IN THE NW ENERGY REQUEST TO ADD JUDITH GAP WINDPOWER TO ITS GENERATION MIX

March 28, 2005, Billings, Montana

            I am Russell Doty. I was lead utility division counsel to the Montana Public Service Commission for almost two years in the mid 1970s. I also was a contract administrative law judge hearing some 60 cases over a five-year period for 13 different public agencies in Minnesota. During that time I wrote several proposed orders in utility rate cases ranging from 10 to 100 pages in length. I speak here today as CEO/General Counsel of New World WindPower LLC, a company I created in December 2004. It can be found on the web at newworldwindpower.com. New World WindPower LLC will be marketing wind power in the Northwestern Energy and other Montana, Wyoming and Colorado service areas. I testify here without monetary compensation from any entity.

I have good news for the Commission and for Montana consumers. This news is summarized in four points.

SUMMARY: First, questions about what capacity is needed to back additions of wind power to a system and the costs resulting from adding wind have been addressed in detail in a definitive accumulation of all major research in the area. The authors of that compilation of research write:

The results to date also lay to rest one of the major concerns often expressed about wind power: that a wind plant would need to be backed up with an equal amount of dispatchable generation. It is now clear that, even at moderate wind penetrations, the need for additional generation to compensate for wind variations is substantially less than one-for-one and is generally small relative to the size of the wind plant. [1]

Second, the ancillary cost of providing power when the wind does not blow is much smaller than has been represented by some and will continue to decrease as we gain more experience with dispatching wind energy. Those costs range from $0.00185 per kilowatt-hour of additional cost to $0.0055/kilowatt hour depending on whether the amount of wind on the system (penetration) is 3.5% or 29%. [2] 

While there are load, internal generation capability, climate and other differences between the utility systems in the accumulated research, that experience creates at least a prima facie case for confidently asserting that similar studies on Northwestern's system would produce similar results.  Extrapolating from these studies and adding these reported ancillary costs to the base $31 per megawatt bid cost of power in the Northwest Energy system would create a cost of wind power for Northwestern consumers in the range of 3.247 cents to 3.327 cents per kilowatt hour of electricity at the proposed 8% saturation.

Third, the capacity factor of windmills is not a measure of how often the wind blows. To use it in that context introduces specious arguments into the discussion of what using increasing amounts of wind in a generation mix ultimately will cost Montana consumers.

Fourth, the costs utilities have been charging to persons who have specifically ordered electricity generated by the wind have come down significantly. For example, the effective price of wind power charged to Xcel's Colorado customers has been reduced from $0.025 cents per kilowatt-hour to $0.0076 cents per kilowatt-hour. This has helped drive the cost of wind power down for other utility customers.

            CAPACITY NEEDED TO BACK WIND: In addition to the compilation of studies on the effect of wind on a generating system mentioned in the summary above, I am including the most recent study I could find on the amount of capacity needed to back up wind. [3] I ask the Commission to take administrative notice of these studies and to find that they create a prima facie case for the proposition that the proposed addition of wind power to NW Energy system will be cost effective for Montana Consumers. This newest study comes at the behest of the Minnesota Public Service Commission, which directed the Minnesota Department of Commerce and Xcel energy to update projections based on Xcel Energy's Northern Service Area, which encompasses much of Minnesota, the Dakotas, Wisconsin and Michigan. Xcel and its predecessor were one of the first utilities in the US to have a significant amount of wind energy installed on its system.

In November I visited the Buffalo Ridge area in southwestern Minnesota and personally witnessed more than 500 utility scale windmills that span a 50+ mile area. The data for the study was taken from those windmills and used to project the effect on Xcel's system of increasing the wind production attached to it to 15% of the energy flowing over it coming from the wind. Xcel has a peak system demand of about 9000 MW, projected to become 10,000.  To put it in perspective, that is more than 4 times the capacity of the entire Colstrip complex. Xcel meets most of that peak load with internal generation and buys about 800 MW on the open market.

Without wind in the generation mix, in order to handle load variations, the Xcel system needs 60 MW of excess capacity (over system peak load). The reason it does not need more than that when its base load capacity is down for maintenance, etc. is because it can obtain the needed power from the grid or other sources it owns.

The study found the addition of 1,500 MW of wind generation to the Xcel control area increases the system regulation requirement by only 8 MW, from 60 to 68 MW to handle all of the variables that take place when the wind does not blow and the wind speed fluctuates. The term "regulation requirement" is a scientific reference to what has to be done to regulate the power on Xcel's system from second to minute so that the quality of that power remains in the 60-hertz range. So, on this short-term time scale, in the regulation timeframe (the very fast fluctuations over the range of seconds and a few minutes), wind variation looks very similar to the load variation that Xcel is already handling with its 60MW regulation reserve. That reserve only needs to be increased slightly when adding wind since adding the wind only increases the combined regulation-scale fluctuation slightly. It is emphasized that this is NOT dealing with the wind variation over the timescale of hours (load following time frame) or days (unit commitment time frame)--just the over the few minutes necessary to shift to any of the alternative fuels on the large grid and regional transmission organization serving the utility (regulation time frame). The costs of dealing with wind variation over those load following and unit commitment time frames was also addressed by the most recent Xcel study discussed below.

ANCILLARY COSTS OF ADDING WIND POWER TO AN ELECTRIC SYSTEM:  Nobody disputes the fact that wind speeds cannot be predicted with high accuracy over daily periods, and the wind often fluctuates from minute to minute and hour to hour. Consequently, electric utility system planners and operators are concerned that variations in wind plant output may increase the operating costs of the system. This concern arises because the system must maintain balance between the aggregate demand for electric power and the total power generated by all power plants feeding the system. This is a highly sophisticated task that utility operators and automatic controls perform routinely, based on well-known operating characteristics for conventional power plants, sophisticated decision-support algorithms and systems, and a great deal of experience accumulated over many years. In general, the costs associated with maintaining this balance are referred to as ancillary-services costs. [4]

            There are typically three time scales of interest in calculating total ancillary costs, which correspond to the operation of the utility system and the structure of the competitive electricity markets:

1) Unit-commitment horizon of 1 day to 1 week with 1-hour time increments (Table 1, Col. E)

2) Load-following horizons of 1 hour with 5- to 10-minute increments (intra-hour) and several hours (inter-hour) (Table 1, Col. D)

3) Regulation horizon of 1 minute to 1 hour with 1- to 5-second increments. (Table 1, Col. C)

Each of these time frames has special planning and operating requirements and costs.

            The composite table of ancillary service costs of major utilities studied to date is contained in Table 1 below. [5]   I have also added to that table from the original study, a column converting the total per MWhr cost to total per kilowatt-hour costs and adding the costs found in the most recent Xcel study to the row at the bottom of the table.

TABLE 1

			$/MWh
			Time Frame
	A	B	C	D	E	F	G
1	Study	Relative Wind Penetration (%)	Regulation	Load Following	Unit Commitment	Total in $/MWh	Total in cents/KWh
2	UWIG/Xcel	3.50	0	0.41	1.44	1.85	0.00185
3	PacifiCorp	20.00	0	2.50	3.00	5.50	0.00550
4	BPA	7.00	0.19	0.28	1.00 - 1.80	1.47 - 2.27	0.00147 - 0.00227
5	Hirst	0.06 - 0.12	0.05 - 0.30	0.70 - 2.80	na	na	na
6	We Energies I	4.00	1.12	0.09	0.69	1.90	0.00190
7	We Energies II	29.00	1.02	0.15	1.75	2.92	0.00292
8	Great River I	4.30				3.19	0.00319
9	Great River II	16.60				4.53	0.00453
10	CA RPS Phase I	4.00	0.17	na	na	na	na
11	Xcel 2004 Study	15.00	0.23		4.37	4.60	0.00460

            In a nutshell, the 2004 Xcel study noted Table 1, line 11) that even at the 15% of electricity coming from the wind currently being called for by SB 415, which is pending in the Montana House of Representatives, there is a very small regulation time frame impact ($0.23/MWh), no significant cost impact in the load following timeframe, and the largest impact ($4.37/MWh) in the unit commitment timeframe (Col E). This clearly suggests that all the attention on the near-real-time issues (imbalance penalties, etc.) is missing the real cost impacts and we should usually be looking at the next day forecasting and scheduling issue as the larger cost impact.  And even then, the cost impact is not all that large on most systems.

            The authors of rows 1-10 of Table 1 offer several other insights and generalizations about their data:

First, the incremental cost of ancillary services attributable to wind power is low at low wind penetration levels; as the wind penetration level increases, so does the cost of ancillary services. Second, the cost of ancillary services is driven by the uncertainty and variability in the wind plant output, with the greatest uncertainty in the unit-commitment time frame, or day-ahead market. Improving the accuracy of the wind forecast will result in lower cost of ancillary services. Third, at high penetration levels the cost of required reserves is significantly less when the combined variations in load and wind plant output are considered, as opposed to considering the variations in wind plant output alone. [6]

            I used the Bonneville Power Administration amounts found on line 4, Table 1 for the additional costs that an addition of 8 % wind penetration would add to NW Energy's costs because it appeared to be closest to the kind of system encountered in Montana. However, even if the only higher ancillary cost found in the 4-8% wind penetration range were used (from line 8, Table 1), the cost of wind power as proposed from the Judith Gap project would still not exceed 3.419 cents a kilowatt hour.

            It should also be noted that the authors of these studies have estimated the additional costs of wind on the high side. For example, one will note from looking online at the data from the original Xcel study found on line 2, Table 1 that the 0.00185 Total additional cost figure assume that the wind forecasting will be wrong 50% of the time.  Currently, wind forecasting is accurate 15-30% of the time in the unit commitment time frame. If one assumes the 15-30% accuracy figure in the cost calculation, it cuts the $1.44 $/MWh figure for the unit commitment time frame on line 2 at least in half.

            This reduced error rate was apparently taken into account in the second Xcel study to obtain cost data that reflected reality more closely. The study authors note, "For the study year of 2010, the cost of integrating 1500 MW of wind generation into the Xcel-NSP control area could be as high as $4.60/MWH of wind energy where the hour by-hour forecast of wind for 16 to 40 hours ahead has a mean absolute error of 15% or less."

Also, there are other things that can be done to reduce costs to consumers of increased use of wind power in the day-ahead time frame. For example, as the 2004 Xcel study authors concluded, "The MISO [Midwest Independent System Operator)] market cases demonstrate that the introduction of flexible market transactions to assist with balancing wind generation in both the day-ahead scheduling process and the day one hour ahead has a dramatic positive impact on the integration costs at the hourly level. For example, in August the hourly cost was reduced by two thirds."

            Given the consistency of these results, given the uncertainty in forecast natural gas and coal gas costs, and given the fact that Xcel spent about $500,000 on its first study, it is not necessary to incur that kind of cost here for additional information that in all probability will not affect the range of costs concerned appreciably. Wind power from Judith Gap will benefit NW Energy consumer. Therefore, I respectfully request the Commission to find that while extrapolation to Montana of the results of these studies is not perfect, it is close enough for purposes of making a prudent and reasonable rate finding allowing the Judith Gap wind energy project into the generation portfolio of NW Energy at this time.

CAPACITY FACTOR EXPLAINED: Recent reported statements about capacity factor demonstrate that some state office holders misuse that concept. Capacity factor measures what we would get from a windmill if it ran full-bore all of the time. It is the amount of energy you get out of windmill divided by what is theoretically possible given the rated design capacity of the windmill if the wind blew all of the time. 

People misusing this concept usually say a windmill has a capacity factor of from 30-38%. It actually can be 42% in excellent wind. That does not mean as Billings State Senator Jeff Essmann concluded in a recent letter to constituents, that "the 62 to 70 percent of the time wind is not providing power" it would have to be provided from more expensive natural gas.

The capacity factor of windmills is not a measure of how often the wind blows. Most wind farms produce some amount of electricity 65 to 80% of the time. They just do not produce at the rate they would if the wind were blowing as hard as it could all of the time.

For example, the engineering studies on the General Electric wind farm at Lamar in Southern Colorado indicate the farm produces electricity 88 to 90 percent of the time (because windmills go around when the wind blows in the 4 to 8 mph range as well as when they twirl in higher wind speeds. Gearboxes, tiltable blades and variable speed alternators and generators allow modern windmills to take advantage of a variety of wind speeds.

To better understand capacity factor let us consider an analogy to a hypothetical small business owning four vehicles, a small car, a van and a small and large truck. All of the vehicles are capable of traveling at speeds up to 65-75 miles per hour. The larger vehicles cost more to operate and use more gasoline. Nobody who drives any of the vehicles drives them at 65-75 miles an hour all of the time even though they have that capacity.

So it is with energy generation sources. Nobody operates them at full capacity all of the time. Coal power plants are shut down for various reasons during the year usually routine maintenance. Nuclear power plants shut down to refuel. Dams provide less electricity in the late summer and early winter because water flows are low and so on.

All the capacity factor means is that if the windmill were run at rated speed in the 20-30 miles per hour range-like a car going 65-75 miles per hour-it would produce 30 - 42 % of the electricity it is rated as being capable of producing. You engineer the windmill to be able to take advantage of the fact that the wind blows harder sometimes and softer sometimes. Likewise, you engineer cars to be able to go 90 mph sometimes and 15 miles per hour at other times.

So when do you drive a small car or use wind power? When you can save money or create less pollution by doing so. The same reason power dispatchers dispatch energy from a dam, when there is water; from a windmill, when there is wind and therefore no fuel or pollution control costs; or from a coal-fired station in the unit commitment time frame, when the wind is not forecasted to blow; or from a natural gas-fired facility only when a lower cost fuel is not available.

Those who say that natural gas power generation will have to be used when the wind does not blow ignore the fact that natural gas is being used now. [7]  A lot of the gas that is being used can be replaced by wind power. They also ignore the fact that our unit commitment time frame forecasting is good enough to allow base load coal plants to replace the energy needed much of the time when the next day is forecast to be calm. And they ignore the overwhelming data from the Western Resource Advocates and other studies that projects that without wind being 21% of the generation mix by 2020, our continued overuse of natural gas will cost consumers in the 7 state Interior Rocky Mountain West $5.3 billion a year too much each year in natural gas costs and $2 billion a year too much in electric costs. [8]

Regardless of designed capacity, you use smaller vehicles when you don't need semi-trailers to get around. You don't blow dust off your dining room table with a stick of dynamite. And you don't continue to produce power with fossil fuel and natural gas when cheaper, clean wind is available.

The 2004 Xcel study found:

While the penetration of wind generation in this study is low [up to 15%] with respect to the projected system peak load, there are many hours over the course of the year where wind generation is actually serving 20 to 30% (or more) of the system load. A combination of good plans, the right resource mix, and attractive options for dealing with errors in wind generation forecasts are important for substantially reducing cost impacts. [9]

            RECENT COST REDUCTIONS FOR WINDPOWER:

            The costs of wind power generally on a system are coming down as the market grows, technology improves, and we get more experience with what those costs actually are. Unlike natural gas costs now hovering at a level more than double what they were four years ago, those wind costs will continue to come down as we blend wind power from the less expensive wind farms like that proposed for Judith Gap with those from the older more expensive wind farms like those on the Columbia River gorge. We can help drive those costs down by simply putting more wind power online and creating a market for wind power.

This has happened in the systems of utilities around the country, some 600 of which offer wind power to customers who want it at a premium. Those premiums are not to be confused with the ancillary costs discussed above. But even the premiums are being reduced. For example, Windsource subscribers on the Xcel Energy Colorado system currently pay $2.50 per 100 kWh block. However, Windsource purchases are exempt from fuel costs and air quality rate riders, resulting in a current net price of about $1.33 per block. That is, those who buy wind do not have to pay for increased fuel costs or the costs of cleaning up coal plants. The net premium is not based on the actual costs associated with the program and the premium fluctuates with changes in the fuel cost adjustment. Based on current rate riders/exemptions the new net premium will be $0.76 per 100 kWh block. Xcel Energy is planning to change the way Windsource charges appear on customer bills. Instead of the $2.50 Windsource line item and reversal of the fuel cost and air quality charges, the bill would say something like:  net Windsource charge = $0.76 per 100 kWh block.

Prior to March of 2004, more than 30,000 Xcel Energy customers and I paid for all of part of our energy from wind generation. The results of this effort have helped to bring the cost of wind power down dramatically. It would be tragic if Montana did not take advantage of that cost reduction now, tragic and anti-consumer if we do not foster the Judith Gap wind project.