A Neural Network application for a ballast water electrochlorination system

One of the principle factors for displacement of the non-indigenous species among marine environments is Ships’ ballast water. Even though only a very small portion of these organisms can survive the rough condition of ballast tanks, these survivors pose a big threat to the new environment they are introduced into. There are numerous proposed ballast water treatment techniques for eliminating diverse effects of ballast water translocation. Chlorine disinfection is one of the proposed chemical techniques for ballast water treatment. A lab-scale chlorine generation system for disinfecting ballast water organisms was used for testing and electrochemical cell design was proposed in our previous studies as well in this current study for artificial neural network applications. This study covers some of these experimental measurements and the Artificial Neural Network (ANN) applications as a non-linear statistical data modeling and a decision making tool. In this study, two different electrochemical cell types are tested to determine the properties of the disinfectant produced under various conditions. These cells are designed to produce chlorine directly from seawater and to be used onboard for ballast water disinfection. Two ANN models are utilized to find relationship between operational conditions and disinfectant quality parameters, i.e., total residual chlorine (TRC) and free available chlorine (FAC). The results which are presented with the ANN tool showed a very good correlation between input parameters and experimental results.