T&D World Magazine
rooftop solar Copyright Alexandra Beier/Getty Images

Distributed vs. Centralized Generation: Battle of the CEOs

Two smart business and technical leaders disagree on the utility of the future. Our advisory panel chewed on the issues and they don’t agree with each other either. But there’s plenty of food for thought here so join the discussion.

Back in March we had quite a discussion about NRG president and CEO David Crane’s announcement that NRG will begin to offer rooftop solar systems to NRG’s nearly 2.3 million retail customers.  (See Rooftop Solar, a Trend that Could Grow). Crane opened a new line of dialogue in the industry (or a big can of worms). In his vision, centralized generation fades away to be replaced by distributed rooftop solar and maybe even gas-powered Stirling engine powered generators for back up and peak augmentation. NRG is a huge national energy company with its fingers in many pies and making acquisitions left and right - when David Crane speaks, folks listen.

So maybe the centralized generation paradigm that’s been around for over a century is beginning to circle the drain??

Not so fast – here comes Joe Welch, the founder and CEO of ITC Holdings, the largest transmission-only company in the U.S. He’s another outspoken industry top dog but with a different opinion.

“Distributed generation wasn’t successful in 1900. It wasn’t successful in 1950. It wasn’t successful in 2000. And it won’t be successful in 2050,” Welch said at the Utility of the Future Leadership Forum as reported in Utility Dive . “I’m not a big fan of seeing distributed generation playing a big role because— from a cost to the customer, from an efficiency of delivery, from an environmental standpoint — it loses on all counts.”

Utility Dive continues quoting Welch: “If we share a generator between the two of us, we’ll find out we need less generation than if each one of us does it alone,” he explained. “Once you get further from the load, you find out that the diversity factor of the load is around five-to-one — which means I can serve five times greater load with one generator than I could if they were located at the customer load.”

“You have to achieve five times the efficiency of a central station plant with a distributed generator to equal the economics of a central station plant,” Welch said. “You can add that all up. You can’t get to the 5:1 efficiency ratio. You just can’t beat it.”

So there you have it – two smart business and technical leaders disagree on the utility of the future. Crane’s approach is a high-level vision without looking too closely at the details. Welch is looking at the details such as diversity effects.

I tossed the discussion out to our diverse advisory panel and they don’t agree with each other either. But there’s plenty of food for thought here so join the discussion.

Paul Mauldin

Distributed Generation Isn’t the Complete Solution

Congratulations to ITC's Welch for standing up to the latest fad fueled by politics and activism that distributed generation is the answer to all our electric problems.

We seem to forget the United States operates the most reliable and cost effective electric system in the world. Moving totally away from the central generating concept would result in just the opposite. Do you really believe distributed generation will produce 99.99...% reliability? Historically, as Mr. Welch reminds us, it has failed to do that and prospectively will not do much better as a major alternative to our current system structure, even with all the fancy electronics available today.

Some claim distributed generation would have some advantages in reducing interruptions caused by storms. This might be true with localized and isolated storms but would not be the case for area wide disturbances such as hurricanes and ice storms. In fact recovery from extreme weather events would probably be more difficult in a large scale area primarily dependent on distributed generation.

As Mr. Welch contends, distributed generation is less efficient, more costly and environmentally less desirable than the central station model we rely on today. True some individual and customized applications of distributed generation may make sense in time, but major reliance on the concept as the ultimate solution is not justified.

Matthew C. Cordaro PhD
Trustee at Long Island Power Authority, Former Utility CEO, University Dean

Redundancy Isn’t so Bad

I think Welch is absolutely correct in two respects: (1) grid-connected DG represents a redundant investment, and (2) it’s generally more expensive than large-scale generation.

Redundancy, however, is a traditional industry staple and not in itself a bad idea. The question is whether the marginal benefit justifies the specific investment, and whether it's the best you can do for a particular value you want to obtain. Also, we are not talking about completely replacing large-scale generation. Rather, I think we are looking for components of a diverse generation portfolio to complement each other. So the promising use cases for DG involve situations where despite a higher cost per installed MW or per MWh, DG beats alternative investments in generation or transmission due to other factors, including a big one: social acceptance.

Where I also differ with Welch is in his application of the load diversity argument. Of course, it is true that this was the crucial driver behind geographic expansion of the grid as we know it. However, what is different today than in 1900, 1950, and even 2000 is the ease with which we manage information and control electronic devices. Smart loads — poised to grow for reasons unrelated to DG per se — in small aggregations can likely exhibit a decent amount of the load diversity that it requires very large numbers of ‘dumb’ loads to produce. Staggering the duty cycles of thermostatically controlled loads is going to be a piece of cake pretty soon, and no consumer has to notice or be worse off for it. Also, it is becoming much easier to discriminate between critical and non-critical loads. So the actual value of DG can no longer be considered in isolation from the loads, and it will be difficult to generalize as the analysis becomes more granular.

Matt raises an interesting point about recovery from large-scale events. I think the impact of DG in this context will depend crucially on the visibility that distribution operators have of their network, and whether they have non-standard operating strategies (e.g. flexible islanding) at their disposal. We are still a ways away from being able to leverage DER for resilience, but that doesn’t mean we shouldn’t work on it.

Alexandra “Sascha” von Meier, PhD
Co-Director for Electric Grid Research, California Institute for Energy and Environment , Adjunct Associate Professor, Dept. of Electrical Engineering and Computer Science, UC Berkeley

Same ol’ Same ol’ is Poor Planning

Sascha is right on target. However, I’d like to emphasize that in a world where concerns over terrorism and the increasing frequency and violence of extreme weather events are growing, the same ol’, same ol’ isn’t much of a plan when it comes to the supply of electricity. New thinking and corporate agility will be required to allay those concerns, i.e., we’re no longer living in a one-size-fits-all universe.  DER of all types will have a significant role to play – especially for critical loads and in local areas – but, as Sascha points out, so will the integration of smarter controls as we transform the grid to meet 21st Century challenges.

Lee Harrison, former editor at McGraw-Hill’s Electrical Week newsletter, editor for Business Week, and researcher with EPRI.

Odds of Doing Harm are Greater than Doing Good

Chiming in as largely in support of what already has been said, although I think it is understood, but unstated, by the authors of the prior statements, I’d like to make it clear that DG per se is not the solution. Deploying DG without properly considering the intentions for deploying DG, and then designing and operating it accordingly, is more likely to cause than solve problems if left to chance. In other words, DG is neither inherently “good” nor “bad,” but capable of doing either. Furthermore, given the relatively greater complexity of deploying DG, the odds of doing harm are greater than doing good. I think that might be one of the worrisome factors for utilities since they are ultimately seen as being the ones responsible for keeping the lights on and costs in check, and yet there are portents that utilities might not be adequately involved in selecting the intentions of deploying DG and in designing and operating it. As has also been said, “smart” technology can go a long way toward assuring generally beneficial deployment of DG, but I worry a great deal of the adequacy of institutional processes and structures in regulation and market design to do their part.

Dr. Merwin Brown
Co-Director, Electric Grid Research
California Institute for Energy and Environment
University of California

DG Works Just Fine in Europe!

Having read the article I would make the following (European-based) comments. We know that individual loads (per metered location) are stochastic. Measuring loads at a given metered household location and with a sample interval of 1 second gives a most interesting and "spikey" - highly stochastic profile - as various electrical items switch on and off - randomly. Discussions with a couple of universities suggest that you could even treat these variations in a similar way to which particles are treated in quantum mechanics.
This leads on to the question: at what point does the "wave function collapse" i.e. at what point does load variation at a point on the network look similar to load variation at another point on the network - due to aggregation effects. "Similar" in this case means a roughly similar load profiles - which could be daily, weekly, monthly or seasonally.
We found that the "wave function collapses" at the level of a 500kVA urban/suburban susbstation feeding around 220 households (or HH equivalent). In the case of rural locations it happens at the level of the primary substation (7.5MVA 33/11kV).
The most popular form of distributed generation is PV. Excess PV from a given location goes on the network and gets used somewhere else. Too much on a given LV feeder, migrates onto the next LV feeder via the busbars and so on and so forth. Too much on a given LV network - then OLTC on the MV/LV transformer pushes power back up the MV network. In meshed networks nothing much happens - in radial networks some work is needed on protection systems. Key point: once aggregation effect kick in distributed generation works just fine. The gist of Mr Welch's argument seems to be that aggregation effects only kick in at the transmission level - well in Europe this is most certainly & demonstrably not the case.
Given the above, Mr Welch's comments regarding "efficiency", at least to a European, look odd. I'd add that National Grid (who operate part of the US transmission network) and the UK network, would likewise find the comments odd. 
Obviously the US network is more heterogeneous than European networks, which don't feature suburban O/H MV with transformers every couple of houses. Perhaps this is Mr Welch's point? If it is, then he did not make it clear. If it is not, then what he says is, frankly, rather strange (I'm trying to be polite).

Mike Parr
Systems Engineer UK DNO: Merseyside & North Wales Electricity Board O&M and new build. Senior Authorised Engineer to 33KV.



Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.