Evaluating demand charges as instruments for managing peak-demand

Reducing peak demand in distribution grids is associated with benefits such as delayed infrastructural investments, decreased losses and a reduced risk of power deficits. One instrument aimed at reducing peak demand is the demand charge, a capacity-based component in a network tariff that intends to encourage users to reduce their peak usage. Using ten years of data from a Swedish distribution network, this study demonstrates that demand charges may be unsuitable for managing the problems they are intended to address. Two critical misalignments in the design of these demand charges are identified: 1) Demand charges are most commonly based on maximum billing demand - a given user's maximum monthly peak - whereas the problem of peak demand overwhelmingly concerns maximum system peaks in the distribution grid as a whole. The lack of coincidence between these peaks suggest that demand charges are, by design, ineffective for reducing peak demand. 2) The peaks which determine a distribution system's maximum capacity requirements are rare, seasonal and largely temperature-driven events, whereas demand charges mainly target users' habitual daily patterns, encouraging daily shifts from peak to off-peak hours. As long as the main driver of network costs, maximum system peaks, are absent in their design, demand charges will neither reflect the costs that users impose on the grid nor provide them with the correct price signals on how to best act.