Development of a Unit Commitment and Optimal Dispatch model used to evaluate the effect of variable Renewable Energy Sources on the power system

Abstract

Abstract. The growing problem of climate change has risen global concerns about the way of using natural resources and has brought several initiatives such as the EU 2030 targets towards the cleaner energy production and the emissions reduction. The high penetration of Renewable energy Sources (RES) is considered as a remedy to this issue. However the higher integration of RES to the system is a rather challenging task for it's different components. Therefore there is a growing need for the introduction of higher flexibility to the system as a countermeasure for the unpredictable and unstable RES generation.

In this report the flexibility potential coming from a portfolio of smart Domestic Hot Water Tanks (DHWT) is investigated. Their ability to provide Active Demand Response (ADR) in order to contribute to the more cost efficient electric system's operation is studied. These systems could allow to modify their electrical load pattern without affecting the final, thermal energy service they deliver, thanks to the thermal inertia of the system. The creation of a general model able to simulate the operation of DHWT is developed. This model is intended to be used in a unit commitment and power dispatch model in order to evaluate the contribution of Demand Side Management (DSM) into the higher RES integration. To accomplish this task first two detailed modeling approaches were followed. The first was was a Rule-Based Control (RBC) strategy able to simulate several thousands individual DHWT. This model was computationally efficient and allowed the simulation of 20000 tanks for 5 days with a sub-hourly operating time scale. The second analytical approach was a linear program (LP) that performed the same computation but for a decreased amount of tanks due to the high computational burden. In both models the ability of a centrally controlled (by an aggregator) portfolio to provide capacity and activation reserves are explored. The two approaches are compared in order to identify benefits and limitations. These models were the main guide towards the modeling of a Virtual Storage Plant (VSP) model. This model is able to follow and simulate the behavior of an increased number of DHWT without much detail. The assumptions made in the VSP model formulation are further explained. The ability of this model to perform in standard conditions and to participate in Demand Response operation are validated through the detailed models. This process lead to the creation of a VSP model that is able to simulate the behavior of a fleet of DHWT. The flexibility that can be offered along with the rate of RES penetration can be investigated by the integration of this model to Dispa-SET, a unit commitment and power dispatch model able to simulate the electric system.

Finally a part of this thesis is related to the improvement of the Dispa-SET simulations and the overall computed system's cost. This was performed by collecting yearly generation data available at Entso-e's website from thousands of generation units in Europe. Afterwards several methods were developed in order to acquire important real-time values for the simulation parameters. These methods are tailor made and are explained in detail. To our knowledge this is the first attempt to collect and process real time generation data for European units.