Multi-objective optimisation of an integrated sugar mill for economic enhancement

This thesis covers the development and implementation of an integrated sugar mill (ISM) model and its subsequent use to perform multi-objective optimisation (MOO) to generate valuable data that assists production staff to optimally operate a sugar cane mill with extensive co-generation capability. A sugar cane mill manufactures crystallised sugar from sugar cane. There is an extensive history around milling and it is a well understood process. The introduction of large scale electricity generation through the combustion of excess bagasse, a renewable energy source made from the fibrous remainder of processed sugar cane, provides new opportunities. Significant, and currently increasing, revenue can be made through generation and traditional operating strategies do not fully take this into account. Steady-state models of Pioneer Mill in Queensland, Australia, a sugar cane mill with substantial co-generation and generation capabilities, are developed. These models can be redeveloped and applied to other mills. Each section of the mill is modelled separately, using a combination of existing models in the literature, building from first principles and using empirical relationships. These models are used to estimate the system parameters representing the state of the sugar mill, using routinely measured operational data. Combining the separate models into an ISM and using these parameters, predictions are made on the performance of Pioneer Mill in response to changes in operating parameters. MOO is applied to the ISM using four objectives: cane throughput, sugar lost, electricity generated and bagasse produced. In the optimisation, sugar lost is minimised while other objectives are maximised. The Pareto-optimal solution, representing the set of solutions where there are optimal trade-offs among objectives, is analysed for guidelines on the characteristics of optimal operation. A simple method of weighting is used to allow production staff to easily select a point from this solution that meets current priorities and determine the operating parameters for optimal operation of the mill.