Study On Groundwater Pollution Tracing In A Chemical Park In Northwest China Based On Simulation-optimization Models

Water is an essential resource for sustaining life and is the foundation of human society’s production and daily life.Groundwater is one of the most important water resources,serving as a primary source for drinking water and irrigation,as well as playing a crucial role in maintaining healthy ecosystems.However,with economic development,human activities are increasingly impacting the groundwater environment,particularly with regard to the pollution generated by industrial production,which is difficult to detect,trace,and expensive to manage.Therefore,this paper uses numerical simulation methods to trace and predict the migration of pollutants in groundwater in a chemical industrial park.Building on traditional modeling methods based on MODFLOW and MT3 DMS,this study uses the Python-based groundwater simulation package FloPy,along with Bayesian optimization methods,Strengthen Elitist Genetic Algorithm(SEGA),and non-dominated sorting genetic algorithm-Ⅲ(NSGA-Ⅲ),to construct a simulation-optimization model for groundwater pollution tracing,targeting the sources of pollution with an unknown location.This study mainly achieved the following results:(1)According to the statistical analysis of the water chemistry test results of monitoring well 41 in the research area,it can be seen that among all the samples,chloride has the highest average concentration,and benzene has the highest concentration in a single sample.Compared with the Class III water quality standards,the most serious average concentration exceeding the standard is ammonia nitrogen,which exceeds the standard by 140.8 times;and the most serious single sample concentration exceeding the standard is also ammonia nitrogen,which exceeds the standard by 4823.1 times.According to the classification of water quality standards,manganese is the most severely polluted among all samples.;(2)By adjusting the release time,concentration,and location of the pollution sources in the numerical model,the measured values of pollutants at the concentration monitoring points were fitted to confirm the location of the pollution sources.The fitting results showed that this method achieved good fitting effects both in time and space.Currently,the groundwater pollution in the study area is still in the early stage,with a small pollution range and high pollution concentration.The maximum concentration of Mn2+ is around 160 mg/L.;(3)A simulation-optimization model coupling genetic algorithm and solute transport model was established,which successfully inverted the location of an unknown pollution source.The simulated pollution history of the inverted source matched well with most of the measured values in time and space.Based on the tracing results,the pollution source and its history in sub-study area 1 were identified.;(4)The results of the prediction of future pollution migration trends show that in the past5 years,the maximum concentration of pollution has decreased rapidly,and by the 5th year,it has decreased to around 36 mg/L.The rate of decline slows down after 5 years,and by the 20 th year,the maximum concentration is still around 10 mg/L.In the short term,the pollution is concentrated within the chemical park,but after 20 years,the pollution plume will migrate to the downstream wasteland outside the park,basically leaving the park area.Although it will not affect downstream residential areas after 20 years,it is highly likely to migrate to residential areas in the longer term,so it is necessary to take remedial measures.The research results of this paper provide references for preventing groundwater environmental degradation,formulating groundwater pollution control plans,and determining liability for groundwater pollution incidents.The research method presented in this paper promotes the application of complex groundwater tracing theory to practical production activities.