Application of parameter inversion method in numerical simulation of multilayer pyrolysis for polypropylene honeycomb panel
This study aims to simulate the pyrolysis process of Polypropylene(PP)honeycomb panels in a cone calorimeter by employing a parameter inversion method.We successfully integrated micro-scale kinetic models and small-scale pyrolysis models to construct a one-dimensional multi-layer pyrolysis model for the honeycomb panel.To begin with,we established three pyrolysis model frameworks under various conditions:a multilayer model coupled with a microscale kinetic model derived from thermogravimetric experimental results,a multilayer model that ignores these results,and a simplified single-layer model.Following this,we built a micro-scale kinetic model based on results from the thermogravimetric analysis and micro-combustion calorimeter test.The kinetic parameters are evaluated based on the Shuffle Complex Evolution optimization algorithm.Subsequently,we employed an innovative parameter inversion method to evaluate the model parameters of the pyrolysis model for the three frameworks,based on cone test results.This method coupled the adaptive Chaotic Particle Swarm Optimization algorithm with the Fire Dynamics Simulator(FDS)platform and was first proposed here.Finally,we explored the impact of different model frameworks on the model's accuracy.By comparing the results of cone calorimetry experiments under varying external heat flow conditions,we confirmed the model's accuracy and the efficacy of the parameter inversion method.Our findings reveal that the multi-layer material model framework can more accurately simulate the pyrolysis process of PP honeycomb panels compared with the single-layer model.Moreover,by incorporating the micro-scale kinetic model,the proposed parameter inversion method can more precisely determine the parameters for the pyrolysis model.Finally,a compensatory effect between kinetic parameters and thermophysical parameters is observed.This result means that acquiring kinetic parameters at the micro-scale beforehand is essential.This process can reduce the number of model parameters that need to be solved and can enhance the precision of parameter inversion.
NETL
NSTL
万方数据
