神经网络算法在确定热物性参数中的应用

Application of Neural Network in Parameter Identification

提出将近年来飞速发展的人工神经网络技术应用于导热领域的研究,通过神经网络算法和导热正问题的结合来求解导热反问题用于确定热物性参数,首先从理论上分析了方法的可行性,并通过数值模拟仿真编写程序得到多组计算结果,验证了算法的正确性,文中提出的算法为热物性参数的测定提供了新颖的方法。

This paper introduces the newly developed technical neural network into the domain of heat transfer. By the combination of neural network and positive heat conduction, we can solve inverse heat conduction problem to measure the identification of parameter. The feasibility of this method is proved in mathematics and the validity is examined by numerical simulation. The algorithm throws new light on the study of parameter identification pr oblem.