14CSER007 - Probabilistic uncertainty assessments in energy-related problems
| Project number | 14CSER007 |
| Project title | Probabilistic uncertainty assessments in energy-related problems |
| Applicant(s) | Dr. S. Dubinkina, Prof.dr.ir. J.E. Frank |
| Abstract | Many important prediction problems in energy production are formulated mathematically as inverse problems for which observational data is available. Two examples are the prediction of wind farm output from weather data, and the estimation of gas and petroleum reservoir structure. Due to imperfect model assumptions, uncertainty in parameters, numerical errors, noisy and incomplete data, and the ill-posed nature of the inverse problems themselves, these problems are computationally complex, nonlinear, large dimensional and highly uncertain. Hence, it is desirable to develop probabilistic methods that provide not only a best estimate of the predicted quantities, but also quanti cation of the associated prediction uncertainties. In this project, we propose a new methodology for probabilistic uncertainty assessment based on simulations of an ensemble of plausible predictions/estimates with probability determined from available observations. We will apply this methodology for energy-related problems such as wind predictions and petroleum reservoir estimations. |
| Project budget | k€ 256 (includes 1 PhD student and material budget) |
| Duration | 2015 - 2018 |