Page 57 - AJWEP-22-6

Statistical analysis of wind energy

P 1 ∞ (v) Hysteresis analysis to detect phase lag or cyclical
3
E = = ρ v f ( ) v dv (X)
A 2 ∫ 0 wind–temperature interactions. 30,37
where E is the power density (W/m²), ρ is the air This methodology aims to elucidate the role of
density (1.225 kg/m³), and f(v) is the wind speed PDF temperature as a modulator of wind patterns in Mongo’s
(Weibull or MEP). Accurate modeling of the wind Sahelian climate, supporting refined assessments of
speed distribution is crucial for realistic energy yield wind energy potential.
estimates. 35
3. Results and discussion
2.3.2. Wind direction analysis using a wind rose
An assessment of wind direction was performed using 3.1. Wind speed analyses
monthly averaged data collected from 2012 to 2022. Table 1 summarizes the monthly and annual mean wind
The directional values were categorized into 16 equally speeds over the 2012 – 2022 period. All annual mean
spaced compass sectors, each covering 22.5°, to ensure values exceed the minimum exploitable wind speed
adequate angular resolution over the full 360° range. threshold of 2.0 m/s, demonstrating that Mongo has
This segmentation enables the systematic quantification a usable, albeit modest, wind resource. The highest
of the directional distribution of the wind over the study annual average wind speed was recorded in 2012, with
period. 4.61 m/s, and the lowest in 2022, at 1.80 m/s. This range
The frequency of wind blowing from each sector is aligns with wind regimes classified as Class 1 under
calculated with Equation XI. International Electrotechnical Commission standards,
generally suited for small-scale or hybrid wind power
n applications. 6,17,23,28
f = N i ×100 (XI) Figure 1 illustrates the monthly wind speed
i
where f is the frequency (in percent) of wind coming evolution over 11 years, presenting a 3D temporal
i
from direction sector i; n is the number of wind direction profile (Figure 1A) and a 2D projection (Figure 1B).
i
observations in that sector, and N is the total number of The temporal signal (Figure 2A) illustrates bursting
valid wind direction records. behavior characterized by short-term spikes in wind
The data were processed using meteorological speed, which may be attributed to local convective
statistical tools capable of handling directional phenomena and regional climate influences. 6,14,18 The
classification and frequency analysis. No filtering based long-term monthly mean (Figure 2B) reveals a seasonal
on wind speed threshold was applied in this computation pattern with higher wind speeds during specific months,
phase to preserve the full angular distribution of wind consistent with similar patterns observed in semi-arid
occurrence. The results of this procedure served as input regions, such as Central Iran and Northern India. 18,19
for constructing a wind rose diagram, which reflects the These temporal and seasonal variations underscore the
statistical dispersion of wind direction at the site. 8,36 importance of temporal resolution in wind resource
assessment and planning. The relatively low but stable
2.4. Temperature analysis and wind temperature wind speeds suggest Mongo’s wind energy applications
relationship assessment should focus on micro or small wind turbine systems
Air temperature affects atmospheric pressure gradients designed for low-wind regimes. 17,23,28
and convective flows, shaping wind dynamics. 28,29
Monthly average temperatures from 2012 to 2022 were 3.2. Probability distribution and wind power
analyzed through: The Weibull distribution parameters, shape factor
(i) Time series analysis to identify seasonal and (k) and scale factor (c), were computed numerically
interannual variability (Equations I – IV) and are presented in Table 2 for annual
(ii) Graphical superimposition of temperature and wind data, and Table 3 for monthly data. These parameters
speed for concurrent trend detection were used to construct the wind speed PDF shown in
(iii) Scatter plots to assess preliminary correlations Figure 3, where the mode centers around 3.25 m/s, close
(iv) Regression modeling (linear and non-linear) to to the mean wind speed. This confirms the adequacy
quantify relationships, evaluated by determination of the Weibull model for this dataset, as supported by
coefficient (R²) studies in Turkey, Iran, and Canada. 12,13,15,18,23

Volume 22 Issue 6 (2025) 51 doi: 10.36922/AJWEP025070039