A model identification approach for the evaluation of plant efficiency

Abstract Regulations and the public expectations on improving efficiency, reducing the carbon footprint and lowering the environmental impact drive the process industry towards improved operation and the development of new technologies. The efficiency of an existing production plant depends on a variety of factors like capacity utilisation, raw material quality, ambient temperature or operational performance. Identifying the influence of these factors on the performance of the plant helps to take suitable measures to drive it towards a more efficient operation. One approach to assess the resource efficiency potential of a plant is the comparison of the actual performance with the best possible operation under the given circumstances. This work presents a surrogate modelling approach for the identification of the best possible operation based on historical data. The surrogate model is compared to a more detailed rigorous model and advantages and possible shortcomings of the surrogate approach are discussed based on real production data at INEOS in Koln.