Subscribe: Contact: Sitemap: Printer Friendly
Subscribe to our RSS feed
Home  >  Research  >  Projects  >  AI for the Smart Grid

AI for the Smart Grid

Reasoning about Discrete and Hybrid Systems

The project develops technology for automating the management of complex systems. The focus is in systems which are discrete in terms of the control actions and control devices, but which are in reality physical and hence better described in terms of continuous variables. A good example is electricity networks, where on-off switches and other discrete control devices are in control of the continuous voltages and currents. Other examples of such systems are water, telecommunications and traffic networks, as well as many kinds of embedded systems.

The research in the project focuses on developing better search and reasoning methods for making the best possible control decisions.This involves automatically determining the current state of the system (state-estimation), predicting the future states of the system under different control actions, and choosing the best actions to take.

The project will demonstrate its research outcomes as one component of an advanced software system for managing electricity distribution or transmission networks.The project started in 2009 and will be completing a pilot phase in 2010.

Application: the Smart Grid

Electricity networks are becoming more complex and difficult to control as the degree of automation increases, co-generation is widely deployed, and electricity use is more actively adjusted to generation capability and costs.

The project develops technology for helping control intelligent electricity networks more effectively. This involves automating routine aspects of control, as well as interpreting and presenting information about the electricity system state to human users such as control room operators in a more effective manner.

Some of the focus areas of the project are:

  • alarm processing and situational awareness in the control room
  • fault diagnosis
  • supervisory control of distribution networks (including fault isolation and supply restoration in outage situations)

Who will benefit?

The end-users of the technology developed in the project are electricity transmission and distribution companies. The project will help in improving the management of electricity networks as the networks become more complex, as well as enable novel ways of controlling networks that are not possible with current technology.

Improved electricity networks will benefit all electricity users, including households, industry and the public sector, through more effective and efficient use of resources, including natural resources and investments in electricity infrastructure, as well as more reliable supply of power.

More generally, the technology can be adapted to other discrete/hybrid supervision problems, such as water networks, intelligent manufacturing systems.

Research Team

Former members: Prof. David Hill

Publications

These are publications from the project and the main publications from its predecessor projects (2006-2009).

Most of the research in the project and its predecessor has been on model-based methods for diagnosis and planning/control, especially based on SAT, constraint-satisfaction and constraint networks.

  1. Jussi Rintanen. Heuristics for Planning with SAT and Expressive Action Definitions.In Proceedings of the International Conference on Automated Planning and Scheduling, AAAI Press, 2011.
  2. J. Rintanen. Heuristics for Planning with SAT. In Dave Cohen, ed., Principles and Practice of Constraint Programming - CP 2010, 16th International Conference, CP 2010, St Andrews, Scotland, September 6-10, 2010, Proceedings. Lecture Notes in Computer Science, Springer-Verlag, 2010.
  3. J. Rintanen. Heuristic Planning with SAT: Beyond Strict Depth-First Search.Twenty-Third Australiasian Joint Conference on Artificial Intelligence,Adelaide, December 7-10, 2010, Proceedings.Lecture Notes in Computer Science, Springer-Verlag, 2010.
  4. J. Rintanen and A. Grastien, Diagnosability testing with satisfiability algorithms, in M. Veloso, ed., Proceedings of the 20th International Joint Conference on Artificial Intelligence, pages 532-537, AAAI Press, 2007.
  5. J. Rintanen, Diagnosers and diagnosability of succinct transition systems,  in M. Veloso, ed., Proceedings of the 20th International Joint Conference on Artificial Intelligence, pages 538-544, AAAI Press, 2007.
  6. A. Grastien, Anbulagan, J. Rintanen and E. Kelareva, Diagnosis of discrete-event systems using satisfiability algorithms, pages 305-310, Proceedings of the 22nd AAAI Conference on Artificial Intelligence (AAAI-07), AAAI Press, 2007. (© 2007 American Association for Artificial Intelligence. All rights reserved. AAAI)
  7. J. Rintanen, Complexity of concurrent temporal planning,Proceedings of the 17th International Conference on Automated Planning and Scheduling, pages 280-287, AAAI Press, 2007.(© 2007 American Association for Artificial Intelligence. All rights reserved. AAAI)
  8. J. Rintanen, Asymptotically optimal encodings of conformant planning in QBF,Proceedings of the 22nd AAAI Conference on Artificial Intelligence (AAAI-07), pages 1045-1050, AAAI Press, 2007.(© 2007 American Association for Artificial Intelligence. All rights reserved. AAAI)
  9. A. Schumann and J. Huang, A scalable jointree algorithm for diagnosability, Proceedings of the 23rd AAAI Conference on Artificial Intelligence (AAAI), pages 535-540, 2008.
  10. A. Ciré and A. Botea. 2008. Learning in Planning with Temporally Extended Goals and Uncontrollable Events. In Proceedings of the European Conference on Artificial Intelligence ECAI-08. Patras, Greece.
  11. P. Kan John and A. Grastien, Local consistency and junction tree for diagnosis of discrete-event systems, Eighteenth European Conference on Artificial Intelligence (ECAI-08), 2008.
  12. A. Grastien and Anbulagan, Incremental diagnosis of DES by satisfiability, Poster in Eighteenth European Conference on Artificial Intelligence (ECAI-08), 2008.
  13. J. Rintanen, A New Approach to Planning in Networks. 18th European Conference on Artificial Intelligence (ECAI-08), IOS Press, Patras (Greece), July 2008.
  14. J. Rintanen, Planning graphs and propositional clause-learning.In Gerhard Brewka and Patrick Doherty, editors, Principles ofKnowledge Representation and Reasoning: Proceedings of the EleventhInternational Conference (KR 2008), pages 535-543, AAAI Press, 2008.
  15. J. Rintanen, Regression for classical and nondeterministic planning.In Malik Ghallab, Constantine D. Spyropoulos, and Nikos Fakotakis,editors, ECAI 2008. Proceedings of the 18th European Conference onArtificial Intelligence. pages 568-571, IOS Press, 2008.
  16. A. Grastien and Anbulagan, Incremental diagnosis of DES with a non-exhaustive diagnosis engine, Twentieth International Workshop on Principles of Diagnosis (DX-09), 2009.
  17. A. Grastien, Symbolic testing of diagnosability, Twentieth International Workshop on Principles of Diagnosis (DX-09), 2009.
  18. Anbulagan and A. Grastien.  Importance of variables semantic in CNF Encoding of cardinality constraints, Eighth Symposium on Abstraction, Reformulation and Approximation (SARA-09), 2009.
  19. A. Botea and A. Ciré. 2009, Decentralized Planning with Temporally Extended Goals and Uncontrollable Events. In Proceedings of the International Joint Conference on Artificial Intelligence IJCAI-09, Pasadena, USA.
  20. L. Blackhall, P. Kan John, A. Grastien, and D. Hill, Diagnosability of hybrid dynamical networks using indicator functions, Twentieth International Workshop on Principles of Diagnosis (DX-09), 2009.
  21. L. Blackhall, P. Kan John, and A. Grastien, Diagnosing structural changes in hybrid dynamical systems, Seventh IFAC Symposium on Fault Detection, Supervision and Safety of Technical Processes (SafeProcess-09), 2009.
  22. L. Blackhall, P. Kan John, A. Grastien, and D. Hill, Diagnosability of networks of hybrid systems, Seventh IFAC Symposium on Fault Detection, Supervision and Safety of Technical Processes (SafeProcess-09), 2009.

Awards

The following papers from the project and its predecessor projects won awards.
  1. M. Wehrle and J. Rintanen,Planning as satisfiability with relaxed E-step plans,In M. Orgun and J. Thornton, eds, AI 2007 : Advances  in Artificial Intelligence: 20th Australian Joint Conference on Artificial  Intelligence, Surfers Paradise, Gold Coast, Australia, December 2-6, 2007,  Proceedings, Lecture Notes in Computer Science 4830, pages 244-253, Springer-Verlag, 2007. The winner of the AI 2007 Best Paper Award
  2. Malte Helmert, Patrik Haslum and Joerg Hoffmann, Flexible Abstraction Heuristics for Optimal Sequential Planning. In Proceedings of the 17th International Conference on Automated Planning and Scheduling, 2007. The winner of the ICAPS 2007 Best Paper Award
  3. Akihiro Kishimoto, Alex Fukunaga and Adi Botea. 2009. Scalable, Parallel Best-First Search for Optimal Sequential Planning. In Proceedings of the International Conference on Automated Planning and Scheduling ICAPS-09, Thessaloniki, Greece. One of two winners of the ICAPS 2009 Best Paper Award