Research on the selection of grouting performance based on the broken degree of entry

Because of the different broken degrees of the surrounding rock in the entry, the grouting materials with different performance should be applied to classify and reinforce the surrounding rock. Based on the BP neural network algorithm, a mutual feedback model of "geological condition-grouting performance" was developed and used to reflect the broken degree of surrounding rock and grouting performance. In this model, 7 geological parameters such as roof rock strength, coal seam strength and floor rock strength were selected as input factors, and 7 d compressive strength, initial setting time and initial viscosity of grouting material were selected as output factors. The experimental data was divided into training set (34 groups) and the test set (8 groups). The results show that the fitting coefficient of the training set reaches 0.999 95, while the relative error of the output factors of the test set is below 13%. The training and testing results are good, which can meet the requirements of selecting and predicting the grouting parameters based on the broken degree of surrounding rock.