Research Projects
Optimization and Planning of Hydro-Power Production
Introduction
A significant portion of electric power in Switzerland is produced by storage hydro power plants, which enable the storage of water inflows to schedule power production at convenient times, i.e. producing energy in most attractive economical conditions, i.e. when electricity prices are high. The storage capacity is however limited and the evolution of influence factors such as electricity prices and water inflows not fully known in advance and volatile. Therefore to take a correct decision about the utilization of the water in reservoir is not trivial at all.
A traditional way of addressing these uncertainties is to rely on tacit knowledge and/or improve the forecast of the volatile factors (e.g. of a Hourly Price Forward Curve (HPFC)). This is however becoming increasingly more difficult with the structural market changes (such as introduction of new products - e.g. power Futures and options or ancillary services), price dynamics (e.g. negative prices) and changing weather conditions.
Project description
The planning of production for storage power plants is usually done in three time horizons, with the following characteristics:
- The long-term planning covers a time range longer than one year, up to several years. Its purpose is to give assistance in e.g. investment planning problems.
- A medium-term planning will produce monthly or weekly reference sea values of the reservoirs in a hydro power plant portfolio as well as water values of the water stored in the reservoirs. The time horizon is one year with a monthly or weekly time step. The key idea here is to incorporate the uncertainty in market prices and water inflows in a robust optimization. Solution concepts here are stochastic programming, especially stochastic dynamic programming (SDP) [1].
- The short-term planning finally performs a detailed production plan for the power plant (portfolio). Because of the complexity of such a plan the time horizon consists only the next day(s) with a time step of an hour or a quarter of an hour. Solution concepts here are non-linear optimizations and evolutionary strategies (ES), taken into account estimations of the future market prices (HPFCs) and boundaries or water values out of the medium-term optimization.
For a medium-term planning, ancillary services, namely secondary control, was integrated in a SDP approach. The offering of secondary control is modeled as first-stage decision whereas the actual production of energy is the recourse action. A non-linear two-stage mixed-integer stochastic program is obtained and solved for a yearly time horizon, in order to estimate water values as well as optimal decisions about offering of secondary control [2].
Within the short-term context a software tool was created. A simulation as well as a deterministic and heuristic (ES) optimization can be performed on a hydro pumped storage power plant. Input and output of data is given in Excel files.
Also proposed was a modeling approach for an unified short- and medium-term optimization. Future power products as well as HPFCs can be considered in a natural way. A two-stage mixed-integer stochastic program with variable time periods is obtained, used to estimate water values as derivative of the profit-to-go function [3].
Goals of the project
The goal of this project is a prototype software which assists power producers with an optimal medium-term production strategy for typically Swiss hydro storage power plants. Key aspects of the software is the integration of the market and regulatory conditions in Switzerland (e.g. ancillary services) as well as the consideration of uncertain factors (e.g. water inflows, market prices). With this software it shall be not only possible to plan the energy production more profitable and more robust but also to plan it with a minimized and known risk.
References
[1] John W.Labadie;“Optimal Operation of Multireservoir Systems: State-of-the-Art Review”; Journal of Water Resources Planning and Management; pp 93-111; 2004
[2] Hubert Abgottspon, Göran Andersson; “Approach of Integrating Ancillary Services into a medium-term Hydro Optimization”; summary accepted for SEPOPE 2012
[3] Hubert Abgottspon, Matthias Bucher, Göran Andersson; “Stochastic dynamic programming for unified short- and medium-term planning of hydro power considering market products”; accepted for PMAPS 2012
