Abstract: To classify the surrounding rocks of superlargespan caverns of 20~70 m, a novel method is proposed. This method uses the area weighted average to determine the basic quality indicators of the surrounding rock during the preliminary classification stage and increases the span influence correction coefficient during the detailed classification stage. The value of the span influence correction coefficient is determined based on the relationship between the stability coefficient of the tunnel and the basic quality index of the surrounding rock, along with the tunnel span. Additionally, a threedimensional numerical simulation of the entire excavation process and strength reduction method are used to calculate the stability coefficient of caverns with spans of 20~70 m in grade Ⅰ~Ⅳ surrounding rocks. Through analysis, an empirical formula relating the stability coefficient of the cavern to the basic quality index and span of the surrounding rock is obtained. The results reveal the following insights: (1) As the span increases from 20 to 70 m, the stability coefficient of the tunnel decreases by 46.54% to 50.67%. (2) When the tunnel span remains constant, the smaller the basic quality index values of the surrounding rock correspond to poorer tunnel stability. The span influence correction coefficient is defined as the ratio of the relative deviation of the tunnel stability coefficient from the reference span (20 m) to the grade difference of the reference span tunnel stability coefficient. This correction coefficient for span influence is 0~1.0, where the most adverse effect of increasing the span on the stability of the surrounding rock is equivalent to decreasing one grade in the grade of the surrounding rock. 

Key words:

superlargespan flat caverns, correction coefficient for span influence, stability coefficient, classification of rock mass, strength reduction method