The Complexity of Smart Grids
- Written by Antonello Monti and Ferdinanda Ponci
As generally envisaged, the smart grid is not only a complicated system made of many parts, but also a complex system—one in which overall behavior cannot not be directly inferred from the behavior of the individual components, and one that no single entity can control, monitor and manage in real-time. Because it is complex, distributed control is more than a convenience; it is required.
The origin of the complexity of smart grids of the future can be summarized in two main factors, which are not completely independent: the interdependence between heterogeneous infrastructures, and the distributed nature of monitoring and control functions.
Let us begin by summarizing the types of interdependent heterogeneous infrastructures, the points of coupling and the emergence of complexity.
The presence of intermittent energy sources, such as renewables, which may not be reliably predicted and dispatched, requires the presence of energy storage (or of other fast dynamic sources) for balancing the variability of generation. But massive deployment of new dedicated energy storage units may not be feasible and may not be necessary either.
Instead, to take best advantage of resources and address practical limitations, we need to adopt a broader concept of energy storage that embraces non-electrical energy networks, such as gas and heat, as well as distributed storage resources, such as plug-in electric vehicles. The electrical system following from that vision will eventually be coupled with the gas, heat and traffic systems.
That is not the only complication. The prediction of wind and solar generation depends almost entirely on weather forecasting, which is but one of several important factors in load prediction.
Business models for operation of these interconnected systems encompass the creation of new markets for reserves and reactive power, the establishment of Virtual Power Plants (VPPs), virtual storage systems, dual demand-side management systems and so on. Thus, the economic organization of the power system represents yet another type of infrastructure and another point of coupling between the aforementioned infrastructures.
Finally, the coherent operation of all these infrastructures, and the unprecedented interactions between generation, transmission and distribution sections of the power system, make communications more critical than ever before. The communications infrastructure acts as the glue holding together the previously mentioned coupled infrastructures. One of the main efforts in this direction is the European project Finseny.
Given this framework of interdependent grids in which a very large number of actors play significant roles, the distribution of control and monitoring functions is a necessity. The energy system envisioned is not only a complicated system made of many parts, but is also a complex system—one in which overall behavior cannot be directly inferred from the behavior of the individual components, and one that no single entity can control, monitor and manage in real time.
That is to say, the net effect of the interdependencies is largely unknown and unforeseeable in the absence of a clear view of the coupling points, ways to model them, and of data and measurements. Moreover, with no way of predicting overall behavior there is no simple way of controlling net behavior either.
The vision of the smart grid as a complex system has practical effects for those involved in its creation, in terms of education and training, tools, and methods and design.
Interdisciplinary study (and research, for that matter) have long been buzzwords in academia. Nonetheless, the instruments to make research and education interdisciplinary are extensively lacking, as the legal and organizational background is not ready for it. The integration of the disciplines of electrical engineering alone, and of closely related fields, is in itself a challenge.
The creation of truly interdisciplinary areas, with related official degrees, conflicts with a structure that was developed for very different (and sometimes avowedly opposite) educational needs.
- What is the right curriculum for smart grid engineers?
- What should the ideal new hire of an energy service company know?
- What topics will the engineer have to keep up with?
Academic education is not the only competence-building environment affected by these challenges. Professional education and training at all levels are equally affected. Field experience—not just for those being educated but also for the educators themselves—is an important element of the whole picture.
By tools we mean: the numerical tools to carry out analysis of complex smart grids, the technologies for de-risking new devices, and algorithms supporting the transition from numerical to in-field testing.
These tools are primarily numerical simulators, and Hardware in the Loop (HIL) and Power Hardware in the Loop (PHIL) testing environments. (An example of their application to the interdependent heterogeneous systems part of a smart grid can be found in "Multi-Physics Test Bed for Renewable Energy Systems in Smart Homes" by C. Molitor et al., which appeared this year in IEEE Transactions on Industrial Electronics, Issue 99).
In a nutshell, these tools should support:
- a multi-physics, multi-technology approach to allow for representing all kinds of dynamic interdependences
- a multidisciplinary approach, for different users working on different aspects of the same simulation scenario, with universally understood knowledge
- dynamic and reconfigurable model-level definition, enabling different users to interact with the simulation schematic focusing though on different details and obtaining results in a reasonable amount of time
- high-level graphic visualization to support system analysis for the different disciplines, providing both preferred individual visualization options and the ability to synthesize a "system-picture"
- uncertainty propagation, from the sources of uncertainty (for example, renewable generation, loads, or prices) through the discipline borders to the entire system.
Methods and design
The lack of methods for representing and analyzing the smart grid as a complex system, and the lack of performance metrics for such a complex system, result in lack of methods for designing holistic controls and for designing components fit for this environment.
The complex smart grid must be designed to manage uncertainty and inconsistencies, to be resilient, and to degrade gracefully when necessary. All the interdependent systems that the smart grid comprises must have these features and use them in a coherent way. This design challenge may be summarized as "joint design."
Having described the factors that make the smart grid a complex system and spelled out some of the consequences, we see that they affect a variety of worlds from education and training, to the manufacture of components, development of numerical tools and the philosophy of designs. An enormous amount of innovation is required in technology and tools, together with a deep change of mindset for the transition to the smart grid to occur. The bright side is that a successful undertaking in the energy sector in the aforementioned directions may then be carried to other sectors, with widespread benefits for all.