Renewables and Simulating the Post-Carbon Grid

No one knows how much a renewable post-carbon grid would cost, but renewable advocates like Greenpeace, and Rocky Mountain Institute, and the proposed Google Clean Energy 2030 plan, all seem to think it will be cheap. However it is impossible for any planner to know if an all renewable grid will work, and if it will work, how much an all renewables grid would cost.

There are fundamental questions about the cost and reliability of an all or mostly renewables grid that cannot be answered without and adequate computer simulation too, and such a tool is not available now, nor will it be for some time to come. An Argonne National Laboratory Report from the December 2008 National Power Grid Simulator Workshop, offers a realistic picture of the current status of renewables based grid planning. For example renewables planners often offer massive expansions of the electrical transmission system as a remedy for grid reliability problems created by renewables. The report states:

The national power grids are composed of many components, each with its own range of vulnerabilities to both natural and man-made events. The most apparent components are the transmission lines that stretch unprotected across open countryside. Natural threats to these lines include wind, ice storms, and fires; potential man-made threats include rifle fire, bombs, chaff, and aircraft collisions. Isolated power line outages can be quickly repaired. The disruption of large numbers of lines or those located in special locations (e.g., across rivers and wide gaps), however, can be time-consuming to resolve. . . . Renewable power genera- tion depends largely on local conditions (e.g., wind or sunlight availability). The best locations for production are often not where the greatest power demands are. Power transmission, then, could be a limiting factor in the growth of distributed, renewable energy generation.

The report also states

Tools are needed to simulate large additions of renewable and distributed power to the grids. Models must deal with the intermittency of these sources, optimize the transmission and resource mix, quantify important metrics (e.g., economics, security, and environmental impact), and assess alternative solutions.

Yet no research tool currently exists which would permit anything like adequate research on the national transmission system that would be required by a mostly renewables generating system. Despite the lack of a research tool that would determine the practicality and cost of an all renewables electrical transmission system, renewables advocates continue to assure us that electrical transmission reliability posses no problem for a future renewables based grid. Would a future renewables based grid be stable? The report states,

There is, in general, a lack of a simulation capability and associated data for national-level studies. For instance, large-scale dynamic transient stability models do not exist, though large-scale linear load flow models do.

The report also notes

Power grid models can be classified along three dimensions: operations, planning and evolution, and disruptions. . . . Most models today are limited to a single axis of this taxonomy. The three dimensions described above, though, are not independent. Issues raised along one dimension (e.g., an evolution toward more renewable energy generation) affect issues along the other axes (e.g., changing the consequences of a wide-area weather event).

It is clear then that plans for a renewables dominated grid are highly premature in the absence of a tool or tools that can accurately asses the reliability and cost of such a grid, and compare that assessment with an assessment of other options, for example the modeling of a nuclear dominated grid. Some assessments suggest that the cost of a nuclear electrical generation system, often said to be high, might be actually far lower than the cost of a renewables dominating electrical system.

Wind on Brave New Climate

Barry Brook's blog "Brave New Climate" is one of the best climate/energy blogs. Barry is an Australian climate scientist who is generally clear thinking about climate/energy issues. Barry is a fan of the Integral Fast Reactor. I believe that the LFTR represents a safer, more flexible, lower cost technology that possesses enormous potential. Other than our sometimes raucous disagreement on most favored nuclear technology we seem to agree on most issues.

During the last few days Barry has conducted a debate on wind issues under the title "Does Wind Power Reduce Carbon Emissions?" Barry's position is derived from a study titled, "Cost and Quantity of Greenhouse Gas Emissions Avoided by Wind Generation," by Peter Lang. Lang is a retired engineer who has over 40 years experience on a wide range energy projects and issues, including managing energy R&D and providing policy
advice for government and opposition. The money quote from Lang states:

“These calculations suggest that wind generation saves little greenhouse gas emissions when the emissions from the back-up are taken into account.

Wind power, with emissions and cost of back-up generation properly attributed, avoids 0.058 to 0.09 t CO2-e/MWh compared with about 0.88 t CO2-e/MWh avoided by nuclear. The cost to avoid 1 tonne of CO2-e per MWh is $830 to $1149 with wind power compared with $22 with nuclear power. If the emissions and cost of back up generation are ignored then wind power avoids about 0.5 t CO2-e/MWh at a cost of about $134/t CO2-e avoided. Even if the costs of and emissions from back up generation are ignored, wind is still over six times more costly that nuclear as a way to avoid emissions.

A single 1000 MW nuclear plant (normally we would have four to eight reactors together in a single power station) would avoid 6.9 million tonnes of CO2 equivalent per year. Five hundred 2 MW wind turbines (total 1000 MW) would avoid 0.15 to 1.3 million tonnes per year – just 2 to 20% as much as the same amount of nuclear capacity. When we take into account that we could have up to 80% of our electricity supplied by nuclear (as France has), but only a few percent can be supplied by wind, we can see that nuclear can make a major contribution to cutting greenhouse emissions, but wind a negligible contribution and at much higher cost.“

Lang's conclusions are truly devistating to the argument that wind electrical generation represents a major solution to the problem of global warming.

In the course of the debate on Barry's blog, Mark Jacobson's work on base wind was touted. Barry responded:

Jacobson’s work on distributed wind and vehicle-to-grid backup have been savaged by Charles Barton at Nuclear Green. . . .

I’d be interested (sincerely!) in knowing where Barton is wrong.

Barry thus challenges wind defenders to answer my criticisms of the Archer-Jacobson wind system. There were a few feble attempts to answer my challenge, which mainly offered alteration to the Archer-Jsaconson rules. For example Fran Barlow suggested

Whatever Jacobson proposes it seems to me that the overbuild for wind need be no higher than the CF would imply for 100% of nameplate.

So assuming a CF of, say 35% (the starting point for feasibility IMO) the overbuild should be no more than 2.84 so that taken together the farm’s reticulated components produce the output almost all the time. Only on those occasions wherea) there was a: decline in output below the anticipated output

AND

b) demand implied the anticipated output

would redundant capacity be brought online. Ideally the sites in question would be highly predictable on 2 hours notice.

Order of call would be

a)demand management measures

and/or

b)pumped hydro/V2G

and/or

c)NG

Note that NG need not be a fossil fuel — waste biomass from ADs or syngas from CSP usaage are options

Barlow's analysis simply assumes that average wind speed over the entire 17 site array would be constant, and simply by multiplying the generating capacity by the inverse of the capacity factor would give you a reliable supply of electricity. The rub is of course that the average wind speed over the entire array is not constant. Thus more wind capacity replication than Barlow calculates is required to create the base load level of reliability.

My reading of Archer-Jacobson is of course influenced by the fact that their analysis was conducted for Southern Great Planes wind resources. The most likely time for there to be a wind problem would be during Summer days, during periods of peek power demand by Texas electrical consumers. Supplying these demands is very much a quality of life issue, and on very hot summer days, when wind generation can drop to an absolute minimum, a significant public health issue. Demand management would be an extreme measure in these situations. My critique of Jacobson on the rationality of a V2G system was part Brook's challenge, and Barlow cited the use of V2G technology without defending that technology against my critique. She also advocates pumped hydro, but fails to offer a convincing analysis of how pumped hydro would provide a low cost solution to the problem of summer winds on the Southern Great Planes. Finally Barlow falls back on hydro-carbon solution including natural gas. The Natural gas solution of course does what we are trying to avoid, that is adds to the atmospheric CO2 levels. The other Barlow solutions, "waste biomass from ADs or syngas from CSP usage," involve further replications and added expenses. Needless to say Barlow does not stop to ponder the questions that her solutions raise.

Barry Brook has generated a necessary conversation on the limitation of wind, a conversation that will need to be repeated over and over again during the next few years.

How much would an all renewables electrical system cost?

The Gore plan, the Google plan, the energy writings of Joe Romm, the views of the Internet site Gristmill, and other self proclaimed energy authorities, all maintain the view that an all renewable grid is possible.  Some time  ago I attempted to evaluate the theory of reliable wind suggested by Mark Z. Jacobson.  Jacobson argued, based on empirical data from 17 sites in the southwestern Great Planes, that wind generation could be made reliable by building grid links between those sites.  Jacobson found that the linked sites could be expected to produce at least 20% of their rated capacity 80% of the time.  Jacobson further argued that this reliability approached that of base generated electricity.  My analysis, using a 2008 wind cost estimate of $2500 per KW, and evaluating the Google energy plan, found that the 380 wind GWs called for by the Google plan would cost $900 billion to install,    This estimated installation cost did not include the expansion of the grid that would be needed to transmit the electricity from the windmill array to consumers.   I found that the linked wind array could be counted on to produce about 80 GWs of electricity 80% of the time.  The linked wind system, however, had a serious flaw.  It could not deliver power on hot summer days when electrical demand peaked.  

I undertook a comparison between the Google wind proposal and  an alternative scheme to build nuclear reactors at the same combined cost as the wind array.   I estimate at $900 Billion dollars would purchase 112 one GWe reactors.  At a .90 capacity factor, the reactors could be counted on to deliver 101 GW of electricity at any time.  Or a little over 20% more electricity at any given time than the wind array.  Unlike the wind array, the reactors could be counted on to deliver electricity at close to maximum capacity on hot summer days.  

In addition I offered evaluations of the cost of wind with three energy storage plans.  The use of batteries, to store wind generated electricity, the use of pumped storage, and the use of Compressed Air Energy Storage (CAES).  My CAES study was in turn based on  "The Economic Impact of CAES on Wind in TX, OK, and NM," by Ridge Energy Storage & Grid Services L.P, for the Texas State Energy Conservation Office. in 2005. I assumed .40 wind capacity factor and that 40% of the energy output from the CAES system would come from the burning of natural gas, a standard assumption for CAES systems. The Ridge Energy study showed that electricity could reliably dispatched on a 24 hour a day basis from a CAES system on a 24 hour a day basis, even during low wind summer days, demonstrating the viability of a Wind-CAES system, However, the energy output of CAES systems is .80 of energy inputs. This suggests that there are considerable in efficiencies in the use of wind generated electricity by the wind CAES system, and that at least 30% of the electrical input is lost to system inefficiencies. Ridge Energy estimated that the capital cost of a CAES system would run @$765 per KW, an exceedingly modest sum, but one which should be examined. The capital cost for source wind array combined with the CAES system is in fact much higher.

I stipulated a cost for new West Texas wind of $2250 per name plate KW in 2009. This price was at the low end of 2008 windmill costs in North America. Since the capacity factor of West Texas runs around .40, the adverage output West Texas wind producer can expect to pay $5625 for every KW of electrical producing capacity. Since only 70% of the electricity entering the CAES facility reaches the consumer, the wind producer must increase his wind generating capacity by 30% to compensate for the energy loss. Thus the price of the wind generated electry entering the CAES facility must compensate the wind producer for something like a $8000 capitol investment for every average KW sold by the CAES facility. When added to the $765 per KW capital investment in the CAES facility, and the cost of natural gas used with CAES technology, we get a very ugly picture, of the cost of wind generated electricity.

In my pumped storage study, I reviewed the cost of the Northfield Mountain Pumped storage facility in New England. I calculated the cost of the 1080 MW facility at 3.7 billion 2008 dollars using the 1972 cost and a standard conversion table. I noted that an estimated 2008 cost for a reactor of similar capacity would be around $5 billion. The pumped storage facility had the ability to deliver power for 10 hours at a tome, while the reactor could be expected to deliver power continuously at least 90% of a year.

In order to produce electricity for the pump storage facility, a wind generating array would have to be built. The cost of that array would be paid for when electricity from the pump storage facility was sold. Pump storage operates at 75% efficiency. That is 25% of the energy input is lost before electrical output. Thus assuming a very generous West Texas capacity factor of .40 for the wind array with a rated output of 1400 MWs operating 24 hours a day would be required to fill the pump storage facility. Lets assume costs at the low end of the 2008 range for windmills, say $2250 per KW. Thus the wind array required to fill the pump storage facility full would cost $3.150 billion. That would give us a figure of close to $7 Billion to be financed by the sale of peak electricity from the pumped storage facility. Seven billion dollars is a l;arge investment for electricity that would be only available for 10 hours a day. Since as reactor capable of producing a similar amount of electricity 24 hours a day could be had in 2008 for 2 billion dollars less, the reactor is the better deal.

Finally, I examined battery storage with wind. Battery storage appears to be the most expensive electrical reliability/storage systems. After producing an estimated cost of Wind + battery storage, i then looked at the cost of a non-storage backup system for wind, a conventional nuclear reactor. The reactor was actually less expensive than a combination windmill battery backup system In addition the nuclear system would be so reliable that wind generation could be dispensed with and the system rely entirely on nuclear power.

Post-carbon electrical generating systems require reliability. During the last few months I have produced 4 case studies of the cost of making wind generated electricity reliable through the use of different technologies. Renewable advocates often complain that nuclear power is too expensive. My assessments show that reliable wind capital costs would more expensive than the capital costs of nuclear generated electricity in any of the noted cases. Facility costs for PY and ST power would be considerably higher per KW than wind in any of the noted cases. Given that the capacity factor for Southwestern Solar is not much higher than .20, it seems likely that reliable solar would be even more expensive than reliable wind, however, since I have not studied the economies of solar storage systems this is impossible to confirm.

My current research has not focused on other hidden cost of renewables generation. These include the cost of new transmission lines that are required by renewable generation systems, to be born by rate payers, the cost of federal and states tax based subsidies, the cost of keeping grid voltage stable. None of my case studies would support the contention that the cost of reliable wind would be competitive with conventional nuclear as a source of reliable electricity.

My conclusions have been acknowledged by some of the more sober minded supporters of the renewables paradigm. It is my contention then conventional nuclear power will cost less than reliable renewable electricity in a post carbon grid, and that National energy priorities ought to be rethought in light of the evidence that conventional nuclear power is the lower cost option. If conventional nuclear power is too expensive, then renewables are even more expensive. Thus we need to find a lower cost electrical generation option.