Parameter Retrieval for Combustible Pyrolysis Based on Machine Learning

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.