高聚物流变性能参数预测系统的研究与开发

【作者】 刘守纪；

【导师】 马万珍；

【作者基本信息】 内蒙古工业大学 ， 材料加工工程， 2005， 硕士

【摘要】 高聚物的流变性能对于正确设计模具、选择加工机械、确定加工工艺条件以及进行高聚物熔体充模、流动分析和模拟而言是不可缺少的物性参数。但是高聚物的流变性能受剪切速率、温度和压力等因素的影响较大,各因素之间呈复杂的非线性关系,很难用准确的流变方程来描述。而且高聚物的种类繁多,不同厂家生产的同一种高聚物也会因为其相对分子量分布不同而导致其流变性能差异甚大,这就给高聚物加工企业设计模具、确定加工工艺等造成不便。因此,准确测定高聚物材料在不同加工条件下的流变性能数据已成为急需解决的问题,而所有条件下的流变性能数据都用测量的方法得到是不切实际和没有必要的,有效的办法是仅通过几次有限的测试,就能够预测出一种高聚物的流变性能数据。随着人工神经网络等非线性科学技术在各行业中的广泛应用,其优越的性能为解决此问题提供了新的途径。作者在前人经验的基础上将人工神经网络与经典流变学理论相结合,选择了广泛用于解决函数逼近问题的BP神经网络模型来预测高聚物的流变性能,并采取了一些措施来改进BP网络的结构。在总结了建立神经网络的方法与原则后,结合VB软件和MATLAB软件的优点混合编程,开发了一套可以用于预测高聚物流变性能的集成应用软件。经实验验证,软件的预测效果比较满意,基本上可以用来指导工程应用和生产实践。更多还原

【Abstract】 The rheological property of high polymer is an indispensable factor for designing mould, choosing processing equipment, deciding craft of processing correctly and filling mould, flow analysis and simulation of the melting body of high polymer. But rheological property of high polymer is affected largely by shear rate, temperature and pressure etc, and the factors take on complicated non-linear relations each other, so it is difficult to describe this exactly with equation of rheology. Further more, high polymer is various and the same kind of production produced by different manufactory has quite different rheological property due to the difference of distribution of their relative molecular weight, which cause the inconvenience for designing mould and ascertaining processing craft of the processing enterprise of high polymer. Therefore, it has become an urgent problem to be solved immediately to measure accurately the rheological property data of high polymer under different condition. However it is impractical and unnecessary that the rheological property data are obtained by tests under all conditions, the effective method is only using limited tests to predict the rheological properties data of high polymer only. With wide use of non-linear science and technology in various fields, such as the artificial neural network and so on, has offered the new way for solving the problem because of its superior advantages. Integrating the artificial neural network and classical rheology theory on the basis of forefathers’ experience, author chose the BP network that is used extensively to solve the problem of approaching function to predict the rheological property of high polymer and took some useful measures to improve the structure of BP network. After summarizing the method and principle of setting up the neural network, integrating the advantage of VB and MATLAB, author has developed an integrative internet application that can be used to predict the rheological property of high polymer by method of hybrid programming. Validated by the experiments, predicting result of the software is more satisfying and it can be used for instructing project applications and production practice.更多还原