VAGE – IMPROVING THE VALUE OF VARIABLE AND UNCERTAIN POWER GENERATION IN ENERGY SYSTEMS
The energy sector is under transformation. The share of variable power generation, such as wind power and photovoltaic (PV), is increasing rapidly. Their output is dependent on weather and therefore much more variable and uncertain than the output from more conventional power generation. Variability and uncertainty brings challenges to power system operators and lowers the value of wind power and PV for the overall energy system and therefore also for the society at large. Variability decreases value since it causes periods with surplus electricity and periods with high net demand (demand minus the generation from wind power and PV – i.e. what other power plants need to provide for). Uncertainty decreases value since decision making under uncertainty is more difficult. Uncertainty leads to suboptimal decisions concerning e.g. when to store energy and when to start up power plants.
VaGe project objective is to improve operational decision making in the power systems when considering the variability and uncertainty of wind, solar, water inflow, heat and electricity demand, their correlations and possible sources of flexibility. Decision making under weather related variability and uncertainty is improved in two different time scales: 1) short-term power plant unit commitment and dispatch decisions (look-ahead up to 36 hours) and 2) medium-term optimization of storage use, consumer resources and other slow processes (look-ahead up to two weeks). More information, i.e. better and more comprehensive forecasts, and energy system flexibility can mitigate variability and uncertainty. Due to systemic interactions, it is important to assess all relevant sources of flexibility.
WORK PACKAGE (WP) 1: FORECASTS AND UNCERTAINTY
The first aim in WP1 is to develop methods for providing realistic weather dependent forecast uncertainty estimations at medium-term time scales. The development work rely on the existing data from global forecasting system, ECMWF-ENS. WP1 will develop a calibration model based on state-of-the-art statistical methods in order to further improve the forecast quality and their uncertainty estimates in general. The second aim in WP2 is to develop new methods to estimate the short-term forecast uncertainty that cannot be provided by global forecasting systems. Here the uncertainty estimate will be based on i) model error estimates by using stochastic physics and ii) boundary condition error by using forcing data from ECMWF-ENS. The new methods will be applied in an ensemble numerical weather prediction model HarmonEPS. Finally WP1 will develop automated conversion tools for casting forecasts and their uncertainty into energy terms.
WORK PACKAGE (WP) 2: DEVELOPMENT OF MULTI-SCALE OPTIMISATION METHODS
The operation of power systems has been typically planned using forecasts up to 36 hours ahead. With increasing uncertainty and utilization of flexibility from time constrained resources (e.g. building heating or batteries) this time scale will not be long enough for cost efficient system operation. However, it is computationally difficult to extend the time horizon in power system optimization models, since these models can already be slow to solve. VaGe tries to address this problem by developing multi-scale optimization methods where it is possible to consider the longer time horizon (up to two weeks using WP1 data) by simplifying and aggregating temporal and geographical elements in the model. At the same time, the first day or so is kept at high resolution for accurate unit commitment and dispatch that will be much better informed by the longer time horizon uncertainty and variability. This can result in better allocation of resources and consequently improved cost efficiency.
WORK PACKAGE (WP) 3: IMPROVING THE VALUE OF WIND POWER AND PV IN ENERGY SYSTEMS
The last work package will utilize the data and the tools developed in the previous work packages. It will test and explore hypotheses about the possibilities of enhanced forecasts and models. It will also investigate market designs that would better accommodate the use of longer time horizons in operational planning.
THE GENERAL OBJECTIVE OF THE PROJECT IS SPLIT INTO:
- Improve the uncertainty estimates of weather related power generation on both medium-term and short-term time scales
- Improve the representation and modelling of weather related uncertainties within the energy system optimization models – including a new model for the medium-term
- Find solutions to mitigate variability and uncertainty utilizing better forecasts as well as flexibility from biomass, consumer participation and electrification of heat and transport