Price and capacity competition in balancing markets with energy storage

Abstract

Energy storage can absorb variability from the rising number of wind and solar power producers. Storage is different from the conventional generators that have traditionally balanced supply and demand on fast time scales due to its hard energy capacity constraints, dynamic coupling, and low marginal costs. These differences are leading system operators to propose new mechanisms for enabling storage to participate in reserve and real-time energy markets. The persistence of market power and gaming in electricity markets suggests that these changes will expose new vulnerabilities.We develop a new model of strategic behavior among storages in energy balancing markets. Our model is a two-stage game that generalizes a classic model of capacity followed by Bertrand-Edgeworth price competition by explicitly modeling storage dynamics and uncertainty in the pricing stage. By applying the model to balancing markets with storage, we are able to compare capacity and energy-based pricing schemes and to analyze the dynamic effects of the market horizon and energy losses due to leakage. Our first key finding is that capacity pricing leads to higher prices and higher capacity commitments, and that energy pricing leads to lower, randomized prices and lower capacity commitments. Second, we find that a longer market horizon and higher physical efficiencies lead to lower prices by inducing the storage to compete to have their states of charge cycled more frequently.

Planning a decentralised and bi‐directional market‐based management system

“…The integrating of multi‐energy, decentralised control and the DRP scheduling of the centralised grid has been extended to aspects of load forecasting [13], spot market balance [9, 10, 21] and power and load management [3, 5], with new challenges and resolutions. The DBMBMS is a model which provides a set of management information to implement agent revenues such as a load management service and optimal time and cost services.…”

“…The results confirmed that the ESS capacities and locations could actively control the market prices and lead to arbitrage advantages. Moreover, the consolidated operation and planning of the wind energy and ESSs in the electricity market were investigated in [13]. Depending on the technological properties of the ESSs and forecasting the energy costs used in this model, the system aims to define the hourly form of the power control of this mixed system.…”

Instrumentation Strategies and Instrument Mixes for Long Term Climate Policy

“…But is it indeed impossible to have markets cleared and funded based on usual market mechanisms? There is not much research on this subject, but see for some key aspects (Taylor, Mathieu et al 2015), who however start from current US circumstances; (Koliou, Eid et al 2014) who analyze current restrictions on effective trading; (Verzijlbergh, De Vries et al 2014), who analyze the need for grid congestion pricing; (Verzijlbergh, Grond et al 2012) analyze the importance of linking electrical vehicles to the grid; and (Brancucci Martinez-Anido, De Vries et al 2012), who analyze the importance of enlarging EU internal international markets with increasing renewables penetration. Lifting the assumption that electricity only markets cannot function adequately then requires a view how well-functioning markets might be developed, without the brown-outs and unacceptably extreme market prices as envisaged by (Baron 2015).…”

“…As this opens the route to price manipulation maybe, this is an issue to investigate. If futures trading would be part of the market as created, then also there the question is how to regulate that part of the market to avoid fraudulent manipulation, see (Taylor, Mathieu et al 2015). The issue of capacity payment or market based investment funding is left open here.…”

A novel method for landslide displacement prediction by integrating advanced computational intelligence algorithms

“…The conventional ANN models can process nonlinear problems, but the selection of network structure and parameters is mainly dependent on experience. In addition, some other drawbacks of the traditional ANNs models are associated with excessive tunable parameters, slow learning rate, high possibility of entrapment in local minima, long computational time, and over-tuning 35 . Extreme learning machine is an algorithm based on single-hidden layer feed-forward neural network, which has been reported with good prediction capacity 36–40 .…”