Real-Time Construction Schedule Analysis of Long-Distance Diversion Tunnels Based on Lithological Predictions Using a Markov Process

Abstract

Schedules are major concerns in construction planning and management. The mutual interferences among construction activities of long-distance diversion tunnels (LDDTs) are complicated because of their crisscrossing and intensive layout, which will affect scheduling to some degree. Especially, lithology has a conclusive influence on underground construction. But the lithology determination methods of current construction simulation, such as the trend of extrapolation from observations, do not fully consider lithology’s inherent characteristics and the transition probability between different lithological classifications. In the present study, the authors propose a hierarchical simulation model coupling the critical path method (CPM) and cycle operation network (CYCLONE) to arrange the intricate construction sequence of LDDTs. Additionally, a Markov prediction model is embedded into the CYCLONE model as a specific component to consider lithological uncertainty and estimate construction parameters of each cycle automatically. Based on the proposed simulation model, a real-time construction schedule analysis system is built. A case study project is analyzed to demonstrate the feasibility of the methodology and to highlight its capabilities. For this project, a reasonable planned schedule is proposed, an optimal equipment selection is determined, the excavation intensity is balanced, and it is verified that a simulation model considering lithological uncertainty using a Markov process provides a higher level of accuracy in prediction than traditional simulations. Finally, regulatory measures are presented for site management based on the simulation results and the deviations between the actual and planned schedules.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 25 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.