Congestion management in the Dutch power sector

Abstract

Due to the introduction of a new grid connection policy, transmission system operator TenneT expects congestion to arise on the Dutch transmission grid in the near future. This new connection policy was introduced by the Ministry of Economic Affairs to abolish the discrimination between existing grid users and new entrants, and should improve competition. It allows generators to be connected to the grid directly, without having to wait for transmission capacity expansions that may be required. As this could cause transmission flows as desired by market parties to exceed the available capacity, TenneT must apply congestion management in order to guarantee the safe and reliable operation of the transmission grid. The Ministry decided that basic system redispatch should be used to manage congestion. This method was regarded the most appropriate short-term implementable option available, but has some drawbacks nonetheless. In existing literature it is argued that it potentially leads to high costs, that it is vulnerable to strategic bidding, and that it creates economically sub-optimal outcomes from a grid efficiency perspective. This study has quantitatively evaluated the application of the method in the Netherlands, in terms of congestion costs, their allocation, the incentives it creates, and the opportunities for (and the consequences of) generators bidding strategically. These outcomes were compared to three other congestion management methods (market splitting, market coupling, and the APX-based method1), in order to assess the validity of the proposition that market-based methods, which form the current trend in Europe, lead to better outcomes. Using a quantitative model of the Dutch electricity system the application of all four congestion management methods was simulated. This was done under four different scenarios, each of which was based on extreme conditions that were expected to contribute to congestion in parts of the grid: • Low wind availability in Germany • Cheap natural gas • Green revolution • Code red The simulations revealed that the transmission link between the Maasvlakte region and the Ring is most prone to become congested. However, this study also found that the resulting congestion costs will be low. This is the case because the variable cost levels of production units in the areas upstream and downstream from the congested grid segment were found to be very similar. A deviation from optimal dispatch will therefore result in only slightly higher dispatch costs. Under the most extreme scenario conditions, in which 1292 MW needs to be redispatched from the Maasvlakte to other areas of the Netherlands, net congestion costs were found to be € 231 / hr. On a yearly basis this would be € 2 mln., which is significantly lower than cost estimates found in literature, which expect this cost to be in the order of magnitude of € 10–100 mln. To identify the most appropriate congestion management method for the Netherlands, multi-criteria decision analysis (MCDA) was applied to compare the methods, in a pairwise manner and on the basis of eleven criteria. The analysis revealed that conflicting objectives preclude the identification of a single most appropriate congestion management method. It found that the APX-based method outranks market splitting and market coupling, but it remained inconclusive with respect to the appropriateness of basic system redispatch in comparison with these methods. The policy objectives of the Ministry thus appear to be different from those presumed elsewhere and by existing literature, considering their explicit preference for market-based methods. In order to improve the results of this analysis, the Ministry must reassess its objectives with respect to the conflicting criteria of proportionality, and long-term generator and TSO incentives. Also, additional research should improve the conclusiveness of the model results that were used for MCDA, as this would contribute to a more conclusive recommendation on method appropriateness. In particular, such research should encompass the options for incorporating a renewable energy compensation scheme under market-based methods, and it should, by constructing a more extensive, continuous, agent-based model that is capable of incorporating the strategies pursued by individual generators, provide a broader insight and more detailed data on the extent of congestion and the (resulting) consequences of strategic bidding.