Research on mechanical characteristics and neural network prediction analysis of crushed stone tailings collaborative cementing filling in a gold mine

This article focuses on the crushing of gold mine waste into crushed stones and full tailings below 5 mm. Under the conditions of preparing paste filling slurry, uniaxial compressive strength and tensile strength tests were conducted on the filling material under different parameters, and single-factor and multi-factor fitting analysis and correlation testing were conducted. The correlation between various influencing factors and the strength of the filling body was determined. The influence of various factors on the strength of the paste-filling body was explored. A predictive model was established using an improved MATLAB neural network to predict the effects of slurry concentration X₁, cement sand ratio X₂, sand gravel ratio X₃, and curing age X₄ on the strength (uniaxial compressive strength Y₁, tensile strength Y₂) of the filling material. The results show that there is a multivariate linear function relationship between the strength of the filling body and various parameters. The lime-sand ratio is the main factor affecting the strength of the filling body, followed by the curing age and sandstone ratio, and the slurry concentration is the smallest. The strength of the filling body increases with the increase of filling concentration, curing period, and lime sand ratio, and decreases with the increase of sand stone ratio. The established strength function model of the filling body has strong adaptability and high accuracy in predicting the uniaxial compressive strength Y₁ and compressive strength Y₂ of the gold mine, which provides a basis for the design of the strength demand of later filling mining.