Research On Hybrid Constrained Optimization Algorithm Of Flat Car Locks And Development Of The Detection System

Container transportation has been applied much wider with the rapid development of economy.It is of great significant to improve the transportation capability and integral performance of railway flatcars,which means the whole process of flatcar manufacturing and detecting should be supervised.Currently,the detection of the quality of railway flatcar locks is one of the key technology to ensure high-quality railway transportation.Containers of railway flatcars are installed on the frame of flatcars with locks.In most cases they are set on lug boss which is higher than frame flat.The flatness error on the mating surface of lock and the matching hole can lead to various consequences in loading and unloading process,such as the overturn of containers and the deformation of locks.At present,the traditional measurement method in assembling and detecting flat cars is of low precision due to complicated types.Moreover,the results are analyzed and checked all by manual detection,which not only wastes time but reaches low accuracy.This thesis developed a flatness detecting system for locks of railway flatcars,which is capable of detecting the position of locks accurately and make adjust schemes for unqualified locks effectively.First,the constraints of locks are analyzed in order to establish mathematical model of constraint optimization.Based on Genetic Algorithm and Artificial Bees Colony Algorithm,the constraint optimization algorithm is established to accomplish the calculation of adjustment amounts.Consequently,aiming at avoiding the local optimal,the algorithm is hybridized to improve operation speed so as to check the qualification of locks and provide the easiest scheme for adjus ting unqualified flatcars.Second,with comprehensive consideration of measuring precision,laser distance-measuring module and magnetic railings ruler are used as detecting element.Appropriate materials and components are chosen as well to complete the s tructural design of rotary table.As the hardware core,STM32 development board is developed to finish the measurement of distance and angular value with serial ports and timers.Upper Computer Software is designed by combined programming of VS2010 and Matlab,realizing the function of type choosing,serial communicating,data inputting and processing,qualification checking,adjustment scheme making,user login,picture drawing and report printing.After setting up the flatness detecting system,calibration experiment is conducted to compensate center aberration,which laids a solid foundation for further experiments that tests the precision of rotary table and accuracy of the system.Those experimental data proves not only the qualification of detecting ac curacy and the reasonability,stability and efficiency of the algorithm.The basic function of software is also tested.It is clear to be indicated that the detecting accuracy can reach 1mm and the algorithm result can maintain 100% optimal with adjust assumption fewer than 20 and 95.5% optimal more than 20.This whole system improves the speed and quality of lock detection prominently.