A SIMULATION STUDY OF TUNNEL EXCAVATION

Abstract

The inherent uncertainties in tunnelling which may be due to geological formations, management effectiveness and equipment performance need to be studied through stochastic modelling. The conventional deterministic approach for evaluation of tunnel construction time is inadequate and results in incorrect time estimates which are generally on the lower side. The stochastic approach developed in this study would hopefully' provide more realistic estimates. The present study has been an iterative process of develop-ment of a methodology involving collection of field construction data; its analysis; identification of the needs for more data; further data collection, and repeating analysis; and final valida-tion of the statistical model. The causes for variations in tunnel advance rates have been identified as changes in job and management conditions. The classification systems have, therefore, been developed for both these conditions dividing them into three -.categories, viz., good, fair and poor. The job conditions have been classified on the basis of six simple parameters. These are the geologic structure, rock strength, contact zones, ROD, joint spacing and joint orientation. All these parameters are normally evaluated during the investiga-tion stage. The ranges of values for advance per round (APR) have also been identified for the three job conditions. Rating values for different parameters affecting the management performance are proposed to classify the management conditions quantitatively. Further, it was noted that the points of inflexion on the curves of cumulative relative frequency(CRF) iv versus equivalent monthly progress (EMP) indicate changes in the management conditions. For each set of job and management conditions, there are two significant components of cycle time which are the actual work-ing time (AWT) and the breakdown time (BDT). These have been found from statistical analysis of data to belong to log-normal and Weibull distributions respectively. The cycle time corresponds to a given advance per round (APR). The analysis for APR did not show a well defined statistical trend. Therefore, weightad average values were adopted. Tunnel excavation for Pandoh-Baggi tunnel has been simulated on the basis of the classification for job and management conditions and the statistical analysis. The input to computer simulation model included nine statistical parameters for each of the nine sets of job and management conditions. All these 81 parameters have been evaluated from the field data of 3695 excavation cycles. The simulated values were found to be very close to the actual production records which indicated that proposed distributions for AWT and BDT and the classification of different job and management conditions for tunnelling project conform, by and large, to real job conditions. Finally, a matrix of coefficients has been developed to evaluate the expected monthly progress under different sets of job and management conditions, where computer simulation may not be practicable. The coefficients of this matrix are applicable for full face tunnelling using drill and blast method, on a tunnel of any diameter. The effect of tunnel diameter is taken...