Algorithm for cavity flow in a new-born goaf and experimental verification

Abstract

Prevention and control measures of spontaneous combustion of coal and gas accumulation in a goaf require an accurate description of its gas flow state. However, the commonly used fluid dynamics in porous media is not suitable for the new-born goaf with fracture cavity combination, multi-scale, and large blocks. In this study, we propose a cavity flow algorithm to accurately describe the gas flow state in the new-born goaf. The genetic algorithm (GA) is used to randomly generate the binary matrix of a goaf caving shape. The difference between the gas flow state calculated by the lattice Boltzmann method (LBM) and the measured data at the boundary or internal measuring points of the real goaf is taken as the GA fitness value, and the real goaf caving shape and the gas flow state are quickly addressed by GA. The experimental model of new-born goaf is established, and the laser Doppler anemometry (LDA) experiment is carried out. The results show that the Jaccard similarity coefficient between the reconstructed caving shape and the real caving shape is 0.7473, the mean square error between the calculated wind speed and the LDA-measured value is 0.0244, and the R2 coefficient is 0.8986, which verify the feasibility of the algorithm.