Euan M. Davidson

Multi-Agent Systems for Power Engineering Applications—Part I: Concepts, Approaches, and Technical Challenges

This is the first part of a two-part paper that has arisen from the work of the IEEE Power Engineering Societys Multi-Agent Systems (MAS) Working Group. Part I of this paper examines the potential value of MAS technology to the power industry. In terms of contribution, it describes fundamental concepts and approaches within the field of multi-agent systems that are appropriate to power engineering applications. As well as presenting a comprehensive review of the meaningful power engineering applications for which MAS are being investigated, it also defines the technical issues which must be addressed in order to accelerate and facilitate the uptake of the technology within the power and energy sector. Part II of this paper explores the decisions inherent in engineering multi-agent systems for applications in the power and energy sector and offers guidance and recommendations on how MAS can be designed and implemented.

Multi-Agent Systems for Power Engineering Applications—Part II: Technologies, Standards, and Tools for Building Multi-agent Systems

This is the second part of a two-part paper that has arisen from the work of the IEEE Power Engineering Societys Multi-Agent Systems (MAS) Working Group. Part I of this paper examined the potential value of MAS technology to the power industry, described fundamental concepts and approaches within the field of multi-agent systems that are appropriate to power engineering applications, and presented a comprehensive review of the power engineering applications for which MAS are being investigated. It also defined the technical issues which must be addressed in order to accelerate and facilitate the uptake of the technology within the power and energy sector. Part II of this paper explores the decisions inherent in engineering multi-agent systems for applications in the power and energy sector and offers guidance and recommendations on how MAS can be designed and implemented. Given the significant and growing interest in this field, it is imperative that the power engineering community considers the standards, tools, supporting technologies, and design methodologies available to those wishing to implement a MAS solution for a power engineering problem. This paper describes the various options available and makes recommendations on best practice. It also describes the problem of interoperability between different multi-agent systems and proposes how this may be tackled.

Distribution power flow management utilising an online Optimal Power Flow technique

This paper describes the current connection regime for distributed generation (DG) in the U.K. and presents a novel application of the optimal power flow (OPF) technique for automatic power flow management (PFM) to manage thermal constraints in distribution networks. OPF formulations have been used, in an offline mode, as a power system planning tool for several years. The novel implementation of OPF for “corrective” PFM in an online operational mode, for MV distribution networks, is presented and tested in this paper. The authors demonstrate, through simulations conducted on a commercially available substation computer, that such an application of OPF can represent first on, last off generator connection agreements that reflect the current principles of access in the U.K. Two case study networks, a 33 kV and an 11 kV, provide the basis for assessment of the OPF-based PFM algorithm in terms of computation time to arrive at a solution in the event of a network thermal excursion and the level of DG curtailment necessary to meet network thermal limits. Assessments are made and fully discussed of the suitability for an OPF-based approach for distribution network management within an online network control scheme including discussion of the important consideration of control robustness.

Automating power system fault diagnosis through multi-agent system technology

Fault diagnosis within electrical power systems is a time consuming and complex task. SCADA systems, digital fault recorders, travelling wave fault locators and other monitoring devices are drawn upon to inform the engineers of incidents, problems and faults. Extensive research by the authors has led to the conclusion that there are two issues which must be overcome. Firstly, the data capture and analysis activity is unmanageable in terms of time. Secondly, the data volume leads to engineers being overloaded with data to interpret. This paper describes how multi-agent system technology, combined with intelligent systems, can be used to automate the fault diagnosis activity. Within the multi-agent system, knowledge-based and model-based reasoning are employed to automatically interpret SCADA system data and fault records. These techniques and the design of the multi-agent system architecture that integrates them are described. Consequently, the use of engineering assistant agents as a means of providing engineers with decision support, in terms of timely and summarised diagnostic information tailored to meet their personal requirements, is discussed.

Issues in integrating existing multi-agent systems for power engineering applications

Multi-agent systems (MAS) have proven to be an effective platform for diagnostic and condition monitoring applications in the power industry. For example, a multi-agent system architecture, entitled condition monitoring multi-agent system (COMMAS) (McArthur et al., 2004), has been applied to the ultra high frequency (UHF) monitoring of partial discharge activity inside transformers. Additionally, a multi-agent system, entitled protection engineering diagnostic agents (PEDA) (Hossack et al., 2003), has demonstrated the use of MAS technology for automated and enhanced post-fault analysis of power systems disturbances based on SCADA and digital fault recorder (DFR) data. In this paper, the authors propose the integration of COMMAS and PEDA as a means of offering enhanced decision support to engineers tasked with managing transformer assets. By providing automatically interpreted data related to condition monitoring and power system disturbances, the proposed integrated system offer engineers a more comprehensive picture of the health of a given transformer. Defects and deterioration in performance can be correlated with the operating conditions it experiences. The integration of COMMAS and PEDA has highlighted the issues inherent to the inter-operation of existing multi-agent systems and, in particular, the issues surrounding the use of differing ontologies. The authors believe that these issues need to be addressed if there is to be widespread deployment of MAS technology within the power industry. This paper presents research undertaken to integrate the two MAS and to deal with ontology issues

Exploiting Multi-agent System Technology within an Autonomous Regional Active Network Management System

This paper describes the proposed application of multi-agent system (MAS) technology within AuRA-NMS, an autonomous regional network management system currently being developed in the UK through a partnership between several UK universities, distribution network operators (DNO) and a major equipment manufacturer. The paper begins by describing the challenges facing utilities and why those challenges have led the utilities, a major manufacturer and the UK government to invest in the development of a flexible and extensible active network management system. The requirements the utilities have for a network automation system they wish to deploy on their distribution networks are discussed in detail. With those requirements in mind the rationale behind the use of multi-agent systems (MAS) within AuRA-NMS is presented and the inherent research and design challenges highlighted including: the issues associated with robustness of distributed MAS platforms; the arbitration of different control functions; and the relationship between the ontological requirements of Foundation for Intelligent Physical Agent (FIPA) compliant multi-agent systems, legacy protocols and standards such as IEC 61850 and the common information model (CIM).

The Use of Constraint Programming for the Autonomous Management of Power Flows

This paper describes a novel network agnostic approach to the real-time management of power flows within MV distribution networks with a number of embedded generators. The approach in question uses the artificial intelligence technique of constraint programming to autonomously make decisions about how to manage a number of generators as part of a larger autonomous regional active network management system called AuRA-NMS. The results of testing using commercially available IEC 61850 compliant substation computing hardware, connected to a real- time simulator, are presented for two different case study networks: an interconnected 33kV network and an 11kV radial network. These results demonstrate the feasibility of the approach both in terms of its applicability to different network topologies and its real-time performance. The paper concludes by discussing the methodology for future testing which aims to assess the robustness of the approach against a number of other proposed approaches and the plans for field trials of the current prototype.

A toolset for applying model-based reasoning techniques to diagnostics for power systems protection

This paper describes a software-based toolset which aids the application and development of model-based reasoning systems for validation and diagnosis in power systems applications. The validation of the operation of power systems protection equipment using disturbance recorder data is used as an example throughout the paper.

Reducing Distributed Generator Curtailment Through Active Power Flow Management

This paper presents an evaluation of the main characteristics of two power flow management (PFM) methodologies against a traditional inter-trip approach typically used by distribution network operators. The two PFM algorithms were developed, by the authors, for real-time operation with an aim to implement them in distribution networks with growing penetrations of renewable DG. The first PFM approach is modelled as a constraint satisfaction problem (CSP), while the second is based on an optimal power flow (OPF) approach. These PFM algorithms are deployed on real substation hardware to simulate the monitoring and control of MV distribution network power flows through DG real power regulation. Multiple scenarios are presented to the closed-loop PFM test environment to demonstrate the algorithms ability of detecting and alleviating thermal overloads and recognizing when the constraint has passed. The main objective of this paper is the quantification of the resultant curtailment levels, for the two approaches, which are compared to that of a traditional inter-trip scheme for the same circuit overload duration. The results demonstrate that taking an active approach to managing power flows can significantly increase the output of DG units in a thermally constrained network.

Distributed voltage control in AuRA-NMS

This paper presents real time test results arising from the application of a case based reasoning technique for voltage control of a section of existing UK 11kV network. The test network includes two distributed generation schemes. The objective of the case based reasoning technique is to maintain voltages within statutory limits while also maximizing the DG access to the network. The control algorithm is embedded on a commercially available hardware platform designed for installation in power system substations. The case based reasoning technique employs the following control actions to achieve its objectives: DG real power control, DG reactive power control and transformer tap change control. The voltage control technique is tested during over-voltage conditions and for situations where only partial sensor data is available.