智能优化在铝合金低频电磁半连续铸造中的应用

Application of Intelligent Optimization in Low Frequency Electromagnetic Semi-continuous Casting for Al Alloy

针对铝合金大尺寸扁锭成型裂纹倾向大、工艺参数不易找准的问题,建立了基于RBF的电磁半连续铸造神经网络模型,以铸锭裂纹量化值最小作为优化目标,以训练后的RBF网络作为评价函数,在工艺指标控制范围内,采用改进后的遗传算法对铝合金电磁半连续铸造过程的工艺参数进行了优化计算,获得了最优工艺参数值:铸造速度为52mm/min,铸造温度为724℃,扁锭宽面冷却强度为134L/min,扁锭窄面冷却强度为22L/min,电磁强度为11749A.匝,电磁频率为27Hz。按照该最优工艺参数值进行了真实试铸,结果表明,铸锭成品率比优化前提高了20%。

ANN(artificial neural network) model was established based on RBF(radial base function)and genetic algorithm to solve high crack tendency in large aluminum alloy ingot and un-rational technological parameters determination.The optimized design of technological parameters for aluminum alloy electromagnetic semi-continuous casting was conducted by means of modified genetic algorithm through taking minimum crack value as optimized end and taking trained RBF network as estimated function,and optimized technological parameters were presented.The experiment was performed based on the optimized technological parameters.The results indicate that the qualified rate of ingot is increased by 20%.