Research On The Filling Properties Of Heavy Truck Crankshaft Forging Mould And Structure Optimization

The heavy truck crankshaft, as one of the important components of locomotiveengines, in five parts of the engine (the body, crankshaft, connecting rod, camshaftand cylinder head) is the component which is the most difficult to ensure the quality.Traditional craft, casting method commonly used to produce crankshaft, it is able to geta more ideal appearance, but can not obtain a relatively ideal mechanical properties,metal streamline, micro-grain structure and so on, thereby seriously influencemechanical behavior and service life of heavy truck crankshaft. Compared withtraditional casting process, forging forming process can get forging which have bettermechanical properties, the distribution of metal streamline more reasonable andmicrostructure more fine; for heavy truck crankshaft usually need for multi-processforming, because it had complex structure and the most typical high gluten and sheet;but crankshaft balance block distribution in different positions, structure differences andother reasons, which make the top of some balance block can not be filled in the forgingprocess. So it is necessary to study on the filling performance of heavy truck crankshaftforging die.This paper will focus on some research about some balance blocks which are hardto be filled and forging flash is too large in heavy truck crankshaft die forging process toanalyse the effect of mould structure on the filling performance of heavy truckcrankshaft forging die. Combined with the actual production status and the forgingsimulation software Deform to simulate the process of heavy truck crankshaft forgingforming process, and analysis the effect of flash slot structure form、taper angle of dieimpression、 fillet radius of die and the pre-forging rod journal place wad height on thefilling performance of heavy truck crankshaft forging die; On this basis, using BP neuralnetwork and genetic algorithm to optimize mould structure parameters, a group of theinput data were obtained first by numerical simulation technology and the data whichcontain upper ramp slope δ, wedge flash slot bevel slope β, the second taper angleγ and the rod journal place wad height ζ, taking the minimum unfilled distance L ofheavy truck crankshaft balance block corresponding mould cavity as the finaloptimization objective function, then using the nonlinear fitting ability of BP neuralnetwork to establish a nonlinear relationship between input and objective function toobtain corresponding network training results, and the training results as the fitness value of individuals in a population of genetic algorithms. Finally, in order to obtain theoptimal results and their corresponding conditions, we will use the program of geneticalgorithm to find out the optimization solution. And using the optimized pre-forging diefor practical production, through the practical production test found that afteroptimization the top of balance block forming good and material utilization increased to81.4%. The way which combine the finite element simulation aided design with BPneural network and genetic algorithm, has changed the situation which was the selectionof die structure parameter total depending on experience, improve the accuracy ofmould design, shorten product development design cycle, save the cost of moulddevelopment, all of this are helpful to improve the market competition of enterprise.The results and laws obtained in this paper can be applied to optimize design othersimilar long axis products, and can provides effective technical support for thedevelopment of new products and optimization design.