Power System Operation Planning and Wind Power Curtailment : Efficient Methods for Power System Scheduling and Integration Studies of Variable Renewable Energy

Abstract: To reduce carbon dioxide emissions, variable renewable energy (VRE) sources are replacing conventional fossil-based power plants for electricity generation. Due to the variability and uncertainty of weather dependent VRE sources, there can be situations when it is not possible to accommodate all the available VRE production, and VRE sources have to be curtailed. The need to curtail VRE production can arise, among other reasons, due to low electricity demand, lack of transmission capacity, or when operational security requires conventional units to remain online to provide intertia or reserves for balancing production and consumption. Managing VRE curtailment and integrating VRE sources into power systems in an efficient manner require tools for long-term power system planning and short-term power system operation planning, e.g., day-ahead scheduling which is handled by electricity markets. This thesis develops tools for both long-term and short-term power system planning, with a focus on estimating the need for VRE curtailment in future power systems and methods that achieve efficient operation of power systems by allowing VRE curtailment. Regarding long-term power system planning, an open source dispatch model for the Nordic power system, ODIN, is developed and used to assess the future need for VRE curtailment arising from wind power expansion mostly in the north of Sweden. Regarding short-term power system operation, further developments of the previously proposed power-based formulation for unit commitment (UC) are made, extending the formulation to include reserves which can better deal with wind power variability and uncertainty, as well as contingencies such as line and generator outages. Also, different situations when VRE curtailment can be efficient and lead to reduced system costs and carbon dioxide emissions are investigated, and an open source model for generating realistic wind power production scenarios for use in UC formulations is developed. Finally, a power-based version of ODIN is implemented to investigate the benefits of using the power-based formulation for production cost models used for long-term power system planning.The methods and models developed in this thesis can contribute to more efficient long-term planning and short-term operation of power systems, particularly in the Nordic region. Excessive VRE curtailment should be avoided through efficient long-term planning, but in the short term the flexibility of VRE production should be used to operate the power system in a way that minimizes system costs.