评价微弧氧化膜层性能的参数主要有膜层厚度、表面粗糙度和硬度,影响氧化膜层性能的主要因素有电流密度、脉冲频率和占空比。利用正交实验数据建立了BP神经网络,对微弧氧化膜层性能进行了预测,对能量参数进行了优化。结果表明:所拟建的BP网络稳定,网络预测当电流密度为10A/dm2、频率为500Hz、占空比为15%时,膜层的厚度、硬度最大,表面粗糙度最小,且其值与实验结果吻合。

The parameters evaluating MAO coating performance contain thickness, roughness and hardness of film. The parameters effecting film property contain contain current density, freduency and duty cycle. Based on the data of orthogonal experiment, the BP neutral network was designed to predict the coating performance and optimize the energy parameters. The results show that BP network has good stabilization, the network predicts that current density is 10A/dm2, freduency is 500Hz and duty cycle is 15%. The MAO coating has maximum thickness and hardness, minimum roughness, the predictive value is the same as experimental value.