Available online at www.sciencedirect.

com
Available online at www.sciencedirect.com

Available ScienceDirect
ScienceDirect
Availableonline
onlineatatwww.sciencedirect.com
www.sciencedirect.com
Procedia Manufacturing 00 (2018) 000–000
ScienceDirect
ScienceDirect 
Procedia Manufacturing 00 (2018) 000–000
www.elsevier.com/locate/procedia
www.elsevier.com/locate/procedia
Procedia Manufacturing 32 (2019) 786–793
Procedia Manufacturing 00 (2017) 000–000
www.elsevier.com/locate/procedia

The
The 12th
12th International
International Conference
Conference Interdisciplinarity
Interdisciplinarity in
in Engineering
Engineering

Wind
Wind Speed
Speed Data
Data Analysis
Analysis UsingUsing Weibull
Weibull and and Rayleigh
Rayleigh
Distribution
Manufacturing Functions,
Engineering Case
Society Study:
International Five Cities
Distribution Functions, Case Study: Five Cities Northern
Conference Northern
2017, MESIC Morocco
Morocco
2017, 28-30 June
2017, Vigo (Pontevedra), Spain
Hicham Bidaouia,b,
Hicham Bidaoui *, Ikram a,b, *,
Ikram El
El Abbassi
b
Abbassib,, Abdelmajid
Abdelmajid El
El Bouardi
a
Bouardia,,
Costing models for capacity optimization inb Industry 4.0: Trade-off
0F

b
Abdelmoumen Darcherif
0F

Abdelmoumen Darcherif
between
Department ofused
a
a capacity
Physics, Faculty of sciences, and operational
Abdelmalek efficiency
Essaadi University, Tetouan,Morocco.
Department of Physics, Faculty of sciences, Abdelmalek Essaadi University, Tetouan,Morocco.
b
b ECAM-EPMI, Graduate School of Engineering, Cergy-Pontoise, 95052, France.
ECAM-EPMI, Graduate School of Engineering, Cergy-Pontoise, 95052, France.
A. Santanaa, P. Afonsoa,*, A. Zaninb, R. Wernkeb
Abstract
a
University of Minho, 4800-058 Guimarães, Portugal
Abstract b
Unochapecó, 89809-000 Chapecó, SC, Brazil
The aim of the present paper is to evaluate and discuss the Energy wind potential of five major cities in Northern Morocco. The
The aim of the present paper is to evaluate and discuss the Energy wind potential of five major cities in Northern Morocco. The
theoretical analysis is based on stochastic models of Weibull and Rayleigh using Probability Density Function approach. Various
theoretical analysis is based on stochastic models of Weibull and Rayleigh using Probability Density Function approach. Various
statistical indicators such as the determination coefficient (R²), Chi square error (χ²), root mean square error (RMSE) and mean
statistical indicators such as the determination coefficient (R²), Chi square error (χ²), root mean square error (RMSE) and mean
Abstract
bias error (MBE) are considered. Then, a numerical simulation of the potential electrical power is carried out using the Enercon
bias error (MBE) are considered. Then, a numerical simulation of the potential electrical power is carried out using the Enercon
E103/2350 wind turbine model. The results show that Weibull is more accurate than Rayleigh, especially for Tetuan and Al-
E103/2350
Under wind turbine
of model. The results show that Weibull is more accurate than Rayleigh, especially forinterconnected,
Tetuan and Al-
Hoceimathe concept
cities. "Industry
Concerning 4.0", production
electrical energy processes
production, only the siteswill be pushed
of Tetouan to begave
and Tangier increasingly
interesting values.
Hoceima cities.
information Concerning
based electrical
on a real energyand,
time basis production, only the
necessarily, sitesmore
much of Tetouan and Tangier
efficient. In this gave interesting
context, values.
capacity optimization
goes
© beyond
2018
© 2018
2019 The the traditional
Authors.
The Authors. Published aimby of capacity
Elsevier Ltd.maximization, contributing also for organization’s profitability and value.
© The Authors. Published
Published by by Elsevier
Elsevier Ltd.
Ltd.
Indeed,
This is anlean management
open access and the
article under continuous
CC BY-NC-ND improvement approaches suggest capacity optimization instead of
license (https://creativecommons.org/licenses/by-nc-nd/4.0/)
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/)
Selection and peer-review
maximization. under responsibility of the 12thand
International Conference Interdisciplinarity in Engineering.
Selection and peer-review under responsibility of the 12th Internationalmodels
The study of capacity optimization costing is anInterdisciplinarity
Conference important research topic that deserves
in Engineering.
contributions from both the practical and theoretical perspectives. This paper presents and discusses a mathematical
Keywords: Weibull distribution; Rayleigh distribution; Wind speed; Probability density function; Wind rose.
model forWeibull
Keywords: capacity management
distribution; Rayleighbased on different
distribution; costing
Wind speed; models
Probability (ABC
density andWind
function; TDABC).
rose. A generic model has been
developed and it was used to analyze idle capacity and to design strategies towards the maximization of organization’s
value. The trade-off capacity maximization vs operational efficiency is highlighted and it is shown that capacity
1. Introduction
1. Introduction
optimization might hide operational inefficiency.
© 2017 The Authors. Published by Elsevier B.V.
In recentunder
years, world’s energy
of theconsumption has been increasing due toEngineering
economic and social development in both
In recent years,
Peer-review world’s energy
responsibility consumption has been
scientific committee increasing
of the due to economic Society
Manufacturing and social development
International in both
Conference
modern and emergent countries. The fast growing energy demand is particularly visible in Africa. While the world
2017.
modern and emergent countries. The fast growing energy demand is particularly visible in Africa. While the world
energy production reached 13 790 Mtoe in 2015, the African continent saw its consumption rose to 8% versus 0.6%
energy production reached 13 790 Mtoe in 2015, the African continent saw its consumption rose to 8% versus 0.6%
Keywords: Cost Models; ABC; TDABC; Capacity Management; Idle Capacity; Operational Efficiency

1. Introduction
* Corresponding author.
* The
Corresponding
cost author.
of idle capacity is a fundamental information for companies and their management of extreme importance
E-mail address: bidaouihicham@gmail.com
E-mail address: bidaouihicham@gmail.com
in modern production systems. In general, it is defined as unused capacity or production potential and can be measured
in several©ways:
2351-9789 tons
2018 The of production,
Authors. available
Published by Elsevier Ltd.hours of manufacturing, etc. The management of the idle capacity
2351-9789 © 2018 The Authors. Published by Elsevier Ltd.
This is an Afonso.
* Paulo open access
Tel.:article under
+351 253 510the761;
CC BY-NC-ND
+351 253license (https://creativecommons.org/licenses/by-nc-nd/4.0/)
This is an open access article under the CC fax:
BY-NC-ND 604 741
license (https://creativecommons.org/licenses/by-nc-nd/4.0/)
Selection
E-mail and peer-review
address: under responsibility of the 12th International Conference Interdisciplinarity in Engineering.
psafonso@dps.uminho.pt
Selection and peer-review under responsibility of the 12th International Conference Interdisciplinarity in Engineering.

2351-9789 © 2017 The Authors. Published by Elsevier B.V.

Peer-review
2351-9789 © under
2019responsibility
The Authors. of the scientificbycommittee
Published Elsevier of the Manufacturing Engineering Society International Conference 2017.
Ltd.
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/)
Selection and peer-review under responsibility of the 12th International Conference Interdisciplinarity in Engineering.
10.1016/j.promfg.2019.02.286
 Hicham Bidaoui et al. / Procedia Manufacturing 32 (2019) 786–793 787
2 Hicham Bidaoui et al. / Procedia Manufacturing 00 (2018) 000–000

in 2014. While, energy demand rose only by 0.5% in non-OECD (Organization for Economic Co-operation and
Development) countries. and decreased by 0.3% in OECD [1].
Wind is one of the most important renewable energy sources. Indeed, the wind energy is well developed,
controlled and accessible. However, in order to increase wind farm’s electrical production performance in a given
location, a deep analysis of wind characteristics is required [2], [3]. Recently, researchers have shown increased
interest in wind energy potential in several regions of the world using miscellaneous probability density functions
(PDF).

Fig. 1. Map of the locations of measurement stations.

In this paper, Weibull and Rayleigh PDFs are used to describe wind speed probability in five cities in Northern
Morocco (Fig. 1); they are the most widely used in literature, as well as their good accuracy in describing and
predicting wind speed distribution in many regions of the world PDF.
The objective here is threefold i) to find the theoretical model closest to the experimental reality for extrapolation
to other geographical areas in the southern Mediterranean, ii) to determine among the five cities those with the best
potential Wind turbine, iii) and to use the models in the sizing of wind farms according to the geographical
specificities of the site.

Nomenclature

v wind speed (m/s)

vm mean wind speed (m/s)
v0 wind speed at reference height of measurement z0
vz wind speed at reference height of measurement z
R2 determination coefficient
χ2 chi-square error
RMSE sum of root mean square errors
MBE mean percentage error
k shape factor
c scale parameter (m/s)
788 Hicham Bidaoui et al. / Procedia Manufacturing 32 (2019) 786–793
Hicham Bidaoui et al. / Procedia Manufacturing 00 (2018) 000–000 3

2. Materials and methods

2.1. Dual-fed antenna design Weibull probability and cumulative distribution function

It’s the widely two parameters distribution function used in literature for describing wind speed data in many
regions. The variations in wind velocity are characterized by two functions: “The probability density function” and
“The cumulative distribution function”. The first function indicates the percent of time for which the wind flows
with a specific wind speed. It’s mathematically expressed as follow [4], [5]:

k −1   v k 
 k  v 
f w (v) =     . exp  −    (1)
 c  c    c  

The cumulative distribution function gives the percent of time that the wind speed is less than or equal the wind
speed v0, it’s expressed by the integral of the probability density function.

 v k 
F (v ≤ v0 ) =
1 − exp  −    (2)
  c  

2.2. Rayleigh probability and cumulative distribution function

The Rayleigh distribution, which is a special case of Weibull distribution with a fixed shape parameter value k=2,
is defined by [6], [7]:

πv  π  v 2 
f r (v )
= exp  −    (3)
2vm2  4  vm  
 

The Rayleigh cumulative distribution function is given by:

 π  v 2 
1 − exp  − 
F (v ≤ v0 ) =   (4)
 4  vm  
 

2.3. Extrapolation of wind speed at different hub height

The observed data used in this study were measured at 10m height. While most of commercial wind turbine have
different hub height. The Hellman’s exponential law was used to extrapolate wind speed data to adequate height to
our study (110m) [8], [9]:

α
 z 
vz = v0   (5)
 z0 

It’s assumed to be 0.143 in coastal regions and 0.2-0.3 in forested area, or evaluated by the following equation:

  z2  
α=
0.37 − 0.088ln(c0 )  / 1 − 0.088ln   (6)
  10  
 Hicham Bidaoui et al. / Procedia Manufacturing 32 (2019) 786–793 789
4 Hicham Bidaoui et al. / Procedia Manufacturing 00 (2018) 000–000

2.4. Statistical performance indicators

To determine the performance of the used models referring to the observed data, common statistical indicators
were considered herein [10]–[12]:
• The determination coefficient (R2),

n n
R2 =
1− ∑
( x − y)2 /
i =1
∑( y − y)
i =1
2
(7)

• The sum of root mean square errors (RMSEs),

n
RMSE
= ∑ ( x − y)
i =1
2
/n (8)

• The chi-square error (χ2),

n
χ2
= ∑ ( x − y)
i =1
2
/ x
 (9)

• The mean percentage error (MBE).

n
MBE = ∑ (( x − y ) / x)
i =1 (10)

3. Results and discussion

Fig. 2 gives the mean wind speed on the selected sites at 10 m height for twelve months. It’s clear that Tetuan
and Tangier presents higher values, the mean wind speed varies from 6.4 m/s to 7.7 m/s for Tangier during the year
and it exceeds 5.8 m/s for Tetuan except in October. For both Nador and Al-Hoceima, the mean wind speed varies
between 3.8 m/s and 5.3 m/s with advantageous values for Nador with a minimum of 4 m/s as it can be seen in this
Fig. 1. Larach presents the lowest values of mean wind speed between November and February; however, it has
better values than Nador and Al-Hoceima from May to September, where exceeds 5.3 m/s.

Fig. 2. Monthly mean wind speed of the studied location.

 Hicham Bidaoui et al. / Procedia Manufacturing 00 (2018) 000–000 5
790 Hicham Bidaoui et al. / Procedia Manufacturing 32 (2019) 786–793

Analysis of mean wind speeds gives an idea of wind in a selected region; however, a deep analysis of continuous
observed wind speed data is necessary to evaluate and describe wind resources. To carry out such study, Rayleigh
and Weibull models were used here, and the results are present as follow:

3.1. Wind speed distribution

The histograms fitted by the used PDFs for the studied sites are plotted in Fig. 3, in order to compare predicted
wind speed frequencies by Weibull and Rayleigh models to observed wind speed data. It’s clear that Weibull
distribution function fitted the observed data much better than Rayleigh distribution function for the selected sites.
Moreover, the difference between the two distribution functions is related to the shape parameter k, it’s equal to 2
for the Rayleigh distribution function, and while the calculated value for Weibull distribution function is slightly
different to Rayleigh case the curves of two distributions are different.
Table 1. Statistical performance indicators of Rayleigh and Weibull distributions for the selected sites.

R2 RMSE MBE χ2
Weibull Rayleigh Weibull Rayleigh Weibull Rayleigh Weibull Rayleigh
Tangier 0,998 0,901 0,0004 0,0031 0,000 0,010 0,0025 0,3200
Tetuan 0,998 0,874 0,0005 0,0038 -0,032 -0,010 0,0317 0,0082
Al-Hoceima 0,982 0,872 0,0022 0,0058 0,560 0,090 0,1360 0,3250
Nador 0,997 0,941 0,0008 0,0034 -0,061 0,310 0,0439 0,0396
Larach 0,999 0,937 0,0004 0,0037 -0,060 0,370 0,0278 0,0209

Furthermore, the statistical indicators of models performance are calculated from the observed data and the
adjusted distribution data, using equations 7-10. The results from the goodness of fit evaluation for both Rayleigh
and Weibull distributions are shown in table 1. It can be seen that R2 is sufficiently high for both Rayleigh and
Weibull distributions in the given locations. However in terms of comparison, the Weibull R2 values are more
interesting with a median equal to 0.998 in front of a median equal to 0.901 for Rayleigh R2 values. In terms of
RMSE, MBE (%) and the chi-square indicators which are very small in this application, the results show interesting
values for Weibull distribution respectively in the following location rank: Larach, Tangier, Tetuan, Nador and Al-
Hoceima; while for Rayleigh distribution the ranking is as follow: Tetuan, Larach, Nador, Tangier and Al-Hoceima.

Fig. 3. Annual probability and cumulative distribution functions (a) Al-Hoceima; (b) Tangier; (c) Tetuan; (d) Larach; (e) Nador.
 Hicham Bidaoui et al. / Procedia Manufacturing 32 (2019) 786–793 791
6 Hicham Bidaoui et al. / Procedia Manufacturing 00 (2018) 000–000

3.2. Wind rose

Wind rose diagram summarizes a very important information in practice for a given location which is the wind
direction. To Analyse this parameter of wind data direction was carried out using WRPLOT view software [13]. Fig.
4 shows the wind rose graphs for the considered sites at 10m height. These polar diagrams are divided on 16 sectors
covering an angle of 22.5° each one. For Al-Hoceima, The highest prevailing wind direction originated 33° east and
west direction (clockwise) with 24% and 19.4% of occurrence. For Tangier and Tetuan, with approximately similar
tendency observed for both location, the dominant wind direction was from 56.5°, north and south direction
respectively with 20.6%, 18.8%, 17.2% of occurrence for Tangier and 21%, 18.1%, 17.2% of occurrence for Tetuan
site. For Nador and Larach sites, the most frequently observed wind direction was 57° south and north direction
(clockwise) with respectively 27.5% and 25% of occurrence for Nador and with 17% and 19.7% of occurrence for
Larach.
For Al-Hoceima, Larach and Nador, wind rose diagram indicates that the less frequently and weakest wind
direction originated from the north.

Fig. 4. Wind rose. (a) Al-Hoceima; (b) Tangier; (c) Tetuan; (d) Larach; (e) Nador.

3.3. Simulation of 2.35 MW wind turbine production

In this part, the E130/2350 wind turbine power curve was used (Fig. 5) to simulate the possible electrical power
output after predicting wind speed data through the Weibull PDF parameters: c and k. Since these parameters were
performed at 10 m height and while the turbine hub height is defined in a range of 98 m and 138 m, the c and k
parameters were extrapolated to 110 m (the chosen hub height for this simulation) using the Hellman’s exponential
law Eq. 5.
Fig.6 (a) presents the monthly mean power production for each location, it is clear that Tangier and Tetuan wind
speed data gives interesting monthly mean power production values, higher than 850 kW for all months except
October in Tetuan case. While for Al-Hoceima, Nador and Larach the maximum of monthly mean power to produce
by the E130/2350 generator doesn’t exceed 34% of the generator rated power. In addition, the monthly percent of
time at zero output (a specific time where the wind speed are less than or equal the turbines cut off wind speed)
presented in Fig. 6(b) doesn’t exceed 12% in Tangier and Tetuan sites except October of Tetuan. For the other
location, it exceeds 15% in Larach and Nador sites and 22% for Al-Hoceima case, which make the E130/2350
generator installation without interest in such locations.
792 Hicham Bidaoui et al. / Procedia Manufacturing 32 (2019) 786–793
Hicham Bidaoui et al. / Procedia Manufacturing 00 (2018) 000–000 7

Fig. 5. Wind rose. (a) Al-Hoceima; (b) Tangier; (c) Tetuan; (d) Larach; (e) Nador.

Fig. 6. (a) Monthly mean production of Enercon E103/2350; (b) Monthly percent of time at zero output.

4. Conclusion

Based on the results obtained here, the Weibull distribution function fitted the best to observed wind speed
probability distribution in the selected locations, this is confirmed by the analysis of PDFs graphs in Fig. 3 and
statistical performance indicators (Table 1). Afterwards, numerical simulations using Enercon E103/2350 model of
the potential electrical power output were performed. The results presented in Fig. 6 have shown the utility of
installing such generators on Tangier and Tetuan sites, which allow exploiting more than 850 kW every month with
less than 12% as maximum of time at zero output. Whereas in the case of the other sites, different generators must
be selected to enhance the feat of the kinetic energy of the available wind and reach better performance. The next
step of this work consists to analyze the effect of different wind turbine characteristics on wind power output in a
selected location with taking in consideration the investment costs.

References

[1] World Energy Balances 2017: Overview, International Energy Agency, (2017).
[2] D. Mazzeo, G. Oliveti, E. Labonia, Estimation of wind speed probability density function using a mixture of two truncated normal
distributions. Renew. Energy, 115 (2018) 1260-1280.
 Hicham Bidaoui et al. / Procedia Manufacturing 32 (2019) 786–793 793
8 Hicham Bidaoui et al. / Procedia Manufacturing 00 (2018) 000–000

[3] V. Katinas, M. Marčiukaitis, G. Gecevičius, A. Markevičius, Statistical analysis of wind characteristics based on Weibull methods for
estimation of power generation in Lithuania. Renew. Energy, 113 (2017) 190-201.
[4] T. Aukitino, M. G. M. Khan, M. R. Ahmed, Wind energy resource assessment for Kiribati with a comparison of different methods of
determining Weibull parameters. Energy Convers. Manag., 151 (2017) 641-660.
[5] F. Fazelpour, N. Soltani, M. A. Rosen, Wind resource assessment and wind power potential for the city of Ardabil, Iran. Int. J. Energy
Environ. Eng., 6 (2015) 431-438.
[6] S. Ali, S.-M. Lee, C.-M. Jang, Statistical Analysis of Wind Characteristics Using Weibull and Rayleigh Distributions in Deokjeok-do Island
– Incheon, South Korea. Renew. Energy, (2018).
[7] M. K. Rastogi F. Merovci, Bayesian estimation for parameters and reliability characteristic of the Weibull Rayleigh distribution. J. King
Saud Univ. - Sci., (2017).
[8] M. Irwanto, N. Gomesh, M. R. Mamat, Y. M. Yusoff, Assessment of wind power generation potential in Perlis, Malaysia. Renew. Sustain.
Energy Rev., 38 (2014) 296-308.
[9] S. S. Chandel, P. Ramasamy, K. S. R. Murthy, Wind power potential assessment of 12 locations in western Himalayan region of India.
Renew. Sustain. Energy Rev., 39 (2014) 530-545.
[10] M. H. Ouahabi, F. Benabdelouahab, A. Khamlichi, Analyzing wind speed data and wind power density of Tetouan city in Morocco by
adjustment to Weibull and Rayleigh distribution functions. Wind Eng., 41 (2017) 174-184.
[11] Z. H. Hulio, W. Jiang, S. Rehman, Technical and economic assessment of wind power potential of Nooriabad, Pakistan. Energy Sustain.
Soc., 7 (2017).
[12] J. Wang, J. Hu, K. Ma, Wind speed probability distribution estimation and wind energy assessment. Renew. Sustain. Energy Rev., 60 (2016)
881-899.
[13] Lakes Environmental | WRPLOT View [free] - Wind Rose Plots for Meteorological Data. [Online]. Available:
https://www.weblakes.com/products/wrplot/index.html. [Accessed: 12-May-2018].