摘要
采用冷压烧结方法制备了不同含量碳纤维(CF)及颗粒状氧化硅(SiO2-P)协同增强的聚四氟乙烯(PTFE)复合材料,利用MPX-2000摩擦磨损试验机测试了不同载荷、滑动速度下PTFE复合材料的摩擦磨损性能,并利用人工神经网络(ANN)建立了摩擦系数、磨损量与材料组成及测试条件之间的非线性关系模型。结果表明,采用SCG算法、5-[15∶10∶5]3-1网络结构的ANN网络模型可以有效预测PTFE复合材料的摩擦磨损,数据的预测值与试验值的误差在10%以内。
Polytetrafluoroethylene（PTFE） composites filled with different carbon fiber（CF） contents and SiO2 particle（SiO2-P） were prepared by the method of cold press sintering in this paper.Their tribological properties under different loading and sliding speed were tested on a MPX-2000 friction and wear tester.Then the non-linear relationship model of material composition and experimental conditions vs friction coefficient or wear loss was established by means of artificial neural network（ANN） based on experimental data.Results indicate that the ANN using SCG algorithm and 5-[15∶10∶5]3-1 network structure can predict the tribological properties of PTFE composites effectively.All the errors between predicting results and experimental results are below 10%.
出处
《合肥工业大学学报（自然科学版）》
CAS
CSCD
北大核心
2010年第9期1308-1310,共3页
Journal of Hefei University of Technology：Natural Science
基金
国家自然科学基金资助项目(51005123)
关键词
人工神经网络
聚四氟乙烯
复合材料
摩擦磨损
artificial neural network（ANN）
polytetrafluoroethylene（PTFE）
composite
friction and wear
