Essays on Demand Response in Electricity Supply Chains

DURAN, ASLIGUL SERASU

doi:https://doi.org/10.21985/n2-xx68-g152

Work

Essays on Demand Response in Electricity Supply Chains

Public

Download PDF

Unlike other products or services, electricity cannot be easily stored, and its delivery cannot be delayed by keeping customers in queues. The electricity supply has to be instantly available to serve a continuously varying demand. Traditionally, electric utilities would match their supply to the demand by adjusting the production output of their power plants. However, the traditional model of vertically integrated electric utilities are rapidly being deregulated and replaced by competitive wholesale and retail electricity markets. The competitive pressures, the modernization of electricity grids and technological innovations have created new solutions to the demand and supply matching problem, such as demand response (DR) which represents the explicit action taken by the consumers to change the time and the amount of their electricity use from the grid, either by altering their normal consumption patterns or running on-site generators. In this collection of academic papers, we study some of the operational questions surrounding the integration and implementation of demand response programs in retail and wholesale electricity markets. Chapter 1 and Chapter 2 cover completed research, while Chapter 3 covers a working project. In Chapter 1, we focus on the downstream supply chain, i.e. the retail electricity market, and investigate a price-based incentive to engage residential customers in demand response, namely time-based pricing. Electricity retailers around the world have been introducing such time-based pricing programs in the recent years. We empirically evaluate the short-term effects of time-based tariffs on the electricity demand, consumer welfare, retailers and the environment. To do that, we build a structural estimation model of household electricity demand and analyze a data set from an Irish field experiment, consisting of the half-hourly electricity consumption of over three thousand households, combined with the wholesale price, system load and generation data. Using the estimates from the structural model, we conduct a counterfactual study to explore various questions of practical importance. Our empirical analysis reveals that focusing on the peak-load reduction metric, one can design a flexible time-of-use (TOU) tariff that is simple and predictable yet performs as well as real-time pricing (RTP) given a fixed time horizon for evaluation. The annual electricity bills of consumers decrease only slightly when they switch from the flat rate to time-based tariffs, but there can be significant volatility in month to month bills under time-based tariffs. In contrast, the more flexible a tariff in terms of pricing, the less volatility it creates in retailer's profits throughout the year. Finally, switching from the flat rate to time-based tariffs would not change CO2 emissions from electricity generation in Ireland significantly. Our findings have several managerial implications. We find that time-based tariffs are effective in peak load reduction. However, the most appropriate time-based tariff depends on the context. If the goal is mitigating demand spikes over very short time spans, e.g. hours, then RTP is the most effective one. If the performance-relevant time horizon is longer, e.g. a month or a season, then a carefully designed TOU tariff with predetermined rates can be just as effective as RTP. Consumers and retailers are largely unaffected by time-based tariffs which suggests that their adoption may be harder under opt-in policies, compared to opt-out policies. From an environmental perspective, our result that CO2 emissions do not increase facilitates the adoption of time-based tariffs. In Chapter 2, we turn our attention to the upstream supply chain, and model the decision making of a DR provider who competes alongside generators in an electricity market that includes a short-term spot market of energy combined with a long-term forward market for capacity. We consider infinitesimally small, identical generators that have constant marginal costs of capacity investment and production. On contrary, the DR provider is modeled as a larger player that has two types of resources; behind-the-meter generation and load-reduction. We compare three scenarios of DR presence in the market: (i). A baseline case where only generators compete; (ii). DR enters all capacity as an emergency resource; (iii). Load-reduction is an emergency resource and behind-the-meter generation enters the spot market. Contrary to conventional wisdom that suggest the entry of DR will reduce wholesale market prices, we find that the average expected wholesale spot market price and the total per unit energy cost of consumers increase after the entry of DR. We show that free capacity market revenues in combination with inadequate non-compliance penalties during emergency conditions can incentivize DR providers to overcommit in the capacity market creating supply shortages and endangering the system reliability. Finally, we show that there exists a penalty mechanism that aligns the incentives of the DR providers and the policymakers such that the behind-the-meter generation is offered as front-of-the meter. Finally, in Chapter 3, we investigate the true short-term marginal costs of DR providers that participate in wholesale electricity spot markets. An increasing number of DR providers are participating in wholesale energy markets, and the actual costs of these providers are not well-known. This creates a challenge in determining effective compensation policies for such resources, and makes it harder to evaluate the true benefits of DR to electricity customers on the retail-side of the market. We propose a methodology to recover the parameters of the true cost functions of the DR providers by comparing the optimal offer decisions to actual price and quantity offers realized in the PJM market under different compensation schemes. We first model an oligopoly and characterize the optimal behavior of multiple profit-maximizing DR providers having asymmetric capacities, taking the generator decisions as exogenous. As a next step in our analysis, we plan to use the actual offer decisions of DR providers under varying compensation schemes in the spot market to recover the parameters of the cost functions of the DR providers that made these offers.

Creator