Flow Based Market Coupling

Abstract

The current market coupling algorithm used in the Nordic power market is the Net Transfer Capacity (NTC) model. It maximizes social surplus according to the transfer capacities between bidding areas, which are provided by the TSOs. The model does not account for the actual physical power flow in the grid.

Flow based market coupling (FBMC) gives a potential better utilization of the power grid because it to a greater extent accounts for the physical properties of the grid. A simplified grid model is provided to the market. The grid model contains Power Transfer Distribution Factors (PTDFs), which describe how a power injection in a node influences the lines in the grid. Since the Nordic power market is divided into bidding areas, the PTDFs must be aggregated in order to reflect how an injection to an area influences the lines in the grid. A Generation Shift Key strategy is used for this aggregation. The GSK strategy defines how the node-to-line PTDFs should be weighted in order to obtain equivalent area-to-line PTDFs.

There is no straightforward, theoretical way of determining a GSK strategy. Consequently, several GSK strategies have been developed, giving rise to the question of how to find good GSK strategies. The task of this thesis is to develop a method that makes it possible to compare GSK strategies. Thereafter, the thesis seeks to investigate the different GSK strategies that are considered to use in the Nordic power market.

The method is based on the deviation that occurs between the predicted power flow and the actual power flow. The idea is to use data from the previous day to estimate the flow of the current day. The estimated flow is compared to the actual flow, resulting in a flow deviation. Small flow deviations indicate that the applied GSK strategy is good. For each GSK strategy, the flow deviation is calculated for each critical network element (CNE) in the Nordic grid.

The thesis also studies how detailed one should apply the GSK strategies. One can use one strategy for all CNEs in the Nordic power market, but one may also apply different GSK strategies in different areas or for different CNEs.

The results showed that applying the optimal strategy of each area lead to a decrease of the global flow deviation. Similarly, using the optimal strategy for each CNE reduced the flow deviation of each area. Thus, using GSK strategies applicable for each area or CNE improves the results. Especially SE3 and SE4 experienced a large benefit using the optimal GSK strategy. Some GSK strategies were found to be generally better than others, where GSK 6 turned out as the best. GSK 7 and GSK 8 were not so good. NO1 and DK2 are examples of areas that are not sensitive to which strategy is applied. Some areas, like NO4 and NO5, experienced large flow deviations. This is due to certain CNEs where the flow is hard to predict.

The results of flow deviations were also divided into high load and low load. There were no major changes in the results compared to the results where all hours were included. The same strategies were best and worst. However, it was a trend that areas with large load were better predicted in high load hours.