Framework for dynamic generation and evaluation of excavation schedules for hard rock tunnels in preconstruction and construction

Projects with uncertain product specifications make it difficult for planners to achieve on-time completion within budget. This challenge is especially important for the excavation of hard rock tunnels because (1) hard rock tunnels include inherent uncertainties about rock mass properties (RMPs) and (2) excavation of these tunnels represents a sizable portion of such projects. Given the characteristics of hard rock tunnel projects, construction planners (CPs) for these projects frequently encounter cost overruns and schedule delays. To overcome this challenge, CPs for hard rock tunnels must be able to make informed decisions about resource-loaded excavation schedules and schedule adjustment policies (SAPs) in preconstruction and construction. These informed decisions require CPs to (1) take into account multiple sets of RMPs, (2) generate a cost-optimal schedule for every set of RMPs, and (3) evaluate expected total excavation costs (TECs) of schedules for multiple sets of RMPs, under given SAPs. However, currently CPs have difficulty making informed decisions about schedules and SAPs because they lack formal representations of estimation rationales and methods for schedule generation and evaluation required to support informed decisions. Thus, in preconstruction and construction, CPs cannot generate multiple cost-optimal schedules for multiple sets of RMPs in a timely fashion and they cannot evaluate the expected TECs of excavation schedules under given SAPs in a timely and consistent manner. In addition, existing ontologies cannot sufficiently account for multiple reference and non-reference sets of RMPs, transition costs and durations among construction methods at the method level of detail, times required for decision-making, and SAPs. Moreover, the existing methods cannot use those representations to generate cost-optimal schedules and evaluate the TECs of the schedules. To address those limitations, I developed a dynamic excavation schedule generation and evaluation framework (DESGEF) for hard rock tunnels. The DESGEF consists of (1) an ontology to represent estimation rationales of TECs for schedules and SAPs and (2) a methodology to generate cost-optimal schedules for multiple sets of RMPs and evaluate expected TECs for excavation schedules under given SAPs. To develop the ontology, I extended the existing construction method models. To develop the methodology, I specialized the existing dynamic programming and earthwork risk analysis methods for the excavation of hard rock tunnels. I validated the completeness of the ontology by interviewing 5 experts in tunnel construction projects. The validation results demonstrate that the ontology represents estimation rationales more completely than the current practice and existing approaches. In addition, to validate the speed, closeness to cost-optimality, and consistency of the methodology, I conducted Charrette tests with 12 test subjects. The validation results demonstrate that, in preconstruction and construction, the test subjects with the methodology (1) generate closer-to-cost-optimal schedules more rapidly, and (2) evaluate TECs of schedules and SAPs more rapidly and consistently than the subjects without the methodology.