Simulation-Optimization Of NAPL Contamination Remediation In Porous Medium

With the development of the oil industry,oil leak happens while producing,processing,transport and usage.The leaked fuel and organic chlorinated solvents,which is called non-aqueous phase liquids(NAPLs)will migrate from unsaturated zone to groundwater and exist in insoluble liquid form.Dense non-aqueous phase liquids(DNAPLs),which weighs more than water,will migrate downward to the bed rock and form a DNAPL pool by gravity,while few DNAPLs stay in unsaturated zone during migration.Compare with off-situ remediation technologies,in-situ remediation technologies have small disturbance to soil and don't cause secondary pollution.Simulation study of NAPL contaminant remediation with on-site remediation technologies could save experiment cost and provide theoretical bases for field remediation.AMALGAM(A Multi-algorithm Genetically Adaptive Multi-objective Method)is very effective to solve groundwater simulation-optimization problem as it adopts a new.Multi-phase multi-objective optimization problems need more attention as they are computationally expensive and consider more parameters relative to single phase multi-objective problems optimization problems.The diversity of in-situ remediation technologies enables multiple selection for NAPL contaminant remediation of soil and groundwater.DNAPLs was selected as a representative of NAPLs.Considering the advantages and disadvantages of different technologies as well as their suitable conditions,the paper choose surfactant enhanced aquifer remediation(SEAR)and in-situ air sparging(AS)to study the remediation process of DNAPL contaminant.Use co-injection of air and steam to avoid the re-condensation of DNAPLs on the water resisting floor and improve the remediation efficiency in AS remediation study.Besides,this paper presents the multi-objective optimal design of NAPL contaminant remediation based on AMALGAM.The main conclusions were drawn as follows:(1)The results of simulation in both homogeneous and heterogeneous media using UTCHEM were consistent with sandbox experiments.Surfactant flushing can remove most of DNAPL contaminant efficiently.The DNAPL contaminant stay in migration path was almost removed while the removal efficiency of the DNAPL contaminant on the water resisting floor and lenses is not very good as there are less contact area with surfactant solution.The DNAPL contaminant is more likely to be removed in heterogeneous medium than homogeneous medium.Increasing discontinuous ganglia increased DNAPL contaminant removal through solubilization by micellar,attributing to a larger contact area between surfactant solution and ganglia than pool.(2)The removal efficiency increased as the concentration of surfactant solution increased and/or the injection rate of Surfactant solution increased.The concentration of surfactant solution influences the removal efficiency of DNAPL contaminant on the discontinuous ganglia only,while the injection rate of surfactant solution influences removal efficiency of DNAPL contaminant on both discontinuous ganglia and pool.(3)While remediation with AS technology,DNAPL contaminant move downward to the right side of fine sand.The removal efficiency of DNAPL contaminant in coarse sand is higher than in fine sand.Couse the migration rate of DNAPL contaminant hysteresis steam front,DNAPL contaminant re-condensed on the water resisting floor in AS remediation with steam injection only.Co-injection of air and steam could avoid re-condensation phenomenon and increase the removal efficiency of DNAPL contaminant.(4)AMALGAM performs better than single evolution algorithm as it employs a diverse set of optimization algorithms simultaneously for population evolution,adaptively favoring individual algorithms that exhibit the highest reproductive success during the search.AMALGAM has a faster convergence speed and gets a perfect Pareto curve with less generations.For complicated multi-phase multi-objective optimization problems,AMALGAM with SP-UCI performs better than AMALGAM without SP-UCI.The Pareto curve of AMALGAM with SP-UCI is smoother,which demonstrate that SP-CUI is efficient and robust to intricate high-dimensional problems.