Managing Distributed Electricity Generation Sources in a Decentralized Architecture

Md Salman Nazir

Department of Electrical and Computer Engineering, McGill University, Montreal, Canada

Email Addresses:

Research Notes: Received 04 November 2011; received in revised form 28 February 2012; online published 16 July 2012


Large scale integration of distributed generation (DG) sources require significant restructuring in operations and planning of electric power. While a more decentralized architecture is envisioned, operators may need to overcome several technical and economic challenges to adapt to such paradigm shift. This paper highlights some of those key challenges and discusses the importance of efficient forecasting and agent based modeling in managing DGs in a decentralized architecture.

Keywords – Distributed generation, renewable, decentralized, agent-based control, forecasting

I. Introduction

The electricity industry is going through a massive restructuring years with significant interest and technological advancements in distributed generation (DG) sources, such as small to medium scale solar photovoltaic, wind, micro-hydro, fuel cells, etc. The innovations in these technologies and industry adoption are mainly driven by environmental concerns [1, 2]. However, in an electricity system that up to now has been very centralized, the economic as well as technical challenge will be to optimally integrate the increasing number of decentralized small generation units. By understanding and properly addressing the technical and economic challenges, it will be possible for the system operators to ensure optimal planning for expansion of generation, transmission and distribution facilities, ensure secure reliable supply of electricity while securing the interest of the end user.

Due to fast growth in ‘active’ distribution networks, power systems need to adjust operating under decentralized control architecture rather than centralized control architecture [1]. In this paper, the possible roles of intelligent forecasting tools and agent-based control are highlighted to manage the future power systems.

II. Characteristics of DG Sources

Chambers [3] defines DGs as relatively small generation units of 30MW or less, while Ackermann et al. [4] define DGs in terms of connection and location rather than in terms of generation capacities. They define DG sources as electric power generation units connected directly to the distribution network or on the customer sides of the meters. Residential or industrial customers, distribution network operators or even traditional power generation companies can own DG units. Table 1 lists the different DG technologies and indicates their respective characteristics.

 Table I: Characteristics of distributed generation sources [1].

Stand by Capability Cogeneration Weather driven
Reciprocating engines Yes Yes No
Gas turbines Yes Yes No
Micro turbines Yes Yes No
Fuel Cells Yes Yes No
Photovoltaic No Yes Yes
Wind No No Yes

Many DG sources use cheap fuel (or ‘free’ in case of wind and solar) and some can act in co-generation mode (cogeneration of heat and electricity). Depending on the type of fuel used and its impact on the environment, many of the DG sources (wind, solar, hydro, biomass-fuelled plants, etc.) are considered as renewable sources. The growth of such distributed sources can ensure larger portion of renewable energy in the overall generation mix. However, sources such as wind and solar are highly weather driven and thus the control is more challenging.

 III. Benefits and Issues Related to Distributed Generation

A. Benefits

Environmentally friendly and lower emission of carbon-dioxide and other toxic gases, these are the main factors driving the growth of renewable distributed sources. Regulatory bodies and policymakers of many countries have put favorable policies in place to promote the growth of such clean energy. Secondly, since many DG sources (except wind farms) can be located close to the loads, thus transmission losses are greatly minimized and the need for investing in expensive transmission lines is reduced. According to International Energy Agency (IEA) [5], on-site production could lead to cost savings in transmission and distribution of about 30% of electricity cost. Thirdly, IEA [5] recognizes the importance of DG in providing reliability services for the power systems. Moreover, in countries that operate under deregulated electricity markets, DG sources can serve as a hedge against price fluctuations since customers also have a flexible mechanism to respond to market conditions rather than simply being the ‘price taker’.

B. Issues

Besides the typical high financial cost during initial installment, there are additional challenges that hinder the rapid integration of renewable DG. Since power systems traditionally operate in centralized mode, the new shift to distributed sources leads to some complications. There have been major concerns related to uncertainty in forecasting weather-driven natural sources such as wind and solar. For example, wind power outputs are highly variable and often experience large positive or negative ramps, i.e. output fluctuations [6]. Forecasting these ramps and allocating back up resources to cover such ramp events in an economic and reliable manner is a challenging task. In the case of high wind velocity, the wind farms need to curtail wind, i.e. spill wind by changing turbine blade angles, since the system operators may not be able to fully compensate for rapid ramp ups or downs of wind due to technical or economic reasons. However, since it is intuitive not to waste such ‘free’ energy, development of highly efficient forecasting tools are enforced. Similarly, solar outputs can fluctuate sharply and hence proper forecasting tools for optimal storage sizing is essential.

Moreover, large scale deployment of decentralized sources of electricity may lead to instability of the voltage profiles since bidirectional power flows and complicated reactive power flows may arise [1]. When connecting DG to grid, operators need to ensure voltage stability and reliability which requires careful and detailed studies.

IV. Overcoming Challenges

A. Forecasting Tools

Since the weather driven renewable sources of electricity introduce uncertainty in the model, these uncertainties must be characterized efficiently and the controllable part of the generation sources must be flexible enough to mitigate the risk associated with uncertainty and variability of the renewable sources [7]. Advanced probabilistic forecasting tools can contribute greatly by efficiently modeling the characteristics of such renewable sources. Large collection of data from various sources (meteorological sites, wind firms, solar plants etc) with high degree of resolution (minutes/ seconds) is required to build efficient forecasting tools.

Moreover, machine learning algorithms and Kalman filtering techniques can be very useful in learning the dynamic characteristics of variable sources and predicting the future states of the network. System operators can benefit from such tools in predicting the level of backup resources that need to be allocated to ensure power systems can operate reliably even during large ramp events in weather-driven sources.

B. Decentralized Agent-based Control

In a decentralized architecture, the role of communication systems and automation is crucial. From customer household/industrial appliances to DG sources including the traditional large power plants, everything can be thought of as agents. Efficient coordination among all these agents is the key to ensure a functional and reliable power system.

Distributed artificial intelligence concepts can be implemented to model an efficient agent-based system [8]. In the agent-based systems, various layers of information, such as the technical conditions (voltage levels in the local network, frequency, outage information etc) and the market conditions (price, spikes, etc.), need to be monitored and communicated with other agents so that agents can negotiate and coordinate. Control algorithms dictate the actions of the agents while respecting all the technical constraints of individual agents and of the network.

Agent-based models are already being used in applications in the area of electric wholesale markets, health care, transportation, etc [9]. These frameworks can be extended to model agent-based operations of power systems. The utility service area can be represented by a sample of utility customers. In the agent-based model, the customers are characterized by end-use equipment holdings, end-use electricity use and hourly loads, demographic and other variables [10], while the generating units can be characterized by their size, operations, flexibility, generation patterns etc.

In a highly flexible market platform, the actions of agents should be supported by economic justifications. The agents representing power suppliers and customers act as self interest maximizing entities, while an operator agent can act to make sure the technical constraints and the local policies are always met during negotiations. Efficient coordination among the agents is realized through a market mechanism that incentivizes the agents to reveal their policies truthfully to the market [11], thus establishing a functional agent-based reliable power system. The dynamic behavior of the agents and the overall system can be better characterized over time with availability of data.

V. Conclusions

With the trend leading towards renewable, decentralized, and highly fluctuating distributed generation sources, there is a tremendous challenge regarding the stability of future power grids. Hence, agent-based decentralized control and advanced forecasting tools can add great value in realizing large scale integration of distributed resources in electric power systems.

VI. Acknowledgements

The author likes to acknowledge Professor François Bouffard, Dept. of Electrical and Computer Engineering, McGill University.

VII. References

1. G. Pepermans et al., 2005, Distributed generation: definition, benefits and issues/ Energy Policy, 33, 787–798.
2. Renewables Global Status Report: 2009 Update. Renewable Energy Policy Network for the 21st Century (REN21), Paris, 2009.
3. Chambers, A., 2001. Distributed generation: a non-technical guide. PennWell, Tulsa, OK, p. 283.
4. Ackermann, T., Andersson, G., Soder, L., 2001. Distributed generation: a definition. Electric Power Systems Research 57, 195–204.
5. IEA, 2002. Distributed Generation in Liberalised Electricity Markets, Paris, p. 128.
6. Erik Ela and J. Kemper, 2009. Wind Plant Ramping Behavior. National Renewable Energy Laboratory, USA.
7. F. Bouffard and M. Ortega-Vazquez, 2011. The value of operational flexibility in power systems with significant wind power generation. IEEE Power and Energy Society General Meeting.
8. T. Sandholm. Distributed rational decision making. In G. Weiss, editor, Multiagent Systems: A Modern Approach to Distributed Artificial Intelligence. MIT Press, 2000.
9. Tesfatsion, L.S., Judd, K.L., 2006. Handbook of computational economics. Agent- Based Computational Economics, 2. Elsevier/North-Holland (Handbooks in Economics Series).
10. J. Jackson, “Improving energy efficiency and smart grid program analysis with agent-based end-use forecasting models,” Energy Policy, vol. 38, no. 7, pp. 3771–3780, 2010.
11. S. Lamparter, S. Becher, and J.-G. Fischer, “An agent-based market platform for Smart Grids,” in Proceedings of the 9th International Conference on Autonomous Agents and Multiagent Systems: Industry track, Richland, SC, 2010, pp. 1689–1696.

Leave a Reply

You must be logged in to post a comment.