Optimization of the hydrotesting sequence in tank farm construction using an adaptive genetic algorithm with stochastic preferential logic

Abstract

In the construction of tank farms there is a requirement for the tanks to be hydrotested in order to verify that they are leak proof as well as proving the lack of differential settlement in the foundations. The tanks will be required to be filled to a predetermined level and then to maintain this loaded state for a certain period of time before being drained. In areas such as the Middle East, water for hydrotesting is not freely available as sea water is often not suitable for this purpose, so fresh water needs to be produced or transported to the construction site for this purpose. It is therefore of major benefit to the project to schedule the hydrotesting of the tanks in such a manner as to minimize the utilization of hydrotest water. This problem is a special case of the resource-constrained project scheduling problem (RCPSP) and in this research the fitness differential adaptive genetic algorithm previously developed by the authors has been modified to enable the solution of this real world problem. The algorithm has been ported from the original MATLAB code into Microsoft Project using Microsoft Visual Basic for Applications in order to provide a more user friendly, practical interface.