Prediction of penetration rate of a small diameter shield TBM based on adaptive neuro-fuzzy inference system and linear regression

초록

Penetration rate of shield TBM is a very important factor used to estimate the construction period and cost. However, the prediction procedure of penetration rate is complicated because there are many factors to consider such as rock mass condition and TBM operating parameters. Therefore, it has become a good alternative to develop a simple prediction model of penetration rate based on some rock mass properties and mechanical measurement data. For this reason, 32 data were collected from the field in this study, and rock quality designation, deformation modulus, Thrust, and revolution per minute were selected as input parameters for model development. Using these variables, two prediction models such as multivariate linear regression and adaptive neuro-fuzzy inference system were developed. In addition, both models were verified and compared by various evaluation indices. The developed two models are expected to be applied for prediction purposes under a similar condition to the field.