Artificial Intelligence
Machine learning and optimisation to guide energy system decision making.
Machine learning to accelerate system design studies
Our machine learning models enable us to increase the speed of engineering analysis by orders of magnitude.
We achieve this by training our models to predict the outcome of simulation studies. For example, we have trained our state-of-the art machine learning models to replicate the outcome of computationally intensive computational fluid dynamics and finite element simulations in a fraction of the traditional calculation time.
The trained model can then be used directly in the system design process to make design decisions which enable much more rapid iteration on potential solutions.
Abbass, A., Rafiee, A., Haase, M., and Malcolm, A., “Geometric Convolutional Neural Networks: A journey to surrogate modelling of Maritime CFD”, In: IX International Conference on Computational Methods in Marine Engineering (MARINE 2021), 2021.
Remaining useful life of energy system assets
Remaining useful life (RUL) is the length of time a machine is likely to operate before it requires maintenance.
RUL prediction requires historical data showing how long it took for similar machines to reach failure. We have developed accurate RUL prediction architectures based on the state-of-art deep learning transformers that can account for data collected from different equipment and a large numbers of sensors.
The outcome is an optimised maintenance regime, reducing operational costs and improved overall project returns.
Optimisation to deliver superior wind farm performance
A classic problem in wind farm energy systems is optimisation of the turbine placement. There can be various objectives, such as maximising the energy production by reducing the wake effects.
The wake effect, caused by wind turbines impacting each other's performance, has long been a challenge. With the high dimension search space arising from numerous turbines, finding an optimal layout is a challenging problem.
By strategically repositioning turbines, we can significantly reduced wake losses and increased overall energy production. In this example, we optimsed the layout configuration of 50 turbines to achieve maximum total energy over period of several years of wind data (ranges of speed and direction).
Subscribe for updates
+61 404 605 227
Level 1, 22 Stirling Highway, Nedlands, WA6009, Australia
Causal Dynamcs Pty Ltd
ABN 21 664 625 521