pg(s) 27-36 Karst can
cause a litany of problems for a windpower project, and it is
good practice to evaluate karst risk before proceeding with a
proposed project. Windpower projects involve widely-spaced
structures with small footprints that can cost $2 million to $5
million each. Financial viability can prove difficult, so it is
important to find useful, inexpensive procedures for evaluating
karst risk. The karst-risk-review process we have used can be
split into the two categories outlined below. Desktop studies:
Search for relevant literature Review aerial-photo and map,
and analyze lineament Search for existing well and boring
logs Survey local experts-landowners, U.S. Geological Survey,
state geological survey, cavers, etc. Field studies: Perform
site reconnaissance Conduct pit tests if bedrock is shallow
Drill-A normal geotechnical investigation includes one boring
per turbine, while karst investigations may include multiple
borings per turbine Use a downhole camera-May be useful in
evaluating extent of voids and convincing clients of risk.
Conduct geophysical studies Effectively communicating with
developers is critical. They want to know the location of the
problem sites and may ask, If there is a cave, what is the
chance that a turbine will fail? The geo-professional needs to
do the following effectively: Explain the short-term
(collapse) and long-term (settlement) risks, and mitigation
options Explain the uncertainty Negotiate liability Costs
of investigation and mitigation Get developers to determine
how much to spend while understanding how much incremental-risk
reduction they will receive The discussion of karst risk should
be ongoing and investigations may proceed on a step-by-step
basis as new information is gathered. It's important to
determine whether to investigate all sites underlain by a
potentially karstic unit or try to rank the sites based on risk
before focusing the investigation on those with potentially
higher risk. Per-turbine karst investigation costs can easily
reach $20,000 and more, so investigating each site in a
100-turbine development can be a significant commitment. When
possible, start karst evaluation early, manage available cash
with a stepwise approach, and communicate.
13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 (settlement) risks, and mitigation options Explain the uncertainty Negotiate liability Costs of investigation and mitigation reduction they will receive The discussion of karst risk should be ongoing and to investigate all sites underlain by a potentially karstic unit or try to rank the sites based on risk before focusing the turbine karst investigation costs can easily reach $20,000 possible, start karst evaluation early, manage available cash with a stepwise approach, and communicate. Introduction seismicity risk, karst risk is not addressed by the Federal may or may not be addressed by local building codes. are often chosen for wind farm sites. In these areas, there is a limited frame of reference for observing subsidence, fewer eyes observing the ground, and, normally, no reason for anyone to care about sinkholes. A sinkhole in downtown Miami gets more attention than a sinkhole in rural Texas. Karst can lead to a wind turbine tilting and even toppling. Abstract Karst can cause a litany of problems for a windpower footprints that can cost $2 million to $5 million each. can be split into the two categories outlined below. Desktop studies: Search for relevant literature lineament Search for existing well and boring logs Survey local expertslandowners, U.S. cavers, etc. Field studies: Perform site reconnaissance Conduct pit tests if bedrock is shallow DrillA normal geotechnical investigation includes one boring per turbine, while karst investigations may include multiple borings per turbine Use a downhole cameraMay be useful in evaluating extent of voids and convincing clients of risk. Conduct geophysical studies Effectively communicating with developers is critical. They want to know the location of the problem sites and may ask, If there is a cave, what is the chance that a following effectively: EVALUATING KARST RISK AT PROPOSED WINDPOWER PROJECTS William J. Bangsund Barr Engineering Co., 4700 W. 77 St., Minneapolis, Minnesota 55435 USA, BBangsund@barr.com Kenneth S. Johnson Oklahoma Geological Survey, 1321 Greenbriar Dr., Norman, Oklahoma 73072 USA, firstname.lastname@example.org 27
NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE stepping forward toward viabilitymore funding becomes available. The additional funding affects the stepwise so the early karst evaluation phases can be completed inexpensively, and the more expensive phases are done later when more funding is available. If possible, the karst professional needs to educate the geotechnical investigation and foundation design are packaged with the construction. In these cases, the issues and implications of karst may come as a surprise, at a point when there is no turning backthe turbines have typically already been purchased. Once in construction, a client has reaction, until faced with what karst evaluation can cost. 28 the turbine to be out of tolerance and lead to expensive widely spaced for optimum performance (see Figure 1), so each proposed turbine location may need to be evaluated independently for karst risk. An installed turbine can cost $2 million to $5 million, so the liability is high. Figure 1. Typical wind farm. Note widely-spaced wind turbines in a remote setting. Figure 2. Typical turbine section and major forces. Spread footings are most common. Hub heights 80100 m (but can go up to 120 m); foundation width 15-22 m; foundation embedment 2-3 m; overturning moment: 35,000 kN*m 110,000 kN*m; dead load: 1,850 kN 5,100 kN forces: the wind load, dead load, lateral load, and overturning moment. Turbines have relatively low dead loads but relatively high overturning moments. While there are several types of foundations that can be used, the most common by far is the spread footing shown on Figure 2. The discussion in this paper generally assumes and relates to the use of spread footings. Note that the ground strength rarely affects the foundation diameter. years are spent completing the development phase. relatively speculative and available funds are limited. NCKRI SYMPOSIUM 2 Figure 1. Typical wind farm. Note widely spaced wind turbines in a remote setting. forces: the wind load, dead load, lateral load, and overturning moment. Turbines have relatively low dead loads but relatively high overturning moments. While there are several types of foundations that can be used, the most common by far is the spread footing shown on Figure 2. The discussion in this paper generally assumes and relates to the use of spread footings. Note that the o verturning moment is such a significant factor that ground strength rarely affects the foundation diameter. Commercial scale typically include 10 100 turbines. Employing a common foundation or foundations to address site specific conditions, the become untenable. Figure shows the basic timeline for building a typical wind farm. Once a promising site is identified, several years are spent completing the development phase. hase it is still relatively speculative and available funds are limited. along the development process stepping forward t oward viability more funding becomes available. The additional funding affects the karst evaluation process K arst evaluation should be stepwise so the early ka rst evaluation phases can be completed inexpensively and the more expensive phases are done later when more funding is available. If possible, the karst professional needs to educate the developer and work with to use funds efficiently Note that many developers structure so that the geotechnical investigation and foundation design are packaged with the construction In these cases the issue s and implications of karst may come as a surprise at a point when there is no turning back t he t urbines have typically already been purchased. Once in construction, a client has little patience is the common reaction until faced with what karst evaluation can cost. This paper will address : the typical karst investigation methods the ways karst risk can be mitigated the issues that must be addressed in communicating with the client Figure 2. Typical turbine section and major forces. Spread footings are most common. Hub heights 80 100 m (but can go up to 120 m); foundation width 1522 m; foundation embedment 2 3 m; overturning moment: 35,000 kN*m 110,000 kN*m; dead load: 1,850 kN 5,100 kN
13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 Field studies: Site reconnaissance Pit tests Drilling (may include downhole camera and downhole mapping methods) These methods are listed, approximately, in the order of increasing cost. Because of their cost, drilling and geophysics are usually not undertaken until late in especially from geological surveys, is often extremely Literature searches are commonly used on all manner of geologic studies, and there is no need to discuss them further here. One example of something that may be karst occurrence. For example, much of southeastern Minnesota is underlain by carbonate bedrock, but in This paper will address: the typical karst investigation methods the ways karst risk can be mitigated the issues that must be addressed in communicating with the client Investigation Methods to evaluate karst risk while keeping costs under control. We have followed a commonly used program (Fischer et al. 1987; Roux, 1987; Tonkin & Taylor LTD, 2011). Not every tool is necessary or appropriate for every site: Desktop studies: Literature search Existing well and boring logs search Survey of local experts 29 Figure 3. Timeline for developing and operating a typical wind farm.
NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE and epikarst development associated with the deeper karst is commonly why lineaments are expressed on the should be conducted, where appropriate, to identify terrains have relatively thick soil covers unrelated to the bedrock that can obscure bedrock lineaments. Lineament analyses have limited or no application in these areas. There is more than one type of karst, and investigations and mitigation must be appropriate to the local whose experience in Oklahoma with evaporite karst was invaluable in evaluating evaporite karst risk at the addition to geological surveys, other geologic experts can include landowners and speleological societies. Well logs are a valuable source of information. More and online. Some examples include: method with a long history. Maps often show the locations of karstic features, especially springs and sinkholes (Figure 4). USDA Natural Resources Conservation Service soil mapping also includes sinkholes and other karst features for many areas and is available nearly staff can review aerial photographs and topographic maps for apparent karstic features. Today, much of this information is available online, but it is still important to look for historic aerial photographs so the site can be viewed from different perspectives relative to the season and time of day. Modern methods such as interferometric synthetic aperture radar and digital elevation models may be particularly valuable. 30 Figure 4. Map of proposed wind farm development area showing mapped karst features and one example of lineation of features. Figure 5. Map of a Scurry County, Texas Wind farm project area showing mapped lineaments. Labeled dots are proposed turbine locations.
13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 anomalies must then be determined through drilling. Risk characterization has a number of questions: formation? Are there any known karst features in the region? Are there karst features at the proposed turbine sites? The results at each stage of evaluation are used to determine if more investigation is required and, if so, the scope of the next phase. shallow carbonate or evaporite bedrock and no evidence of karst from the desktop phase or reconnaissance. The lack of evidence does not mean there is no risk. The question then is, how much investigation is required? Lineament analysis has been used to identify areas with higher potential risk. Then, intense investigation can be completed in these areas. If no subsurface voids are found, it may be acceptable to forego further karst investigation in other areas. Risk mitigation mitigation must be applied. More than one method of are several ways of mitigating karst risk: Move the turbines at risk. It may be possible learned to include alternative locations early in the process for this type of outcome. Depending on the number of sites that are eliminated and the number of alternate sites, the cost may range from practically nothing to the loss of the investment and revenue related to the net lost sites. Conduct detailed investigation. Some developments may have very limited constraints risk alternative locations may not be available. A developer can then decide to do more intensive uncooperative because landowners may be concerned about the effect of karst on their land value, and cavers are often reluctant to share private mapping with outsiders Site reconnaissance is important for the general characterization of the area. It may also identify karst features near or at individual turbine sites. Classic are important because so much cost and risk can be to quarries is especially valuable even if outside the Where bedrock is shallow, test pits can be useful in evaluating the bedrock surface and investigating the nature of depressions to determine whether or not they are related to karst formation. equal to the width of the turbine foundation, and the depth is chosen based on the vertical stress induced by the foundation (Das, 2010). Karst investigations may include multiple borings per turbine. The question is, how many are required to assess karst risk? Advanced geotechnical modeling can provide an indication of the size of void verses depth that may be problematic. vary across the proposed wind farm, requiring multiple models. The cost of drilling multiple borings per turbine quickly increases the cost of investigation. This can be especially useful in convincing the client that there is a risk. Although not used by these authors, laser The use of geophysics in karst evaluations is well studied and reported, and it is regularly addressed at karst conferences (Beck and Wilson, 1987; Beck and tools, but it rarely attains a useful depth of penetration; below grade. In fact, most geophysical methods lack the 31
NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE Dont build the project. Developers typically best approach may be to move on to the next one. This means losing the investment to that point, so this is not done lightly. There is often great pressure to move forward despite the evidence of karst. As noted previously, the earlier that karst risk can Risk communication The cost of failure of a single turbine can range from hundreds of thousands of dollars (slight but unacceptable differential settlement) to millions of dollars (extreme tilt to catastrophic collapse). It is therefore important to communicate the cost implications to the client as risk of karst is the apportionment of risk amongst the Karst risk and risk apportionment is a very important conversation. follow through with the level of investigation needed to completely characterize the risk or carry all the liability per turbine, which does not offset the potential for a lost $5 million turbineespecially when that risk is multiplied by tens or hundreds of turbines. Therefore, it is important to educate the client about karst and karst risk to the extent that the client can carry the bulk of the risk and can make informed decisions regarding the degree of risk and how extensive the risk characterization will be. Effectively communicating with developers is critical. They want to know the exact location of the problem sites and may ask, if there is a cave, what is the chance that a turbine will fail? The developers cases, the issue is cover collapse or soil piping, not cave collapse. It is also important to communicate the inherent uncertainty of karst risk and the cost of reducing the uncertainty. investigation of a proposed turbine location to see if moving the turbine a short distance can reduce risk. This method of mitigation can add tens of thousands of dollars and may not be successful. Provide thick soil cover to mitigate the risk of subsidence In some areas, thick soil unrelated may provide an effective bridge over bedrock karst features, and soil thickness may be preliminarily determined based on existing mapping and drilling logs. Eventually, each proposed turbine site should be drilled to determine actual soil thickness. needs to be answered. There may be precedents. expressions in Minnesota occur where there is less than 15 m of glacial cover. The Minnesota of karst features was not found for areas with similar for terrace and dune deposits (Johnson developer need to come to their own conclusion. Since a typical geotechnical investigation for foundation design includes borings at each proposed turbine site, this mitigation method is Use construction methods. Most turbine spread foundations are relatively shallow (~2 the foundation can be placed on piles that are supported on rock below the karst zone. This may require additional investigation of the bedrock for the design of a pile foundation. Another option is to grout the underlying voids full to eliminate the potential for collapse. One advantage with grouting is that you can complete the detailed investigation to identify voids at the same time as the mitigation is being completed. Another possible construction method not encountered by these authors is to construct a foundation that bridges the risk zone. While a typical spread foundation is likely capable of bridging a small gap, the normal design process does not evaluate that possibility. Such a design consideration basis. Constructed mitigation adds hundreds of thousands of dollars to the cost of each turbine. Note that implementing constructed mitigation often means that detailed karst characterization is no longer required. 32
13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 Project Examples where karst risk was evaluated mainly by the senior author. Following are some brief descriptions of a few of these sites. North Central Iowa indicated that the bedrock is dolomitie (as opposed to limestone), with which karst development is linked in this exists over most of the site to mitigate risk (Figure 6) and did liability: Ignore the issue This is clearly unacceptable. Add a disclaimer. The disclaimer will state that it is impossible to completely know what is underground. This is a typical practice. Keep the investigation and evaluation of karst out of scope. In other words, pass the buck. Educate the client. workload and risk. 33 Table 1. Project Summaries. NA = Project did not advance Site Location No. of turbines Built? Lit Search Remote Sensing/ Lineament Experts Recon Drill Geophys ics Comment Arizona 1 62 No Yes Yes Yes NA NA NA progressed past desk top phase Arizona 2 62 Yes Yes Yes Yes Yes Yes Yes Developed area was reduced Iowa 79 Yes Yes NA Yes Yes Yes Yes Kansas 100 Yes Yes Yes Yes Yes Yes Yes Minnesota ~140 No Yes Yes Yes Yes NA Yes Unbuilt as of spring 2012 New York ~90 No Yes Yes Yes Yes NA Yes early. Develop er kept looking for a different answer Ohio 175 Yes Yes No Yes Yes Yes Yes Oklahoma 1 129 Yes Yes No No No Yes Yes Due to constraints and schedule, investigation Oklahoma 2 ~90 No Yes No Yes Yes NA NA Dune cover Watonga Pennsylvania 24 Yes Yes NA Yes Yes Yes Yes Expensive mitigation Texas 1 160 Yes Yes Yes Yes Yes Yes Yes Field investiga tion was limited based on linea ment analysis Texas 2 242 Yes Yes Yes No Yes No Yes 260 Yes Yes Yes Yes No Yes Texas 4 28` Yes Yes Yes Yes Yes Yes Yes cal modeling
NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE 34 Southwest Pennsylvania Literature review indicated, and site reconnaissance confirmed, that karst features were present in the area. Karst was associated with particular stratigraphic units, so areas of relative risk could restrictions on where development could take place, and those limitations took precedence over karst risk. before any subsurface investigation was completed. Once drilling began, numerous subsurface voids were found beneath most of the proposed turbine enough detail, so multiple drill holes were completed at turbine locations that were at risk. Although not budgeted for, the developer ended up installing deep pile foundations at some sites and grouting voids in others, at great expense. South Central Minnesota The client was a contractor bidding on constructing Figure 6. Cross section of wind project in North Central Iowa showing depth to bedrock. Thicker soil=less risk. Figure 7. Map showing relative risk for a wind farm in southwest Pennsylvania.
13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 35 areas of Minnesota (Figure 4). The contractor was continued to try to bring it to fruition for several years. Northwest Oklahoma Investigations in Blaine County, in northwestern Oklahoma, evaluated potential problems that gypsum karst may pose for the proposed Watonga Windpower depths ranging from 10 to 45 m below ground level. The Blaine is overlain by the Permian Dog Creek Shale and by unconsolidated Quaternary sands, clays, and gravels that may obscure karst features. Field showed that there is no direct evidence of gypsum gypsum beds was appropriate risk mitigation: where gypsum is 25 m below ground level or deeper, the risk related to gypsum karst is low, and where gypsum beds are less than 25 m deep, risk was medium to high. A map (Figure 9) was prepared showing areas of low, medium, and high risk related to gypsum karst. Figure 8. Cross section of shear wave velocity showing a sinkhole underlying a proposed wind turbine site in southwest Pennsylvania. Boring blow count decreased with depth. Figure 9. Risk categories at Watonga Windpower Project, based upon depth to the Shimer Gypsum at top of the Blaine Formation (Johnson et al., 2013)
NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE 36 Conclusions Karst can lead to dramatic tilting and even toppling of a wind turbine. Subtle differential settlement of turbine foundation can cause the turbine to be out of tolerance, requiring remedial action. There are many tools available for evaluating karst risk desktop methods and field methods with widely ranging costs from reconnaissance to intensive drilling. The right tools at any given phase of a windpower development will be based on the management discussions with the client. References Special Publication No. 122. Proceedings of the Alabama. ASCE. Beck BF, Wilson WL, editors. 1987. Karst hydrogeology: Engineering and Environmental Applications. Proceedings of the 2nd Multidisciplinary Conference on Sinkholes and the Environmental Impacts of Karst. Orlando, Florida. Rotterdam (Netherlands): A.A. Balkema. Beck BF, Stephenson JB, editors. 1997. The Engineering Proceedings of the 6th Multidisciplinary Conference on Sinkholes and the Engineering Missouri. Rotterdam (Netherlands): A.A. Balkema. Das, BM. 2010. Principles of geotechnical engineering. Planning & design consideration in karst terrain. In: Engineering and Environmental Applications. Proceedings of the 2nd Multidisciplinary Conference on Sinkholes and the Environmental Impacts of Karst. Orlando, Florida. Rotterdam Department of Natural Resources. County Atlas and potential problems in siting wind turbines in Blaine County, Oklahoma. NCKRI Conference Proceedings. fracture trace and occurrence of groundwater in of sites proposed for development in the dolomite karst regions of southern Africa. In: Beck BF, Engineering and Environmental Applications. Proceedings of the 2nd Multidisciplinary Conference on Sinkholes and the Environmental Impacts of Karst. Orlando, Florida. Rotterdam Tonkin & Taylor, Ltd. 2011. Puketoi Wind Farm Prepared for Mighty River Power Ltd. Monroe Counties, West Virginia. Map ECS 17.