Engineering, 2013, 5, 104-107
doi:10.4236/eng.2013.51b019 Published Online January 2013 (http://www.SciRP.org/journal/eng)
Annual Maximum Loads Estimation Modeling for
Kingdom of Bahrain
Isa S. Qamber
Deanship of Scientific Research, Univer s ity of Bahrain, P. O . Box 32038, Isa Town, Kingdom of Bahrain
Received 2013
Abstract
The present paper proposes the impact of the air temperature on electricity demand as expected. The annual maximum
load is recorded versus the years starting by the year 2009 . At present, t he graph fittin g was ap plied with some mathe-
matical and computational tools considering the lower values, the higher values and the average values of the annual
maximum loads of Kingdom of Bahrain. For the three scenarios, the models are obtained by curve fitting technique. As
well, the model of actual load s is obtained finally which has mostly the closest values obtained.
Keywords: An nual Maximum Load, Cu rve Fittin g, Load S cenarios.
1. Introduction
For technical operations and control of Power Systems,
the knowledge of the evolution of electric power load is
impo rt ant . The r e aso n b e hi nd t hat is t he electric de mands
must meet secure and reliable operation at all times.
The electric power load scheduling for the next years is
related to the operation of the Power Systems. The his-
torical data of the maximum load is needed to estimate
the coming future maximum loads. In the present paper,
the historical data for the Kingdom of Power was consi-
dered and based on that the estimated and expected
maximum loads were obtained.
Power System operation, at the moment, faces new chal-
lenges associated with non-controlled power independent
producers, such as wind farms, photovoltaic plants and
small-hydro power plants (SHPPs). At the moment, a
large amount of the independent renewable po wer gener-
ation is characterized for being a type of electric power
production difficult to control: thus, on one hand, gener-
ally there is some kind of intermittency in the power re-
sourc e, which is no t c o ntr ol la bl e, while on t he o the r ha nd ,
there are as many operation strategies as managers of
renewable power stations.
Short-term forecasting of power production, for each
kind of power plant, is a key matter for the Power Sys-
tem, si nce such s hor t-ter m forecasting is an esse ntial tool
for e nsuri ng po wer sup ply, p lanni ng of reser ve pla nts, or
inter-power-systems electric energy transactions, or
helping to solve power network congestion problems.
A lot of research activity has been carried out, during the
last year s, i n the sho rt -term forecasting of power produc-
tion. The forecasts of load variables for the next years,
has helped in the improvement of short-term forecasting
models based on statistical models.
2. Literature Review
As it is well known that finding an appropriate load fo-
recasting model for a specific electrical network is not an
easy task. As previous studies carried out for such type
of purpose, Almeshaiei and Soltan in their paper [1] they
present a pragmatic methodology that can be used to
construct a model for electric power load. The metho-
dology they followed is based on decomposition and
segmentation of the load time series. T he study was car-
ried out on a daily real data from the State of Kuwait,
where the electric network was used as a case stud y. The
methodology was followed in the present paper is mainly
based on principles of time series segmentation and de-
composition. As well, some additio nal statistical a nalysis
was followed to aid the decision making based on the
adopted forecasts such as probability plots.
Emovon et al in their paper [2] are concerned with the
optimization of the workforce scheduling for solving
mai ntenance problems. The program they have written
was written in Quick Basic. The program software de-
signed to produce a seven day schedule for organization
operating a seven day week. Hence organization oper-
ating a five day schedule wishing to change to a seven
day schedule the researchers found the software very
useful. The Quick-Basic computer programme was
based on Alfares algorithm for solving schedule problem.
Their data collected from Afam power station, Nigeria

I. S. QAMBER
105
which was used as input data. The test result s s how t he
software is capable of determining workforce size and
assigning workers to day-off pattern. The seven-day
schedule produced savings of 11% maintenance labour
cost when compared with the 5-day schedule currently
being practiced by the Power station.
From the data collected by the authors [2], from Afam
Power Station Nigeria it’s on record that the first major
gas turbine station built in Nigeria is the Afam Ga s Tur-
bine Po wer Statio n. T he power station is located in the
Niger Delta because of the large reserve of natural gas in
the region.
The software used by Emovon et al [2] is capable of de-
termining workforce size and assigning workers to
days-off pattern for Afam power station and other or-
ganization like airli ne, police station and resta urant o per-
ating a 7 day a week. The software does not require
spec ia liz ed t ra ini ng u nli ke i nt e ger p ro gr a mmin g so ftware.
Also, the so ftware was tes ted with data from Afa m pow-
er station Nigeria. The tested results show that the
software is capable of generating a seven day schedule.
The seven day Schedule is more efficient and cost effec-
tive t ha n the f ive d a y wor k sc he d ule . The aut ho r s made
a comparison and they made the comparison between the
existing fi ve-day schedule practiced by Afam power sta-
tion Nigeria and the seven-day schedule generated by
this software. The seven-day schedule is expected to
produce savings of 11% maintenance labour cost an-
nually.
Jiang et al [3] in their paper examine a new time series
method for very short-term wind speed forecasting.
The time series forecasting model is based on Bayesian
theory and structural break modeling, which could in-
corporate domain knowledge about wind speed as a prior.
Besides this Bayesian structural break model predicts
wind speed as a set of possible values, which is different
from classical time series model’s single-value predic-
tion This set of predicted values could be used for vari-
ous applicatio ns, such as wind turb ine predictive control,
wind power scheduling. The proposed model is tested
with actual wind speed data collected from utility-scale
wind turbines. The authors [3] highlight in their paper
that in the recent years, many time series models are de-
veloped to deal with nonlinearities in time series that
cause problems for traditional linear models. Among
these models, structural break models have become in-
creasingly popular and achieved promising computation-
al results in estimation and prediction of economic time
series data. To test the efficacy of the proposed approach
[3], real wind speed time series are used. The wind
speed data used in this paper was collected by the ane-
mometer installed on the top of each wind turbine’s na-
celle from a wind farm located in the east coast of
Jiangs u provinc e, China. Thoug h the data was sa mpled
at a very high frequency by a wind turbine’s SCADA
system, it was averaged and stored at 10-s or 10-min
intervals (referr ed to the 10-min o r the 1 0 -s average data)
in wi nd ind us tr y. The 10 -s ( hi g h fr eq uenc y) d at a s hows
the d ynamic nat ur e o f the win d tur b ine , while the 10-min
data reflects rather steady state of the turbine. The data
used in this research is collected over a period of two
weeks. The proposed forecasting method is tested on
both 10-s and 10-min average data and compared with a
persistent forecasting model. The authors in their paper
[3] introduced a new time series model of forecasting
very short-term wind speed. The forecasting model
integrates the concepts of structural breaks and
Bayesian inferences, which allows prior information
about the wind speeds to be incorporated into the model
and somehow boosts forecasting performance. For very
short-term wind speed or power forecasting (e.g. the fo-
recasting ti me step is 10 s), persistent model is ver y hard
to beat. The proposed method is tested with real-world
wind speed time series and its forecasting performance is
compared with a benchmark persistent model and other
popular forecasting approaches.
Three scenarios have been carried out by Qamber [4] to
calculate the predicted maximum annual load for the
kingd om of B ahrai n wi th the obje ctive o f for mulatin g an
expansion plan for a future generating system. The
results of the three scenarios were obtained and com-
pared using a comprehensive analysis. The maximum
annual load was calculated at average rates of 6.79% in
the more reasonable scenarios using the MATLAB
package following the curve-fitting polynomial tech-
nique.
Hahn, et al. [5] used various models and methods to pre-
dict future load demand. These various models and
methods help decision-makers in the electricity sectors in
facilities planning and an optimal day-to-day operation
of power plant. The authors conclude that finding an
appropriate approach and model is at core of the decision
process and a decision maker in the energy sector has the
need of accurate forecasts since most of the decisions are
necessarily based on forecasts of future demands, where
the selection of an appropriate model is the one of the
first decision to be made.
Doege, et al. [6] realized power markets have been
restructured since 1990s worldwide with electrical power
nowadays being traded as a commodity. T he liberaliza-
tion and with it, uncertainty in gas, fuel and electrical
power prices, requires effective management of produc-
tion facilities a nd their financia l c ontracts.
3. Results and Discussions
Because of random fluctuations of the obtained data of
maximum annual load for Kingdom of Bahrain over the
years starting by the year 2009, the curve fitting tech-
nique is applied to find out the estimated load values for

I. S. QAMBER
106
the coming years in previous study [4] and the best fit
equation. Then curve fitting analysis identifies the
trend of the data for the individual data points. The main
reason behind that is to build a described model. A sim-
ple models are obtained to optimize the maximum annual
load for the coming years. Table (1) shows the loads un-
der study.
Table (1) The Loads (MW) vs Year
Year
Author-
ity
Scena-
rio1
Scena-
rio2
Aver-
age
Actual Peak
Load (MW)
2009 2194 2436 2377 2336 2437
2010
2325
2601
2508
2478
2708
2011 2453 2778 2646 2626 2500
2012
2588
2967
2792
2782
2948
2013 2731 3168 2945 2948 ?
2014
2881
3383
3107
3124
?
2015 3039 3613 3278 3310 ?
Fig. (1) shows the lower value of annual maximum load
in (Mega Watt) as a function of the years, where the re-
sult is a straight line. This type of result has already
been observed in the following form:
PL (χ) = 140.17857 χ - 279437.71 (1)
where: PL(χ) is the lower value load (MW)
χ is the year
Fig. (1) Lower Val ue Load vs Year
The straight line corresponds to the data obtained
through the calculation made/obtained by the Electricity
and Water Authority of Kingdom of Bahrain, as well as
done in both scenarios carried out in previous study [4].
The results confirms the validity of equation (1).
Fig. (2) shows the higher value of the annual maximum
load in (Mega Watt) as a function of the years, where the
result is a straight line. This type of result has already
been observed in the following form:
PH (χ) = 195.89286 χ - 391144.14 (2)
where: PH (χ) is the higher value load (MW)
χ is the year
The straight line corresponds to the data obtained
through the calculation made/obtained by the Electricity
and Water Authority of Kingdom of Bahrain, as well as
done in both scenarios carried out in previous study [4].
The results confirms the validity of equation (2).
Fig. (2) Higher Value Load vs Year
Fig. (3) shows the average peak value load (annual
maximum load) in (Mega Watt) as a function of the
years, where the result is a straight line. This type of
result has already been observe d in the following form:
PAPV (χ) = 162 χ - 323143.43 (3)
where: PAPV (χ) is the average peak value load
(MW)
χ is the year
Fig. (3) Average Valu e Loa d vs Y ear
The straight line corresponds to the data obtained
through the calculation made/obtained by the Electricity
and Water Authority of Kingdom of Bahrain, as well as
done in both scenarios carried out in previous study [4].
The results confirms the validity of equation (3).

I. S. QAMBER
107
Fig. (4) Actual Peak Load vs Year
From the three figures (1), (2) and (3) it is interesting to
note that these types of maximum annual loads are rais-
ing against increasing of the years for a reasons. These
reasons are the increasing of populations and the in-
creasing of the industrial companies (factories), except
for a conditions such as temperatures in some cases (un-
expected) it might be decreases during summer (as it
happened during the year 2011 in Bahrain). This cause
less demand o f electricit y whi ch is un-expected as shown
in Fig. (4). The best straight line that fits the points of
figure (4) is:
PAPL (χ) = 132.5 χ - 263743 (4)
where: PAPL (χ) is the actual peak load (MW)
χ is the year
4. Conclusions
Through the curve fitting technique, t he algorithms have
been found by feeding the data for each case using the
computer simulation. The used simulation is highly
efficient. It was shown that the annual maximum load
increases every year except for exceptional case when
during summer the temperature is decreases as un-usual
case (Year 2011 in Kingdom of Bahrain). The relation-
ships found during the carried study has the following
shape:
f (χ) = a χ - b
where this reveals that summer has the fastest annual
maximum temperature rate at growth (except some un-
usual cases, e.g. the year 2011 where the temperature
dropped down). It will known that the temperature has
impact on electricity demand.
5. Acknowledgemen ts
The author would like to express his thanks to the Uni-
versity of Bahrain for the preparation of the facilities to
make this research possible.
REFERENCES
[1] E. Almeshaiei and H. Soltan, “A Methodology for Elec-
tric Power Load Forecasting”, Alexandria Engineering
Journal , Vol. 50, 2011, pp. 13 7-144.
[2] I. Emovon, M. T. Lilly and S. O. Ogaji, “Design of Soft-
ware for Maintenance Workforce Scheduling (Acase
Study of Afam Power Station, Nigeria)”, International
Journal of Engineering, Vol. 4, No. 5, 2012, pp. 235-244.
[3] Y. Jiang, Z. Song and A. Kusiak, “Very Short-Term
Wind Speed Forecasting with Bayesian Structural Break
Model”, Inter na t ional Journal of Renewable Energy, Vol.
50, 2013 , pp. 63 7-647.
[4] I. Qamber, “Maximum Annual Load Estimation and
Network Strengthening for the Kingdom of Bahrain”,
AGJSR, Vol. 28, No. 4, 2010, pp. 214 -223.
[5] H. Hahn, S. Meyer-Nieberg and S. Pickl, “Electric Load
Forecasting Methods: Tools for Decision Making”, Eu-
ropean Journal of Operational Research, Vol. 199, 2009,
pp. 902-907.
[6] J. Doege, M. Fehr, J. Hinz and H. Luthi, “Risk Manage-
ment in Power Markets: the Hedging Value of Production
Flexibility”, European Journal of Operational research,
Vol. 199, No. 3, 2009, pp. 936-943.