Loading...
HomeMy WebLinkAboutAlaska Wind Power User Manual 1st Edition 1980j | Semen ee ty Alaska Power Authori LIBRARY COPY Otc cereecr Poscccevecs Ex Raa ee) ere WIN 007 pi " SaLLniows oniand ONV tt il LNAWLYvd3d lh a avis FIRST EDITION — AUGUST 1980 Report No: AK- RD- 80-2 I PROPERTY OF: taska Power Auihority 334 W. 5th Ave. ,whorage, Alaska 99501 ALASKAN WIND POWER An Introductory User's Manual Report No. AK-RD-80-2 by Dr. Tunis Wentink, Jr. Professor of Physics Geophysical Institute University of Alaska Fairbanks, Alaska prepared for The Research Section Department of Transportation and Public Facilities 2301 Peger Road Fairbanks, Alaska 99701 August 1980 First Edition GE /&6 Aw GY no. ale C.2 PREFACE AND ACKNOWLEDGEMENTS The author has long searched for an opportunity to "translate" for the Alaskan layman his extensive studies of Alaskan wind characteristics and experience with wind power applications. These studies were necessarily largely statistical and often contained mathematical detail not pertinent to the potential user of a wind machine in rural Alaska. However, the results clearly pointed out that wind power in Alaska can be a significant renewable energy source, with appreciable positive impact on the life style of many Alaskans. The welcomed opportunity to write this manual finally came, from the State of Alaska's Department of Transportation and Public Facilities. Early in the preparation of this manual it became clear that limitations of the text were necessary, and that many key facts were best presented without proof. A minimum of mathematics is a goal we chose, but found difficult to follow; energy is of necessity a field requiring numbers. In our overview, applied wind power involves three major areas. These are the determination of wind characteristics (where and how much wind), the design and development of wind machines useful in the winds (the how-to problem) and the prediction and use of the actual power such machines could produce in Alaskan wind regimes (how much available wind energy). We do not address the design and construction of wind machines since these are active and rapidly evolving technological areas researched elsewhere. Thus, the approach here is to determine how machines, supplied by others, will behave and produce energy in Alaska. If the reader obtains enough information from this manual to ask the appropriate further questions so that he can properly select a windmill from a reliable source, and if it operates at his location within his estimated (not guessed) expectations, the purpose of this manual will have been achieved. A preliminary version of the manual was issued recently as a Scientific Report by the Geophysical Institute, to DOTPF for review and to the public as advance information to solicit comment. Subsequent minor revisions and the addition of figure 13, 14, and 17 have clarified our intent, and hopefully enhanced the utility of the manual; the assistance of Matthew Rickard and Firmin Murakami of DOTPF in making these useful corrections and additions is gratefully acknowledged. The patient assistance of Sheila Finch in typing the many versions of the manuscript is also greatly appreciated. oe Milind Sr Tunis Wentink Jr. Professor of Physics Geophysical Institute FOREWORD This first edition of "Alaskan Wind Power" was prepared by Professor Tunis Wentink of the Geophysical Institute, the acknowledged expert on Alaska's wind energy potential. It is intended to be a working manual to informationally aid those individuals involved in the conception, planning, design, construction, and use of public facilities in Alaska. Yet the implications of this manual go beyond the specific interests of the Department of Transportation and Public Facilities, and anyone interested in the appropriate application of wind generated electricity in Alaska should find something of interest. , With the knowledge concerning the wind resources of the state continually expanding and the technology of wind machines rapidly developing, it is not possible to compile any user oriented manual which does not soon become obsolete. For this reason we are calling this a "first edition" with the intention implied to upgrade the information contained as needs dictate. At this time there is no fixed schedule for subsequent revisions or editions, however by writing to the address listed below your name will be placed on a mailing list and you will be kept informed of updates as they develop. In preparing the first edition we have made an effort to make it useful to a wide variety of potential users who will approach the subject of wind power with diverse points of view and varying levels of technical expertise. Your feedback concerning how well the manual fits your needs will be welcomed and should prove helpful to compilation of future editions. Research Section Cis Department of Transportation and Public Facilities 2301 Peger Road Fairbanks, Alaska 99701 eroy ¥/ ar Chief Of Energy and Buildings Research ACKNOWLEDGEMENTS AND NOTICE This report was prepared largely for Alaskan wind power users and planners. The writing of this document was sponsored by the State of Alaska, Department of Transportation and Public Facilities, Division of Planning and Programming, Research Section. It was made possible only by accumulation of considerable background technical information and computational results obtained by the Geophysical Institute under contracts from the U.S. National Science Foundation, the U.S. Energy Research and Development Administration and the U.S. Department of Energy; similar support came from State of Alaska funds appropriated to the Geophysical Institute. Wind machine characteristics obtained from companies and the U.S. National Aeronautics and Space Administration were used herein as examples of contemporary machines. However, this use constitutes neither our endorsement of these devices nor any guarantee by the manufacturers of proven performance. Thus, neither the State of Alaska, nor the United States, nor any of their agencies, nor any state or federal employees, nor any of their contractors, subcontractors or their employees, make any express or implied warranty, or assume any legal liability for the accuracy, completeness, or usefulness of the information in this document, or ~ represent that its use would not infringe privately owned rights. TABLE OF CONTENTS PREFACE I. Introduction General Remarks Early History of Windmills How Much Wind? Can WECS Use be Practical and Profitable? Organization of this Manual Measurement Units 7AMOOWDY II. Basic Concepts and Definitions A. General Remarks B. Basic Relationships C. Windmill Power Characteristics III. Height Effects A. Wind Speed Adjustment to Selected Heights B. Appropriate Heights IV. Where Are the Winds? A. Wind Speed Averages for Alaska B. Wind Data Records for Alaska C. Why No Wind Data? V. Wind Speed Variations and Measurement A. Wind Speed Variations B. Wind Measurements VI. Power in the Wind and Power Extracted A. General Relations B. Power in the Wind C. Power Output of a WECS VII. What Windmills? A General Remarks B. Types of WECS few Specific WECS D Various Approaches to WECS Power Output Estimation E Buying a Windmil] VIII.WECS Survival W A. -WECS Lifetimes” B. Towers C. WECS Killers D Other Potential Problems IX. Energy Storage or Alternatives A. Energy Fluctuations B. Storage C. No Storage; Synchronous Inverters D. System Efficiencies Xe Information Sources A. B. Technical Literature on Recent Windmills Alaskan Data XI. Economics of WECS A. B. C. References General Remarks Wind vs. 0i1 (Fixed Loan Period) Wind vs. Oi] (Payoff Time) ii Page 89 hl 93 94 95 oo 10. i. 12. 13. 14. 15. 16. 17. LIST OF FIGURES Average Qutput Power vs. Average Wind Speed for a 15 kW-rated WECS. A Medium Size WECS; 40 kW-rated Prototype. Probable Distribution of Wind Speeds for Various Average Wind Speeds. Power Characteristic of a 40 kW-rated WECS. Predicted Average Output Power from a 15 kW-rated WECS at Various Hub Heights vs. Average Wind Speed at 10 meters. Ratio of Predicted Average Power Output at 10 m and 30 m Hub Height for a 15 kW-rated WECS vs. Average Wind Speed at 10 m. Alaskan Locations Rated for Annual Wind Power at 30 m Above Ground (map). Hourly Wind Speeds at Three Heights from a Meteorological Tower at Ft. Greely, Alaska. Daily Average Wind Speeds at Three Heights from a Meteorological Tower at Ft. Greely, Alaska. Monthly Variation of Average Wind Speeds and Distribution Shape for a Southern Alaskan Island (Middleton). Monthly Variation of Average Wind Speeds and Power Flux for an Ice-prone Alaskan Coastal Site (Nome). Typical Wind Speed and Direction Sensors. Windmill Schematic Representation. Siting Considerations. Block Diagram of Ways WECS DC Power may be Used, or Converted to AC Power. Block Diagram of Synchronous Inverter Used with a WECS and AC Utility. WECS System Efficiencies iti Page 5 13 15 22 24 35 41 42 43 43 45 56 73, 74 78 84 87 3. n DO wm Oo Oo Ve LIST OF TABLES Page Summary of a 40 kW-rated WECS Power Characteristic. 7 Average Power vs. Average Wind Speed for a 40 kW-rated WECS. V7 Values of (h,/h,)? for Conversion of Measured Wind Speeds to WECS 20 Hub Heights. Values of (hy/h,)3? or Pi/P. of Winds at Various Heights. 20 Locations of Sites Treated in this Manual. 26 Annual Average Wind Speeds at Three Heights for Alaskan Sites. 31 Average Power Flux of Winds vs. Mean Wind Speeds. 49 Some Characteristics of WECS Considered in this Report. 59 Power Characteristics of WECS Considered in this Report. 60 Average Power of WECS vs. Average Wind Speeds. 62 Energy Productivity and Savings for a 40 kW-rated WECS 96 iv CHAPTER I INTRODUCTION A. GENERAL REMARKS Since 1973 the author has been involved with the study of Alaskan winds and the extraction of wind energy for useful applications. Rural electrification by wind power, at least as an important energy supplement, has always been a major objective. Our previous projects have resulted in several major reports |~10* However, for various reasons, these reports often do not serve to answer the many requests we steadily receive, by visit and by other contact methods, from Alaskans about using wind energy for their often diversified needs. Legislators often have inquired on behalf of their constituents. This document is aimed chiefly at these potential users and also those whose actions could accelerate the resurgence of wind power use in Alaska. The feasibility of making this manual sufficiently detailed to serve as a design manual for architects and engineers was carefully considered, but we concluded that such detail was not feasible. However, in the form of a user's manual it should provide some guidance to Alaskan wind system designers. More importantly, it may serve as an information link between planners, engineers and their clients. The usual questions posed by a potential windmill] user are similar to the following: 1. Can I use windmills (or wind energy conversion systems, WECS) at my location? (The question really is how much wind (speed) do I need?) ay How large a WECS do I need to meet my power requirements? 3. How much wil] WECS cost? Where do I buy one? *These are reference numbers; see the references section for the numbered full description. 8. How do I store the energy, or what do I do when the wind stops? Is it too cold for WECS at my location? Will icing be a problem? How Jong will a windmil] last? How do I measure my available winds? (Many have no quantitative knowledge of local wind speeds, or commonly overestimate their winds). Where can I learn more details about the above? As we deal with questions like these, it must be kept in mind that we are discussing Alaskan situations, where already high energy costs support the case for WECS. However, each situation has unique problems or advantages. B. EARLY HISTORY OF WINDMILLS The word windmills originally applied to devices that used wind power to drive millstones for grinding or milling grains. Modern tech- nology has coined more general terms like wind energy conversion systems (WECS); present usage applied to smal] WECS (below 100 kW capacity) yields the abbreviation SWECS. Wind turbine generator (WTG) is the term preferred by some. We will use the terms windmills, wind machines or WECS, independent of the size or shape. We briefly outline early history to make two points. One is that there is a long record of successful use of windmills, to do large amounts of work. The other is that this practical application involved excellent engineering centuries ago, passed on and pertinent in con- temporary design and use of WECS. Confirmec information on windmills and their application to solving some of man's problems seems to have started with sparse references to Persian (Iran and Afghanistan) windmills of about 950 A.D. Early design and application writings by the Dutch appeared around 1300 A.D., cl imaxed by the detailed (and feasible) plan of J. A. Leeghwater in 1641 (reprinted 17 times up to 1838). This plan was for a large drainage effort using 160 windmills. Ironically, when the ultimately successful project was started in 1848, the prime movers were coal-fired English steam engines. These engines permitted faster pumping. Thus, the slow slide to near oblivion of the windmill] began with the advent of suitable fossil-fueled machines. Modern problems with fossil fuels, especially oil, may reverse this slide. Serious windmill] engineering literature continued to appear spas- modically, with electricity-producing WECS first tried about 1890. A flood of information began around 1974, continuing unabated to the present. Some good examples of the many worthwhile and delightful recent 12 and Torrey!3. These should be books are those by Stokhuyzen, !! Spier of interest to anyone interested in windmills. Torrey's book takes the reader from the Asiatic and European histories through the surprisingly considerable past and probable future use of wind machines in America. C. HOW MUCH WIND? How large a wind speed (V) is needed to make WECS useful and eco- nomically worthwhile? In answering this question we will deal continually with the more appropriate average wind speed (v)", since the behavior of *The bar over a symbol indicates an average (or mean) value of the quantity involved. a WECS can be quickly described through the relationship between the V and the average power output (Py) of a wind machine. Figure 1 is a typical curve relating Py and V, for a medium size (15 kl!) machine. Each WECS will have such a curve, which may or may not be available from the manufacturer. Leading the reader to understanding of curves (or values) for specific machines, as displayed in Figure 1, is one major objective here. For instance, if V were 12 mph, Pay would be about 3.0 kilowatts for the machine involved in the example. Modern electricity-generating WECS usually cut-in (begin rotating) at about 7 miles per hour (mph), but the average wind speed is what counts. AV of 12 mph is often used as a criterion for useful WECS use. However, average wind speeds must be interpreted carefully. A few very quiet (wind-wise) summer months can off-set in the yearly average very windy winter months, when conditions creating the largest energy demand occur. (See the remarks on the “summer dip or minimum" in Chapter V. ) Depending on the user's needs and plans, lower V can prove to be of interest. For instance, at the village of Flat (well inland, where the Prospects for wind power are normally poor), the published V data give an annual V = 9.4 mph, not too promising. However, a small WECS operated there successfully for many years (10 or more) prior to about 1965, providing quite dependable battery charging and light for the village store. All in all, a V of 10 mph, measured at about 30 ft. or less, in a general location is desirable for further serious consideration of a WECS installation. Local prospecting may reveal windier spots. Also, as discussed in Chapter III, the V at the hub height of the blades is the real criterion. Increased tower height is often worthwhile, since the V increases with height above ground. 4 GRUMMAN 15 KW V, MPH Figure 1. AVERAGE OUTPUT POWER vs. AVERAGE WIND SPEED FOR A 15 kW-RATED WECS (Each point from measured winds for various months at some 50 Alaskan sites). D. CAN WECS BE PRACTICAL AND PROFITABLE? The potential WECS user needs to define his energy situation through some key questions: 1. Is there sufficient wind? 2. How much energy is needed? 3. What is the user willing to pay for this energy? (The economics of WECS is discussed in Chapter XI). A yes to (1) usually means a variety of WECS can be used. Most all WECS will operate in a given wind environment, some better than others. The answer to (2) determines the WECS size and cost. Then (3) becomes the critical question. It may turn out that an analysis of costs, including shipping and erection, may indicate an equivalent or lesser expenditure in dwelling insulation and improving generators already in place may be a better choice than a WECS, at least for a few-year period. However, given the apparently non-reversible trend in fossil fuel costs, WECS seem to be the way to energy efficiency in many parts of Alaska. Question (1) above is detailed in Chapter IV, dealing with "where is the wind?". In general, the answer concerning practical WECS use ranges from an unqualified "yes" for most of coastal Alaska through "maybe" in the southernmost often-shielded panhandle region (the Alexander Archipelago, the region between Cape Fairweather and Annette Island), to "no" in many interior locations. However, even in the interior there are pockets of potentially useful wind (e.g., Big Delta, Summit and some hills, like those around Fairbanks). In determining (2) above, the kind of energy (E) load needs to be specified; for instance, lights, communications devices (TV, radio or village satellite stations), motors (pumps, refrigerators, tools), appliances and heaters. The time (t) factor should also be kept in mind. While we deal here mostly with power (P), recall that energy involves power and the time power is generated (and used); i.e. P is the rate at which E is generated. Thus, in the commonly used units of kilowatts (kW) and kilowatt-hours (kWh) | E(kWh) = P(kW) x t (hrs). (1) For example, an oven requiring 2 kW (P) that is operated for 3 hours uses 6 kWh (E). Too frequently kWh and kW are confused. Most electricity user's bills are for energy, not power. Question (3) above is an involved story. Generally, wind electric power can be expected to compete best (and well) with oi], perhaps with coal if coal transport is a problem, and most definitely not with hydro- electric power already available. Of course, there is a difference, regarding (3), in trying to reduce dependence on expensive imported fuel for power plants already locally available or in not having any energy source excepting wind . E. ORGANIZATION OF THIS MANUAL A comprehensive document to cover the many aspects of wind power is far beyond the limits imposed here by funding and time. There are numerous publications that cover in more detail some of the topics discussed here. Thus, additional information sources must be covered, as in Chapters V and X. We hasten to add that this report is intended to_be a broad treatment, yet containing enough information to answer the questions posed in this chapter. A particular objective is to address the special needs of the Alaskan planner and WECS user. We deal with wind characteristics more than some others do, since the Alaskan winds and their understanding by the user ultimately deter- mine the practical utility of WECS. Windmills in Alaska have a Jong and successful, but sporadic history. The real question is will they function at a particular location? Some repetition will be noted; this is intentional, to stress the particular questions we hope to impress on the potential purchaser and user. Thus, basic concepts introduced early in Chapter II are expanded on in Chapter V, after discussion of the important questions of height effects, in Chapter III, and the general distribution of Alaskan winds, in Chapter IV. Power in the wind and how much can be extracted, presented in Chapter VI, lead to the major question of Chapter VII, what windmills? Then threats to WECS and their survival, energy storage or alternatives and some economic considerations are considered in later chapters. Only one picture of a WECS, Figure 2, is presented herein. The presumption is that the reader has seen one or more windmills or at least a few of the many windmill pictures found in the news media, films, magazines and trade publications or vendor literature. These verify the general conclusion that WECS come in a great variety of shapes and sizes, as discussed in Chapter VII. We are concerned mostly with the fundamentals common to these WECS variations. Figure 2 was selected as an example of a contemporary (1979) WECS, and of about the largest size applicable to Alaskan village use (author's opinion). Figure 2. A MEDIUM SIZE WECS (Kaman Co. 40 kW-RATED Prototype; from reference 23). F. MEASUREMENT UNITS A consistent system of units (English or metric) in a given publication is always desirable. However, herein units from both systems were used, for convenience and to make an important point. Convenience involved using some figures and tables prepared earlier for various purposes. The important point is to alert the reader that mixed units are frequently encountered in wind and WECS literature. For instance, the National Weather Service uses mostly knots for wind speed measurements and then, frequently, MPH in later tabuliations, both in the same serial publications. Also, different WECS manufacturers issue specifications using mixed units in their own brochure, or differ from units in literature of others. The current trend is to give values for each quantity in both unit systems. Movement toward the metric system has been slow. A proposed systematic set of units (metric) for wind energy devices is given in a recent Federal Register issue.24 Some appropriate conversions are: m (meter) x 3.28 = ft ft x 0.305 =m knots x 1.15] = MPH knots x 0.514 = m/second MPH x 0.869 = knots MPH x 0.447 = m/second m/s x 2.237 = MPH s = seconds 10 CHAPTER II BASIC CONCEPTS AND DEFINITIONS A. GENERAL REMARKS The characteristics of the winds, and the continual change in the windspeed need description. Also, in considering the performance of any wind machine there are several terms that need careful definition. Thus, we first spell out some important terms and their symbols (convenient abbreviations). Note that we distinguish between wind characteristics and windmill behavior in this chapter. ls Instantaneous wind speed (V). V is the speed at any given time. (In practice the V reported usually represent a one minute average reading, since even in the short term V usually fluctuates). 2. Average or mean wind speed (V). V is the regular numerical average of the many different V readings, usually for a con- siderable time period (say a day, week, or longer). The bar over a symbol means that it is an average. Thus, a daily average for four V readings of 10, 12, 8 and 2 mph is 10 + 12 + 8 + 2) = 8mph = V. 3. Cube of v (3). This means V multiplied by itself three times. Thus, for V = 8, V9 = (8)9 = 8x 8x8 = 512. In the same way the cube of an average wind speed (Vv?) is calculated. 4. A wind regime or wind distribution. This means a collection of the many V that make up the average wind speed pattern at a given location. It is usually described briefly by V and the so-called curve shape. Such a distribution gives the chance or probability of the wind blowing at a given V during some period (but, does not predict that a specific V will occur at any given time in that period). 1 Figur: 3 shows how the wind distributions change wiia increasing V. These smooth curves are the idealized summaries of the many hour-to-hour and longer period variations actually encountered in actual wind measurements. B. BASIC RELATIONSHIPS The importance of the terms defined above is in the following: 1. The instantaneous power of a wind involves v3, 2. The average power (or power per unit area) of a wind regime depends on wv, and to a lesser extent on the distribution curve shape. 3. The instantaneous power output of a WECS can be zero above and below certain wind speeds, vary with v3 in a given V range and be constant at rated power in another V range. 4, The relationship between the average power output of a WECS and the average wind speed (V) is often a linear one over an appreciable range of V. In mathematical terms (just standardized abbreviations using symbols instead of words) a more complete statement of item (1) above is p=2 ay, (2) 0 2 where P is the power in the wind with speed V, A is the area of the "Sheet" of wind involved and p is the density (weight per volume unit) of the moving air (i.e., the wind). This is the basic expression for all wind power work. In Chapter VI numerical values based on this equation are given, along with other results related to items (2), (3), and (4). Various authors, such as reference 22, derive equation (2). 12 el Probability of V, in % Figure 3. WECS RANGE a V, knots PROBABLE DISTRIBUTION OF INSTANTANEOUS WIND SPEEDS FOR VARIOUS AVERAGE WIND SPEEDS. These curves are typical. Other wind distributions which have been adopted change these shapes, but not significantly for our purpose here. References 8 and 25 treat various distributions in depth. The arrow indicate the operating V range of typical WECS. The expression indicates that both the area swept out by a wind machine and the wind speed are important. For a circle (the swept out disc) the area depends on the square of the radius (r) (the blade length) through the relationship A = mre (t = 22/7 or 3.142). Thus, doubling the radius will increase power available to the windmill four times. Also, doubling the wind speed will increase the input eight times. Both doublings achieved together would increase the power in the wind, fed to the machine, 32 times. Of course, if the V drops a factor of two the P will decrease eight times. This V-cubed effect along with the usual fluctuations in a wind regime result in wide swings in machine output power during WECS operation. C. WINDMILL (WECS OR WTG) POWER CHARACTERISTIC The curve giving the variation of power output (Pi) from a wind machine as the V varies is usually called the power characteristic. Such a characteristic, for the WECS of Fig. 2, is shown in Fig. 4. It involves several key wind speeds. These are: 1. V (cut-in) or Pe This is the V at which wind pressure on the blades overcomes the inertia and frictional forces of the machine and the blade assembly (the rotor) begins turning. In some WECS electricity production is delayed until a V slightly greater than the V of first rotation is reached. 2. V (limiting), V or V (rated), Vp lim As V increases, Pu increases rapidly until the rated power (Pp) is reached. This wind speed (Vp) is usually set by the generator capacity. At Vp the rotation enead is limited, mostly by blade feathering, so that the output is usually independent of further increases in wind V, within limits. 14 POWER, KW & HP 60 50 4o 30 20 10 Figure 4. KAMAN 40 KW WTG MAX. POWER VS. WIND SPEED ROTOR POWER WTG OUTPUT MPH 6 M/S / V FEATHER (CUT-OUT) 8 10 WIND SPEED AT HUB, MPH & M/S POWER CHARACTERISTIC OF 40 kW-RATED WECS. The output curve is for the WECS. The rotor curve is for 26 M/S the blade assembly before further machine losses are considered. 15 KAMAN AEROSP CORPORAT ACE ON 3. V (furling), Ve or V (cut-out), a At some V the generator output can no longer be controlled by feathering, so that the generator and/or the mechanical structure can be endangered. Then the machine is shut down, by turning the rotor plane 90 degrees out of the wind or by a brake arrangement. This V for shut down is called Voor (The furling or cut-out is usually automatic; re-setting often is not. An important consideration in the choice of a windmill] is whether the return from shut down to operating conditions is automatic or not! For instance, if strong winds shut down a mill at midnight, is the owner prepared to wait and then turn on the unit manually, sometimes by some strong yanking on a re-set cable?) From the example of Figure 4 one can expect the results tabulated in Table 1, for the V at the rotor hub height. Thus, this 40 kW-rated machine is capable of yielding Pu of 40 kW. In practice, as V changes, the actual Py will range from zero to 40 kW. In terms of V (not V), which is the key parameter over a period of time, the average machine output power (Pry) frequently will be well below 40 kW and will never reach the rated 40 kyl. Some typical average values that might be expected for this WECS are given in Table 2; in that table we use V up to 30 mph but a V of 15 to 20 mph is already indicative of a very windy and perhaps inhospitable location. For instance, Cold Bay and Amchitka Island, two notoriously windy Alaskan locations, have yearly V of 19 and 21 mph, respectively. The methods for determining the Puy vs V relationship for any WECS are covered in Chapters VI and VII. 16 TABLE 1 SUMMARY OF A 40 KW-RATED WECS Pu vs. V CHARACTERISTIC 10 to 20 20 to 60 Above 60 (Recall Figure 4) TABLE 2 Remarks No rotation. Rotation begins; no power taken out. Power delivery begins (about 4 kW). Power ranges from 4 to 40 kW. Power constant at 40 kW. No output (and rotor stopped, or at least turned out of the wind and rotating slowly). AVERAGE POWER (Pry) vs. AVERAGE WIND SPEED (V) FOR A 40 kW-RATED WECS (corresponding to Figure 4) V (mph) 8 10 i 12 13 15 18 20 25 Py kW)? % of rated Py° 2.9 - 6.2 oo 723.16 7.1 - 9.4 16 - 24 9.6 - 10.9 240. 27 eels 23 30 - 3] 13.6 - 14.7 34 - 37 15.8 - 19.3 40 - 48 18.4 - 24.8 46 - 62 19.7 - 27.5 49 - 69 21.5 - 32.1 54 - 80 a: The ranges depend on the fact that for a given V the shape of the wind distribution can differ. 17 See p. 40 for more on this. This page is intentionally left blank. 18 III. HEIGHT EFFECTS A. ADJUSTMENT OF WIND SPEEDS TO SELECTED HEIGHTS One of the most pertinent questions for the potential user is what is the average wind speed (V) at the installed rotor height (h;) of the intended WECS? Which WECS is selected depends largely on the power desired and, of course, the economic capacity of the user. This hy will range from about 30 to 100 ft. for the WECS we discuss here. Heights here are those above ground level, and not for above sea level (i.e., not altitude). Available data measured by an anemometer at height ha usually correspond to he of 10 to 60 ft; 33 ft is supposedly a standard anemometer height but in practice he depends on the actual measuring station. A key relationship involves the ratios of the instantaneous V and h, = p (V;/Vq) = (hg/hg)?. (3) This is the so-called height power law; p is an exponent or mathematical power (not to be confused with our P for wind or WECS power). The value of p is usually taken as 1/7 (=0.143) or sometimes 1/5 (= 0.20). Thus, we are dealing with 1/7 or 1/5 roots, similar to the 1/2 power in taking v7 are given in Table 3. square roots. Values of the quantity (hs/h,) The importance of the above equation is that since V5 depends on Ve and the h for a chosen p value, the cube of the Mia used in calculating wind power depends on h and 3p. A p of 1/7 is usually a conservative value. Example: Data from anemometer he = 30 ft, Vs = 10 mph. WECS hub h; = 70 ft. Assume p = 1/7. Then, from Table 3, (70/30)!/7 = 1.13 So, if We = 10 mph, V; = 10 (1.13) = 11.3 mph. 19 Anemometer height, he 10 20 30 40 50 60 70 100 *Any height units, as long as h; and hy units are the same. Anemometer height, he 10 20 30 40 50 100 TABLE 3 VALUES OF (hy 7h, )P FOR CONVERSION OF MEASURED WIND SPEEDS TO WECS HUB HEIGHTS* (p = 1/7) Tey, 1.06 0.96 0.93 0.91 0.89 0.84 VALUES OF (hj/n,)2? or P./P. OF WINDS AT VARIOUS 1.60 1.19 0.88 0.80 0.60 1.22 1.10 1.04 0.97 9.94 0.92 0.88 -81 -39 47 9) 68 WECS hub height, h, 1.26 1.14 1.08 97 95 o Oo fo 91 TABLE 4 1.20 1.13 1.08 1.05 0.95 1.22 J.15 1.10 ].07 0.97 WECS hub height, h; (p = 1/7) 50 = 60 1.99 2.16 1.48 1.60 1.24 1.35 1.08 1.19 1. 1.08 0.74 0.80 20 70 2.30 1.71 1.44 1.27 1.16 0.86 80 2.44 1.81 1.52 1.35 1.22 0.917 1.24 VW 1.12 1.09 0.99 h 2.56 1.91 1.60 1.42 1.29 0.96 100 1.26 1.19 1.14 1.10 1.08 1.05 100 2.68 1.48 1.35 Equation (3) and Table 3 also may be applied to a good approximation* to the ratio of the average wind speeds. Thus, one can substitute V for V in equation (3). After that substitution, cubing all terms and rearranging the expression gives 373 3 V,> = V,> (hy/n,)°P (4) Thus, proceeding as before, if a ¥. of 9 knots was measured at he = 20 ft and were applied to a WECS at 60 ft, from the tables Ve0/V20 = 1.17 and so Yeo = 10.5 kts. Also, Pi/P. = 1.60 or Pe = 1.60 Ps from Table 4. The foregoing discussion leads to some practical conclusions. First, higher towers result in greater wind speeds and usually greater WECS energy productivity. See Figure 5. But, there are several con- siderations that will determine the actual height used. At about $25 to $40/ft of tower (currently), economics may limit the selected height; the extra energy produced may not be worth the additional tower (including foundation) cost. Also, there is a possible detrimental physical effect that can actually result from too high a tower. This effect involves the fact that the wind power distribution (e.g., the distribution peak) shifts to higher V as the average V increases. Depending on the WECS, if the cut-out V is too low, the WECS might be shut down during appreciable periods of high speed winds. Figure 6 shows this diminishing effect of the WECS power output as V increases, for a specific machine. For example, if V were 10 mph at h = 33 ft (10 m), raising the hub height to 98 ft (30 m) would increase the WECS output by a factor of about 1.8 *Some authors use more complicated expressions, in which p also is said to depend on V. These expressions have been measured for only one Alaskan location, and anyway do not significantly influence our conclusions. al {0 {5 | 20 25 KTS 534 6 8 40 42 {14 m/s VAT h=10m Figure 5. PREDICTED AVERAGE POWER OUTPUT FROM A 15 kW-RATED WECS AT VARIOUS HUB HEIGHTS vs. AVERAGE WIND SPEED AT 10 m. 22 (80% increase). However, if in a very windy place V at 33 ft was 20 mph, raising the WECS to 98 ft would result in an increased WECS output of only about 25%. B. APPROPRIATE HEIGHTS All in all, unless large machines requiring high towers because of blade length are involved, tower heights of 50 to 75 ft seem best for most Alaskan locations. Blade tip clearance above ground should be at Jeast one blade length and in any case preferably more than 25 ft. This clearance is for safety reasons and to minimize interference from highly variable air currents near the ground. Additional practical canments on towers are given in Chapter XIII. 23 v2 45 kW WECS 10 m/s 5 40 45 20 25 MPH V at h=10m Figure 6. RATIO OF PREDICTED MEAN OUTPUT POWER AT 10 AND 30 m vs. MEAN SPEED AT 10m (For Prototype 15 kW WECS at 4 stations and all individual months of the year. See Table 5 for station identification). CHAPTER IV WHERE ARE THE WINDS? A. GENERAL REMARKS The main objective of this chapter is the presentation of a table giving the average wind speed (V) at 77 Alaskan locations at selected heights. The locations and V are given in Tables 5 and 6. The height above ground is important, as discussed earlier. Also, the wind data almost always were measured at heights different from those appropriate to planned WECS installations. In our table of V for various stations the measured data were adjusted to reference heights of 10 meters (33 ft), 30 m (98 ft) and 50 m (164 ft). Figure 7 is a map of the wind power potential at 30m shee ground level. Height adjustments for Table 6 were made using p = 1/5 in equation (3), contrary to p = 1/7 often recommended. Actually, p = 1/7 is often quite conservative but using either will not significantly alter the ranking in Table 6 or change a judgement of "yes" or "no" for a given site. The results in Table 6 may or may not be typical of the user's intended WECS location. Many locations involved airports or anchorages, sometimes specially selected. Topographic shielding sometimes was intentional (e.g. seaplane landings at Ketchikan and Sitka) or necessary (e.g. Dutch Harbor). None of the sites listed have been measured for optimum WECS location, and local surveys within the areas might be desirable. The sources and limitations of the data are also discussed. In Particular, some readers will not find their locations listed here but may want to search further. Thus we deal with the known records and also the reasons for no data. 25 LOCATIONS OF SITES TREATED IN THIS REPORT SITE Adak Amchitka Anchorage; Int. Apt. Anchorage; Elmendorf AFB Anchorage; Merrill Field Aniak Annette Isl. Atka Attu; Naval Air Unit Attu; Alexia Pt. Barrow Barter Isl. Bethel Bettles Big Delta Cape Lisburne Cape Newenham Cape Romanzof Cape Sarichef Cape Thompson Chuginadak Is]. Cold Bay Cordova TABLE 5 OUR SYMBOL Ad Am Anc An AI Ak Atn Ata Ba BI Be Bt BD CL CN CR cs CT Ch CB Co OUR 14,34 ] 16 27,28 23 38 15 17 49 13 20 21 22 32 45 10,36 ,37 26 ] STATION# GENERAL , Aleut. Aleut. C, SE CemSE CSE Int., Alex. Aleut. Aleut. Aleut. Ci, N Cc, N Cc, SW Int: Enters C, NW C, SW C, SW Aleut. C, NW Aleut. A Pen. C, SE LOCATION? SW COORDINATES 3 Ny W_USUALLY? 5] 5] 61 61 61 61 55 52 52 52 Zl 70 60 66 64 68 58 61 54 ~68 52 55 60 53 23 15 13 10 40 02 13 176 179 149 149 150 159 131 174 48 E173 50):E173 18 08 47 54 00 53 39 47 35 10 12 12 20 156 143 161 151 145 166 162 166 164 166 162 169 145 38 15E 48 50 01 42 35 12 10E 19E 47 5 48 31 44 08 04 02 54 00 42 50 30 Craig Driftwood Bay Dutch Harbor : Eielson AFB; See Fairbanks Elmendorf AFB; See Anchorage Fairbanks; Int. Apt Fairbanks; Ladd Field Fairbanks; Eielson AFB Fort Greely; see Big Delta Fort Yukon Galena Gambel] Golovin Gulkana Homer Indian Mt./Utopia Ck. Juneau Kaltag Kaktovik; see Barter Is]. Kenai Ketchikan King Salmon Kiska Kodiak Kotzebue Koyuk Cr 0B DH Ei Em Fb Ei BD FY Ga Gb Go Gu Ho IM Ju Ka BI Ke Kt Ki Ko Kz 27 52 43 43 46 39 33,40 24 29 Alex. Aleut. Aleut. Int. Int. Int. Int. 55 29 53 53 64 64 63 66 64 63 64 62 59 66 58 64 60 55 58 Si 57 66 64 59 53 49 51 41 59 35 44 46 33 09 38 00 22 18 34 21 40 59 44 52 133 09 166 166 147 147 147 145 145 156 171 163 145 151 153 134 158 151 131 156 177 152 162 161 5] 32 52 35 05 43 18 56 45 01 28 29 42 35 43 15 39 39 33E 31 38 06 Lonely Point McGrath Middleton Is]. Moses Pt. Nenana Nikolski Nome North East Cape Northway Nunivak Isl. Ocean Station PAPA Oliktok Platinum Point Barrow; see Barrow Point Hope Point Lay Point Spencer Port Heiden Port Moller Prudhoe Bay St. Matthew Is]. St. Paul Isl. Shemya Sitka Skagway Solomon Lo Mc MI MP Ne Ni No NC Nw Nu OSP 01 Pt Ba PL PS PH PM PB SM sP Sh Si Sk So 28 41 42 12 19 30 47 25 48 1 35 26 C,N Int. Isl., C, W Int. Aleut. C, W TS lis ints, C, SW C,N C, SE C, WW Cc, NW C,W A Pen. A Pen. C,N Isl., TSl ss Aleut. Alex. Alex. C, W W SE SE SW SW 70 62 59 64 64 52 64 63 62 60 70 59 7] 68 69 65 56 56 70 60 57 52 57 59 64 54 58 28 43 33 55 30 19 58 23 00 30 01 18 21 40 15 57 00 15 21 08 43 02 27 35 153 155 146 162 149 168 165 168 141 166 145 149 161 156 166 163 166 158 160 148 172 170 174 135 135 164 10 37 19 05 05 47 26 56 58 12 00 54 47 47 47 00 2] 39 31 20 42 16 O6E 2] 19 24 Sparrevohn Tatalina Tin City Ugashik Umiat Umnak/Cape AFB Unalakleet Valdez Wainwright Wales Wiseman Yakataga Yukutat Ta Tc uC Un Wa Wi Ya Yu 29 44 51 18 31 50 Int., SW Int., W C,W A Pen. Int., N Aleut. CW uC, SE C, NW C,W Int., N C, SE C, SE 61 62 65 57 69 53 63 61 70 65 67 60 59 06 43 34 35 24 23 54 08 36 37 26 05 31 155 155 156 157 152 167 160 146 159 150 150 142 139 34 57 55 30 08 54 47 14 54 13 i3 30 40 NOTES: ue Computer set (site) identification #. More than one # means more than one wind data set used in our analyses. Aleut. = Aleutian Island Chain Is]. = Island not part of Aleutian Chain A Pen. = Alaska Peninsula Alex. = Alexander Archipelago (extreme SE, below Lat 59°) Int., x = Interior with additional locator x C, x = Coastal, x where x = N or NW if above Lat. 66° = Wor E if between Lat. 66-62° = SW if between 62-58° (Bering Sea if coastal) = SE if between 61-59° (Gulf of Alaska if coastal) If no x, an interior site within 100 mile radius of Fairbanks. The directions are roughly those from Fairbanks. E.g., 51 53 176 38 means 51° 53'N latitude, 176° 38'W longitude; 179 15E is 179° 15'E longitude. 30 Table 6 ANNUAL AVERAGE WIND SPEED AT THREE HEIGHTS FOR 77 ALASKAN SITES (Listed in decreasing order) LOCATION NOTE OS PAPA* 1; Amchitka Is]. Cape Thompson Amchitka Is]. Tin City Shemya Is1. Cold Bay Wales Kiska Is1.* Gambel1 Nikolski Cape Sarichef St. Paul Isl. Cape Romanzof Nunivak Is]. Umnak/Cape AFB* Chuginadak Is1.* Port Heiden Pt. Spencer St. Matthew Is1.* N. E. Cape 2 2 AVERAGE ANNUAL WIND SPEED At h=10m At h = 30m At h = 50m MPH m/s MPH m/s MPH m/s 26.7. 11.9 33.3 14.9 36.8 16.5 23.9 10.7 29.8, . 13.3 33.0: 44.8 22.6 10.1 28.1 12.6 31.2¢. 13.9 20.8 9.3 25.9. 11.6 28.7 12.8 20.5 9.2 25.5 11.4 28.3. 12.7 20.4 9.1 25.4 11.4 28.2 12.6 19.0 8.5 23.7. 10.6 26,2. , Vis7 18.6 8.3 23.2 10.4 25.7 | | aS 18.3 8.2 22.8 10.2 29.2 | | Ta 17.9 8.0 22.3 10.0 24.7 Wl 17.3 7.7 21.5 9.6 23.8 10.6 16.8 735 20.9 9.3 23. 2°. F054 16.5 7.4 20.5 5,3 ant. * TOE 16.0 7ut 18.9 8.9 22.0 9.9 15.9 7.1 19.8 8.9 22.0 9.8 15.6 7.0 19.5 8.7 21.6 9.6 15.6 7.0 19.5 8.7 21.6 9.6 13.2 6.8 19.0 8.5 21:0 9.4 15:1 6.7 18.8 8.4 20.8 $.3 14.8 6.6 18.5 8.3 20.5 9.1 14.6 6.5 18.2 8.1 20.2 9.0 31 Table 6 (Cont'd) AVERAGE ANNUAL WIND SPEED At h=10m At h = 30m At h = 50m LOCATION NOTE MPH m/s MPH m/s MPH m/s Pt. Lay 14.5 6.5 18.0 8.1 20.0 8.9 Cape Lisburne 14.3 6.4 17.8 8.0 19.8 8.9 Adak Is]. 14.4 6.4 17.9 8.0 19.8 8.9 Platinum 14.4 6.4 17.9 8.0 19.8 8.9 Middleton Is]. 14.1 6.3 17.6 7.9 19.5 8.7 Barter Is]. 14.0 6.2 17.4 7.8 19.3 8.6 Oliktok 13.8 6.2 17.2 Tal 19.0 8.5 Solomon* 13.3 6.0 16.6 7.4 18.4 8.2 Kotzebue 12.9 5.8 16.1 Tat 17.8 8.0 Prudhoe Bay 12.9 5.8 16.0 7.2 17.7 TY Pt. Hope 12.8 Saif 15.9 7.1 17.7 7.9 Moses Pt. 12.7 Said 15.8 7.1 17.5 7.8 Atka Is1. 12.5 5.6 15.6 7.0 17.3 7.7 Cape Newenham ~ V.6 5.6 1.5 6.9 17.2 taf Bethel 12.4 5.5 15.4 6.9 17.1 7.6 Barrow 12.0 5.4 15.0 6.7 16.6 7.4 Wainwright 12.0 5.4 15.0 6.7 16.6 7.4 Attu (Alex. Pt. )* 11.7 5.3 14.6 6.5 16.2 7.2 Lonely Pt. 11.7 5.3 14.6 6.5 16.2 7.2 Golovin 11.4 5.1 14.2 6.4 1537 7.0 Dutch Harbor 3 11.0 4.9 13.7 6.1 15.2 6.8 King Salmon 11.1 4.9 3.8 6.2 15.3 6.8 32 Table 6 (Cont'd) AVERAGE ANNUAL WIND SPEED At h= 10m At h = 30 m At h= 50m LOCATION NOTE MPH m/s MPH m/s MPH m/s Koyuk* 10.9 4.9 13.6 6.1 152) 6.7 Annette Is]. 10.8 4.8 13.4 6.0 14.9 6.6 Unalakleet 10.7 4.8 13e3 6.0 14.8 6.6 Port Moller 10.8 4.8 13.5 6.0 14.9 6.7 Driftwood Bay 4 10.4 4.7 1320 5.8 14.4 6.4 Attu (NUU) 4 10.4 4.7 13.0 5.8 14.4 6.4 Skagway 10.3 4.6 12,8; -5.7 14.2 6.4 Nome 9.8 4.4 1252 55 13.6 6.1 Kodiak 9.9 4.5 12.4 $.5 1337 6. 1 Yakataga 9.7 4.3 W220 5.4 a2 6.0 Big Delta 9.6 4.3 11.9 5.3 13.2 S29 Geen eescniemaneen Approximate limit for WECS use Craig 4 8.2 3.7 10.2 4.5 3 5.0 Juneau 8.1 3.6 10.0 4.5 WET 5.0 Kaltag 8.1 3.6 10.0 4.5 f,1 75.0 Ft. Yukon 8.0 3.6 10.0 4.5 1,7 4.9 Ketchikan 7.8 3.5 9.8 4.4 10.8 4.8 Yakutat 7.6 3.4 9.4 4.2 10.4 4.7 Kenai 7.5 3.4 9.3 4.2 10.3 4.6 Indian Mt. 7.4 Sao) 9.2 4.1 10.2 4.5 Bettles 7.4 3.2 9.2 4.1 10.2 4.6 Umiat 7.0 ae 8.7 Sed 9:6 4.3 Gulkana 6.8 3.0 8.5 a8 9.4 4.2 Aniak 6.5 29 8.1 3.6 9.0 4.0 33 iy Ree temp oe oe cay Table 6 (Cont'd) AVERAGE ANNUAL WIND SPEED At h= 10m At h = 30m At h = 50m LOCATION NOTE MPH m/s MPH m/s MPH m/s Homer 6.6 2.9 8.2 3.7 9.1 4.1 Sparrevohn 6.4 2.9 8.0 3.6 8.8 4.0 Nenana 5.9 2.7 7.4 3.3 8.2 3.7 Sitka 3 5.9 2.7 7.4 3.3 8.2 a7 Tataline 5.9 2.6 7.3 3.3 8.1 3.6 Galena 5.8 2.6 7.3 3.3 8.0 3.6 Anchorage 5.7 29 7.1 Sie 7.8 335) McGrath 5.5 2.5 6.9 3.1 7.6 3.4 Cordova 3 5.2 2.3 6.4 2.9 7.1 3.2 Northway 4.9 2a2 6.1 257 6.7 3.0 Valdez* 4 4.8 Cae 6.0 2.7 6.7 3.0 Fairbanks 4.8 2.1 5.9 2.7 6.6 2.9 Wiseman* 4 4.4 2.0 5.4 2.4 6.0 207 Eielson AFB 3.7 1.7 4.7 221 5.2 2.3 NOTES: lis * Height unknown; 10 m assumed for measured data. 2. Ocean Station PAPA not an Alaskan site, but 6 months V (Oct.-March) compared with Amchitka for same period. 3. Severe shielding of anemometer site known. 4. Probable shielding; details unknown. 5. Approximate lower limit for WECS use if located at 50 m or less above ground; (if V = 12 mph is the criterion). 34 sé 720 116° 7 160° +f 188" 164° 160° 176° we Figure 7. ALASKAN LOCATIONS RATED FOR ANNUAL WIND POWER AT 30 m ABOVE GROUND (Good means V is 12 mph or greater). B. WIND DATA KECGRDS FOR ALASKA We treat here the major sources of Alaskan wind data; the documents cited may be of interest also for the reader seeking further information. The National Climatic Center (NCC) at Asheville, NC lists 639 entries (data descriptions, but no data) in their Index of Original Surface Weather Records, Alaska Stations (1sowr) !4, This covers data acquired from 1898 to 1975. Several entries may apply to the same Jocation served in subsequent years by different agencies (Army, Navy, Air Force, civil agencies like the National Weather Service, etc.). Also, separate entries often are made for small (up to a few miles) location shifts in the named location. Setting aside these multiple entries for a given location, one finds about 338 distinct Alaskan locations with known weather records. However, about one out of three of such measured sites did not measure (or at least record) wind data, for a variety of practical and policy reasons. Thus, IOSWR indicates that there are known wind data for some 191 Alaskan locations. Changery!® (1978)* lists in his National Wind Data Index Report the stations in IOSWR yielding wind records and gives valuable auxiliary information not included in IOSWR. Then the question of practically useful (summarized) wind data arises; the large amount of raw station readings too frequently has not been summarized, and these are stored in various archives (at NCC, or the Weather Bureau Office/Anchorage). There are 345 wind summaries listed in Index-Summarized Wind Data by Changery and others !® (1977), 16 another very useful guide. Changery and others ~ sometimes list several *Qur abbreviations use shortened but standard forms of references. Thus, Changery (1978) means a publication in 1978 by Michael J. Changery. The full title, as given in the references section, is M. J. Changery, National Wind Data Index, Final Report #HCO/T1041-01, National Climatic Center, Asheville, N.C. 28801, December 1978. 36 wind summaries for a given location; these summaries vary in date, quality and format, depending on the originating agency. For instance, some old summaries give only an annual speed average, omitting the desirable monthly breakdown. Again setting aside the multiple summaries for a site, Changery and others!® ist about 164 distinct locations having available wind summaries (as of mid-1977). All these deal with data availability but not the data. Went ink2?>© and Reed!” have published a good fraction of the summarized data for Alaska. The original weather data records for Alaska are on file at the National Weather Service in Anchorage, Alaska. Copies of these records are at NCC and most all] are also at the office of the State Climatologist. The latter is at the University of Alaska Arctic Environmental and~ | \// : Information Data Center (AEIDC), also in Anchorage. The AEIDC now is ’ the best starting point for Alaskans seeking wind data for specific locations*. C. WHY NO WIND DATA FOR A GIVEN LOCATION? We recorded "no wind speed results" for many locations, despite extensive search, for various reasons. These reasons ‘are one or more of the following: 1. No wind records exist. (a) Too few people at a location so no one took measurements. (b) Wind data were not taken because of the weather station class (e.g., some took only temperature and precipitation). This, unfortunately, applies to about | out of 3 Alaskan weather stations. *See also Chapter X for a brief description of a pending Alaska Wind Atlas, by Wise and Wentink. 37 (c) Wind data were taken for the convenience and safety of aircraft or used for local forecasts, but were not recorded and saved. This unfortunate situation still exists at many smal] Alaskan airports and for some new automatic weather stations. The wind data are still in the raw form in various archives, but have not been summarized; i.e., have not been extracted from the original forms, microfilms or data tapes so as to give V and other useful data. Such summarization is a long and expensive process ($150 to $1500 per station from NCC), depending on whether the original data needs to be transcribed or has been digitized and so is ready for computer processing. Wind summaries exist, but we did not consider them because of the time and cost involved, or the data pertained to a short period (say for only a few months). 38 CHAPTER V WIND SPEED VARIATIONS AND MEASUREMENT A. WIND SPEED VARIATIONS Variations in V and so in wind power over different time scales must be considered, despite emphasis on long term V in estimating the worth of a WECS at a given site. These variations also relate to what energy storage or other power supplies might be needed. The V will usually vary over the course of a day, due to changes in atmospheric heating by the sun; this is the so-called diurnal change. Also, the day-to-day or week-to-week variation in V depends on local weather. A fairly smooth month-to-month change gives rise to a yearly cycle, with quite reproducible effects (on average) from year to year; i.e., the V of a given month often will be similar from year to year. The yearly cycle often shows a pronounced dip during the summer at coastal Alaska locations, and sometimes a peak in interior locations. Thus, a WECS can be useful several months of the year, and not worthwhile during other months. Hence, one needs to consider monthly V. In judging wind data the 12 mph V criterion is only approximate. First, the actual minimum acceptable value depends on the WECS. Also, V measurement errors can be about 5%. Additionally, the V of a given year can fluctuate about 15% around the many-year average. For individual months the V can vary up to 50% from the long term average for that month. Thus, depending on the WECS, one should consider V of 10 mph marginal and 12 mph or more adequate. Users will be familiar with the wind trends at their locations, so we will not belabor these points. Some figures below show some typical V and V variations. Figure 8 shows the hourly data for winter winds at Ft. Greely, Alaska. Figure 9 gives the daily averages for the same location. Both show the height effect discussed earlier. 39 Figure 10 displays the yearly cycle and the pronounced summer dip in V at Middleton Island, in the Gulf of Alaska. Such a large change in V results in great changes in V? and in the average wind power (P/A). This last term is discussed in depth in the next chapter. The shape of the wind speed distribution curves (recall Fig. 3) for the same V can also vary; a flatter curve indicates a site with more variable winds than one with a pronounced peak. Such variation evidently depends on terrain and local weather; the practical influences on shape need further study. Differences in V distribution have a significant effect on the average wind power P at a site. Differences in the shape of V distribution curves are described mathematically using a "shape factor" known as k; tne effects of this on average wind power P are described mathematically with a related factor 8,25 f(k). These are discussed in detail elsewhere. Values of k at various sites range from 1.2 to 2.6. When the V distribution curve is unknown, a value of k=2 is often assumed; the corresponding value of f(k) is 1.91. 3 For Alaskan winds , however, an average of k=1.8 has been found ; the corresponding value of f(k) is then 2.14. The importance of these values on estimates of available wind power at a given site is discussed in more detail in the next chapter. There is another empirical effect we noted for some ice-prone Alaskan coastal] sites. For instance, as in Figure 11, at Nome the summer dip in V is not marked; however, there is still a major summer decrease in the average wind P/A. This effect seems to correlate with the formation and melting of the sea ice, which in turn appears to change the k of the wind distribution. Thus, in cases involving seasonal ice a user needs to be cautious in estimating summer wind power. 40 Lv FT. GREELY, AK. (WINTER winds) A 4 1 I a Hl al ! a WAN as Luh Figure 3. Time —> (Bars normally 1 hour apart) HOURLY WIND SPEEDS AT THREE HEIGHTS FROM A METEOROLOGICAL TOWER AT FT. GREELY, ALASKA (1976). ev V (DAILY) KNOTS Figure 9. DAY (O=18 NOV, 1976) DAILY AVERAGE WIND SPEEDS AT THREE HEIGHTS FROM A METEOROLOGICAL TOWER AT FT. GREELY, AK. ev V,KT MIDDLETON ISLAND V,=11.9KT i= 28FT. / NOME V,=9.6KT f= 59FT. 500 20 40a 15 , 300 a > = ~~ 2 nN 10 200 5 100 0 J oF M AM JS JS A S O N OD J oF MAM J J A S O N OD MONTH MONTH Figure 10. MONTHLY VARIATION OF V and k for Figure 11. MONTHLY VARIATION OF V, k and A SE ALASKAN ISLAND P/A FOR AN ICE-PRONE COASTAL SITE. B. WIND MEASU .-Mi.NTS Most users will need to measure their local winds. Estimates are usually uncertain. Most people over-estimate wind speeds, especially averages, at a given location. Long-term measurements, at a known height and close to the location of the intended WECS, are usually necessary. Trained and experienced observers can deduce wind conditions from tree or debris motion, stunted or deformed trees, etc., but we put these aside as being relatively scarce observations. However, Jong-time residents of an area can be invaluable in pointing out particularly _windy spots in local situations; except in flat open plain regions, local wind prospecting for optimum WECS sites is very worthwhile and almost mandatory. There is a variety of available wind speed measuring devices (usually called anemometers). We do not normally recommend the hand- held type, especially since a sustained measurement program at heights of 30 ft or more is usually needed. However, the hand-held instrument can be useful to spot check an anemometer calibration, if held close to the anemometer. Anemometers usually are either of the propeller-driven type, or the more conventional and considerably less expensive multi-cup (3 or 4) configuration. See Figure 12. The propeller, or one variety of the cup-driven type, drives a small, usually direct-current generator (thus, a mini-WECS) that feeds a calibrated voltmeter or strip-chart recorder. Alternatively, when read visually (on a meter) no auxiliary power is needed, but readings should be taken at least every three hours, ona regular basis. Another cup-driven type, the integrating (summing) or so-called "run-of-wind" anemometer, produces no voltage, but drives a switch that 44 Wind Cup Anemometer \ Ee Wind Cup Anemometer with Wind Vane Propeller-Type Wind Anemometer Figure 12. TYPICAL WIND SPEED AND DIRECTION SENSORS (From USDOE Report; reference 18) 45 then pulses a counter (odometer) powered by a battery. The difference in counter readings (i.e., miles of wind through the cups) divided by the time between readings yields the average wind speed. These summing types are well-suited to the amateur user, especially since one or two daily readings are often adequate. Users of the integrating type should monitor the battery voltage, since the mechanical counters often employed consume appreciable power, i.e., low batteries can result in too-low average wind speed conclusions. Some units having low current optical read-out instead of mechanical counters are also available. Purchased anemometers usually are accompanied by adequate installation instructions. In general, the instruments should be installed at least 10 ft. above a roof, and there only when necessary. Preferably, anemometers should be on a separate mast at least 30 ft. high and clear of trees and other possible wind obstructions. In the case of the presence of trees the anemometer should be at least 200 ft. from any trees, or, if ina forested area, more than 15 ft. above the tops. Measurements in interior Alaska showed that there is practically no measureable wind within thick tree stands, even when significant winds blew above the tree top canopy. 46 CHAPTER VI POWER IN THE WIND AND POWER EXTRACTED A. GENERAL RELATIONS Here we will expand on the ideas of Chapter II, and then discuss one very important objective, the estimation of the power produced by a WECS. As before, one must distinguish between power in the wind and the power from the WECS. A set of basic relationships is given below. Much more detail, of interest mostly to designers and students of wind characteristics, is given in reference 25. B. POWER IN THE WIND (No WECS involved here) We first add some detail to the basic equation (2) given earlier. A wind with instantaneous speed V will have an instantaneous power density or flux P/A (for a vertical or "sheet of wind" area A) of a 3 PY/A = C, vv. (5) C depends on the units of A and V, and the air density. The average power flux P/A in a wind regime (the collection of V values over a long period, say a month or year) is not what one might expect from the above equation, i.e., PA u OQ < but is pl 7 PA = |Ga f(k) V. (6) C, depends on the air density and the units of the various parts of the expression. The value of f(k) depends on the temporal speed distribution of the winds, i.e., the shape of the so-called frequency distribution 47 curve. For Alaskan winds f(k)* = 2.14 is a good value for estimation Purposes, and it is used herein. Then, for A in square meters, V in miles per hour (mph), C. is 0.05472 if Py is in watts (w) and equation ° (3) becomes P/A (w/m?) = (0.05472) (2.14) V° (mph?) or P/A = 0.1171 V. (7) So, a rough "rule-of-thumb" is that the average power flux of a wind regime is 1/8 of the cube of the average wind speed, for the units given above. Table 7 gives results using equation (7). One must be careful in applying equation (7) to engineering situations. While the mean wind power flux is useful as a guide to determining the wind power potential of a site, the PJA is appreciably greater than PA/A. This latter term is the WECS average power flux, treated further below. First at most only 59.3% of PUA can be extracted by an ideal (100% efficient) single disc WECS, and then only if the WECS operates at all V in the wind spectrum. This is a fundamental limitation of nature (like the Carnot Law for steam engines). The derivation of this limit can be found in Reference 27. In practice, the cut-in and cut-out V of a WECS make some of the wind spectrum unusable. This limited range further reduces the power that can be extracted. Furthermore, for a rotor the efficiency or power coefficient at various V is variable, ranging from zero at low V to some maximum value (0.3 to 0.4 at best) and then back to zero, for increasing V. This power coefficient depends on the ratio of the rotational speed of the blade tips to the wind V. These factors all result in the fact that a good windmill (about 80% efficient compared to 100% for an ideal mill) will extract at_most less than 50% of the average power in the wind. *Recall p. 40 and see references 8 or 25 for more on this. 48 MEAN SPEED V KNOTS OSS a SOKCOBDIIAADADAHDSPRWHWWNNM] HOO GY DDONNNADAIAE HS COCROD-—UNWDEONHADANDAWOUDANNWOF ODF NWDHROMDAHANWOW Nm i NW AVERAGE POWER FLUX OF WINDS vs. MEAN WIND SPEED TABLE 7 (Use shape factor k = 1.8 for Alaskan winds) MPH —i apt iond VSSSCGBRBIVARAARRGONN = HOSE OSErNNAIAAT OSMONMNWOMNOHDONONIDMNIOMIOMS MOMDOMNIMWIDMADHNOWOS ee ee ee NMP Nn aouw 23.0 23.5 24.0 24.5 25.0 P/A in watts/me k= 1.4 21 28 36 46 57 71 86 103 122 143 167 194 223 254 289 327 367 412 459 510 565 623 685 751 822 896 975 1059 1147 1240 1338 1441 1549 1662 1781 1905 2035 2171 2312 2460 2614 k=1. 17 23 30 38 47 58 70 84 100 118 137 159 183 209 238 268 302 338 Sui, 419 464 512 563 617 675 737 802 870 943 1019 1100 1184 1273 1366 1464 1566 1672 1784 1909 2021 2148 6 49 k =] 15 20 25 32 40 50 60 72 86 101 118 136 157 179 203 230 259 290 323 359 397 438 482 529 578 631 686 745 807 873 942 1014 1090 1170 1253 1341 1432 1527 1627 1731 1839 8 13 18 23 29 36 44 54 65 77 90 105 122 140 160 182 206 231 259 289 32] 355 392 431 473 517 564 614 666 722 780 842 907 975 1046 1121 1199 1281 1366 1455 1548 1645 2.0 k= 2.4 am 15 20 25 31 39 47 56 67 78 91 106 122 139 158 179 201 225 251 279 309 34] 375 411 449 490 533 579 627 678 732 788 847 909 974 1042 1113 1187 1264 1345 1429 C. POWER OUTPUT OF A WECS It is important to remember that the foregoing remarks dealt directly with the power (or energy content) of the wind. Now the question of the energy extraction by a WECS must be considered. In the following this is done, in terms of average power delivered at the output of the wind- driven turbine (generator). No further losses, such as those due to inverters, transmission lines, or other power-conditioning devices between the WECS and the load are considered. We define the average power produced by a specific (model, size, etc.) WECS as Pus One can define an analog to the wind flux PUA by PY/A where in the windmill] case A is the disc area swept out by the blades. There is no simple theoretical relationship between PO and Pays except that, of course, the larger the B, the larger the expected Py A general trend to be expected is that the Pu will vary proportionally : with V over an appreciable range of V; at high V the Py tends to increase less with increasing V and then flatten out. Figure 5, given earlier, shows this typical behavior. Let us review results so far by estimating the upper limit of what Py can be expected from a selected machine in a given wind environment. For convenience the 15 kW-rated WECS used to produce Fig. 5 is used. Assume V is 12 mph, at hub height of 10 m (33 ft). (This height does not enter into this calculation, but refers to the final checking of the estimate). The WECS has a 25 ft. disc (12.5 ft. radius (r), or blade jength). Then we get: (a) From equation (7) or Table 7, PA = 202 watts/me at V = 12 mph. (b) From A = nr’, swept area A = 49] ft? or 45.6 m. = 2 (1 m = 3.28 ft; 1m = 10.76 ft?) (c) P, = (202 watts/m°)(45.6m2) = 9215 watts or 9.22 kW. At best, 59.3% of 9.22 kW can be extracted by an ideal (100% efficient) WECS, thus Pu (upper limit) = 0.593 (9.22 kW) = 5.46 kW. (d) If the actual WECS has 75% of the efficiency of an ideal WECS, Py (expected) = 9.75 (5.46 kW) = 4.10 kW (at 10m height). From Fig. 5, based on a much more detailed calculation, this WECS can be expected to yield Py = 3.7 kW (at 10m height) so the above estimation Procedure is justified. The efficiency estimate is a major uncertainty, but the method is usually valid in letting a user "back into" the WECS size appropriate for his local V and average electrical power need. 5] CHAPTER VII WHAT WINDMILLS? A. GENERAL REMARKS WECS range in physical size, usually described by the tip-to-tip blade length (or the swept-out disc diameter), from a few feet to several hundred feet. Several 125 ft. WECS are being evaluated by the U.S. Department of Energy (USDOE). USDOE. has installed for testing a 200 ft. disc unit, and plans to build and test a 300 ft. disc machine. These WECS range in rated (maximum) power from 200 kW to 2500 kW, respectively. However, we do not consider these large machines further here, and confine our remarks to machines rated at most up to about 50 kW. Then the diameters range from 14 to about 65 ft. Our general approach is based on the author's opinion that large WECS (discs near 100 ft. or more) are not practical in Alaska, at least at this stage of technology. A complex of smaller units (often called a windmill "farm") is the best scheme if more power is needed. Specifically, the reasons are: 1. A complex provides back-up if one or few of the WECS fail. Also, wind power expansion by adding units is easier, mechanically and economically. 2. Large WECS shut down at wind speeds (30-40 mph) commonly prevailing for long periods in Alaska. Smaller WECS with extended speed range are more desirable. 3. Communities where WECS may be used best usually do not have the personnel and heavy equipment necessary to erect and service the large units. 4. The variable and difficult soil conditions in much of Alaska, especially during thawing, make stable foundations a serious problem. 53 5. The survivability of large machines in Alaska is questionable, until actual testing in extended very cold periods is accomplished. Blade materials and lubricants could be problems. At best, USDOE testing of large units in Alaska probably will not occur for several years, due to their program priorities. B. TYPES OF WECS There are many types of WECS, and often various models of each type, available or in the development stage. The main axis of rotation may be vertical or horizontal, and in the latter case there are up-wind or down-wind types. We put aside discussion in depth of the Darrieus, a type of vertical axis wind turbine (VAWT), despite the extensive work and advanced state of satisfactory development. This neglect (of ours) is because the Darrieus is not self starting, as the winds pick up. It also tends to stall in high winds (unless auxiliary devices are incorporated). The Darrieus seems best suited for utility grid operation, where adequate power is always available to start rotation; the induction motor at the bottom of the axis is used to start the unit and then becomes a generator synchronized to the utility line at suitable wind speeds. The USDOE Operations Office at Albuquerque, NM can supply abundant detail on the Darrieus and other kinds of VAWT. Horizontal axis WECS are called up-wind types when the wind strikes the rotor before reaching the tower. Heading into the wind is controlled by a fin on a tail boom behind the generator, as for a weather vane. The down-wind type receives wind passing around the tower, and has no tail. The WECS in Fig. 2 and large USDOE WECS mentioned above are examples of down-wind machines. Up-wind machines are braked by cocking the tail 54 boom 90 degrees out of the wind. Down-wind machines are braked by mechanical devices, or sometimes by flaps deployed at the blade tips. Brake systems must be manually operable so that the machine may be stopped for maintenance or in case of malfunction. Nearly all WECS in operation today are of the two or three bladed propeller type. A schematic view of the major components of a typical propeller type WECS is shown in Figure 13. As shown in the figure, the speed of the rotor is stepped up by a gearbox, and the higher rpm output in turn drives the generator. The rotors of some of the smallest WECS rotate fast enough that they can directly drive the generator, and no gearbox is necessary. The gearbox, generator, and ancillary equipment is mounted inside a housing, which is attached to a turntable base. This base is mounted on the tower cap, with provisions for the precise leveling of the machine. Power is carried from the (moveable) Minit to the (stationary) cables leading down the tower via a set of slip rings. Overspeed control. on a WECS of this type is performed in a number of ways. On some WECS this is done by "feathering" the blades, i.e. pivoting them on their axis to change their pitch. Feathering is usually actuated by centrifugal force, which overcomes the force of a spring at speeds above the desired limit. Overspeed control can also be accom- plished with flaps of various sorts which, when deployed, create enough drag to slow the rotor. These, too, are sommbtnes actuated by centrifugal force or, alternatively, by hydraulic or electric drive mechanisms. An anemometer and wind direction indicator may be required for the operation of automatic controls or merely for the monitoring of machine 55) CoN MW LFW Ye ROTOR BLADES 9, ROTOR HUB LOW SPEED SHAFT 10, BRAKE SYSTEM GEARBOX lL. HIGH SPEED SHAFT 12, GENERATOR B. POWER LINES 14, NOT SHOWN: OVERSPEED CONTROL SLIP RINGS CINSIDE HOUSING) LEVELING ADJUSTMENT TOWER CAP TOWER ANEMOMETER HOUSING (NACELLE) WIND DIRECTION INDICATOR TAIL VANE/ YAW CONTROL FIGURE 13. WINDMILL SCHEMATIC REPRESENTATION 56 performance. These are usually mounted on a boom attached to the tower, but may be placed on top of the windmill housing itself. C. SPECIFIC WECS We will use as examples WECS in the approximately 3 to 40 kW-rated range (14 to 64 ft. diameters). The availability of the machines and prices are changing so rapidly that a detailed list of available windmills is not meaningful. A recent compilation!? lists some 30 WECS manufacturers, with up to about 7 machines of various ratings offered by each manufacturer. See Chapter X for suggested information sources of commercial units. The intent here is to indicate the range of expected performance of typical WECS. Thus, the following units are considered: 1. Jacobs 3 kW-rated; a 1930-50 vintage well-tested and efficient WECS, often used as a standard of comparison (coded for tables as J). is Dakota Wind and Sun 4 kw-rated, a modern modification of the Jacobs (coded Da). Sis Dunlite 3 kW rated; a frequently used old but well-tested machine still in production (coded Du). 4. Windworks, 8 kW-rated; the prototype of a machine developed for USDOE. The final model released probably will be rated near 10 kW (coded W). 5. Grumman Windstream 25, 15 kW-rated; the prototype of a machine field tested and now in redesign (coded G). 6. Kaman 40 kW-rated; a prototype developed for USDOE and in test (Dec. 1979) (coded K). This model seems typical of what modern developments in blade material, blade design and control technology can offer in the wind power field. Also, the 32 ft. 57 individual blades and 75 ft. (sectional) 3-legged tower seem to be within the shipment, installation and service capabilities at most Alaskan locations. Table 8 lists the blade diameter, number of blades and key speed points of the power characteristic of these machines. Table 9 gives the preliminary power characteristics over the appropriate V range. In this report we stress that all computations and conclusions are the author's responsibility. USE HERE OF SPECIFIC MODEL MACHINE NAMES AND POWER CHARACTERISTICS DOES NOT CONSTITUTE PRODUCT ENDORSEMENT BY THE U.S. GOVERNMENTS AND STATE OF ALASKA, THE UNIVERSITY OF ALASKA OR THE AUTHOR. THE POWER CHARACTERISTICS USED MUST BE CONSIDERED ONLY APPROXIMATE AND OFTEN PRELIMINARY; OUR USE OF THESE DATA, WHILE FURNISHED BY RELIABLE SOURCES, DOES NOT IMPLY ANY MANUFACTURER RESPONSIBILITY, TEST VERIFICATIONS OR GUARANTEE OF PERFORMANCE. Brand names are used to facilitate discussion, not to rate the brands. 58 TABLE 8 SOME CHARACTERISTICS OF WECS CONSIDERED IN THIS REPORT V in mph = Diam. Rated 7 b ie ag : f (n)* - (ft) P kW oh! Ri co nists Jacobs* 14 3 T 23 60 - (3) u Dunlite 13 3 10 25 509 - (3) u Dakota 14 4 8 24 60 - (3) u Windworks 33 8 8 20 40 - (3) d (3)ed 33 10 8 22 45 165 Grumman 25 15 8 26 50 130 (3) d Kaman 64 40 10 20 69 125 (2) d a: (n) = number of blades; u = up-wind type, d = down-wind type b: ci = cut-in i c: R_ = for rated power; varies - 2 depending on literature e: co = cut-out (shut-down) f: s = survival speed, with recommended tower g: used in our calculations; some reports give 80 mph * Deduced from literature; Jacobs never published a power characteristic. 59 TABLE 9 APPROXIMATE POWER CHARACTERISTICS OF WECS* CONSIDERED IN THIS REPORT V, KTS Vv, MPH Du J Da W G K 2 ge ee ee 6.5 UES 0.25 7 8 0.40 0.30 8 9.2 0.20 0.48 0.65 0.6 8.7 10 423 9 10.4 0.30 0.60 1.05 1.0 10 1/55 0.04 0.48 0.72 1.65 es) 7.9 11 Vea 0.15 0.64 0.86 2.40 1.7. 10.9 12 13.8 0.26 0.83 1.00 3.40 2.2 14.4 13 15 0.38 1.06 1.14 4.45 3.0 18.4 14 16.1 0.52 1232 1.28 5.60 337, 23-0 15 FES 0.64 1.63 1.47 6.85 4.6 27.8 16 18.4 0.78 1.97 1.68 7.75 555i O92 7 19.6 2.37 1.90 7.95 6.6 38.8 17.4 20 8.00L 40.0L 18 20.7 1.06 2.81 2.18 7.8 18.4 hae 3.00L 19 21.9 122 2.50 9.1 20 23 1.40 ~ 2.86 10.6 21 24.2 Sees 12.0 22 2555 3.68 13.8 22.6 26 4.00L 15.0L 24 2716) 2.42 26 29.9 2.87 27 Sire 3.02 28 32e0 8.15 29 33.4 3:25 30 34.5 3.30 Sal 35.8 3.36L 34.8 40.0 4.0S 38.9 44.8 3.36S 8.0S 52.1 60 15.0S 40.08 60 69 3.0 *Actual characteristic will depend on particular machine of a given model, and also on the type and condition of the load. L: Start of power limiting S: Shut-down or furling. 60 Table 14 y.ves the expected Py vs. V for some typical maui.ines. Brand names are used to facilitate discussion, not to rate the brands. The discussion should be considered in terms of wind machines that might have rotor sizes and power characteristics resembling the examples used. This table, one of the most important of our tables, illustrates several points especially pertinent in comparing WECS (especially sales literature). These points are: 1. No WECS average power output reaches the rated power, and the former is considerably less. 2. At the same V the percentage of the rated power achieved varies considerably among different WECS. For instance, at V = 12 mph the WECS (left to right in Table 10) produce Py/Pp of 12, 28, 23, 36, 19 and 32%, respectively. This is due mostly to differences in matching blade sizes to generator sizes and to differences in overall efficiencies of the WECS. 3. Rated power alone is not necessarily a good criterion of ex- pected performance. The entire power characteristic needs to be considered, especially the V range for operating at Vp: The wind regime is also important, as expected. For example, the 8kW - 33 ft. rotor Windworks appears to be better than the 15 kW - 25 ft. rotor Grumman, for V = 13 mph or less. In higher V regimes the Grumman should be superior. 4. The cut-in speed is less important than the rate at which the WECS output increases, as V increases, to arrive at the limiting power. The Vp is an important consideration in WECS choice. The first limiting Vp is being decreased in modern designs through the use of lighter blades and improved rotor and gear box performance. Thus, the Windworks and Kaman 61 29 TABLE 10 AVERAGE POWER Pa OF SOME WECS vs. AVERAGE WIND SPEED V WECS Du? a Da° Da® wd of Kf Rated Pps kW 3 3 4 4 8 15 40 V(cut-out), mph 50 60 60 40 40 50 60 vy, Vv Pup kW mph kts 8 7.0 0.09 0.28 0.38 0.38 1.1] 0.85 4.47 10 8.7 0.21 0.55 0.64 0.64 1.98 1.71 8.52 i 9.6 0.29 0.69 0.78 0.78 2.42 2.23 10.6 12 10.4 0.37 0.84 0.93 0.92 2.84 2.78 12.7 13 11.3 0.47 0.98 1.08 1.07 3.23 3.36 14.6 15 13.0 0.68 1.24 1.38 1.34 3.9] 4.5] 18.1 18 15.6 0.99 1.56 1.78 1.65 4.64 6.11 22.4 20 17.4 1.19 1.74 2.01. 1.79 4.94 7.02 24.6 25 207 1.55 1.87 2.42 1.89 5.18 8.65 28.1 30 26.1 1.73 2522 2.59 1.81 4.96 9.39 29.2 a: Dunlite . d: Windworks b: Jacobs-like; recall footnote * of Table 8 e: Grumman c: Dakota; note different Vco f; Kaman cypes have Ve of 20 mph (design goals) while the olcer Dunlite and Grumman (before pending redesign) have Vp near 25 mph. 5. Differences in cut-off speed in the usually used range of 40 to 60 mph are significant only in higher V regimes, above about 15 mph. For instance, for the Dakota WECS the difference in Veg in the table begins to be significant in Pay above V of 15 mph. In practice there is a more pertinent aspect when any Veo? selected by designers, is reached; this aspect is the time a WECS remains shut down in the high winds. In the calculations for Table 10 there was no provision for down time after the winds dropped below Vos Some machines restart automatically, while others do not. D. VARIOUS APPROACHES TO Pu There are various ways for the user to estimate the Py for a given machine at his location. There are different degrees of accuracy, but all should give predictions well within the needs of a user. This is particularly true since one cannot predict accurately what the actual winds will be at the time of his need. Again, all we can deal with are averages. The approaches are: 1. Use_known Vand the WECS suppliers information for Py vs. V. However, some suppliers give only the P vs. V curve, not the Py vs. V data. Then you can't use this approach. The experimental problem in obtaining a measured comparison of Puy and V, or even P vs. V, is in establishing the V at the WECS blades. This proves to be a difficult task. The P vs. V curve is often computed theoretically, or 63 deduced eventually from long term measurements using several anemometers at different locations around the WECS tower. Use known Vand Equation (7) or Table 7. Estimate the average power flux in the wind, at V. Then Proceed, as in the example of page 50, to compute the power potentially extracted by the WECS. Don't forget to consider the WECS area swept out, and that only 59% of the wind power can be extracted in the ideal case. Assume a usually conser- vative WECS efficiency of 70%. Then, since (0.59) (0.70) = 0.41, Pu should be about 41% of the average power in the wind, Pas w Use detailed wind statistics and the P vs. V characteristic. (Some readers may prefer to skip this section and go to section E of this chapter). The actual wind speed distri- bution curve for a given location is combined with the P vs. V characteristic of the WECS. This combination method is described in reference 25 and Figure 1 therein. The method is relatively simple, but tedious and is best done on a computer. The result is that all the wind V values for a period (say, a month) are used to get a Pu value for the period. The many individual statistics also yield a V, so the entire computa- tion boils down to a single pair of values, Pu and V, and so one point on a graph. Figure 1 of Chapter I was computed this way. Each point came from measured winds for one of the twelve months, or the year, for about 50 different Alaskan locations. 64 4. Use synthetic wind curves and the P vs. V characteristic Often a site may have a known V, but the detailed wind sta- tistics are not available. Then a useful approach is to make up a synthetic (artificial) wind speed distribution curve based on the known V, and then proceed as in (3) above. Indeed, that is largely what reference 25 is about. E. BUYING A WINDMILL (Lightning Rod and Windmill] Salesmen) The purchase of a windmill] should involve caution and prudence, as for any expensive item. In earlier days in rural Jaren there were many humorous, but too-often justified, warnings related to the question- able ethics of the traveling lightning rod salesmen, and the quality of his products. Then, and now, the same applies to windmill salesmen. Indeed, a popular ploy in earlier days was for one salesman to inspect and "adjust" the controls of a competitive windmill already installed, followed by a replacement pitch when the machine became erratic. Today, with renewed interest in alternative energy sources, one can expect both reliable and questionable sales approaches. There are indeed reliable and ethical large and small manufacturers. However, many small companies, under-funded in a rush to capitalize on the energy crunch, may exercise poor quality control. Also, they may lack sufficient engineering experience. In particular only a few models may have been built, with inadequate testing of any unit. Misleading information and shoddy workmanship or poor materials are not uncommon. For instance, the rated (maximum) power of a unit may be stressed, while in practice at many locations only one-fifth to one-half (at most) of that power may be expected, on average. Another point relates to a lively market in used windmills. For instance, the old Jacobs units, quality machines with a long history of 65 reliability and excellent performance, are available as used items. Properly reconditioned, as many are, they are still good power sources. Unfortunately, these machines are often on the market in various states of partial and slip-shod restoration, proving to be expensive sources of frustration but of little or erratic power. Blade sets should be matched, for balance; some sets are not. All one can do is to investigate the machine source, get other customer references if possible, be careful and obtain some meaningful guarantee from the vendor. Other items to be considered are: 1. Guarantee Materials and parts may be guaranteed; but, does the purchaser's installation of the WECS instead of by the vendor invalidate the guarantee? Does the vendor supply and require a specific tower and tower cap? In the case of possible advanced payments required, will delivery be guaranteed so as not to miss the short Alaskan construction season? 2. Test History How many machines of the model have been built? Small companies may only produce a few (about 3) units before drastic changes during normal evolution of a product. Has the model in question been tested? Is the test data available? Is an adequate instruction manual available? A prospective customer can often buy such a manual before purchase of the WECS. Has the model been tested in, or at least designed to cope with, very cold climates? Lubrication can be a problem, especially in cold starts or in yawing as the wind direction shifts. Has the _ model been tested to show adequate sealing from wind-blown moisture or sand? This sealing relates to salt water Spray, Snow or dust entering both the generator and hub mechanism. 66 3. Installation Instructions Does the vendor supply adequate instructions for initial raising (and perhaps lowering) of the machine? In remote areas without heavy equipment this can be a problem for WECS rated above about 3 kW. 4. Inventory of Parts on Shipment Arrival This is a non-trivial suggestion. For instance, it is no fun to attempt to instal] a WECS in a remote area, and find a piece of aluminum irrigation pipe instead of the required steel tail boom! Also, missing smal] items, like special bolts to fix the blades to the rotor roots, with a source hundreds of miles away, can be a problem! Other practical considerations are treated in the next chapter, dealing with WECS survival. 67 This page is intentionally left blank. 68 CHAPTER VIII WECS SURVIVAL A. WECS LIFETIMES As for any mechanical item involving an appreciable investment, the expected lifetime of a WECS is an important question. Historically, many old mechanical windmills have stood for centuries. In modern times, electricity-producing units have lasted for decades. Indeed, the present commerce in WECS includes offerings of reconditioned excellent smal] units originally used in agricultural America in the 1930-1950 period. One such unit survived for many years at Admiral Byrd's South Pole station. On the other hand, some individual modern machines have lasted only a few months or days! In the following we examine some of the failure modes of tIECS, and precautions to extend wind machine lifetimes. Present trends in design goals are for expected lifetimes of 20 to 30 years. B. TOWERS The presumption here is that the tower used at least meets the individual manufacturer's mechanical load specifications for the WECS involved. Important practical considerations are provision for hoisting and perhaps lowering the WECS, working on it safely when it is in position and for routine inspection and maintenance. Good attached ladders, a safe working platform and always the use of safety belts are mandatory. A removeable ladder section at the base and/or a secure fence may be wise, to protect unauthorized people (especially children). The user should carefully plan a course of action involving the hoisting mechanism, whether the blades will be attached on the ground or at tower top, the presence of a hoisting davit and its initial removal 69 and later re-installation if needed. The working platform should be easily removed, or have provisions for lowering to provide adequate clearance of the blades. The blades can flex surprising amounts (several inches) during rotation in high speed and/or varying winds. Both self-standing three or four-legged well-trussed towers and single-legged guyed towers are usable. The author prefers the self- standing versions, resulting from long experience with towers generally and the installation of a WECS in remote Alaska. In any event the foundations (footings) must take into account the peculiar and varied soil conditions in Alaska, especially during break-up. Soil conditions are pertinent to site access, site use, and tower and guy wire footings. The large Alaska area and variety of climates do not permit generalizations, other than that a wide range of soil types and conditions exist. These should be taken into account for transportation of heavy equipment to a site, and for towers. Soils range from predominantly unconsolidated gravels (e.g., coastal spits like Pt. Hope) through clay- gravel mixtures to fine silts, sometimes involving discontinuous or continuous permafrost of often uncertain depth. Gravels may be frozen or not. Test borings should be made early for a potential windmill in- stallation, especially for planning of hole-making equipment as well as permanence of footings and guy anchors. A more detailed discussion of soils is outside the intent of this manual. Tower manufacturers (not vendors) should be consulted; they can recommend safe tower bases when the customer's WECS size and weight and soils are described. Massive concrete slabs resting on the surface may be useful, although we prefer heavy separate concrete masses, for the individual legs, buried in the ground. eateue posts frozen in permafrost is not unreasonable. 70 Guyed :ov-rs can be used. However, careful and frequent vnecking of guy tension is necessary, especially if the guy anchors can shift during weather changes, as during break-up. Any tower tilt or motion is almost certain to cause WECS blade tilts and vibration outside design limits and probable eventual failure. We know of one instance where guy Joosening during an Alaskan thaw resulted in the loss of two smal] WECS (on separate towers). Maintenance of the tower is simple, but necessary. A WECS, properly designed and installed, still sometimes vibrates considerably; so does the tower. For example, in the case of a WECS shut down, when it is turned on again (especially after the wind has shifted), the initial gyrations, mechanical surges and noise when it is heading into the new wind direction are often awesome. Routine maintenance should include periodic tightening of all tower bolts and of the tower cap. These should be checked also after any significant gale or storm. C. WECS "KILLERS" The major causes of damage to, or loss of, a properly installed WECS seem to be: 1. Gusts Modern WECS (and tower) usually are rated for mechanical survival, when properly shut-down, at wind speeds of around 125 mph. Some WECS have survived much higher equivalent wind forces. A more severe and damaging wind condition is a gust. Gusts are extreme wind speed bursts building up over a period of a few seconds, thus exceeding WECS design limits before usual automatic shut-down devices are effective. One of the most vulnerable parts of a rotor is the hub-blade (or root) attachment. Gusts may snap these roots. 71 Site choice may invite gusting problems. For instance, at Council, Alaska a WECS was located at the intersection of two valleys. While the mill was operating normally in the wind of one valley, strong transverse gusts down the other applied sudden turning forces that destroyed most of the machine (blades and tail boom). Figure 14 illustrates some guidelines for windmill siting to minimize turbulence and gusting problems. 2. Uneven de-icing or blade unbalance Any turbine can be destroyed rapidly by blade or rotor shaft un- balance. Initial installation should include careful attention to any manufacturer's instructions on blade balancing. Obviously damaged or mismatched blades or hub components should be replaced immediately (usually al] blades are replaced in sets). In the case of blade ice-up, a rotating wind machine may vibrate, but wil] usually shed the ice due to normal blade flexure. However, a shut-down unit that then ices up should have uniform thawing of all the blades before start-up; otherwise, main bearing failure is highly probable. Another possible failure mode could arise from ice or wind-packed snow accumulating on or behind the hub mechanisms. There are often devices (weighted rods, springs, etc.) that control speed by blade feathering on the hub, in addition to the blade roots. There are several WECS on the market having no hub covers but having sufficient mechanical protrusions that seem to invite ice accumulation on, and perhaps jamming of, the control devices. We know of no reported infor- - mation on this potential problem. Similarly, heavy icing of the yaw mechanism should be monitored. Otherwise, the WECS might not follow drastic wind direction shifts. This could lead to blade failure. 72 Rough, hilly terrain can produce severe turbulence near the ground. The WECS on the right is well placed; high enough to be in the high winds and avoid turbulence. The WECS in the middle is poorly sited - wind speeds will be too low, turbulence too great. The WECS on the left, on a gentle slope, may be sited adequately, particularly if prevailing winds are from left to right in the picture. When winds blow from right to left, however, this WECS will be in the wind shadow of the hill. a WECS should be sited away from obstructions such as trees or buildings, like the WECS on the right. The WECS on the left will experience turbulent winds. A rule of thumb holds that the windplant should be at least 30 feet higher than any obstructions within 100 yards. FIGURE 14. SITING CONSIDERATIONS 73 A sea cliff, like other topographical obstructions, can cause severe turbulence in its vicinity. The tower at left is not high enough to aviod this turbulence. A higher tower should be used or, like the WECS on the right, siting should be further away from the cliff. A valley may be a good site for a WECS if it is oriented with the Prevailing winds. The WECS located at the junction of two valleys, however, may experience abrupt changes in wind direction which, if severe enough, may damage the machine. FIGURE 14 (CONTINUED). SITING CONSIDERATIONS 74 (a) PossiLle blade protective layers, like teflon, often are suggested to minimize icing. Some WECS sources offer teflon coated blades, at a premium. There is no firm evidence that such fluorocarbon coatings work in de-icing. In fact, the leading manufacturer of teflon (DuPont) laments this fact, in view of the potentially large market (air-plane propellers, towers, utility lines, etc.) for effective de-icing coatings; DuPont said "practically, it just doesn't work!" The most effective protection for blades seems to be an outdoor urethane paint, or just keeping metal blades smooth and free of corrosion. 3. Control component failure Various devices designed to shut down a WECS and keep it inoperative during winds above the furling speed can fail. Then, rotation at wind speeds above the design limit can make for rapid destruction or at least bearing and generator damage. In our experience, at Ugashik a total loss seemed to be traceable to a composite failure: the WECS turned on under wind conditions when it should have remained shut down, due to a control link failure, and then the tower cap loosened because of excessive vibration. Vertical lift forces then lifted the complete unit off the tower. D. OTHER POTENTIAL PROBLEMS While normally not WECS "killers", the usual problems associated with poor quality control can result in inoperative or low-output WECS. A present problem is that sometimes "reconditioned" old WECS have not in fact been reconditioned. Let the buyer beware! No extensive compilation of WECS failures seems to exist. However, the USDOE Rocky Flats (Golden, Colorado) WECS test facility has started, 75 from tests there, to study the accumulation of failure causes for smal] machines. Poor welds, cold-soldered joints, electronic component failures and poor bearings have been observed. Improper key-way lengths in the hub (repeated in an entire limited production run of one manufacturer) and control cable breakage were also found. Blade failures due to metal fatigue have occurred within a day of installation. Elsewhere (Alaskan Peninsula, at Nelson Lagoon) the problem of wind-driven moisture through improper seals caused control and generator problems. Serious salt-water corrosion of the aluminum blades also resulted. The very desirable use of solid-state microprocessors (containing transistors or "chips") for fine control in some WECS presents a problem, certainly solvable if recognized. The microprocessors (specialized but simple mini-computers) help anticipate early the need for control action and so enhance the chances of proper operation. The solid-state devices may not function at the very low temperatures in parts of Alaska, unless auxiliary heaters, preferably powered by a separate power source, are Provided. A WECS purchaser should ask the vendor about this problem. 76 ENERGY STORAGE OR ALTERNATIVES A. ENERGY FLUCTUATIONS A WECS without auxiliary electrical equipment usually will produce variable power and voltage as the wind V fluctuates. The current can be direct current (DC) or alternating current (AC). While AC at a con- trolled frequency of 60 Hz (60 cycles/sec) is usually desired, the fre- quency of a WECS can also vary widely with V. Separately, there te the problem of maximum use of the energy produced, solvable either by storage or by diverting otherwise unused energy to some waiting load. Figure 15 sketches several possible ways for the optimum energy use. B. STORAGE 1. Batteries and Inverters Conventional* lead-acid batteries (a bank of several) may be a reasonable route for the individual (or small dwelling) WECS user. As the V and WECS output vary, the charging rate can swing widely but the battery bank voltage changes only very slowly. Thus a constant DC is available. A subsequent DC to AC inverter will yield regulated AC. The batteries and inverter costs can amount to an appreciable fraction (20 to 40%) of the WECS cost. The user should consider seriously using the DC directly, if possible, although modern appliances with DC motors are hard to find at present. This situation may change as the energy crunch continues. DC motors are more tolerant of voltage changes than are AC units. *We consider car or truck batteries conventional. These do suffer from limits on discharging too completely, say below 50%. There are better but more expensive batteries, as for electric automobiles, that permit discharge down to zero and have longer total lifetimes. 77 BZ FANS (STACK ROBBER) UNREGULATED DC OIL STORAGE TANKS WATER HEAT, TREATMENT BATTERY REFRIGERATORS REGULATED DC Ha FeTi Hy OIL, COAL SPONGE INVERTER AC SYNCHRONOUS UTILITY INVERTER REGULATED AC Figure 15. BLOCK DIAGRAM OF WAYS WECS DC POWER MAY BE USED OR CONVERTED TO AC POWER There is ur inherent and unavoidable 15% inefficiency in u%ing batteries, since more voltage is needed for charging than is available when discharging. Also, DC to AC inverters must be selected cautiously since most such devices have high (90%) efficiencies at full capacity, but low (to 20%) at light loads. Thus, one should select his inverter size (electrical rating) to match his average load; i.e., not too large, to avoid excessive loss in the inverter itself. Also, some inverters consume power even under no load and so Jong idle loads may run down the batteries. Others feature a special low-powered monitoring davice to avoid this drain. Note: the inverters discussed so far are devices that use the DC power for their operation. The synchronous inverter discussed later is a different device. A protected and warm battery area is desirable. However, battery freezing is a potential problem in Alaska; but it is often avoidable with simple precautions. One inexpensive expedient is to place the batteries in contact with bare ground, covered by a simple insulated box and in turn heaped over with a few feet of snow. The combination of the relatively warm ground (even though frozen) and the snow insulation will usually keep the battery from freezing unless the battery is over half (approximately) discharged. Indoor installations require ventilation. Frozen batteries sometimes escape damage, but should be thawed slowly before recharging; in any case low temperatures impair circulation of the electrolyte in the plates and also cause lowered output voltage. An old but excellent source of information on batteries is the book by Vina129, 79 2. Wind-Derived Hydrogen and Oxygen, and Fuel Cells The subject of hydrogen production by WECS generated electricity is included only because of numerous questions we received. It is a feasible technique for large scale future energy and chemical in- stallations. Eventually it may have strong use in cities and petrochemical manufacture. However, the costs and problems of making the hydrogen and storing it.are not compatible with the needs and resources of most of the Alaskan WECS users we address here. In recent years there have been considerable discussion and numerous proposals for storing wind-derived energy by means of electrically-decomposed water (electrolysis). In such schemes (sound in principle, costly in practice) the electrolysis of water (H,0) yields hydrogen (Ho) and oxygen (05). Both, or at least the Hos are then stored. When electricity is needed the Ho and ) (or the 05 from air) are combined in a fuel cell* (roughly, an electrolysis cell in reverse). The products are H,0 and DC electricity. Many fuel cells combined in series can produce adequate voltages; the sizes already developed range in power from watts to megawatts. We do not dwell on fuel cells, but on the most pressing problem, practical (especially economic) storage of the hydrogen. Also, we ignore safety problems in handling Ho » since Ho is no more dangerous than gasoline. *A modern technology "development", first demonstrated in 1832! But, - you still can't buy one off the shelf, and getting one specially made costs a fortune. 80 The mejor problem with storing Hy is the inherent bzsic difficulty and resulting expense in compressing it; it is completely unrealistic to consider compressed Ho at the individual or small community level. However, there are under development promising hydrogen "sponges", materials that react chemically with, or adsorb (soak up on the internal surfaces), huge amounts of Ho. Adsorbtion happens at near room temperatures, and desorption (release) occurs with increase in temperature (to about 200°F). Two important examples of Hy sponges being studied are iron- titanium alloys and zeolites (special earths). These can “soak up" almost their own weight of Ho. Relatively small volumes of material (a few cubic feet) can adsorb large volumes of gas. For instance, a cubic foot of the iron-titanium alloy will store Ho that when released and burned will yield about 250,000 BTU; the same volume of gasoline when burned produces 830,000 BTU. The gasoline, of course, is destroyed; the sponge can be recharged with more Ho. 3. Pumped Water Storage A frequent question concerns the feasibility of using WECS to replenish the reservoir feeding a hydroelectric power plant, especially during off-peak load hours (e.g., during the night). In principle it is technically feasible. Economically the feasibility is generally doubtful, and each situation needs to be considered in detail. Usually the WECS power would be better used to power other loads, like water heaters. Under ideal conditions the efficiency of the water-driven turbines is about 90%; for the pumping to return the water from the hydro-plant to the reservoir the efficiency is near 80%. Thus the round trip energy efficiency is at best 72%, independent of the auxiliary 81 electrical power conditioning needed between the WECS and the water pumps. The actual overall efficiency is not promising. However, there is an alternative route that should be examined carefully. At some locations there may be unconnected reservoirs near the height of the reservoir feeding. the penstock. Fluids pumped sideways require much less energy (to overcome friction) than when pumped upwards. As an example, say 10 kW of wind power were expended to pump a given quantity of water against a 100 ft. head; this assumed head includes frictional and lift energy re- quirements. The water goes into the hydro-plant reservoir and then falls 400 ft into the Pelton wheel. Thus the 10 kW of WECS output is effectively converted to about 29 kW at the hydroplant: 400 10 kW x 0.8 x oq x 0.9 = 29 kW. This admittedly simplified example does show that WECS power multipliers can be obtained in this way. Hence, especially at small hydropower installations, as in the panhandle region of Alaska, careful terrain surveys should be made to find unused lakes or auxiliary watersheds with suitable connection (by WECS) paths for hydroplant feeders. One obvious question is why not use the vel l=taytel, old, many- bladed, water pumping windmills directly, instead of the several step route of windmill-electricity-electrical driven pump? Direct mechanical pumping is indeed feasible and often very economical, provided that the wells are not too deep and the winds are not too strong. A principal limitation on direct mechanical pumping is that cavitation (bubbles) in the pump limits the pumping speed; the 82 direct pul-2* usually shut down at V of 25 mph. Another limitation is that the mechanical pumper must be placed where the well is, not necessarily at the best wind site. C. NO STORAGE; SYNCHRONOUS INVERTERS When a source of regulated AC is available, as from a utility or a local diesel-driven generator, a WECS can be exploited without energy storage. This approach uses a synchronous inverter (SI) which can be a key unit in a village hybrid wind-and-oil-fueled generating system. In brief, the SI is an automatic electronic switch controlled by a 60 Hz utility line so that the varying DC WECS output is converted to AC and fed to the load at a 60 Hz rate and constant AC voltage. See Figure 16. A more technical description of an SI is given later. 1. Wind Power - Load Combinations Various combinations of variable wind and user load that explain the modes of operation of an SI are given below. All the electricity to the load passes through the SI. (a) No wind and appreciable user load: AC utility supplies the full demand load. (b) Usable wind and normal user load: WECS is the main power source. AC utility supplies only a few percent of the power (that required to drive the SI). If the load exceeds the WECS output, the excess demand is automatically satisfied by the utility. (c) Usable wind and small user load: WECS power not being used feeds back, through the SI, to the utility line and so is available for a load elsewhere. Actually, the electric meter used to bill the utility's customer can be run backwards, or the utility can buy this excess energy. 83 v8 Figure 16. AC "UTILITY" SYNCHRONOUS INVERTER Awees > FLoap BLOCK DIAGRAM OF SYNCHRONOUS INVERTER USED WITH A DC WECS AND AC UTILITY (c) Utility failure: This is a problem area not infrequent, especially in rural Alaska. There are two consequent difficulties: (i) The WECS power (voltage and frequency) is no longer controlled, and this situation can damage some of the load devices. So, the WECS is shut down as in Git). (ii) Without proper precautions the WECS voltage fed back into the utility line presents a safety hazard to utility personnel working on the lines or other devices of the utility. Present rules, very variable according to the utility involved, usually require an automatic disconnect of the WECS system from the line when the utility fails. This is done by a simple fail-safe relay in the SI. Stand-Alone System The situation (i) above can be circumvented by a so- called stand alone WECS-SI system. A small AC standby generator could be switched in automatically when the utility failed, thus providing the synchronizing power for the SI. The kinds and size of switching devices, size of the standby generator and costs depend on the WECS model and power levels involved. A WECS purchaser should learn from the WECS vendor how the machine can be matched to a SI, and what the power consumed by the recommended SI will be. The answers depend on the generator in the WECS and the SI selected. Generally, with proper matching, the inverter can be 95% efficient at full 85 load rating. ‘The no-load power draw of the inverter can be about 1/2%. These values make for an energy-efficient system. 3. Technical Description of the Synchronous Inverter An SI is basically a combination (bridge circuit) of four solid-state electronic devices (gated rectifiers) called thyristers. The gating or control (on-off) voltages for the thyristers are supplied by the AC synchronizing (utility) jJine. The WECS DC power passes through the bridge to the user load at the synchronizing frequency. The voltage wave shape at the load is a good 60 Hz AC sine wave; the current wave supplied by the WECS is a series of pulses at 60 Hz with a shape depending on the relative values of the WECS power output and the required load power. Any additional current required by the load comes from the utility grid in normal fashion. Excellent detailed literature on the subject is available. The "Gemini" SI from Windworks2! are popular units. Prices of SI currently are about $200- 300/kW (f.0.b. mfgr) for units in the 4 to’ 8 kW rated range. D. SYSTEM EFFICIENCIES The overall efficiency of a WECS will depend on its configuration (see Fig. 17). Rotor efficiency may be about 45% over much of the operating range of windspeeds; generator and gearbox losses are small, around 10%. Therefore about 40% of the wind's energy may actually be converted to electricity at best. Power storage and conditioning have their own efficiencies; batteries around 80%, solid state DC/AC inverters up to 90%, synchronous inverters as much as 95%. The efficiency of all these components varies with operating conditions. The figure shows efficiencies of WECS systems under favorable conditions. 86 A: GRID CONNECTED WIND SYSTEM ROTOR BLADES SYNCHRONCUS 5 INVERTER E5= 90% ac ELECTRICITY OVERALL SYSTEM EFFICIENCY = .45 x .90 x .90 = 36% B: INDEPENDENT DC WIND SYSTEM ROTOR BLADES BATTERY E3= 80% oc ELECTRICITY OVERALL SYSTEM EFFICIENCY = .45 x .90 x .80 = 32% C:; INDEPENDENT AC WIND SYSTEM BATTERY AC ELECTRICITY INVERTER E, =85% OVERALL SYSTEM EFFICIENCY = .45 x .90 x .80 x .85 = 27% FIGURE 17. WECS SYSTEM EFFICIENCIES 87 This page is intentionally left blank. 88 CHAPTER X OTHER INFORMATION SOURCES A. RECENT TECHNICAL LITERATURE There are many government and commercial sources of information on windmill and wind power accessories. Some, with addresses given in section 4, are: 1. U.S. Department of Energy (USDOE). USDOE (Division of Distributed Solar Technology) has taken over responsibility for the Federal Wind Energy Program. In this capacity earlier results from the National Science Foundation (NSF) or the now defunct Energy Research and Development Administration (ERDA) are often distributed under USDOE direction. The Technical Information Center of USDOE (TIC) may supply documents, or assist in locating suggested references. In practice, the actual documents usually are obtainable for a reasonable fee from the U.S. Government Printing Office (USGPO) or the National Technical Information Service (NTIS). Actual titles or document numbers often need to be specified. One excellent booklet available through USGPO is Wind Machines, by F. R. Eldridge®?. It is highly recommended, especially as a well- illustrated survey of many possible windmill] types. 2. Solar Energy Research Institute (SERI SERI, a contractor to USDOE, is another information supplier. They have issued a good Wind Energy Information Directory (1979), a guide to other publication sources. We recommend this as a guide to anyone before he begins the often tedious search for wind power information. In practice SERI may refer the requestor to NTIS, but can often make recommendations of specific documents. They also provide lists of manufacturers. 89 3. WECS and Wind Neasuring Equipment Sources; trade journals Many manufacturers and vendors supply good, practical literature on WECS, anemometers, etc. Nominal charges are not unusual, because of heavy demand and smal] size of many of the companies. The Wind Power Digest (WPD), a magazine of the American Wind Energy Association (AWEA), deals mostly with the smaller machines and accessories. It: is a good source of vendor and manufacturer addresses. New machines are often described, along with the P vs. V characteristics. WPD also sells a 1980 Wind Energy Directory. The recent tabulation in Solar Age!? is also excellent. Rockwell International Corp. operates a WECS test facility (Rocky Flats) near Golden, Colorado, for USDOE. They issue test reports on various machines from time to time. 4. ADDRESSES TIC: Technical Information Center U.S. Department of Energy Oak Ridge, TN 37830 USGPO: Superintendent of Documents U.S. Government Printing Office Washington, D.C. 20492 NTIS: National Technical Information Service 5285 Port Royal Rd. Springfield, VA 22161 SERI: Solar Energy Research Institute 1617 Cole Blvd. Golden, CO 80401 WPD: Wind Power Digest (Also, 1980 Wind Energy Directory) 109 E. Lexington Elkhart, IN 46514 AWEA: American Wind Energy Association 1609 Connecticut Ave. N.W. Washington, D.C. 20009 90 B. ALASKAN |: IND DATA Elsewhere in this manual several references to Alaskan data sources have been made. The Geophysical Institute (GI) or the Arctic Environ- mental Information and Data Center (AEIDC) usually were cited. Another noteworthy pending publication (late 1980) is the USDOE - sponsored Regional Wind Energy Resource Assessment - Alaska Region. This AEIDC-GI joint effort will result in a wind energy atlas for Alaska, a considerably expanded version of Chapter IV. It will deal with the winds for some 180 sites and estimates for unmeasured locations. 91 CHAPTER XI ECONOMICS OF WECS A. GENERAL REMARKS Writing this chapter was like wrestling with an octopus! There are so many variables, even in a stable economic situation. Galloping inflation renders any firm judgements on oi] and WECS costs and loan interest rates quickly obsolete. Also, there are distinct differences if the WECS purchaser is a private individual or non-profit cooperative or a utility having other costs like paid operators and dividends to stockholders. Thus, we examine the general features of an analysis that will] let the non-profit WECS user estimate the cost of electricity generated by wind power at his location. The cost of oi] is an important variable. Indeed, the central feature of the following brief analysis, described more in reference 26, is that in many Alaskan locations the installed cost of a WECS may be paid back through oi] savings in about five years or less. This is a short time for payoff of an energy system. 1. Escalation of Fuel Oi] Prices Compounding of oi] costs with time is a compelling argument for alternate energy sources like WECS. There is a general consensus that in the foreseeable future oi] costs will rise at a rate (r, as a fraction, not %) above the general inflation rate. We use a 15% yearly increase (r=0.15), perhaps already optimistic. The formula for compound interest is used here, Fy(1+r)”. (1+r)" means the sum (l+r) multiplied n times. i is the initial amount, with growth in interest or cost as the situation calls for. 93 For instance, say oil cost ie ($/gal) during the first of a 10 year period. Then during the second year the cost per gallon will be EG (itr)!, for the third year Ea Ger)’, and for the tenth year Fo (i+r)9, Thus, oi] at $0.75/gal escalating at 15% will cost 0.75(1+0.15)? = 0.75 (3.52) = $2.64/gal ten years from now. The average cost for the period would be $1.50/gal. If r = 0.20, the comparable costs would be $3.87/gal and $1.95/gal, respectively. B. WIND vs. OIL (Fixed loan period) The energy output (E) for a WECS operating for t years in a given wind regime (V) leads to the oi] volume equivalent (OVE). The latter is the oil displaced (saved) by WECS use. Thus, since there are 8760 hrs/yr; 8760 t Pu = E. For example, if the annual Pu = 12.6 kW, for a 10 year period, E = (8760 hrs/yr)(10 yr)(12.6kW) = 1,100,000 (or 1.1 x 10°) kWh. We note that even at a cost of 6.3¢/kWh (a cheap rate now), this E is equivalent to $69,500. The oi] volume equivalent can be computed from OVE (gal) = 8760 t Py G fs (8) where G is the gal/kWh of a diesel-fueled generator we assume is competing with the WECS. The energy use factor (te) is one or less. We assume an ideal situation, involving a synchronous inverter, so that all WECS energy is used and so i = 1. If the diesel produces 11.5 kWh/gal, G = 0.0870 gal/kWh; this is an efficient diesel. Then equation (8) gives OVE = (8760 h/y)(10y)(12.6kW)(0.087g/kWh) (1) = 96027 gal of oil saved. If the average cost of oi] in the period was $1.50/gal, the oil cost saving would be $144,000. 94 The above indicates that appreciable savings are possible through WECS use. Every location and situation, of course, requires such an analysis. Table 1] illustrates a more systematic analysis for a 40 kW- rated WECS if operated in regimes where V; ranges from 10 to 20 mph. For the 10 year period and the other assumptions listed, the WECS- Produced energy costs range from about 10 to 3¢/kWh, the oi] volume savings are 63,000 to 194,000 gal and the oi] cost savings are $13,000 to $186,000. For the conditions assumed the diesel-produced energy cost would be 6.5¢/kWh during the first year ant 18.1¢/kWh in the tenth year, averaging 11.5¢/kWh over the 10 years. Thus, using WECS where the average wind speed was 12 mph, the production cost of electricity could be halved; at V = 18 mph the cost reduction is a factor of 4, Many readers may dispute the assumptions. We intentionally used low figures for the first cost of oil, the loan interest rate, the oil escalation rate and chose an efficient diesel competitor. Higher rates and poorer diesels make the WECS case better. Higher WECS costs, naturally, weaken the argument. But, costly WECS may be better than a dry, idle diesel! C. WIND vs. OIL (Payoff Time) An alternate approach to the above is to calculate a loan period equal to the time the saving in oil] costs pays off the cost of the installed WECS (including interest). This involves a relationship similar to equation (8) and the factors for paying off a loan as the outstanding balance decreases (as for paying off a house mortgage). This method is described in reference 26. 95 TABLE 1] ENERGY PRODUCTIVITY AND SAVINGS FOR A 40 kW-RATED WECS (Loan period of 10 years) V, mph (h, = 30") 8.3 10 12 15 le? V, mph (h, = 75') 10 12 14.4 18 20 Pry» kW (annual) 8.25 12.6 17.3 23 25.5 Plant factor (%)! 21 32 43 58 64 E, kWh (10°)2 0.72 1.10 1.52 2.01 2.23 E cost?, ¢/klh 9.62 6.30 4.59 3.45 3.11 0i1 saved (10° gal)* —0.63 0.96 1.32 1.75 1.94 Cost of oi] saved (10° $)2 0.83 1.26 1.73 2.31 2.56 $ Saved by WECS (10°$)2 0.13 0.57 1.04 1.61 1.86 Assumptions Loan (L) for installed WECS and synchronous inverter $50,000 Loan rate 6.5%, 10 yearly payments. First cost of oil (F_) $0.75/gal; oi] escalation rate 12%/yr. Diesel competitor; engrgy yield 11.5 kWh/gal. Synchronous inverter allows al] WECS energy to be used; fo =]. % pf rated P actually Produced, on average. 5 10° =] millon; 10° = 1 hundred thousand; e.g., 0.63 (10°) = 63,000. Over 10 years, for L and interest only; this is a minimum cost. Loan paid off costs $69,552 ($6,955/year). Pwnm—- 96 Such an analysis applied to the WECS case of Table 1] shows that the payoff time of the loan is about 10 years (for V = 10 mph), 5.5 years (15 mph) and 4.3 years (20 mph). This magnitude of payoff time, in our estimation, seems to be the most convincing argument for WECS use in Alaska. 97 This page is intentionally left blank. 98 10. REFERENCES Wentink, T., Jr., Surface wind characteristics of some Aleutian Islands, in Proceedings of First Workshop on Wind Energy Conversion Systems, Washington, D.C. (11-13 June 1973), NSF/RA/W-73-006, December 1973. Wentink, T., Jr., Wind Power Potential of Alaska. Part I: Surface Wind Data from Specific Coastal Sites, Geophysical Institute Scientific Report UAG R-225, University of Alaska, August 1973, NSF/RANN Grant GI-43098. Wentink, T., Jr., Wind Power for Alaska, The Northern Engineer, 5, 8 (Winter 1974). Wentink, T., Jr., Summary of Alaskan Wind Power and Its Possible Applications, in Proceedings of Second Workshop on Wind Energy Conversion Systems, Washington, D.C. (9-11 June 1975), NSF-RA- N-75-050 (Mitre Corp. MTR-6970), December 1975. ‘ Wentink, T., Jr., Alaskan Wind Power and Its Possible Applications, extensive Final Report to the National Science Foundation, NSF/RANN/SE/AER74-00239A01/FR/74/4, 29 Febrary 1976, available through NTIS as PB 253-339. Wentink, T., Jr., Wind Power Potential of Alaska, Part II: Wind Duration Curve Fits and Output Power Estimates for Typical Windmills, Geophysical Institute Scientific Report UAG R-240, University of Alaska, July 1976. ERDA Report RLO/2229-T12- 76/1; available through NTIS. Wentink, T., Jr., Alaskan Wind Power, in Proceedings of Third Biennial Conference on Wind Energy Conversion Systesm, Washington, D.C., (19-21 September 1977) (JBF Scientific Corp. Report CONF-77091). Wentink, T., Jr., Alaskan Wind Power Study, in Proceedings of the Conference and Workshop on Wind Energy Characteristics and Wind Energy Siting 1979, Portland, OR, June 1979 p. 243 (Battelle Pacific Northwest Laboratory report PNL-3214). Wentink, T., Jr., Wind Energy Potential of the Bristol Bay Communities, survey report for the Alaska Power Administration (USDOE), through Robert W. Retherford Associates, Anchorage, Alaska, April 1979. Contact Retherford Associates; reference Contract #35-79AP10002.0, Bristol Bay Energy and Electric Power Potential, December 1979. Wentink, T., Jr., Wind Power Potential of the Calista Region of Alaska, survey report for the Bureau of Indian Affairs, Trust Facilitation, U.S. Dept. of the Interior, Washington, D.C., ref. 9K99-9990741, June 1979. 99 11. 12. 13. 14, 15. 16. V7. 18. 19. 20. 2s 22. 23. 24. F. Stokhuyzen, The Dutch Windmill, published (in English) by C.A.J. Dishoeck, Bussum, Holland (1962). P. Spier, Of Dikes and Windmills, Doubleday and Co., Inc., Garden City, NY (1969). V. Torrey, Wind-Catchers, The Stephen Greene Press, Brattleboro, Vt. (1976). Index of Original Surface Records for Stations in Alaska, National Climatic Center (Attn: Publications), Asheville, NC 28801 ($0.50/copy). M. J. Changery, National Wind Data Index Final Report; try first NCC as in reference 14; then NTIS, report HCO/T1041-01 (September 1977). M. J. Changery, W. T. Hodge and J. V. Ramsdell, Index-Summarized ' Wind Data, NCC (as in reference 14) and Battelle Pacific Northwest Laboratories, Richland, WA 99352; report BNWL- 2220/Wind-11, (September 1977). \ J. W. Reed, Wind Power Climatology of the United States, Sandia Laboratories, Albuquerque, NM, Report SAND74-0348 (June 1975); inquire about revisions. Jd. Park and D. Schwind, Wind Power for Farms, Homes and Small Industry, Nielsen Engineering and Research, Inc., Mtn. View, ; A ene RFP-2841/1270/78/4) September 1978 (go to NTIS; 9.25). Wind Products Supplement (an insert) in Solar Age 5, February, 03108) (Inquire of Solar Age, P.O. Box 4934, Manchester, NH 03108). , G. W. Vinal, Storage Batteries, 4th edition, John Wiley and Sons, New York City, NY 1955. Gemini Synchronous Inverter brochure, from Windworks, Inc., Rt. 3, Box 44A, Mukwonago, WI 53149 (1979). F. R. Eldridge, Wind Machines, USGPO stock number 038-000-00272-4 (October, 1975). An expanded second edition is available through Van Nostrand Reinhold Co., via Litton Educational Publishing, Florence, KY (1979). M. A. Bowes and R. Perley, Analysis, etc. of 40-kW Wind Turbine Generator, p. 31 of Proceedings as in our reference 8. Proposed Rules, U.S. Department of Energy Standards for Wind Energy Devices, Federal Register, 44, #247 (December 21, 1979); USGPO or USDOE. 100 25. 26. 27. Wentink, T., Jr., Notes on Wind Speed Distribution Functions . and Alaskan Near-Surface Wind Regimes, accepted for publication in Journal of Wind Technology; printing delayed and preprint available from author at Geophysical Institute, Fairbanks, Ak. Wentink, T., Jr., WECS Economic Payoff Time, internal working paper, Geophysical Institute, Fairbanks, Ak. (revised version, 1979); available from author. Golding, E.W., The Generation of Electricity by Wind Power, E & RN Spon, Ltd., London, 1976. 101