To study the relationship between multiple parameters and stability of downward drift, orthogonal simulation and partial least squares regression were used to analyze factors affecting the stability of downward drift. Safety coefficient and reliability analysis were applied to optimize these parameters. On the basis of that, an industrial test was conducted. The results show that: 1) The factors affecting tensile stress of the loading layer in a descending sequence in terms of significance are thickness of the loading layer, drift width, drift height, two-sided elastic modulus of the drift, elastic modulus of the loading layer and Poisson ratio of the backfill, and the first three factors are more effective than the others. 2) Satisfactory fitness and forecast were obtained through partial least squares regression with errors controlled within 15%. 3) The reliability analysis is more reasonable than the safety coefficient method in optimizing multiple parameters, which can achieve safer, more efficient and economical downward drift exploitation. The satisfactory results of industrial test indicate that partial least squares regression has theoretical and practical values in optimizing multiple parameters for downward drift mining.

downward drift  /  orthogonal test  /  partial least squares  /  safety coefficient  /  reliability analysis