Comparison of Water Quality Index Values and Groundwater Quality Making Use of Machine Learning Methods: A Case Study
B. Vamsi *
Department of Civil Engineering, Lingayas Institute of Management and Technology, Vijayawada -521212, Andhra Pradesh, India.
Soniya Chowdary Anatha
Department of Computer Science and Engineering, Lingayas Institute of Management and Technology, Vijayawada - 521212, Andhra Pradesh, India.
T. Dheeraj
Software Development Engineer, Freelancer and Uzvi Service, Software Technology Parks of India, Vijayawada - 520008, Andhra Pradesh, India.
K. Upendra
Department of Computer Science and Engineering, Lingayas Institute of Management and Technology, Vijayawada - 521212, Andhra Pradesh, India.
T. Gowri Pravallika
Department of Artificial Intelligence and Machine Learning, Lingayas Institute of Management and Technology, Madalavarigudem, Vijayawada-521212, Andhra Pradesh, India.
N. Gnanasri
Department of Artificial Intelligence and Machine Learning, Lingayas Institute of Management and Technology, Madalavarigudem, Vijayawada-521212, Andhra Pradesh, India.
R. Anjanadevi
Department of Computer Science and Engineering, Lingayas Institute of Management and Technology, Vijayawada - 521212, Andhra Pradesh, India.
G. Siri Vennela
Department of Artificial Intelligence and Machine Learning, Lingayas Institute of Management and Technology, Madalavarigudem, Vijayawada-521212, Andhra Pradesh, India.
G. Madhusri
Department of Artificial Intelligence and Machine Learning, Lingayas Institute of Management and Technology, Madalavarigudem, Vijayawada-521212, Andhra Pradesh, India.
T. Hemalatha
Department of Artificial Intelligence and Machine Learning, Lingayas Institute of Management and Technology, Madalavarigudem, Vijayawada-521212, Andhra Pradesh, India.
B. Praveen Kumar
Department of Geology, Andhra University, Andhra Pradesh, India.
*Author to whom correspondence should be addressed.
Abstract
The variation in groundwater quality resulting from societal changes is a matter of concern, as groundwater is considered as one of the most vital sources of water supply among all available resources. Assessing water quality through systematic monitoring provides a basis for determining its suitability for various purposes, including effective water quality management. The present study focused on leveraging a hierarchical reconciliation algorithm (HCA) for forecasting water quality parameters and the Gradient Boosted Tree, Decision Tree and Random Forest models to predict Water Quality Index (WQI) values. This study was based on the experimental results conducted for monitoring and evaluating the quality of groundwater samples collected from two locations namely, Mudirajupalem and Lingayas Institute of Management and Technology (LIMAT), Vijayawada campus, Krishna district, Andhra Pradesh, India for the parameters, such as alkalinity, pH, total dissolved solids (TDS), total hardness (TH) and acidity using standard methods. The pH value was in the range of 8.5-10 for the Mudirajupalem samples, whereas pH value in the range of 6.5-8.0 was obtained in the LIMAT campus samples. It was reported that a range of 460-850 mg/L alkalinity was obtained in the Mudirajupalem samples, whereas a range between 550-750 mg/L of alkalinity was obtained in the LIMAT campus samples. Very high values of TDS were obtained in the Mudirajupalem samples with a range between 931-994 mg/L, whereas TDS values ranged between 199-273 mg/L in the LIMAT campus samples. It was reported that TH ranged between 20-246 mg/L for the Mudirajupalem samples, whereas for the LIMAT campus, TH values were in the range of 0-230 mg/L. Groundwater from LIMAT campus and Mudirajupalem samples showed the WQI values in between 39.59 to 41.01 and 396.73 to 397.09, respectively, which confirms that Mudirajupalem groundwater is not fit for public consumption. The foreseen WQI values were found to be similar with the results obtained via experiments, reinforcing the conclusion that the groundwater in Mudirajupalem is unsuitable for public consumption. Additionally, the HCA, which was employed for forecasting key water quality parameters proved to be effective. This study demonstrated promising results in predicting groundwater quality, opening up opportunities for further research into the use of advanced machine learning techniques to achieve even more accurate long-term groundwater quality predictions.
Keywords: Groundwater analysis, machine learning, water quality index, total hardness, forecasting