From power engineering literature, four main application areas for agents exist. They are, broadly, monitoring and diagnostics, distributed control, protection systems, and modelling and simulation.

The first three of these areas utilise the facets of agents that enhance systems integration, allowing disparate autonomous components to interoperate and share information where appropriate. Agents are used to encapsulate particular tasks or blocks of functionality, which can operate independently if required, but together produce a system that aims to offer more than the sum of its parts. The key agent properties for this are social ability and autonomy, which can give robustness and graceful degradation to partial system failure, and openness and extensibility for the future.

The properties of agents aid modelling and simulation in a significantly different way. While systems integration is important for software deployed on the network, simulation is generally less likely to require openness to other software developers’ agents. Instead, agents give the ability to replicate functionality in many light-weight autonomous actors, which can be used to study potential emergent behaviours of the community. This reflects an agent-oriented modeling of the world, in comparison to agent-oriented software design.

The following sections highlight key literature in each of the four application areas.

Diagnostics

Monitoring and diagnostics is well-suited to an agent approach, as it allows the development of multiple autonomous modules of data processing to refine and improve a diagnosis. Two key applications are condition monitoring and post-fault analysis.

Considering the condition monitoring of plant, data must be collected from sensors, analysed for information, and conclusions presented to engineers. Across complex items of plant or within a substation, this may involve many sensors of disparate types, a range of data analysis techniques, and the need to combine evidence to give a coherent plant health assessment from the results of each technique. Each individual stage of processing is relatively simple and constrained, but the overall system needs to flexibly incorporate the stages most appropriate to a given monitoring situation. Agent-oriented design of this system provides a way of managing the complexity and staged-deployment of this functionality.

For post-fault analysis of faults on the power system, there are again disparate sources of data including SCADA and fault records, which can be analysed individually. However, knowledge of the network structure can more efficiently direct the interpretation of these data sources; for example, by using SCADA analysis conclusions to identify the most pertinent fault records to examine. In a similar way to condition monitoring, an agent-oriented system design allows individual agents to encapsulate the high-level task they accomplish, making a complex system conceptually more simple and extensible.

The area of monitoring and diagnostics is one of the least widely researched areas, with all IEEE journal papers arising from activity at the University of Strathclyde. They are:

  1. V. M Catterson, S. D. J. McArthur, M. D. Judd, and A. S. Zaher, “Managing Remote Online Partial Discharge Data”, IEEE Trans Power Del, 23(4), October 2008, pp 1754-1762.
  2. E. M. Davidson, S. D. J. McArthur, J. R. McDonald, T. Cumming, and I. Watt, “Applying Multi-Agent System Technology in Practice: Automated Management and Analysis of SCADA and Digital Fault Recorder Data”, IEEE Trans Power Sys, 21(2), May 2006, pp 559-567.
  3. S. D. J. McArthur, S. M. Strachan, and G. Jahn, “The Design of a Multi-Agent Transformer Condition Monitoring System”, IEEE Trans Power Sys, 19(4), November 2004, pp 1845-1852.
  4. S. D. J. McArthur, C. D. Booth, J. R. McDonald, and I. T. McFadyen, “An agent-based anomaly detection architecture for condition monitoring”, IEEE Trans Power Systems, 20(4), November 2005, pp 1675-1682.
  5. J. A. Hossack, J. Menal, S. D. J. McArthur, and J. R. McDonald, “A multiagent architecture for protection engineering diagnostic assistance”, IEEE Trans Power Sys, 18(2), May 2003, pp 639-647.
  6. E. E. Mangina, S. D. J. McArthur, J. R. McDonald, and A. Moyes, “A multi agent system for monitoring industrial gas turbine start-up sequences”, IEEE Trans Power Sys, 16(3), August 2001, pp 396-401.
  7. E. E. Mangina, S. D. J. McArthur, and J. R. McDonald, “Reasoning with modal logic for power plant condition monitoring”, IEEE Power Eng Review, 21(7), July 2001, pp 58-59.

Distributed Control

With the growing size and complexity of power system networks centralized control schemes may no longer be sufficient. By employing agents with local decision making capability the possibility arises to distribute out, where possible, system control. Actions such as network restoration, reconfiguration, the dispatch of generation and the management of loads could be managed locally. Distributing out control removes the reliance on a single centralized processing unit, and would remove latency in the case of a dynamic system.

Particular applications which are under investigation include: power system restoration, active distribution networks, microgrid control and the restoration/reconfiguration of shipboard electrical systems. A selection of papers in this area are:

  1. M.M. Nordman and M. Lehtonen, “An agent concept for managing electrical distribution networks,” IEEE Trans. Power Del., vol. 20, no. 2, pp. 696-703, Apr 2005.
  2. H. Ni, G.T. Heydt, and L. Mili, “Power system stambility agents using robust wide area control,” IEEE Trans. Power Syst., vol. 17, no. 4, pp.1123-1131, Nov. 2002.
  3. M. Amin, “Toward self-healing energy infrastructure systems,” IEEE Comput. Appl. Power, vol. 14, no.1, pp. 20-28, Jan. 2001.
  4. K. Huang, S.K. Srivastava, and D.A. Cartes, “Solving the information accumulation problem in mesh structured agent system,” IEEE Trans. Power Syst., vol. 22, no. 1, pp. 493-495, Feb. 2007.

Protection

By mapping agents to protective devices it is envisaged that novel protection schemes can be developed which are fault tolerant, self-coordinating and adaptive to local requirements. This is something like a hybrid use of agent benefits: modelling devices on the network as separate agents, but instead of using the model to simulate events, deploying the agents as integrated software components. The benefits of this approach have not been fully articulated in the literature to date, but papers which advocate this approach include:

  1. S.-J. Park and J.-T. Lim, “Modelling and control of agent-based power protection system using supervisors,” Proc. Inst. Elect. Eng., Control Theory Appl., vol.153, no. 1, pp.92-98, Jan. 2006.
  2. S. Sheng, K.K. Li, W.L. Chan, Z. Xiangjun, and D. Xianzhong, “Agent-based self-healing protection system,” IEEE Trans, Power Del., vol 21, no. 2, pp. 610-618, Apr. 2006.
  3. C. Fukui, H. Kudo, J. Koda, K. Yabe, and Y. Tomita, “A cooperative protection system with an agent model,” IEEE Trans. Power Del., vol.13, no. 4, pp.1060-1066, Oct. 1998.

Modelling and Simulation

The driver for the use of agents for modelling and simulation is the complexity of many power system-related operations. While designing and building the overall system from the top down may present a highly complex challenge, the behaviour of individual actors within the system may be much simpler to model and build. In such situations, a bottom up approach makes building the system far easier.

This gives the added advantage of allowing the study of emergent properties of the system. If all the actors behave in a particular way or with specific goals, it could lead to pathologic behaviour. Alternatively, a particular mixture of actor types may produce particularly desirable outcomes.

A comprehensive web resource including analysis of the benefits of agents for electricity market modelling and active researchers in the field is maintained by Leigh Tesfatsion.

Particular applications within this field are energy market modelling, transmission planning, scenario analysis, and the integration of models within software packages. Key papers include:

  1. H. Li and L. Tesfatsion, “ISO Net Surplus Collection and Allocation in Wholesale Power Markets Under Locational Marginal Pricing”, IEEE Trans Power Sys, to appear, 2010.
  2. S. E. Widergren, J. M. Roop, R. T. Guttromson, and Z. Huang, “Simulating the dynamic coupling of market and physical system operations”, IEEE PES General Meeting 2004.
  3. D. Koesrindartoto, S. Junjie, and L. Tesfatsion, “An agent-based computational laboratory for testing the economic reliability of wholesale power market designs”, IEEE PES General Meeting 2005.
  4. J. Contreras and F. F. Wu, “Coalition formation in transmission expansion planning”, IEEE Trans Power Sys, 14(3), August 1999, pp 1144-1152.
  5. K. Hopkinson, X. Wang, R. Giovanini, J. Thorp, K. Birman, and D. Coury, “EPOCHS: A Platform for Agent-Based Electric Power and Communication Simulation Built From Commercial Off-the-Shelf Components”, IEEE Trans Power Sys, 21(2), May 2006, pp 548-558.
  6. S. D. J. McArthur, E. M. Davidson, G. J. W. Dudgeon, and J. R. McDonald, “Toward a model integration methodology for advanced applications in power engineering”, IEEE Trans Power Sys, 18(3), August 2003, pp 1205-1206.

Citation

If you want to reference material on this topic, please consider citing these papers as appropriate:

  • S. D. J. McArthur; E. M. Davidson; V. M. Catterson; A. L. Dimeas; N. D. Hatziargyriou; F. Ponci; T. Funabashi, “Multi-Agent Systems for Power Engineering Applications—Part I: Concepts, Approaches, and Technical Challenges”, IEEE Transactions on Power Systems, Vol. 22, No. 4, November 2007.
  • S. D. J. McArthur; E. M. Davidson; V. M. Catterson; A. L. Dimeas; N. D. Hatziargyriou; F. Ponci; T. Funabashi, “Multi-Agent Systems for Power Engineering Applications—Part II: Technologies, Standards, and Tools for Building Multi-agent Systems”, IEEE Transactions on Power Systems, Vol. 22, No. 4, November 2007.