 10%compared with the design results of direct location by region.Thirdly,for the optimal connection path of the control cable between the distributed test devices,taking test equipment as node,physical connection relationship of control cables as edge,and test points as node weight.The connection scheme between test equipment is abstracted as multi-branch traveling salesman problem.And for the shortcomings of the standard discrete particle swarm optimization algorithm cann't solve the multi-branch non-return variant traveling salesman problem.Based on the ant colony algorithm,the genetic selection,crossover and mutation mechanisms are added to track the extremum coding by the intersection of particles and individual and population extremum.The pheromone concentration matrix is updated by the particle distance function.The optimized design of the branch reduces the total length of the control cable by 18%compared with the design result of the original connection.Finally,the summary of the whole work in this dissertation is given,and some valuable topics to be researched further are also discussed.