Development of Efficient Pollution Source Identification Model Using ANN-GMS-GA Based Simulation-Optimization Approach

No Thumbnail Available
Journal Title
Journal ISSN
Volume Title
Identification of unknown pollution sources is an important and challenging task for the engineers working on pollution management of a groundwater aquifer. The locations and the transient magnitude of contaminant sources can be identified using inverse optimization techniques. In this approach, an error function is formulated which can be minimized using an optimization algorithm. The error function is the difference between simulated and observed contaminant concentration at observation locations. The observed concentration is measured in the field. On the other hand the simulated concentration can be calculated using aquifer simulation model. As such, there is a need to incorporate the aquifer simulation model with the optimization model. Thus technique is also known as simulation-optimization approach as aquifer simulation model is incorporated with the optimization model. The performance of the source identification model is highly related to the aquifer simulation model. Incorporation of sophisticated numerical simulation model will give better performance, but the model will be more computationally expensive. On the other hand, the model will be computationally less expensive, if approximate simulation model like ANN model is incorporated with the optimization model. However, in this case, there will be a reduction in the predictive performance of the model. Keeping this in view, this study develops four improved methodologies for identification of unknown groundwater pollution sources. In the first approach, the groundwater modeling system (GMS) is linked with optimization model for solving source identification problem. The incorporation of GMS with the optimization model enables to solve bigger real world pollution source identification problems. The main challenge of this approach is the linking of the external aquifer simulator, GMS with the optimization model. This has been overcome by executing GMS in Matlab environment. The main drawback of the approach is that the approach is computationally expensive. For reducing the computational time, the second approach uses artificial neural networks (ANN) model to simulate the flow and transport processes of the aquifer. The ANN model is then externally linked with the optimization model. This approach drastically reduces the computational time of the simulation-optimization model. The problem which was solved in few days can now be solved in few hours. However, most of the time, it yields only the near optimal solution. Therefore, in the third approach...
Supervisor: Rajib Kumar Bhattacharjya