Overview
Rising energy demand and the imminent threat of climate change are critical issues in society today. Thermofluid systems provide the backbone of almost all energy conversion processes for renewable and conventional power generation, as well as heating and cooling systems such as heat pumps and refrigeration cycles. Components typically found in these systems include pipes, valves, nozzles, reservoirs, heat exchangers, evaporators, condensers, cooling towers, pumps, blowers, fans, compressors, turbines, boilers, and reactors.
The operation of these systems is governed by the fundamental principles of the thermofluid sciences including thermodynamics, fluid mechanics, heat and mass transfer, as well as turbo machines. The term fluid refers to the collection of substances known as liquids and gasses. Therefore, the working fluids in thermofluid systems can be liquids, gasses, or two-phase flows. These two-phase flows can include liquid- vapour mixtures such as those found in the boilers of the Rankine water-steam cycle and in the evaporators and condensers of heat pump cycles, or gas-solid mixtures such as those found in the flue gas and ash particle suspensions found in solid fuel furnaces and boilers.
The Applied Thermofluid Process Modelling Research Unit (ATProM) specialises in modelling these components and systems to evaluate novel technologies, improve the efficiency and control of processes, and detect anomalies for condition monitoring purposes.
Vision & Mission
Our vision is to expand knowledge, develop tools, and enhance skills in applied thermofluid process modelling that will contribute to the improvement of the flexibility, availability, and efficiency of industrial energy systems.
Our mission is to:
- Develop process models, simulation methodologies, and machine learning methods for the integrated analysis, design, optimisation, and condition monitoring of thermofluid energy systems.
- Provide training in thermofluid energy systems through postgraduate research and course-based programs, as well as support development of the undergraduate thermofluids curriculum.
Our competitive advantage is:
- Close collaboration with industry and other academics working on thermofluid process modelling.
- Direct access to developers of the relevant software tools and expert knowledge of the fundamentals that underpin the software.
- The ability to combine fundamental thermofluid principles and machine learning techniques to develop accurate and computationally inexpensive numerical tools to address industry needs.
- Access to high performance computing facilities, both in-house via several high specification workstations, and externally via the Centre for High Performance Computing (CHPC).
- A purpose designed postgraduate diploma programme focused on power plant engineering that supports the research focus.
Historical Overview
The ATProM research unit was established in 2017 primarily to host the Energy Efficiency (EE) specialisation of the Eskom Power Plant Engineering Institute (EPPEI) and was formally accredited in December 2017.
Initially the focus was on coal fired power plants to support Eskom, who provided substantial funding up to 2022. This focus included the Rankine water-steam cycle which is the main process common to coal fired, biomass fired and nuclear power plants and includes 90% of all plants in South Africa. The main tool employed at that time was the Flownex® Simulation Environment software, which is based on a one-dimensional thermofluid network modelling methodology, and was co-developed locally by the Honorary Prof Pieter Rousseau.
To support the transition to cleaner energy production, the application focus has since shifted gradually to include biomass boilers, supercritical CO2 (sCO2) Concentrated Solar Power (CSP) plants, and gas turbine Brayton cycles. Flownex® is still used extensively, but the tools and skills have expanded to include Computational Fluid Dynamics (CFD) via the Ansys Fluent software, as well as optimisation and machine learning techniques. These are mainly built on the Python programming platform utilising standard toolboxes which includes PyTorch for machine learning. The funding basis has also broadened to include other local and overseas industrial partners.