Main content area

Disturbance modelling through steady-state value deviations: The determination of suitable energy indicators and parameters for energy consumption monitoring in a typical sugar mill

Mkwananzi, Thobeka, Mandegari, Mohsen, Görgens, Johann F.
Energy 2019 v.176 pp. 211-223
coproducts, crystallization, energy efficiency, heat transfer coefficient, imbibition, models, monitoring, prediction, recycling, steam, sugar industry, sugarcane, sugarcane bagasse, sugars, syrups
The growing emphasis in alternative revenue streams, from the conversion of sugarcane bagasse into energy co-products has prompted a desire by the sugar factories to achieve higher levels of energy efficiency. This study presents a comprehensive overview of the parameters whose steady-state deviation leads to excess energy use in sugar mills. The effect of eight operating parameters on the defined energy indicators was investigated, using a MATLAB simulation of a 250-tonne per hour sugar mill. The captured energy trends were used to develop energy prediction models based on the parameters whose steady-state offsets resulted in excess energy use. High prediction accuracies of over 95% were obtained when the developed energy prediction models were validated for seasonal cane quality and evaporator heat transfer coefficient variations. Increased A-massecuite recycling had a more pronounced effect on the energy used per kilogramme of water in the evaporator unit as compared to the crystallisation unit. Of the eight parameters, elevated imbibition water use and A-massecuite recycling had a cumulative percentage effect of 79% on the overall steam used per tonne of cane and sugar produced. Hence increasing the syrup concentration, decreasing imbibition water use and massecuite recycling are preferred operational strategies for improved energy-efficiency in a sugar mill. Overall, this study strengthens the existing literature by illustrating an approach for developing suitable energy indicators for excess energy use monitoring in sugarcane mills.