Dynamic pricing and carbon intensity in demand response functions

Abstract: The European power sector is facing significant challenges related to investments in grid infrastructure and generation capacity. The continued deployment of intermittent renewables also puts pressure on current grid conditions. Smart grids is seen as a cost-efficient way to overcome these challenges through a more efficient use of current capacity. Demand response is a corner-stone in smart grid development,  and is implemented to introduce flexibility on the demand side. Most demand response programs have used dynamic pricing to incentivize consumers to shift consumption from peak to off-peak hours. In Stockholm Royal Seaport, where a sustainable energy system is envisioned, it has been proposed that dynamic pricing should be complemented with an indicator depicting carbon intensity of purchased electricity. This indicator is based on average emissions, which is one of two fundamental perspectives on assessing environmental impacts of electricity consumption.  The aim of this study was to evaluate whether the approach used to quantify carbon intensity in Stockholm Royal Seaport is appropriate in the context of demand response. To achieve this, a literature review has been conducted regarding potential benefits of demand response, power system dynamics and carbon dioxide allocation methods. A quantitative analysis has also been conducted, where the signal proposed for Stockholm Royal Seaport has been modeled under different timeframes. The results show that the CO2-signal in Stockholm Royal Seaport is constructed in such a way that it is largely affected by hydro generation, which in turn makes it correlate negatively with price. As a result, the CO2-signal would counteract many of the predicted long-term benefits of demand response. Furthermore it seems unlikely that the signal would result in significant short-term emission reductions, since hydro generally is used to balance supply and demand in the Swedish and Nordic systems.  Based on the literature review, it was concluded that marginal emissions would be a more appropriate environmental indicator than average emissions. However, it remains a difficulty to construct a day-ahead control signal based on this perspective because of system complexity and lack of data. Historical marginal carbon intensity was nevertheless modeled in this study using a linear regression model. The results indicate that price itself might be a sufficient indicator of marginal emissions. Finally, a model for a signal based on prognoses of intermittent renewable generation is proposed, where the rationale is that consumers should decrease consumption during hours of low renewable generation. This signal was modeled using data on renewable generation from Denmark since corresponding data in Sweden is not yet available. Results show that it would be possible to construct a rather accurate control signal in this way. There are also reasons to believe that demand response based on this type of signal would result in long-term environmental benefits.