However with efficienct small generators, I agree that isolated power supply again looks to become better, with grid connected DG at the periphery only until the cost of integrating load rises above the cost of manufacturings fuel cells. The former is rapidly increasing. According to NETL's production cost curves, the cost of the latter would decline rapidly with volume production.

The AC transformer is not necessarily the do all and end all. Professor Cuk at CalTech who also owns and operates the Tesla Engineering at Irvine has divised a method of DC-DC voltage conversion which rather efficiently changes voltage with relatively invexpensive solid state materials if large quantities of power are not involved, which should prove even better for the use of DC to power electronics. For many years the DC "transformer" was a DC motor/generator set. Using DC for electronics would certainly eliminate 69 cycle hum.

My look at DC-AC inversion efficiencies by a brief survey on the internet seems to show efficiencies of 85% to 93%, or losses of 7% to 15%. Does the author suggest percentages that also include the rectification of the AC back to DC in the appliances?

It is interesting to note that PlugPower says the system architecture of its residential fuel cell is designed with the idea of excluding the AC-DC inverstion ultimately as well as the natural gas reformer based on the premise that ultimately hydrogen gas will be piped to households.

Thanks for your interesting article.

DC power supply on commercial scale is not yet feasible. We have to be honest in our assessment. Transmission losses in ideal A/C system are 6% only for average distance of few hundred miles. No DC system can match that.(HVDC is viable beyond 500 km.) I am not against on site generation which is efficient. But in developed countries over 90% population live in cities that have 1% geographical area. Or Population density may be well in excess of 100 times than outside. In Delhi when only one in 10 homes operate 1kw avg. gensets during blackouts, -there is unbearable polution. So it may not be feasible to switch to local generation.

Micro turbines develop high frequency which can be easily rectified so DC local supply is fine. Utilities give reliable source of electrcity at fixed voltage 110V or 220V and consumers can plug in any commercially available appliance, all shall work. But Home generation may not be able to maintain supply. And typical consumption pattern is such that for average demand of 1 kw -peak demand may be 20 kw. So consumers have to make big adjustments. Heat devolped by micro tubines may not be required at all times. In the case of non use of exhaust heat, Microturbines will be least efficient. BUT FOR STANDBY / SEASONAL USE they are excellent.

You assessment is off target when we consider that bulk consumption devices will continue to use A/C.( Airconditioners, Frig, Washing Machines, Microwaves, Fans.) You will not like to switch to DC to run LAPTOPS or Telephones. DC supply in bulk will not be feasible in near future. Only major tech. breakthroughs like economical solar cells or very cost of gas warrant switch to DC power on large scale. ---Ravinder Singh

I feel strongly that the integrated efficiency of these DG systems has to include the rectification at the device, but that is not were we differ on the efficiency I quoted for DC to AC inversion. There I have a different beef.

The 85% to 93% efficiency you quoted I think represents the peak efficiencies, but omits a certain kind of threshold loss, or the power consumed by the inverter while it is spooling-up to deliver the AC. I would argue that the shoulders of the day are infact broader than one might casually consider. Therefore, if insolation is the numerator in the ratio, I think that it is a bigger number than is usually assumed. Put another way, I think the 85% to 93% range ususally quoted doesn't represent the total available resource, and that if it did, you'd have to take it down 5% to 10% at least.

On the rectification side of the equation, I am quoting the CEO of PowerOne, who said that the range of efficiencies for power supplies for appliances generally ranges from 60% to 85%, depending on cost.

This is a commercially viable option now. My company has installed the largest commercial solar system in San Diego, CA, which includes both a conventional inverter and a Direct Coupled (r) system to highlight the differences between the two.

Here we are directly connecting the DC output from the solar cells to a large, operating DC load in the building, the florescent lights, without going through the AC inversion and therefore saving those losses. This method increases the efficiency of solar, which has the same economic impact as a cheaper module does, over time.

A switch of distribution components to DC is already happening in some data centers. DC distribution makes sense in computer rooms for load aggregations up to certain sizes. Those sizes are well within the typical loads suggested by the author.

The further development of distributed generation will make these DC sources more available, triggering the opportunity to DC networks (DCLANS?). Such networks can have benefits not realized by AC systems; isolation, readily-available backup power (in batteries, SMES, and ultracapacitors), reduction of potential points of failure.

With the growth of alternative energy systems, I believe we will see significant progress in this direction.

By the way, I don't believe in the"genius" of Edison, except as a businessman. His statement that "genius is 1% inspiration and 99% perspiration" was true for him. His genius was in the use of other people top make inventing/developing into a business. Those small DC generators of Edison were due to a lack of technology and vision, not to a surplus of it.

gk

I do, however, agree with George that a large and predominantly DC load like the computers and modems in a Data Center would be an excellent candidate for a grid-independent DC feed. This would be similar to the powerplants I work in, where nearly all have a separate DC bus (supported by batteries during outages) that operates all the controls and protection equipment. As we become more dependent on technology and communications, Data Centers are critical loads.

Perhaps the more important lesson from Edison is that while his power plant was "only" 8% efficient at producing electricity, it was 50% efficient overall, recovering waste heat for local industrials. This is an inherent benefit of DG and inherently not possible in the central model, as demonstrated by the paltry 33% we get today from the central system. As enticing as the 50%+ efficiencies quoted by Mr. Brand are, it's even better when you consider that these fuel cells can achieve 80% or higher overall CHP efficiency - and indeed, virtually every distributed generator that gets installed nowadays for non-peaking service has this CHP functionality built into its design. With the right regulatory signals, we could start deploying these technologies and return the grid to the overall efficiencies it realized in the 1920s. (And if that sentence reads as counter to your understanding of how businesses and technologies progress over time, you are starting to understand the core problem.)

One last comment and correction: Savage notes that "nearly all DG technologies are intrinsically DC-producing". This claim is often asserted, but inaccurate. According to FERC data (which grossly undercounts the total), there are over 23 GW of end-user sited CHP plants currently in operation, with a median generator size of 2.8 MW. This is most definitely DG, and while it may lack the sex-appeal of PV, wind and fuel cells it is the reality of the DG universe today. It's gas turbines, it's steam turbines, it's engines with waste-heat recovery... and it's almost all intrinsically AC. In the wake of last month's outages, we need to remain keenly aware of the fact that there are no shortage of currently available, currently cost-effective DG technologies out there. There is an implicit danger in focusing only on the next generation, DC-driven approaches, as this tends to drive regulators to treat a regulatory problem as an R&D problem and allocate resources to the wrong (near term) solution.

All that said, I think there is a good case to be made for DC appliances and DC generators in off-grid and/or microgrid applications where the series of inverters and rectifiers are simply unnecessary efficiency penalties - but it will take a lot more severe outages before such applications become anything more than a tiny minority of the DG market opportunity.

The other point is that I would not suggest that the DC building-side network is a grid independent model. In fact, I think one of the more elegant features of what we are doing is that it is completely complemetry to the grid. These systems are grid connected, not interconnected, so we do enjoy the grid's delivery system, but only one-way. Our system is not designed to redeliver power via the grid, although we do contemplate servicing non-DC loads in the building too. I would send you to the diagram on the splash page of our website, which makes this point clearly.

Further, I would reiterate to you as I did to Mr. Clarke-Johnson that the idea I am promoting is for greater connectivity for power - like a router does for information- that includes AC as an input. I am not out to destroy invested capital, but to complement it.

Finally as for the definition of what DG is and what is inherently DC. I like three speeds which allow for small power plants to be called Disbursed Generation (of say 1MW and up, but with only one customer or so) and Distributed Generation to account for the micro generators below that level. Be all that as it may- whether we want to say that anything not centralized is distributed or not - I still like to mix it up on the subject of what is inherently DC. I think that by definition only synchronous generators can produce AC that is useable. Usually it is the case that there is, even in the high-speed generator, a DC state that the electriciy must go through to be properly be made back into usable AC. This happens at very high voltages, and so it's more efficient than low voltage conversions. But these are hair-spliting distinctions, and I certainly grant you that the vast majority of non-centralized power production is done by reciprocating and turbine engines.

One last bit - I think it is a regulatory issue too, but not all grid regulation. I think if we got real federal coordination on the National Electrical Code and the grid that the government could support the development of DC networks in buildings that would bring the country greater efficiency, flexibility, and reliability.