摘要
采用RBF神经网络对固化工艺中缠绕速率、芯模温度和管道外表面温度之间的关系建立模型,采用有限元数值模拟方法对模型进行训练,将训练好的模型用于不同缠绕速率和期望外表面温度下的芯模温度预测。结果表明,采用该方法设计的加温历程可使玻璃钢管道外表面温度控制在期望的范围内。本研究不仅为降低实验成本、提高生产效率提供有效方法,而且为基于理论分析的工艺参数确定和优化提供依据。
A relationship model of winding speed, core mould temperature and outer surface temperature of FRP pipes was built based on RBF neural network and the finite element numerical simulation method was adopted to train model to predict the core mould temperature of different winding speed and expected outer surface temperature. The result shows that the designed heating process can make the outer surface temperature of FRP pipes within the expected range. This research provides effective method for reducing the experiment cost and improving production efficiency, and provides theory basis for the identification and optimization of process parameters based on theoretical analysis.
出处
《纤维复合材料》
CAS
2013年第1期22-23,26,共3页
Fiber Composites
基金
哈尔滨市优秀学科带头人项目(2012RFXXG84
关键词
RBF神经网络
数值模拟
内固化工艺
热芯模
热芯缠绕
高压玻璃钢管道
RBF neural network
numerical simulation
internal curing process
heated mandrol
heated-mandrel winding
high-pressure FRP pipes
