Featured Industry Collaborations
Water Flow Optimisation
Research partners: Mark Wallace, Guido Tack, and Ilankaikone Senthooran
Melbourne Water supplies drinking and recycled water and manages Melbourne’s water supply catchments, sewage treatment, rivers, creeks and major drainage systems. It manages a complex interconnected network of storage reservoirs, tunnels, aqueducts and bulk water-transfer pipelines. Distributing water through such systems requires careful planning to ensure optimal strategies to supply water. This requires consideration of many criteria including minimising operating costs and maximising water storage. We have developed a scenario-based interactive optimisation tool which can take demand forecasts for multiple years into account and incorporates potential drought-response plan actions based on modelled storage levels. The tool provides the flexibility to modify the underlying network configuration for the purposes of representing asset management or system improvement actions. This will enable the rapid and interactive evaluation of different plans for different operating scenarios, fast-tracking decision-making.
Plant Layout Optimisation & Interactive Visualisation
Industry partner: Woodside
Research partners: Mark Wallace, Maria Garcia de la Banda, Gleb Belov, Ilankaikone Senthooran and the Immersive Analytics team
Process plants – for fuels, chemicals and materials – cost billions of dollars to design and construct. Optimizing the layout of the equipment and connecting pipes is an important and complex task that aims to minimize the total cost of the plant while ensuring its safety and correct operation.Unfortunately, the complexity of this task is such that it is still done manually, taking multiple engineers several years to complete. There are no available tools to do this, and current research techniques are too simplistic and either give up optimality or do not scale.
We are working Woodside Energy Ltd. to develop and interactive visualisation and optimisation software that enables engineers to quickly design highly efficient plants, and to fix imperfections by interactive redesign and re-optimisation. Our current solution integrates realistic requirements, such as flexibility, maintenance and support constraints, can be optimally solved (within a given gap) for part of the plant, and can be visually and collaboratively explored via a web-server interface. Further work is in progress for increasing the scalability of the optimiser and incorporating interactivity in the visualiser.
The optimiser is implemented in the high-level modeling language MiniZinc. The use of MiniZinc has both reduced the amount of time required to develop the model and allowed us to easily experiment with different solvers in order to determine the best ones.
Industry partner: Peter MacCallum Cancer Centre, Alfred Health, University of Sydney Faculty of Physics, Sir Charles Gairdner Hospital, Western Australia
Research partners: John Betts and Guido Tack
One of the greatest challenges in cancer radiation therapy is treatment dose planning. This requires careful identification of the specific locations to be treated, and the amount of dose to be delivered at each location. The planner must ensure that sufficient dose is delivered for treatment to be effective, whilst also ensuring that the dose received by surrounding tissue is kept to a minimum. It takes an experienced clinician approximately one hour to plan the treatment for prostate cancer brachytherapy (the permanent implantation of very small radioactive “seeds”).
Working with clinicians and researchers at major cancer research hospitals, radiation physicists and image processing specialists, Monash optimisation researchers have developed software to assist clinicians prepare treatment plans. Using a 3D model of the patient’s prostate created from MRI data, our optimisation software produces a feasible treatment plan in under one minute. As well as offering high computational speed, the optimisation software also enables more sophisticated planning than can be achieved manually. This enables the creation of treatment plans that deliver high radiation doses to regions of the prostate where tumours have been detected whilst also reducing the radiation dose received by the sensitive surrounding tissue.