A case study on Artificial intelligence based cleaner production evaluation system for surface treatment facilities

The metal finishing industry is water intensive. Surveys of South African metal finishing companies indicate that water consumption is as high as 400 L/m of metal surface treated, whilst best available practice can achieve less than 10 L/m. The industry uses hazardous chemicals such as chrome VI, cadmium, nickel and cyanide. If consumption of these chemicals can be optimized, quantities of heavy metals released into the environment will be reduced. In some cases where cleaner production techniques were applied by local companies, heavy metals have been completely eliminated from effluents discharged to municipal sewers, which represents a significant benefit to the urban environment. This benefit was accompanied by significant reduction in the use of chemicals, with a concomitant cost saving and competitive advantage to the companies concerned. A Danish environmental aid initiative promoted cleaner production in the South African metal finishing industry. Local consultants were trained by Danish experts in this field. The general methodology was to conduct an audit of the chemical, water, human resource and environment aspects of the company and compare it to best available practice. Once the review was completed, a detailed feasibility was carried out on systems and equipment required to reduce chemical consumption, water consumption, human resources and environmental impact. Applied to a number of South African companies, these methods have typically achieved reductions of the order of 90% in water use and 50% to 60% in the use of chemicals. There were difficulties in applying the Danish methodology to South African metal finishing companies, as it makes use of quantitative indices derived from the process operations. The companies are often small and technically unsophisticated, and do not have ready access to the process data that is needed. An alternate system is required to simplify the evaluation and optimization process. This paper proposes a case study on a fuzzy logic operator based evaluation system that outputs the cleaner production status of the company. The model is compared to an established cleaner production tool.