Analysis of construction processes with nonstationary work task durations

A construction process is regularly subjected to changes in conditions. These changes may cause individual work task durations to vary over time. In such a case, the process is defined as nonstationary. A significant factor causing variations in work task durations is the change in locations of work areas. Such changes are apparent in linear operations like earthmoving where loading and dumping are performed in different cut and fill areas respectively, and concrete pouring in high-rise buildings as the pouring operation advances to higher floors. Simulation provides an excellent tool for the analysis of construction operations with their dynamic nature. Very little effort has been directed towards simulation analysis of nonstationary operations. The reason is mainly the lack of an efficient tool which allows for the incorporation of nonstationary durations in the simulation model. In this dissertation, a new computer system for the analysis of nonstationary construction processes is introduced. It integrates a simulation system, STEPS, which has been developed recently at the University of Maryland, with a process planning module. The integration allows that a work sequence can be established according to a chosen work plan, which feeds automatically into the simulator. Earthmoving is used as a test-bed for this research. The integrated system allows for the designation of cut and fill areas, with their locations and quantities, and for specifying the sequence of moving earth among various areas. Condition factors may be assigned to cut areas to reflect the effect of different soil conditions on excavation time. The effect of the change of hauling distances, which results in nonstationary durations for traveling work tasks, was studied. It was found that process duration and resource idleness when considering the impact of nonstationarity were significantly different from those obtained when that impact was ignored. An approach to the allocation of resources, which varies the number of resources in relation to the hauling distance, is presented. For different test cases, the use of this approach always resulted in less duration and cost, and higher resource efficiency than results obtained using a fixed number of resources.