APPLICATION OF ARTIFICIAL NEURAL NETWORKS FOR WATER QUALITY PREDICTION

Many domestic and industrial users are concerned about the hardness of water since it effects consumption of soap in laundry and formation of scales in boilers respectively for their applications. In the present study, empirical models based on multiple regression and artificial neural networks are developed to predict the value of hardness with respect to the corresponding values of chloride, fluoride, and calcium contents of the groundwater sample based on a region specific data. A thirty-point data set consisting of data regarding chloride, calcium, fluoride and hardness is taken and is used in developing the physical models for predicting the value of hardness based on the above-mentioned parameters. Initially a Multiple Regression Model is developed using Multiple Regression technique. The accuracy of the model is verified using a tenpoint data set by calculating the Standard Deviation (SD). The SD value in this study found to be high (0.404). Novel techniques such as Artificial Neural Networks (ANNs) can be used to predict the output from the data set with better accuracy than that using Regression technique. Hence, ANNs are used in the present study to predict the hardness of water using the above data base. Back Propagation Network of ANN is used for the study and the results are obtained. The SD value obtained in the ANN model is encouraging (0.00054).