基于机器学习的可燃物热解参数反演研究

Parameter Retrieval for Combustible Pyrolysis Based on Machine Learning

数值模拟是研究可燃物热解过程的重要技术手段,但建立数值模型时需要准确的可燃物热解参数,而通过实验手段获得给定材料的全部参数较为困难。针对此问题,本文提出了一种基于机器学习的热解参数反演技术。首先建立了数值模型;然后建立了基于神经网络及遗传算法的混合参数反演模型;最后使用数值模拟数据进行了验证,结果表明提出的参数反演技术可有效兼顾反演效率和反演精度,能够为准确快速获得热解参数提供一种新的工具,具有较好的应用前景。

Numerical simulation is an important tool to study the process of combustible pyrolysis.But a numerical model requires accurate pyrolysis parameters of combustibles.It is difficult to obtain all the parameters of specified combustibles by experimental measurement.To overcome this limitation,we report a parameter retrieval approach based on machine learning.Firstly,a numerical model is built.A hybrid approach utilizing neural network and genetic algorithm is then proposed to retrieve the parameters.The approach is finally validated by numerical data.Results suggest that the proposed method can retrieve the parameters with high accuracy and efficiency.It provides a new tool to obtain pyrolysis parameters of combustibles.