摘要
在正交试验的基础上,利用BP神经网络的优点对化学镀Ni-W-P的沉积速率进行了预测。以硫酸镍、钨酸钠、次磷酸钠和柠檬酸钠的质量浓度为输入因子,以沉积速率为输出因子,通过结构优化确定了三层结构(4×9×1)BP神经网络模型。以正交试验结果为样本,对神经网络模型的拟合效果和预测能力进行了检验。结果表明:该神经网络模型的拟合效果较好,并且具有较强的预测能力。
Prediction of the deposition rate of electroless Ni-W-P plating was investigated based on the orthogonal experiment and utilizing the advantages of BP neural network.The mass concentration of nickel sulfate,sodium tungstate,sodium hypophosphite and sodium citrate were used as input factors and the deposition rate was used as output factor,and the three-layer(4×9×1)BP neural network model was determined via structural optimization.Orthogonal experimental results were utilized as sample to test the fitting effect and forecasting ability of neural network model.The results showed that this neural network model possesses satisfactory fitting effect and forecasting ability.
作者
宋长斌
SONG Changbin(Nanyang Radio and TV University,Nanyang 473004,China)
出处
《电镀与环保》
CAS
CSCD
北大核心
2019年第4期40-42,共3页
Electroplating & Pollution Control
基金
河南省科技攻关重点计划项目(132102210215)