AESC’s Energy Intelligence Group (EIG) has been developing and applying intelligent agent technology since 1997. During that time AESC has either collaborated on, or served as the prime contractor on seven (7) projects exploring application of intelligent agent technology in the energy industry.


In early projects AESC worked closely with Reticular Systems Incorporated, a developer of intelligent software agent systems, to identify and develop utility industry applications for intelligent software agents. In the process, AESC staff members applied their considerable knowledge of the electric utility industry as well as their background in advanced computing technologies to assist Reticular in these efforts. AESC and Reticular collaborated on research projects for the Department of Energy and the Electric Power Research Institute that laid the groundwork for later efforts.


In anticipation of the dynamic electric markets that was to be the key feature of the newly deregulated California electric marketplace; the California Energy Commission (CEC) awarded AESC a contract in 1998 to explore using intelligent agent technology to control distributed generation (DG) assets. The resulting project, CEC-PIER 500-098-040, "Intelligent Software Agents for Control and Scheduling of Distributed Generation" was successful in demonstrating that agent technology could be used to schedule DG assets in response to market price signals. Demonstration software, developed as part of the project, showed how an agency of Smart*DER agents could collaborate to routinely collect the necessary inputs (i.e., pricing, weather data, etc.) and generate operating schedules that maximized the benefit of DG operation.

The deregulated California electric marketplace was subsequently replaced by a newly regulated market with a strong emphasis on customer demand response and on efficient energy use. Understanding that Smart*DER technology could be useful in this new marketplace, the CEC awarded AESC a follow-on contract, CEC-PIER 500-00-016 "Demonstration of Intelligent Software Agents for Control & Scheduling of Distributed Energy Resources", in 2000 to demonstrate Smart*DER technology in a "real-world" environment. During this project AESC updated and refined the Smart*DER agents in response to the evolving California marketplace and installed a Smart*DER system at the Los Angeles County Sanitation District's Calabasas landfill. The Smart*DER based control operated successfully to schedule DER asset operation in response to a Critical Peak Pricing (CPP) further demonstrating the utility of Smart*DER technology. As a result of this successful effort, AESC is moving forward with commercialization Smart*DER technology.

On a national scale, the evolution of the traditional centralized grid to a Smart Grid that embraces advanced information technologies such as agent technology continues to gain momentum. In a recently completed CEC PIER 500-07-020 "Intelligent Software Agents for Integration of Renewables and Storage" project AESC demonstrated that intelligent agent technology could successfully coordinate the energy production and delivery from wind generation resources in the Tehachapi region of California. During the 2-month demonstration phase of the project a multi-agent system successfully collaborated to gather and process SCE SCADA data in real time using a state of the art Bayseian Belief Network (BBN). The agents then developed recommendations for capacitor bank operation, storage system operation and wind curtailment. Agent recommendations for storage system operation were implemented via agent control of a 100 kW Beacon Power flywheel storage unit that was installed and operated continuously during the demonstration. The agent-based system successfully operated the storage unit in support of frequency regulation during normal operation with "hybrid" operation of the storage unit for short periods to provide voltage/VAR support or line loading support as dictated by local conditions. The project successfully demonstrated that local, dynamic control control using agent technology could utilize storage more fully to address both global and local needs. Powerful BBN capabilities related to operation with missing or unknown data and the ability to train using either actual or simulated data were also demonstrated.


In addition to the CEC-PIER sponsored efforts, AESC also received a 2003 DOE SBIR grant to study application of agent technology to integration of distributed generation into the distribution system. During the project, SBIR DG02-03ER83604 "Distributed Intelligent Agents for Decision Making at Local DER Levels" AESC worked closely with the Center for Networked Distributed Energy at Colorado State University to identify a hierarchy of intelligent agents needed for electric grid system control. At the distribution system level this hierarchy of collaborating agents was termed a "Power Neighborhood". AESC developed demonstration software that showed how a hierarchy of agents, working via a web-based auction process could routinely schedule operation of their DER assets in response to market prices or, if needed quickly reconfigure operation of their DER assets in response to a signal indicating eminent loss of power to their distribution feeder. As a result of this successful effort, AESC is moving forward with commercialization of Power Neighborhood technology.