Implementing predictive dynamic process simulation technology to optimize sugar refinery vacuum pan operation

摘要

One of the remaining horizons in optimizing the sugar refining process is the automation of the crystallization sequence. High concentration sucrose solutions behave non-linearly which makes performing engineering calculations on the sequence complicated. New developments in process simulation software make developing and optimizing the crystallization step more attainable, with engineers able to incorporate process knowledge and data into a simulated model. The future of modelling software could also include a direct communication from the modelling software straight to a refinery's DCS. This report outlines the findings of preliminary optimization calculations of the crystallization sequence, using SysCAD, a modelling software that has valuable and specific sugar unit operations. A model of Louisiana Sugar Refining's plant operations was created in the software using available plant data and correlations developed by the modelling team. Different strike parameters were input to the model, along with cost and profit parameters, and the financial results were recorded for each scenario. Initial results indicate large cost savings associated with better valve tuning, improved insulation, enhanced pan control towards the end of a strike, high inlet liquor brix, and shifting focus to reducing strike times rather than strike cost. Furthermore, the results justify the continued investment and expansion for the project to identify more cost savings and improve accuracy. These scenarios will be tested and translated into refinery boiling sequence improvements, via LSR's Siemens automation system, that have the potential for large cost savings in the refinery. The project also demonstrates the future potential of artificial intelligence and reduction of the human factor in sugar crystallization.