Deterministic and simulation based planning approaches for advance and logistic processes in mechanized tunneling

Many tunnel construction sites deal with major time losses and low productivities. Often the actual productivities fall far behind the planned levels with insufficient or improper jobsite logistics systems being the reason. There are many projects where it is unclear to the responsible staff, which countermeasures would actually improve the situation. Therefore, using better planning tools would allow improving the productivity. The main reason for these deficiencies being so widely spread throughout the industry is the lack of structured planning methods. Furthermore, there is little reference data available to planners, which would be necessary to estimate the actual impact of changes in the logistic system. This thesis systematically analyzes available planning methods and develops a systematic planning approach for mechanized tunneling jobsites that overcomes these deficiencies. The systematic planning approach combines deterministic and simulation based elements. The logistical setup of several tunnel construction sites has been analyzed. The analysis has been used to derive an extensive set of reference data that forms a foundation for future planning. The reference datasets include the timing of all logistics operations on tunnel jobsites and their statistical properties. Subsequently, the most common planning methods, which are currently used by the industry, have been examined for their potential and limitations and a systematic approach for deterministic planning of tunnel logistics has been derived. In a first step, the deterministic planning approach has been extended by spreadsheet based monte carlo simulation methods , in order to reflect the stochastic nature of construction process durations. As process interdependencies play a major role in the formation and propagation of delays within a logistic system, a more flexible environment for simulation modeling has been chosen. Tunnel construction simulation models have been developed within the general purpose simulation framework Anylogic. A systematic approach of data acquisition and abstraction into SysML models has been developed for mechanized tunneling. The SysML models have subsequently been transferred into executable simulation models in Anylogic. In order to capture all effects which influence mechanized tunneling performance, different simulation paradigms, namely discrete event simulation, agent based simulation and system dynamics simulation have been utilized. The practical usability of process simulation for planning tunnel construction logistics is evaluated in detail and compared to deterministic planning methods. Using the developed planning approach, construction managers can improve the productivity of their jobsites and decisions regarding additional investments in logistic equipment can be made transparently.