Optimization of a Groundwater-Monitoring Network For an Industrial Site by The Combination of MCLP and GAMS Techniques (Case Study: Shazand Oil Refinery, Iran)

Abstract In this research, the Maximal Covering Location Problem (MCLP) method was employed to discretize Shazand Refinery Complex in Iran and the General Algebraic Modeling System (GAMS) model was used to optimize boreholes number. In an industrial area like an oil refinery, the contents of storage tanks, leakage history, groundwater flow direction and the facility’s leakage potential can be considered as the weighting factors for the GAMS model. As the fourth refinery of Iran from the production volume value view of point, the Shazand Oil Refinery imposes a high risk of leakage into the underlying aquifer and was chosen for this study to evaluate the applicability of the MCLP and GAMS methods at the refinery field scale. Due to optimization, the initial estimated number of boreholes by the MCLP model for the study area was reduced from 410 to 230. A high density of optimal boreholes was dedicated to the processing unit area and storage yard because of a given priority for these areas that have a history of leakage and those located downgradient of the tanks containing light oil products. The contamination potential map shows that high potential areas correspond with a high density of boreholes located at the north, west and east parts of the refinery site. The network was installed with the drilling of 65 monitoring wells and a round of sampling was carried out. Analysis for oil compounds revealed the aquifer contaminated areas and confirmed the monitoring network efficiency and accuracy of the applied designing method.