014 Hassan Ijise 11

Int. J. Industrial and Systems Engineering, Vol. X, No. X, xxxx 1

Copyright © 20xx Inderscience Enterprises Ltd.

Analysis of wind power potentials at selected airport locations in Canada

Quazi K. Hassan* Department of Geomatics Engineering, Schulich School of Engineering, University of Calgary, 2500 University Dr. N.W., Calgary, Alberta T2N 1N4, Canada Fax: +1 403 284 1980 E-mail: [email protected] *Corresponding author

Ahad Ali A. Leon Linton Department of Mechanical Engineering, Lawrence Technological University, 21000 W Ten Mile Road, Southfield, MI 48075, USA Fax: +1 248 204 2576 E-mail: [email protected]

Navdeep S. Sekhon and Xin Wang Department of Geomatics Engineering, Schulich School of Engineering, University of Calgary, 2500 University Dr. N.W., Calgary, Alberta T2N 1N4, Canada E-mail: [email protected] E-mail: [email protected]

Abstract: Wind power has huge potential as a source of sustainable/renewable energy. In this paper, we analysed wind speed data from 21 major Canadian airports over 1971–2000 to determine the top four locations (i.e. having long-term annual average wind speed greater than 4.83 m s−1 at a height of 10 m) for further analysis. We employed Weibull probability density function to fit with the actual probability density (i.e. relative frequency distribution) of the hourly wind speed data for the year 2005. The parameters of Weibull probability function were found location-specific and then used to predict the probability density from 2006 to 2008, and found strong relations (i.e. r2-values in the range 0.83–0.94) existed with the measured probability density. On the one hand, our simulation also indicated that location of St. John’s, Newfoundland and Labrador could produce the highest amount of total annual power (i.e. 883,993 kWh), and Charlottetown, Prince Edward Island, on the other hand, could produce the least amount (i.e. 344,508 kWh in total, annually).

2 Q.K. Hassan et al.

Keywords: wind speed; modelling and simulation; Weibull probability density function; sustainability; renewable energy.

Reference to this paper should be made as follows: Hassan, Q.K., Ali, A., Sekhon, N.S. and Wang, X. (xxxx) ‘Analysis of wind power potentials at selected airport locations in Canada’, Int. J. Industrial and Systems Engineering, Vol. x, No. x, pp.xx–xx.

Biographical notes: Quazi K. Hassan is an Assistant Professor in Earth Observation for Energy and Environment in the Department of Geomatics Engineering and Centre for Environmental Engineering Research and Education at the University of Calgary. He received a BSc in Electrical and Electronics Engineering from Khulna University of Engineering and Technology, Bangladesh; an MSc in Civil Engineering from University Putra Malaysia and a PhD in Remote Sensing and Ecological Modelling from the University of New Brunswick, Canada. His research interest falls in the area of modelling of environmental issues, remote sensing, renewable energy, natural resource management, etc.

Ahad Ali is an Assistant Professor in the Department of Mechanical Engineering at the Lawrence Technological University. He received a BSc in Mechanical Engineering from Khulna University of Engineering and Technology, Bangladesh; an MS in Systems and Engineering Management from Nanyang Technological University, Singapore and a PhD in Industrial Engineering from the University of Wisconsin – Milwaukee, USA. His research interests fall in the area of manufacturing systems modelling, simulation and optimisation, intelligent scheduling and planning, artificial intelligence, design for reliability, e-manufacturing, lean manufacturing and supply chain management.

Navdeep S. Sekhon is currently an MSc student in the Department of Geomatics Engineering and Centre for Environmental Engineering Research and Education at the University of Calgary. He received a BTech in Agricultural Engineering from Punjab Agricultural University, India and an MEng in Environmental Engineering from Punjab Engineering College (Deemed University), India. His research interests fall in the area of remote sensing, geographical information system and environmental modelling.

Xin Wang is an Assistant Professor in the Department of Geomatics Engineering at the University of Calgary. She received a BSc in Computer Science from Northwest University, China; an MEng in Software Engineering from Northwest University, China and a PhD in Computer Science from University of Regina, Canada. Her research interests fall in the area of spatial data mining, geographical information systems, knowledge engineering and engineering applications using artificial intelligence.

1 Introduction

One of the greatest challenges in the 21st century is to explore ways of reducing the emissions of greenhouse gases into the atmosphere, which potentially could at least slow down the process of global warming. In this context, we need to investigate sources of renewable and clean energy (e.g. wind, solar, water, geothermal, tidal wave, etc.) which release little to no greenhouse gases. Several studies investigated potential of such energy

Analysis of wind power potentials 3

sources at global scale (e.g. Archer and Jacobson, 2005; de Vries et al., 2007; Liu et al., 2009; Lu et al., 2009). It is also demonstrated that renewable energy sources could be able to supply:

1 equivalent electrical power demands at least in USA and New Zealand by 2020 (Sovacool and Watts, 2009)

2 all purpose energy demands at world level by 2030 (Jacobson and Delucchi, 2009), using the current technological developments.

In this paper, our focus is to explore about one of the renewable energies (i.e. wind) in Canadian context.

2 Literature review

In Canada, currently ~1.1% of the county’s total electric demand (i.e. 3,249 MW) is being produced from wind according to Canadian Wind Energy Association (CanWEA, 2009a). According to CanWEA, wind energy can potentially support 20% of Canada’s electricity demands (i.e. 55,000 MW) by 2025 (CanWEA, 2009b). In 2008, Environment Canada released an updated version of Canadian Wind Energy Atlas, covering the entire country at a spatial resolution of 5 km (CWEA, 2009). However, the wind speed and thus the potential wind power within 5 km grid cell may not be identical. It may be the case as the wind speed varies as a function of topographic variability, land use/cover and uneven heating of the earth surface from the sun (Woofenden, 2009).

In theory, prior to implementing any proposed system, one of the best methods is to consider modelling and simulation approaches for predicting its behaviour and sustainability (Claudio et al., 2010; Farr et al., 2010; Gopalakrishnan et al., 2007; Himri et al., 2009a; Liu et al., 2009; Pandian and Ali, in press). In such cases, the application of Weibull probability density function is widely recognised in determining the probabilistic shape of the event of interest (Benatiallah et al., 2010; Jenab et al., 2010; Leithead, 2007). Hence, the objectives of this paper are to:

1 determine the probabilistic shape of the hourly averaged wind speed at four selected airport locations (i.e. Gander, Newfoundland and Labrador (NFLD); St. John’s, NFLD; Charlottetown, Prince Edward Island (PEI) and Regina, Saskatchewan (SK)). The rationale of choosing these locations is to be discussed in Section 3

2 evaluate the sustainability of the wind speed at four selected airport locations as mentioned in objective (1)

3 simulate how much wind power actually can be produced under certain conditions.

3 Modelling and evaluation techniques of wind speed

In selecting the airports of interest that had experienced relatively high wind speed, we analysed the long-term average annual wind speed data for the period 1971–2000 at 21 locations across the ten Canadian provinces (see Figure 1). These data were made available from Environment Canada (2009). It revealed that the four airport locations (i.e.

4 Q.K. Hassan et al.

Gander, NFLD; St. John’s, NFLD; Charlottetown, PEI and Regina, SK) had experienced relatively high average wind speed > 4.83 m s−1 on an annual basis (see Figure 1 for more details). At these locations, we also acquired hourly wind speed data for the period 2005–2008 at 10 m height from Environment Canada (2009). The location-specific data from 2005 were used to determine the shape of probability for the entire range of wind speed. Thus, the obtained coefficients of the shape would then be used to predict the probable frequency distribution for the period 2006–2008, and compare with the actual frequency distribution measured at the location of interest.

Figure 1 (a) Location of the ten provinces in Canada; (b) long-term (1971–2000) average annual wind speed at 21 major airports situated across all of the provinces. Relatively high wind speed (i.e. >4.83 m s−1) was observed at Gander, NFLD; St. John’s, NFLD; Charlottetown, PEI and Regina, SK

Analysis of wind power potentials 5

3.1 Determining the probabilistic shape of the wind speed data

To determine location-specific probabilistic shape of the wind speed data for the year of 2005, we executed the following steps:

1 generated the frequency distribution as a function of wind speed at discrete intervals of 1 m s−1

2 calculated the relative frequency distribution by dividing the counts in each of the intervals by the total number of measurements in the given dataset (see Figure 2 for more details)

3 as most of the cases, Weibull probability density is widely employed to delineate the frequency distribution of wind speed data (Barnsley, 2007; Bowden et al., 1983; Celik, 2004; Leithead, 2007), we also used such an expression:

1

Φ( ) expk kk u uu

c c c

− ⎡ ⎤⎛ ⎞ ⎛ ⎞= −⎢ ⎥⎜ ⎟ ⎜ ⎟⎝ ⎠ ⎝ ⎠⎢ ⎥⎣ ⎦ (1)

where u is the wind speed (in m s−1), Φ(u) is the probability density of the wind speed u (unitless), c is a scale parameter (in m s−1) and k is a shape parameter (unitless). We determined the values of location-specific values of both c and k (see Figure 2 for more details).

3.2 Evaluating the sustainability of wind speed

The sustainability of wind speed at a particular location is critical to make decision in selecting the site for installing wind turbine. For evaluating the sustainability of the wind speed, we performed the following steps:

1 calculated relative frequency distribution (i.e. measured probability density) for each of the location for every year during the period 2006–2008

2 generated location-specific Weibull probability density using the c and k values obtained in step (3) in Section 3.1 to be termed as ‘modelled probability density’

3 compared the measured and modelled probability density using linear regression analysis. The details of these analyses would be discussed in Section 4.

3.3 Estimating wind power

Available wind power at a given location can be determined using the following equation (Leithead, 2007; Twidell and Weir, 2006):

2 3W

12

P R uρπ= (2)

where PW is the available power (in W), ρ is the air density (=1.225 kg m−3 at mean sea level atmospheric pressure at 15oC, R is the rotor radius (in m) and u is the wind speed (m s−1). In practice, the available wind power cannot be converted entirely into effective/usable power. It is due to the following factors:

6 Q.K. Hassan et al.

1 the maximum fraction of available wind power that could be convertible to usable one, termed as power coefficient, CP, could have a maximum of 0.5926 (Betz, 1966). A well-designed wind turbine could have a CP value of 0.4 (Barnsley, 2007)

2 loss of power in the electric generator, known as generator efficiency, Ng, and could have 0.8 as maximum (Manwell et al., 2002)

3 loss of power in the mechanical parts of the turbine (that includes gearbox and bearings), known as mechanical efficiency, Nb, and could have a maximum of 0.95 (Barnsley, 2007).

Upon considering all of the above-mentioned factors, the power to be harvested from the available wind power, PR (in W), could be as follows (Barnsley, 2007):

2 3R P g b

12

P R u C N Nρπ= (3)

4 Results and discussion

Figure 2 shows the relative frequency distribution and the fitting of the Weibull probability density function of the hourly wind speed data for the year 2005 at the locations of Gander, NFLD; St. John’s, NFLD; Charlottetown, PEI and Regina, SK. It revealed how the Weibull probability density fits for the four airport locations as mentioned in Figure 1. The values of both shape parameter (k) and scale parameter (c) were found in the range between 2.2–2.5 and 5.8–6.5 m s−1, respectively, as shown in Figure 2. In most of the cases, the values of scale parameter and shape parameter were found to be location dependent (Bekele and Palm, 2009; Himri et al., 2009b) and our findings also confirmed this notion.

We also conducted linear regression analysis between relative frequency distribution (i.e. measured probability density) and Weibull probability density (i.e. modelled probability density) using the determined scale parameter and shape parameter values as obtained in Figure 2 at the four locations. Figure 3 provides linear regression analysis between measured probability density (i.e. relative frequency distribution) and modelled probability density (i.e. resulted from the fitting of Weibull probability density as shown in Figure 2) for the year of 2005 at the location of Gander, NFLD; St. John’s, NFLD; Charlottetown, PEI and Regina, SK. Strong relations (i.e. r2-values in the range 0.89–0.93, slopes ≈ 1 and intercepts ≈ 0) revealed between them, except at the location of St. John’s, NFLD (i.e. r2 ≈ 0.79, slope ≈ 0.84 and intercepts ≈ 0.009; see Figure 3 for more details). It was due to the fact that the relative frequency in the interval 1–2 m s−1 wind speed was high and not be captured by the Weibull probability function. In fact, the wind speed <3 m s−1 is not important as the cut-in frequency for wind turbine for generating power is considered >3 m s−1 (Barnsley, 2007; Dalton et al., 2009; Jung et al., 2008).

Analysis of wind power potentials 7

Figure 2 Relative frequency distribution of the hourly wind speed data for the year 2005 and fitting of the Weibull probability density at the location of (a) Gander, NFLD; (b) St. John’s, NFLD; (c) Charlottetown, PEI and (d) Regina, SK

To evaluate the sustainability of wind speed and thus the wind power, we executed the steps as described in Section 3.2. Comparisons between measured and modelled probability density for the years 2006–2008 at the four locations of interest are shown in Figures 4–7. In general, we observed strong relations (i.e. r2-values in the range 0.83–0.94, slopes in between 0.87 and 1.04 and intercepts ~ 0) between measured and modelled probability density at all of the four airport locations. On the one hand, among these locations, the location of Gander, NFLD was having the most strong relations (i.e. r2-values in the range 0.92–0.94, slopes ~ 1.0 and intercepts ~ 0; see Figure 4). On the other hand, the location of Charlottetown, PEI showed the least strong relations (i.e. r2-values in the range 0.83–0.86, slopes in between 0.90 and 0.99 and intercepts ~0; see Figure 6). On the basis of the above-mentioned results, we could conclude that the observed relative frequency distribution of wind speed in 2005 at all of the four locations (as shown in Figure 2) were similar during the period of interest (i.e. 2006–2008), and to be considered as sustainable.

In estimating the actual wind power generation, we simulated Equation (3) with the following values: 1 ρ = 1.225 kg m−3 at mean sea level atmospheric pressure at 15°C 2 R = 3 m

8 Q.K. Hassan et al.

3 CP = 0.4 4 Ng = 0.75 5 Nb = 0.90.

We also considered the cut-in and cut-out wind speed for the turbine as 3 m s−1 and 25 m s−1, respectively. As the wind speed data were measured at a height of 10 m, the simulation would provide values at that particular height. Example estimates of actual wind power generation for the year 2008 at the four locations (i.e. Gander, NFLD; St. John’s, NFLD; Charlottetown, PEI; and Regina, SK) are shown in Figure 8. In 2008, the total annual amount of wind power that could be generated with the above specification was:

1 493,649 kWh at Gander, NFLD

2 883,993 kWh at St. John’s, NFLD

3 344,508 kWh at Charlottetown, PEI

4 476,704 kWh at Regina, SK.

Figure 3 Linear regression analysis between measured and modelled probability density for the year 2005 at the location of (a) Gander, NFLD; (b) St. John’s, NFLD; (c) Charlottetown, PEI and (d) Regina, SK

Analysis of wind power potentials 9

Figure 4 Relation between measured probability and modelled probability density at Gander, NFLD location for 2006–2008

Figure 5 Relation between measured probability and modelled probability density at St. John’s, NFLD location for 2006–2008

10 Q.K. Hassan et al.

Figure 6 Relation between measured probability and modelled probability density at Charlottetown, PEI location for 2006–2008

Figure 7 Relation between measured probability and modelled probability density at Regina, SK location for 2006–2008

Analysis of wind power potentials 11

Figure 8 Example estimates of wind power generation for the year 2008 at the location of (a) Gander, NFLD; (b) St. John’s, NFLD; (c) Charlottetown, PEI and (d) Regina, SK

In general, we observed that the summer months could produce relatively less amount of power due to lack of suitable wind speed. During that period, relatively longer hours of sunshine was available and that could potentially be harvested. However, such detailed analysis is outside of the scope of this paper. Note that the actual power generation could be significantly enhanced by:

1 putting the turbine at a higher height (>10 m) as wind speed normally increases exponentially with the height (Hussain, 2002)

2 increasing the radius of the rotor (>3 m)

3 installing more than one wind turbine.

12 Q.K. Hassan et al.

5 Concluding remarks

In this paper, we demonstrated that the probability density of location-specific wind speed at annual scale were relatively stable. The parameters of Weibull probability density derived from 2005 wind speed data would be able to predict the measured probability density to a greater extent for the years 2006–2008 at all the four selected locations of Gander, NFLD; St. John’s, NFLD; Charlottetown, PEI and Regina, SK. Our analysis revealed strong relations (i.e. r2-values in the range 0.83–0.94) existed between the measured and modelled probability density.

Thus, all of these locations could be ideal for generating wind power because of reliable speed of the wind. The maximum wind energy could be generated at St. John’s, NFLD. Further research needs to be conducted in understanding environmental consequences (e.g. impacts of noise, changes in micro climate, habitat suitability in the surrounding areas, etc.) and feasibility of installing wind turbines near the airport in greater details.

Acknowledgements

The authors would like to acknowledge partial funding from the Department of Geomatics Engineering and Schulich School of Engineering at the University of Calgary to Dr. Quazi K. Hassan for this research. We would also like to thank Environment Canada for providing wind speed data at free of cost. We would also be grateful to the anonymous reviewers and the editor for providing useful comments on the earlier version of this paper in enhancing its overall quality.

References Archer, C.L. and Jacobson, M.Z. (2005) ‘Evaluation of global wind power’, J. Geophysical

Research, Vol. 110, No. D12110, 20p. Barnsley, M.J. (2007) Environmental Modeling a Practical Introduction. Florida, USA: CRC

Press. Bekele, G. and Palm, B. (2009) ‘Wind energy potential assessment at four typical locations in

Ethiopia’, Applied Energy, Vol. 86, No. 3, pp.388–396. Benatiallah, A., Kadi, L. and Dakyo, B. (2010) ‘Modelling and optimisation of wind energy

systems’, Jordan Journal of Mechanical and Industrial Engineering, Vol. 4, No. 1, pp.143–150.

Betz, A. (1966) In D.G. Randall (Trans.), Introduction to the Theory of Flow Machines. Oxford: Pergamon Press.

Bowden, G.J., Barker, P.R., Shestopal, V.O. and Twidell, J.W. (1983) ‘The Weibull distribution function and wind power statistics’, Wind Engineering, Vol. 7, No. 2, pp.85–98.

CanWEA (2009a) ‘Canadian Wind Energy Association’, Available at: www.canwea.ca/farms/ index_e.php. Last Visited on 24th December 2009.

CanWEA (2009b) WindVision 2025 Powering Canada’s future: Backgrounders on Wind Energy. Ottawa, Canada: Canada Wind Energy Association, 2008, Available at: www.canwea.ca/ images/uploads/File/Windvision_backgrounder_e.pdf. Last Visited on 26th December 2009.

Celik, A.N. (2004) ‘A statistical analysis of wind power density based on the Weibull and Rayleigh models at the southern region of Turkey’, Renewable Energy, Vol. 29, No. 4, pp.593–604.

Analysis of wind power potentials 13

Claudio, D., Zhang, J. and Zhang, Y. (2010) ‘A simulation study for a hybrid inventory control strategy with advance demand information’, Int. J. Industrial and Systems Engineering, Vol. 5, No. 1, pp.1–19.

CWEA (2009) Canadian Wind Energy Atlas, Available at: www.windatlas.ca/en/index.php. Last Visited on 26th December 2009.

Dalton, G.J., Lockington, D.A. and Baldock, T.E. (2009) ‘Feasibility analysis of renewable energy supply options for a grid-connected large hotel’, Renewable Energy, Vol. 34, No. 4, pp.955–964.

de Vries, B.J.M., van Vuuren, D.P. and Hoogwijk, M.M. (2007) ‘Renewable energy sources: their global potential for the first-half of the 21st century at a global level: an integrated approach’, Energy Policy, Vol. 35, No. 4, pp.2590–2610.

Environment Canada (2009) ‘Environment Canada’, Available at: www.climate.weatheroffice.ec.gc.ca/climateData/canada_e.html. Last Visited on 10th December 2009.

Farr, R., Tannock, J. and Almgren, T. (2010) ‘A parametric approach to logistic control within manufacturing simulation’, Int. J. Industrial and Systems Engineering, Vol. 5, No. 3, pp.313–326.

Gopalakrishnan, B., Banta, L. and Bhave, G. (2007) ‘Modelling steam-based energy supply chain in an integrated steel manufacturing facility: a simulation-based approach’, Int. J. Industrial and Systems Engineering, Vol. 2, No. 1, pp.1–29.

Himri, Y., Himri, S. and Stambouli, A.B. (2009b) ‘Assessing the wind energy potential projects in Algeria’, Renewable and Sustainable Energy Reviews, Vol. 13, No. 8, pp.2187–2191.

Himri, Y., Stambouli, A.B., Himri, S. and Draoui, B. (2009a) ‘Wind energy assessment for Timimoun’, Int. J. Agile Systems and Management, Vol. 4, Nos. 1/2, pp.182–190.

Hussain, M. (2002) ‘Dependence of power law index on surface wind speed’, Energy Conversion and Management, Vol. 43, No. 4, pp.467–472.

Jacobson, M.Z. and Delucchi, M.A. (2009) ‘A path to sustainability energy by 2030’, Scientific American, Vol. 301, No. 5, pp.58–65.

Jenab, K., Kazeminia, A. and Liu, D.S-C. (2010) ‘A computational intelligent approach to estimate the Weibull parameters’, Int. J. Industrial and Systems Engineering, Vol. 6, No. 1, pp.62–78.

Jung, S.Y., Jung, H., Hahn, S.C., Jung, H.K. and Lee, C.G. (2008) ‘Optimal design of direct-driven PM wind generator for maximum annual energy production’, IEEE Transaction on Magnetics, Vol. 44, No. 6, pp.1062–1065.

Leithead, W.E. (2007) ‘Wind energy’, Philosophical Transactions of the Royal Society A, Vol. 365, No. 1853, pp.957–970.

Liu, Q., Miao, Q., Liu, J.J. and Yang, W. (2009) ‘Solar and wind energy resources and prediction’, Journal of Renewable and Sustainable Energy, Vol. 1, No. 4, pp.043105:1–043105:12.

Lu, X., McElroy, M.B. and Kiviluoma, J. (2009) ‘Global potential for wind-generated electricity’, Proceedings of the National Academy of Sciences, Vol. 106, No. 27, pp.10933–10938.

Manwell, J.F., McGowan, J.G. and Rogers, A.L. (2002) Wind Energy Explained: Theory, Design and Application. New York: John Wiley & Sons.

Pandian, A. and Ali, A. (in press) ‘Automotive robotic body shop simulation for performance improvement using plant feedback’, Int. J. Industrial and Systems Engineering.

Sovacool, B.K. and Watts, C. (2009) ‘Going completely renewable: is it possible (Let alone desirable)?’, The Electricity Journal, Vol. 22, No. 4, pp.95–111.

Twidell, J.W. and Weir, A.D. (2006) Renewable Energy Resources (2nd ed.). Abingdon, UK: Taylor & Francis.

Woofenden, I. (2009) Wind Power for Dummies. Indianapolis, IN: Wiley Publishing Inc.