摘要：In this paper, new relations are proposed to predict the gout penetration length into jointed rock mass by means of multivariate linear and non-linear regressions. Numerical analysis was used to obtain the grout penetration length into the rock mass by considering all effective rock and grout properties. The results of numerical analysis were in a good agreement with analytical methods. As a case study Bakhtiary dam and power plant of Iran was used to assess the results of prediction models. Because of many effective properties on grout penetration length and restriction of simple regression models, multivariate regression analysis was used to correlate all parameters with penetration length of grout. Effective rock joint properties as joint hydraulic aperture, orientation, roughness, spacing and trace length along with grout properties including viscosity, yield stress, pressure and grouting time were used in regression analysis. Three possible relations one linear and two non-linear relations were proposed to predict penetration length. The validation of the equations was checked by the F-test (analysis of variance) analysis. The results of all three relations showed good cross-correlations with analytical method and field data. It is concluded that the proposed relations can predict grout penetration length into the rock mass, reasonably.
