From GridLAB-D Wiki
Jump to: navigation, search

Basic Project Information

Project Title:

A Test Bed for the Integrated Experimental Study of
Retail and Wholesale Power Market Designs

Principal Investigator:

Leigh Tesfatsion
Prof. of Econ, Math, & Electrical and Computer Engineering
Iowa State University, Ames, IA 50011-1070
tesfatsi@iastate.edu
http://www.econ.iastate.edu/tesfatsi/

Research Assistant:

Zhao Huan
Economics PhD Candidate
Iowa State University, Ames, IA 50011-1070
huanxjtu@gmail.com

Project Description:

Retail and wholesale power markets are intrinsically interdependent systems. The main goal of this project is to develop an agent-based computational test bed for the study of retail and wholesale power markets as integrated systems operating over realistically rendered distribution and transmission grids.

The specific context for the wholesale power market portion of the test bed is the design recommended in a 2003 White Paper [1] by the U.S. Federal Energy Regulatory Commission (FERC) for common adoption by all U.S. wholesale power markets, referred to below as the Wholesale Power Market Platform (WPMP). Versions of the WPMP design have been implemented (or adopted for implementation) in U.S. energy regions encompassing over 50% of U.S. generating capacity. These energy regions include the Midwest (MISO), New England (ISO-NE), New York (NYISO), the mid-Atlantic states (PJM), California (CAISO), the southwest (SPP), and Texas (ERCOT).

The key design element of the WPMP is a two-settlement system to be managed by an independent system operator (ISO). Roughly, a two-settlement system refers to the combined workings of a day-ahead energy market and a real-time energy market that are separately settled each day by means of Locational Marginal Pricing (LMP). Under LMP, a separate price for power is determined at each point of the transmission grid at which power is injected or withdrawn.

As envisioned in the WPMP, and implemented in practice, the wholesale day-ahead market is structured as a double auction. Load-serving entities (buyers) are permitted to submit hourly demand bids consisting of both fixed and price-sensitive hourly demands, and generation companies (sellers) are permitted to submit hourly supply offers consisting of price-sensitive hourly supplies. In actuality, however, the day-ahead market effectively functions as a single-sided seller auction because the bulk of the demand takes the form of fixed hourly loads (i.e., load profiles) implying essentially vertical demand curves.

A key difficulty is that downstream (retail) markets in the U.S. are still largely regulated with cost-based pricing, so that wholesale load-serving entities in fact have little incentive to submit price-sensitive demand bids. Even in states that have nominally introduced retail competition, the use of extended default service contracts and long-term wholesale procurement contracts reduces market entry and contributes to the persistence of vertical demand curves in wholesale day-ahead markets. As experimentally demonstrated in [2], under this scenario the wholesale generation companies are easily able to learn to implicitly collude on reported supply offers involving higher-than-true marginal costs. The result is much higher market operating costs.

These adverse market performance characteristics suggest the need for an integrated restructuring of both retail and wholesale power markets. Rather than use actual systems as test beds, however, this project proposes to develop an agent-based test bed that seams together two previously developed agent-based test beds:

  • AMES [3], a software platform developed by a team of researchers at Iowa State University for the study of restructured wholesale power markets;
  • GridLAB-D [4], a software platform developed by a team of researchers at the Pacific Northwest National Laboratory (PNNL) for the study of power distribution systems for end-use customers.

The resulting integrated agent-based test bed will permit us to pre-test, through intensive systematic experimentation, how an integrated restructuring of retail and wholesale power markets might best be implemented. Of particular interest will be the systematic experimental exploration of recent "smart grid" proposals for improving the efficiency and reliability of overall power system operations. Planned research topics include:

  1. Potential impacts of increased price-responsiveness of demand (e.g., through advanced metering infrastructure or demand response programs) on the efficiency and reliability of market operations.
  2. Appropriate design of pricing contracts and risk management tools under alternative mixtures of real-time-pricing and flat-rate-pricing.
  3. Potential strategic behaviors of integrated retail-wholesale power market participants, including possible incentives for implicit horizontal and vertical coordination.
  4. Desirable types, sizes, and sitings of distributed energy resources, including both distributed generation and distributed storage.
  5. Impact of integrated retail-wholesale power market restructuring on environmental quality.
  6. Appropriate design of integrated retail-wholesale power market rules and regulations to mitigate environment damage from power plant emissions.
  7. Impact of integrated retail-wholesale power market restructuring on long-run investment in generation and transmission and on the provision of ancillary services.
  8. Efficiency implications of restructurings of the distribution and/or transmission networks to incorporate microgrid and other possible smart grid architectures.

References:

[1] FERC, Notice of White Paper, U.S. Federal Energy Regulatory Commission, April 2003.

[2] Hongyan Li, Junjie Sun, and Leigh Tesfatsion, "Separation and Volatility of Locational Marginal Prices in Restructured Wholesale Power Markets," ISU Economics Working Paper #09009, June 2009: [http://www.econ.iastate.edu/research/publications/viewabstract.asp?pid=13075 ]

[3] The AMES Wholesale Power Market Test Bed homepage: [http://www.econ.iastate.edu/tesfatsi/AMESMarketHome.htm ]

[4] GridLAB-D, PNNL, power distribution simulation software. [http://gridwise.pnl.gov/technologies/GridLABd.stm ]

See also

Market_User_Guide

Market_module

Market Specifications

Controller Specifications

Wholesale_Markets