ENAC Seminar Series by Dr M. Kreitmair
Event details
| Date | 16.07.2019 |
| Hour | 13:00 › 14:00 |
| Speaker | Dr Monika Kreitmair |
| Location | |
| Category | Conferences - Seminars |
13:00 – 14:00 – Dr Monika Kreitmair
Postdoctoral Research Associate, Alan Turing Institute, University of Cambridge, UK
Uncertainty quantification in hydrokinetic energy extraction
Uncertainty is ubiquitous in the modelling of the natural world, with sources of uncertainty ranging from measurement errors, model simplifications, and numerical discretisation error. It is important to account for input parameter uncertainty when making model predictions as this can impact the mean values predicted, standard deviation, and associated optimisation strategies. In shallow flow modelling, a parameter with significant uncertainty is the bed roughness coefficient which accounts for the frictional interaction between channel beds and the flow past it.
This uncertainty, in turn, propagates through to velocity profiles and subsequent metrics, such as power predictions from hydrokinetic energy models and water elevation in urban flooding models. In this talk, we will explore two methods for the propagation of bed roughness coefficient uncertainty in models of tidal stream energy extraction. The first method is a perturbative approach which may be applied to analytic models, and the second is a numerical probability density distribution (pdf) transfer approach, whereby a given pdf in the uncertain parameter is transferred through a numerically generated power surface to produce the corresponding pdf in the output metric, i.e. power.
These methods allow the calculation of expected power and its standard deviation and enable optimisation of tidal turbine deployment under uncertainty. The methods have ready application in other free surface flow problems such as river flow routing and urban and coastal flooding, where they may be used to account for uncertain terrain and terrain cover.
Postdoctoral Research Associate, Alan Turing Institute, University of Cambridge, UK
Uncertainty quantification in hydrokinetic energy extraction
Uncertainty is ubiquitous in the modelling of the natural world, with sources of uncertainty ranging from measurement errors, model simplifications, and numerical discretisation error. It is important to account for input parameter uncertainty when making model predictions as this can impact the mean values predicted, standard deviation, and associated optimisation strategies. In shallow flow modelling, a parameter with significant uncertainty is the bed roughness coefficient which accounts for the frictional interaction between channel beds and the flow past it.
This uncertainty, in turn, propagates through to velocity profiles and subsequent metrics, such as power predictions from hydrokinetic energy models and water elevation in urban flooding models. In this talk, we will explore two methods for the propagation of bed roughness coefficient uncertainty in models of tidal stream energy extraction. The first method is a perturbative approach which may be applied to analytic models, and the second is a numerical probability density distribution (pdf) transfer approach, whereby a given pdf in the uncertain parameter is transferred through a numerically generated power surface to produce the corresponding pdf in the output metric, i.e. power.
These methods allow the calculation of expected power and its standard deviation and enable optimisation of tidal turbine deployment under uncertainty. The methods have ready application in other free surface flow problems such as river flow routing and urban and coastal flooding, where they may be used to account for uncertain terrain and terrain cover.
Practical information
- General public
- Free
Organizer
- ENAC
Contact
- Cristina Perez