How does one predict the rises and falls of unpredictable energy consumption?
Electricity domestic consumption in 2013.
Statisticians dig through data and make predictions based on the past, but engineers take a different approach. By combining electrical and computer engineering, it is possible to develop predictive algorithms based on computing and economics.
Associate Professor Fangxing (Fran) Li of The University of Tennessee at Knoxville spoke at the Special Seminar at the Thayer School of Engineering on April 13th. He discussed his research in developing these algorithms, an essential component in formulating “smart grids” for the present.
Traditionally, the power industry is a vertical monopoly. Companies are financially independent of each other, undergoing scheduled exchanges of generated energy. However, current policies are deregulating the power industry into a wholesale energy market. What were once separate flow lines of energy and profit has turned into a web in which many transmissions occur, with competition present at both the generation and distribution sectors.
As the energy sector undergoes deregulation and separate bodies become connected, components of the energy sector will be more inter-dependent than ever before. Variables such as location of energy, in both production and distribution, demand and incremental cost have more impact on consumers and the energy market than ever before.
Li’s research rests primarily on critical load levels. Critical load levels (CLLs), made possible by these grid connections, are places where locational marginal pricing (LMP), or the incremental cost to produce an extra unit of product, dramatically changes as load varies. These locations, once unpredictable, can now be forecasted.
Forecasting mechanisms are created with complex mathematics; using these functions, the amount by which the LMP changes as well as a quantified risk associated with the forecasted LMP can be predicted.
Li’s research has reaches into the home. With the power of forecasting, consumers are able to play an increasingly active role in regulating cost and consumption with a smart home energy management system (SHEMS).
In his laboratory, with the help of student researchers, he has created a platform to analyze how SHEMS can interact with the energy grid. This platform has generated the basis for an automatic appliance control design, as well as the integration of motion sensor and user history into how a home spends energy, thereby reducing electricity costs and decreasing nonrenewable energy depletion.
Li has also applied his mathematical algorithms in the realms of renewable energy, primarily wind energy. Wind energy is often regarded as the most unpredictable, but his current work aims to understand the impact of high-penetration wind power on aspects of the power system, combining step-change characteristics of LMP vs. Load and wind power forecasting and wind speed forecasting, maximizing revenue and minimizing waste.
Li’s research is a key component of modern energy management. As world population continues to grow, well-informed spending and consuming will become increasingly important. Such smart energy use has immediate and potentially enormous impact in balancing energy need and energy production.
References:
1 F. Li. “Concept and Applications of Critical Load Level (CLL) in Locational Marginal Pricing (LMP) Based Economic Studies.” Thayer School of Engineering, Dartmouth College. Hanover, 13 April 2015.