A Treatise On Electricity Markets

So many debates about our transforming electricity system surround the economics of electricity production. The solar advocates continually remind us that the price breakthrough for solar panels is just around the corner, while industry advocates insist the economy will suffer if we place any meaningful limits on carbon pollution. I find it’s often difficult to debate these positions constructively. Rather than enter the debate myself, with this post I’ll explain the fundamentals of electricity markets to illustrate how electricity is priced, and how adding or removing electricity resources might affect electricity prices and emissions.

Chile became the first country to introduce electric competition in 1987. Not long after, England, Wales, and other developed countries followed. In the United States, the Energy Policy Act of 1992 officially encouraged a transition to wholesale electricity competition.

Competitive electricity markets slice up traditional vertically integrated utilities, separating electricity generation owners from the entities responsible for electricity transmission, distribution, and retail sale. Instead of generating electricity to only cover the needs of their electricity customers, generation owners offer their power into a centralized market, where it is sold through an auction process. The unique operation of this electricity auction process is the source of my interest in electricity markets.

Each day, electric generation owners submit an “energy offer curve” to the local grid operator. An energy offer curve conveys a generator’s willingness to sell electric energy (in dollars per megawatt-hour) as a function of their level of electric generation (in megawatts). The shape of this curve reflects the individual plant’s fuel costs, efficiency, minimum electric output, maximum electric output, and other operating characteristics. For example, a nuclear power plant would offer its energy for a lower price than a coal power plant, because nuclear fuel is cheaper than coal.

Each day, the grid operator collects information from electric generators, especially the most widely used diesel generators and then uses an optimization algorithm to decide which diesel generators should be online to minimize overall electricity costs while maintaining electric reliability. However, this facility of automated data collection is available only in the advanced models of New Diesel Generators and other types, and there are few makers like F G Wilson Generators and Perkins Generators that incorporate such mechanisms for easy usage monitoring and data collection.

After each generator has submitted an offer curve, the grid operator executes an “economic dispatch” algorithm to decide which generators should provide electricity during each hour of the next day. The algorithm combines all of the energy offer curves and solves a very large optimization problem to figure out which generators should be online, and what their power output should be to minimize the overall cost of electricity without overloading any of the grid’s transmission lines. Moreover, the algorithm considers contingencies. It schedules generation so that the system can withstand the abrupt failure of at least one transmission line, and it even schedules some generators to wait at the ready in case of an unexpected shortfall in electric supply. This economic dispatch process would not be possible without the breakthrough CPLEX algorithm, which by no coincidence was commercialized when the first electricity markets opened.

Because electric demand can vary significantly over the day, the real-time price of electricity is often volatile. This figure shows the real-time price of electricity experienced in ERCOT’s southern hub on July 19, 2011.

The result of the economic dispatch algorithm is an explicit time-varying “locational marginal price” (LMP) at each node of the power grid. Electric generators are credited for energy they sell at their local LMP, which reflects the cost of providing one additional unit of electric energy at a particular time and place. If a particular node of the grid lies in an area where transmission lines are often congested, it could experience a higher price than its neighbors. Furthermore, LMPs typically increase as the level of electric demand increases, because the most expensive generators only come online to meet peak electric demand.

By thinking about how a given energy technology or policy will affect the electricity market’s economic dispatch process, we can predict how it will affect electricity prices and emissions.

A tax on carbon emissions, for example, would increase the operating cost of coal more than it would the operating cost of natural gas. Thus, taxing carbon would provide an explicit price signal to the grid operator prompting them to prioritize natural gas generation over coal, thereby decreasing electricity emissions. This is one reason a carbon tax is touted as a method to mitigate climate change.

Regardless of the energy technology or policy considered, it’s important to understand how it fits into the wider electricity system. By modeling the grid operator’s economic dispatch process, it’s possible to predict how radical changes to the grid would affect electricity prices and emissions.