Due to a recent successful application to the Science Foundation Ireland (SFI) Investigator Programme that funded a proposal of mine with title ‘Advanced Modelling for Power System Analysis and Simulation’, I was given the opportunity to contribute to this forum.
In this article, I wish to talk on an evergreen open question in power-system analysis. The question can be formulated as follows: what is best to properly understand the dynamic response of a complex power system? Is it a brute-force time-consuming campaign of simulations or a ‘smart’ simplified model able to capture the essence of the system behaviour?
In the good old times, when there were no computers or computing capability was limited, the answer was simple: researchers had to study simplified models. For this reason, the one-machine-infinite-bus (OMIB) system is ubiquitous in the literature.
This was not bad, however. The lack of powerful computers led to great ideas based on pure physical and/or mathematical considerations. The book Transient Phenomena in Electrical Power Systems by V. A. Venikov (originally published in Russian in 1958 and later translated into English in 1964 by Pergamon Press) is a very good example of how creative a brilliant mind can be with a pencil and a piece of paper.
Power system analysis - size matters
[caption id="attachment_32631" align="alignright" width="300"]
Source: http://www.evwind.es[/caption]
While fascinating, though, it is clear that a power system is not an OMIB. Actually, interconnected continental high-voltage transmission systems, such as the European system (ENTSO-E), are among the most complex systems ever built by mankind. Complex behaviours cannot be captured by using over-simplified models.
It is well known that Einstein used to say: “Everything should be made as simple as possible, but not simpler.” This is actually very true in power system analysis. Increasing the size of the system, thus leading to a large set of nonlinear equations, gives birth to a variety of unexpected transient responses that simply do not occur in tiny networks.
So, in power systems, the size does matter. This simple concept is well understood by system operators. The need for as detailed as possible simulations has led in the 1960s-70s to the construction of transient network analysers – which were basically custom analogical computers tailored to emulate the physical behaviour of a transmission system – and, more recently, to the construction of real-time digital simulators with several processors and outstanding computing capability.
The machines above can be very costly. In recent years, however, the cost of powerful workstations has decreased so much that anybody can simulate a detailed model of a large power system using any of the available commercial and even open-source software tools at a relatively inexpensive price.
Thus, the temptation to substitute thinking with brute-force simulation becomes very sexy, especially for young researchers. The result is easy to guess: young generations of electrical engineers spend less and less time to understand the dynamic behaviour of the grid, and more and more to simply observe what happens looking at simulation results.
This is dangerous, as the ability to find efficient solutions to the security and stability issues that a power system can face largely depends on how deep is the understanding of the physical mechanisms of its transient response.
With regard to this topic, there has been recently a very interesting diatribe raised by Prof Janusz Bialek, Skoltech, Russia, against the “overuse of simulation as a replacement for thinking”. Prof Bialek completed his provocative letter appealing “to fellow power engineers for not switching on their computers for one day and thinking instead about the problems they are trying to solve, the physical fundamentals, and the mathematical model”.
There are also risks in focusing exclusively on pure mathematical aspects. In the diatribe above, David Hirst, inventor and consultant in the area of power system analysis, reminds that, generally, “The best engineering papers present a non-mathematical intuitive view of the world first, and the maths presents a more precise hypothesis for model testing whether this view is valid.” Intuition is thus crucial but can happen only through experience, which in power system analysis means solving thousand of simulations and studying their results.
I believe that both simulation and maths are necessary. Simulations provide the ‘experimental results’, allow testing new techniques as well as proving the correctness of hypothesis and simplifications. I (ab)use the word ‘experimental’ here because, unfortunately, it is just impossible to test, for example, a new technique that prevents transient instability, on a real-world system. No system operator will ever allow that!
Simulations are thus the closest thing to the real-world system to which researchers in the area of power systems can have access. Then good mathematical tools and intuition are also necessary, because simulation results alone are nothing else than a silent set of numbers.
Hence simulation and maths have to be utilised together. This is a classical situation where the powerful Hegel's approach becomes handy. This consists in taking thesis (i.e. simulation) and antithesis (i.e. maths and intuition) to lead to a synthesis, which not only retains the best of thesis and antithesis, but also provides a higher level of understanding of the problem.
As a reviewer of several scientific journals and as an editor of the
Institute of Electrical and Electronics Engineers Transactions of Power Systems and
Institution of Engineering and Technology Generation, Transmission & Distribution for the last few years, I have sadly noticed that the vast majority of manuscripts submitted to scientific journals are either too mathematical with little or no application, or too simulation-oriented with no new insight or theoretical added value.
In these days, it is very rare to see a paper with the right balance of both theory and practical aspects.
If you have read until this point, you may have started asking yourself, “This is ok, but what is the role of the Irish transmission system in this context?” Well, I think that it can actually play a very interesting role, as I will try to explain in the remainder of this article.
Role of Irish transmission system
Being a relatively small island, the Irish system can be modelled in detail with a relatively small set of equations – few thousands of variables as opposed to the hundreds of thousands or millions necessary to model the ENTSO-E system. This size can be properly handled with modern workstations.
The Irish system is self-contained and stand-alone. In fact, while there are connections with other grids – the Irish system is connected through DC cables to Wales and Scotland and a DC connection to France in on the way – these connections are non-synchronous and are actually very interesting from the stability point of view as they can provide regulation that standard ac lines cannot do.
The Irish system has a high penetration of renewables, in particular wind generation, which makes it an ideal system for the study of the transient response and control of networks with low inertia. There are also plans to implement massive load demand response, which means the ability of loads to actively participate to frequency regulation. The reduced size of the system can make this possible in the short term.
Last but not least, people in EirGrid – the Irish system operator – have the expertise and the will to collaborate with academia to investigate new solutions to improve the security and the efficiency of the grid.
In summary, the Irish transmission system can be viewed as a nation-wide smart grid. It is large enough to show most dynamic behaviours that a complex system can show and small enough to be mathematically tractable.
It appears thus that the Irish system can be the answer to the old dilemma among simulation and pure maths and intuition. At least, this is what I suggested in the proposal that has been recently granted by SFI. Whether this will happen or not, however, is a bit early to say.
Clearly, I hope the research will be fruitful as it promises to be and that I will have the opportunity, in five years from now, to write again on this forum and talk about the results of the project.
Author: Federico Milano, PhD, FIEEE, is Professor of Power Systems Control and Protections at UCD