Parameter Design And Optimization Of Gear Group Inautomobile Gearbox Based On Hybrid Intelligent Algorithms

As one of the important components of automobile transmission system, the stand or fall ofgearbox performance determine the dynamic performance of the engine and vehicle performancedirectly. As the most important transmission parts of gearbox, the improvement of gear’s designquality and efficiency has become the key to improve the gearbox quality. Currently, empiricalmethod and analogy method are still used for the most gearbox manufactures in the gear design.As a result, the gearbox products be designed either have a larger transmission parametersand a conservative safety factor cause bulky volume or have a smaller parameters causinginsufficient intensity of transmission parts, shorter life expectancy, which results a poor qualityand a longer product development cycle. It is an effective method to solve the current situation ofthe gearbox development that conduct automatic parameters design and optimization for geargroup of the gearbox. CAD technology and development of intelligent optimization algorithmsprovide a strong technical support for it becoming reality.The gear group of a mini-car gearbox has been used research object in this thesis. Firstly,gear parameters aided design system has been developed using Visual Basic language inaccordance with general processes of gear parameter design and design standard of the gearbox.Automatic parameter design of gear group has been realized. And automated management ofdesign data has been realized through using Access database. Secondly, the parameters optimaldesign has been realized on the MATLAB platform and the algorithms of combining withimprovement genetic algorithm and BP neural network as the hybrid optimal tool by means ofbuilding the optimal design model of gear group and using ActiveX invoke technology. In the optimization process, dynamic penalty function nonlinear programming genetic algorithm hasbeen structured by means of improving the penalty function of genetic algorithm and combiningwith nonlinear programming algorithm. The convergent rate of genetic algorithm has beenspeeded via combining with the inherent rapid and non-linear mapping of BP neural network. Themapping relationship between input parameters and the strength of gear group has been obtainedthrough training. On this basis, parametric automatic drawing of gear parts after optimization hasbeen achieved by using of AutoLISP language to AutoCAD secondary development. Finally,taking mini-car5T80H model gearbox for example carry out parameter design and optimization.The result of optimization achieve the desired requirement. The quality of gear group has lightenby22%.In this thesis, one gear group CAD software that combines rapid development withoptimization, parametric drawing and data management capabilities has been developed throughcombining hybrid intelligent optimization technology with CAD technology, secondarydevelopment technology and database technology. The design efficiency of gear has beenimproved through using the designed CAD system. The hybrid intelligent optimizationalgorithm’s application improve the quality of gear design. Meanwhile, the method can providereference for the design of other key components of gearbox.