The Flexibility of Thermostatically Controlled Loads as a Function of Price Notice Time

Abstract

Due to increased use of variable renewable energy sources, more capacity for balancing and ancillary services (AS) is required. Non-generating resources such as thermostatically controlled loads (TCLs) can arbitrage energy prices and provide AS due to their thermal energy storage capacity. This paper explores the impact of energy/AS price notice time, i.e. the time between when the price is announced and when it takes effect, on the TCL energy consumption and AS capacity bids, and quantifies trade-offs between notice time and flexibility. We first optimize the energy consumption and AS capacity offers at a given notice time, varied from 24 hours ahead to real-time. We then introduce uncertainty in TCL availability, formulate the stochastic optimization problem, and evaluate how the trade-offs change. We find that price notice time impacts TCL profits, but does not significantly affect the total AS capacity offered over the day. However, AS capacity offers are impacted by uncertainty, which is likely to increase with notice time.

Impact of Market Timing on the Profit of a Risk-Averse Load Aggregator

“…where individual chance constraints are used to manage uncertainty in random TCL parameters Y t,ξ = {B t,ξ , P t,ξ , P t,ξ , S t,ξ , S t,ξ }, which are used to compute the random bounds Q t,ξ , Q t,ξ , L t,ξ , L t,ξ , and [8]. Specifically, (10d)-(10g) are equivalent to (30)-(33) of [8], with the exception that (10d) and (10e) capture the SOC deviation for the worst-case reserve activation more accurately by using 1 2 (s t + s t−1 ) rather than s t . The random bounds Q t,ξ , Q t,ξ , L t,ξ , L t,ξ include bound and baseline forecast error but use a method to mitigate the propagation of baseline forecast error, as detailed in [8].…”

“…T HE increasing use of variable renewable energy sources requires more flexible resources that can respond in realtime to supply/demand imbalance. To this end, thermostatically controlled loads (TCLs) have been proposed for energy arbitrage [1], [2] and ancillary services [3]- [8]. By exploiting the inherent thermal inertia of TCLs, their electricity consumption can be varied while still meeting the desired service quality, i.e., temperature range set by the consumer.…”

“…Third, we discuss the policy implications of our results. This paper builds upon our preliminary work [8]. However, here, we model and investigate the impact of reserve contract period, include price uncertainty in the form of price scenarios, and model financial risk-aversion using the conditional value at risk (CVaR).…”

Quantifying flexibility of load aggregations: impact of communication constraints on reserve capacity

“…As ambient temperature decreases, the baseline power of heating TCLs increases. Thus, the symmetric reserve capacity P symm also increases until the baseline power exceeds half of the installed capacity, and then drops again, as expressed in equation (8).…”

“…An increasing use of variable renewable energy sources requires more flexible resources that can provide reserve capacity and respond in real-time to system imbalance. To this end, thermostatically controlled loads (TCLs) have been proposed for energy arbitrage [1,2] and reserve provision [3][4][5][6][7][8][9]. TCLs can provide reserves due to their thermal energy storage capacity constituted by their thermal capacity and indoor temperature dead bands.…”