聚酯玻璃钢大气老化力学性能BP人工神经网络预报模型的建立  

Establishment of Prediction Model for GFRP′s Mechanical Property in Atmosphere Aging Based on BP Neural Network

作　　者：杜武青[1] 刘颖慧[1] 赵晴[2] 王莹[1] 马达兵 

机构地区：[1]北京理工大学珠海学院材料与环境学院 [2]南昌航空大学材料科学与工程学院 [3]中国人民解放军91515部队

出　　处：《装备环境工程》2017年第5期97-101,共5页Equipment Environmental Engineering

基　　金：国家自然科学基金重点项目(50533060)

摘　　要：目的准确地分析并建立起一个老化模型,探究不饱和聚酯玻璃钢在大气环境中的条件与老化性能的变化联系。方法在气象数据的基础上分别分析不饱和聚酯玻璃钢在不同季节不同大气环境下各种力学性能的变化,利用数学建模的人工智能方法-BP神经网络进行建模。结果实际值与预测值有很好的一致性,模型精确度也较高。结论此预报模型可比较精确地评价不饱和聚酯玻璃钢在大气中的老化行为。Objective To analyze accurately and establish an aging model to explore conditions of fiberglass reinforced plastic in atmospheric environment and relations in change of its aging properties. Methods Changes of fiberglass reinforced plastic＇s mechanical properties in different atmospheric environments of different seasons were analyzed based on the meteorological data and a model was established with artificial intelligence method-BP neural network of mathematical modeling. Results The actual value and predicted values were in good agreement and the model accuracy was high. Conclusion The prediction model can evaluate aging behaviors of unsaturated fiberglass reinforced plastic in the atmos- phere accurately.

关 键 词：不饱和聚酯玻璃钢 大气老化 人工神经网络 

分 类 号：TJ04[兵器科学与技术—兵器发射理论与技术] TQ317.6[化学工程—高聚物工业]