Mamdani fuzzy based decision support system for prediction of groundwater quality: an application of soft computing in water resources.

Abstract

Groundwater is a primary source of living which also requires preservative measures for furture generations. Due to the lack of effective management technologies, the wastewater generated by rapid urbanization and industrialization is being disposed untreated, leading to groundwater contamination, caused by infiltration and accumulation. This problem has become more intense in major cities of India. The present work is based on determining the water quality using fuzzy index developed for the Perambalur district, Tamilnadu, India, from where 30 groundwater samples were collected from bore well as well as dug well sources. The research focusses mainly on chemical parameters like total hardness (T.H.), total dissolved solids (TDS.), potential hydrogen (pH), calcium (Ca2+), magnesium (Mg2+), potassium (K), sulphates (SO42-), total nitrates (NO3+ NO2), fluoride (F), bicarbonate (HCO3), carbonate (CO32-) and chloride (Cl2-). These parameters were assessed for fuzzy water quality index (FWQI) model, and the index was designed concerning Mamdani fuzzy inference system. Five FIS models with different linguistic variables were developed based on triangular membership function with the implementation of 189 numbers of rules. Finally, fuzzy model was classified into five categories, such as excellent, good, poor, very poor and not-suitable. Based on the results obtained from this model, 6 samples were classified into excellent, 8 samples into good, 12 to poor, 3 to very poor and 1 to not-suitable. In connection with that, the results of proposed model were compared with the output obtained from the deterministic method.