Design of sustainable coffee processing wastewater treatment system using K-means clustering algorithm

Abstract

Coffee processing wastewater has a COD of 3100-14343 mg per liter and a BOD of 5000-35000 mg per liter that will cause water pollution and may contaminate the water ecosystem if it is dumped directly into the environment. The literature shows that coffee processing wastewater requires special treatment to reduce the negative impact on the environment. This study develops a wastewater treatment system model capable of identifying sources of emissions and pollutants to increase the effectiveness of pollutant reduction. This study aims to analyse the coffee processing wastewater treatment system, define the most significant attribute, and develop a coffee processing wastewater treatment design using a K-means clustering method. A relief-feature selection method was used to analyze the most significant attribute of coffee processing wastewater related to certain specific wastewater treatment needed. This study used an unsupervised machine learning technique to develop clustering based on the most significant attribute of coffee processing wastewater using the K-Means method. The K-means clustering result shows five coffee processing wastewater treatment clusters based on BOD and Acidity. For clusters with BOD < 25.000 mg per liter: biological and chemical wastewater treatment will be applied through coagulation, adsorption, filtration, bio-filtration, and anaerobic microbial processes to separate pollutants. For clusters with BOD >25.000 mg/l: physical wastewater treatment will be applied through electro-coagulation.