Predicting tunnel boring machine utilization using discrete event simulation models for hard rocks

Abstract

Tunnel Boring Machines (TBM) are perhaps the most efficient method for tunnel excavation, as they have become the dominant mode and method of tunneling around the world. One of the essential steps for justification of the use of TBMs as well as planning the completion time and estimating the cost of projects involving TBMs is estimating the machine performance in given ground conditions. The main performance indicator for TBMs is the average daily advance rate (AR) which is product of rate of penetration (ROP), utilization (U). While there are various models for estimation of machine ROP with acceptable levels of accuracy and reliability, the models for prediction of machine utilization are often not sufficiently accurate and do not include many of the critical input information. This is because the predictive models for estimation of machine utilization do not have the capability to deal with many of the site-specific issues, including variation in geological settings, TBM and backup specifications, and site set up and logistics. Improving the accuracy of models for TBM performance prediction by estimation of utilization has been the focus of many ongoing studies in mechanized tunneling. It requires accounting for many factors including project geology, machine design and configuration, and operational issues. Simulation of tunnel activities considering TBM to be a tunneling factory can be a reliable method for estimation of TBM utilization factor. In this study, modeling tunneling activities and downtimes using discrete event simulation approach is adopted to predict TBM utilization. A few mechanized tunneling projects in hard rock have been considered for development of the model and verification of the results. A database template using visual basic for application (VBA) and data from available projects is established where the project and machine specifications along with the various geotechnical unit details serve as the input parameter for the simulation. The simulation model, CSM2020 is prepared using Arena© simulation software. The results of modeling were verified by comparing the estimated utilization with that of the recorded TBM utilization at the jobsites, broken down by formations and reaches of tunnel. This allowed for verification of the capabilities of the modeling concept to incorporate the impacts of geology and tunnel length. Sensitivity analysis of the model relative to certain activities allows for evaluating the ability of the simulation system to account for each activity and its impact on machine performance within a given geology and at certain reach of the tunnel. Parametric studies were conducted relative to various input parameters to the simulation model including geology, machine specification, and site set up to see the response of the model to various input parameters and to assess their impact on the utilization. The results were compared to some case histories for verification and proved that the modeling concept is functional. To apply the DES model to different projects, data recording, screening, and analysis protocols have been developed to allow for adjustment of the models with additional onsite input data. This approach was tried by both Arena© and the MATLAB code and site data for 8 different projects. The impact of interdependencies of various tunneling activities on machine utilization were also assessed to identify the critical activities that might act as bottlenecks in the excavation process. In parallel, a code was developed in MATLAB to allow for performing des modeling with better control of the process and data structure for individual activities and resources. The code also has the advantage that it does not require Arena© license to run and can be made into an executable program. In general, use of CSM2020 des model for utilization estimation prove to be a promising approach which can be further developed to include additional features and be expanded to variety of tunneling operation, while it can be further developed to be used as part of a system for optimization of tunneling operation and selection of TBMs with a quantitative approach to evaluate various features of the machine and backup system.