Comparing Costs between Dispatchable and Non-Dispatchable Electricity Generation is Pointless

Posted in Uncategorized at 6:35 am by Administrator

The development of renewable energy sources is taking place in all regions of the world and now attracts well over $200 Billion in investment annually.  The various available technologies compete for these investment dollars and are evaluated based upon a number of commonly used financial measures.  Such evaluations are made more complex due to various financial incentives including capital grants, feed-in-tariffs, production tax credits, capital depreciation schemes and other financial factors, which vary greatly between jurisdictions.

Two of the technologies that have the greatest potential to enable the transition to a sustainable energy environment invariably appear to be very expensive relative to wind and solar power and, as a result, have attracted very little investment.  These are geothermal and hydro-kinetics.

The financial measures used to make these comparisons were initially developed to evaluate the investment opportunities between various dispatchable generation technologies (hydro, coal-fired, natural gas-fired, and nuclear plants).  These measures do not provide a reasonable basis for comparing the value of dispatchable vs. non-displatchable   renewable energy sources.


Problem #1: Capital Expense (CAPEX)

The NREL 2016 Annual Technology Baseline lists CAPEX for Wind Farms that range from $1,737/kW to $2,109/kW for TRG 1 to TRG 10 resource grades respectively.

There is a very fundamental problem with references to CAPEX based upon the nameplate capacity of a wind development – a problem which many decision makers and almost all popular media reports fail to recognize.  That problem is capacity factor.

For example, a 200 MW wind farm that costs $400 Million to construct will have a reported CAPEX of $2,000/kW.  But if that wind farm has a capacity factor of 50% it will generate, on average, only 100 MW.  Therefore, the effective CAPEX is $4,000/kW.

As a result, the capacity factor for a given development is a critical factor in determining the effective CAPEX for a wind farm.  The table in the NREL report lists weighted average Capacity Factors of 51% for TRG1 to 12% for TRG 10 developments.

But what is the most likely capacity factor that can be expected from a wind farm developed in an area with excellent wind resource?  The average for the Texas fleet, as reported by ERCOT, is 36% and Texas has some of the best wind resources in the country.  Based upon that capacity factor the 200 MW wind farm described above would generate, on average, only 72 MW and the effective CAPEX would be $5,555/kW.

The situation with solar is far worse.  The maximum capacity factors for utility scale solar developments, even in Southern regions of the country, are on the order of 28%.  Solar capacity factors vary greatly based upon latitude, season, and local weather conditions.  In Germany, with the second largest solar fleet in the world, the average annual capacity factor is 11%.

Assuming a CAPEX for utility solar developments of approximately $2000/kW (from the NREL ATB) the effective CAPEX for a solar development will range from $7,000/kW to $20,000/kW depending upon the capacity factor of the particular development.

Geothermal capacity factors (again from the NREL ATB) range from 80-90% with CAPEX ranging from $4,000/kW to $6,000/kW.  Taking into account the difference in capacity factors geothermal CAPEX values are as good or better than onshore wind or utility solar.  In other words, it takes the same level of up-front investment (or less) to produce a kW of electricity from geothermal as it does from on shore wind or utility solar.

There are no utility scale hydro-kinetics developments yet deployed but based upon numerous successful pilot projects it is estimated that the CAPEX for such projects would be 8,000-10,1000/kW with very high capacity factors.

Taking into account capacity factors the CAPEX for geothermal and hydro-kinetics developments is comparable to that of wind and solar projects.


PROBLEM #2 – Levelized Cost of Electricity (LCOE)

Another commonly used comparative measure is the Levelized Cost of Electricity.  There are so many issues with LCOE calculations that it is almost not worth discussing this measure.  Unfortunately, LCOE values are used widely and indiscriminately so the problems with this measure need to be addressed.

A graph of LCOE (again taken from the NREL ATB) demonstrates the uncertainty inherent in these calculations.  The figures presented for geothermal are particularly disturbing, ranging from less than $50/MWh (EIA) to almost $400/MWh (NREL).

Without undertaking a “deep dive” into the NREL figures it is likely that the extremely high value for LCOE for geothermal presented by NREL derives primarily from financing costs.

The financing costs for generation sources that have very long service lives are difficult to evaluate and depend to a large extent on a number of assumptions.  This problem has been discussed at length with regards to hydro dam developments.

The general approach is to amortize the cost of an installation over the service life.  This is reasonable for wind and solar developments that are projected to have service lives of 20-30 years.  However, geothermal installations can operate effectively for an indefinite period of time.  The facilities in Lardarello, Italy have been in operation for almost a century.  The Geysers development in California has been generating electricity for a similar period of time.

Financing any capital cost over more than 30 years results in the vast majority of total cost being represented by interest payments as shown below:

It would be more realistic to assume that the capital cost of a project was paid off using a “30 year bond” equivalent.  This results in a step change LCOE consisting of a relatively high value while the debt for the facility is paid off after which there are only operating costs which are typically very low.  The multi-generational LCOE using this approach is low and, to some extent, unknown as it will continue to decline as long as the facility operates.  As an example, the graph below displays how LCOE changes after the capital costs have been paid off.


Problem #3 – The value of Peak Demand Availability

Almost every comparison of costs from different generation sources completely ignores the value of electricity generated at times of peak demand.  This demonstrates a fundamental lack of understanding about the mechanics of the electricity industry.

For most hours of most days there is more than enough generation capacity available in any system.  That is by design.  A theoretical display of demand and supply is shown below for a region in the Southern U.S.:

The inability to dispatch wind and solar generation during those peak demand hours greatly diminishes the value of those assets to the utility.  In fact, the utility will have to provide redundant generation assets in order to meet peak demand and, as a result, most utilities see wind and solar generation as a problem as much as an asset.

In Germany, neighbouring countries including Norway, Sweden, and France provide redundancy, whether they like it or not (they don’t).  In Texas, the rapid development of natural gas peakers has been used to provide redundancy for wind generation.

The need to provide redundant generation is not addressed in comparisons of costs between dispatchable and non-dispatchable generation sources.  Wind and solar are essentially given a “free ride” under the assumption that “some other generation source” will be available when wind and solar are not.

For wind generation it might be reasonable to factor in the cost of energy storage to cover periods of calm.  But what does the “100 year calm” look like?  Certainly it would be multiple days, probably a week or two.  The amount of energy storage required to cover that length of time without significant wind generation would be enormous and the cost to provide such energy storage would be many, many times the capital cost for the wind generation facility itself.  To get some sense of the enormity of this problem read the blog post by Euan Mearns calculating the amount of storage needed to get through a period of calm winds in the U.K. in 2015.

For solar in sub-tropical regions (south of about 35 degrees latitude) the problem is manageable because of the diurnal cycle of solar radiation and the fact that peak demand in this region typically occurs in the summer.  Facilities such as the Solana Concentrated Solar Power plant in Arizona handle this by incorporating thermal energy storage.

North of 35 degrees latitude solar cannot be backed up in any reasonable way.  The capacity factors decline precipitously in the winter months, exactly at the time when electricity demand peaks because of increased heating and lighting requirements.  This trend will only become more extreme if we de-carbonize the economy by reducing the use of natural gas for heating purposes. Even if batteries or other energy storage technologies advance by an order of magnitude there would simply not be enough solar radiation to simultaneously provide power during the day and recharge the energy system for use at night.



Cost comparisons between dispatchable and non-dispatchable electricity generating sources are simply not useful.  They are qualitatively different resources.  Such comparisons have been very detrimental to the development of geothermal energy which represents a reliable and renewable energy source.

The question should not be “is geothermal cost competitive with wind and solar?”

The question should be “how do we develop geothermal resources as cost effectively and to the largest extent possible?”

And a second question should be “how to we commercialize hydro-kinetic technology so that it can play an important role in the transition to a sustainable energy environment.”

41 total views, 1 views today

Leave a Comment

You must be logged in to post a comment.