为了提高WC-12Co涂层质量,采用曲面响应法对等离子喷涂WC-12Co涂层的工艺参数进行优化,以涂层显微硬度为评价指标,设计了以电流、氩气流量和喷涂距离三因素的Box-Behnken实验模型.利用方差分析三因素的显著性及交互作用,采用BP神经网络建立3×9×1的神经网络模型,并与回归模型预测结果进行比较.通过实验方法对优化参数进行验证,同时分析了不同喷涂距离对涂层组织与性能的影响.研究表明:回归模型复相关系数R2为0.9799,BP神经网络的复相关系数R2为0.9991;神经网络的平均相对误差为0.46%,低于多项式回归模型的平均相对误差1.56%.喷涂距离对涂层显微硬度影响最为显著,最优工艺参数为:电流I=390 A,氩气流量QAr=2500 L/h,喷距d=130 mm,能够预测的最大硬度为1336.9HV0.5.%To improve the quality of WC-12Co coatings,the WC-12Co coating was deposited on Q235 substrate by atmospheric plasma spraying(APS)technology under different process conditions and micro-hardness values of coating were measured as the response values. The mathematical model was established between the influence factors such as spraying current,plasma gas flow,spray distance and the response values of micro-hardness. The significance of single factor and interaction effects were also discussed, and BP neutral network(3×9×1)was applied to compare with the results of polynomial regression model. The optimized parameters were verified by experimental method, and the effects of different spraying distance on the microstructure and properties of the coatings were analyzed.Results showed that the complex correlation coefficient of polynomial regression model was 0.979 9,and that of BP neutral network model was 0.999 1.The average relative error of neural network was 0.46%, which was lower than that of polynomial regression model of 1.56%. Influence of spraying distance on the micro-hardness of coating is the most significant.The maximum micro-hardness of 1 336.9 HV0.5could be obtained under the optimal parameters of the current of 390 A,flow of argon of 2 500 L/h and spray distance of 130 mm.
展开▼
