Pakistan Journal of Nutrition 9 (4): 387-391, 2010

ISSN 1680-5194
© Asian Network for Scientific Information, 2010

Statistical Process Control on Production: A Case Study of Some Basic

Chemicals Used in Pure Water Production
Abubakar Usman and Nasir Mu’Azu Kontagora
Department of Mathematics/Computer Science, Federal University of Technology, Minna, Nigeria
Department of Integrated Science, Niger State College of Education, Minna, Nigeria

Abstract: Statistics is the art of making decisions about a process or population based on an analysis of
the information contained in a sample from that population. In any production process, regardless of how
well designed or carefully maintained it is, a certain amount of inherent or natural variability will always exits,
Such variability like background noise. When the variability is small, we usually consider this an acceptable
level of process performance or the process is within the process control. A process that is operating in the
presence of assignable causes is said to be out of control. A data set collected from randomly selected
packaged water producers referred to as “Pure water producers” and subjected to laboratory test of the of
the level of some basic chemicals used in the production of pure water, such as pH, conductivity (µS/cm;
Lead (Pb); Aluminum (Al) and Chloride (Cl). The test shows that most of the chemicals used are out of
process control. This is danger to health of the consumers.

Key words: Production, process control, tweaking, tasteless, assignable causes, variability, chemicals

INTRODUCTION C Monitor and maintain the process on target.

Production processes must perform consistently C Determine when a process needs “tweaking”
overtime to be capable of meeting production and (adjusting) and when it does not.
design requirements. Statistical process control is a C Establish process stability and detect process
methodology to monitor and benchmark a process to changes so that corrective action can be taken.
improve its variability, stability and capability. According C Improve quality and productivity by improving the
to Douglas Montgomery (1985) states that Statistical process, which reduces product inspection, scrap
methods play a vital role in quality assurance. They and rework at the end of the line.
provide the principal means by which product is
sampled, tested and evaluated and information in those Background of the study: Pure water is an odourless,
data used to control and improve the production tasteless liquid. It has a bluish tint, which may be
process. In any production process, regardless of how detected, however, only in layers of considerable
well designed or carefully maintained it is, a certain pressure. Accessibility and availability of fresh clean
amount of natural variability will always exist. The natural water is a key to sustainable development and an
variability “background noise “is the cumulative effect of essential element in health, food production and poverty
many small, essentially uncontrollable causes. When reduction (Adekunle, 2004). However an estimated 1.2
the background noise of a process is relatively small, we billion people around the world lack access to safe
usually consider this an acceptable level of process water and close to 2.5 billion are not provided with
performance. adequate sanitation (Third World Water Forum on Water,
There is variability that usually arises from improperly 2003). The standard industrialized world model for
adjusted machines, operator error and/or defective raw delivery of safe drinking water and sanitation technology
materials. Such variability is generally large when is however, not affordable in much of the developing
compared to the background noise. This variability that world. Thus, given the renewed global commitments
is not part of chance cause pattern is referred to towards the Millennium Development Goals (MDG)
“assignable causes”. marked for 2015, the importance and contribution of
Veronica and Patrick (1997) stress that control charts locally sourced low cost alternative drinking water
are among the most important tools in Statistical schemes to sustainable access in rural and semi-urban
process control. They were developed in the 1920's by settings of developing nations cannot be over
Dr. Walter Shewhart a scientist at Bell laboratories, the emphasized (UNDESA, 2004).
research arm of American Telephone and Telegraph. Statutorily, portable water supply in Nigeria had been by
Statistical control can be used to: the Government Owned Public Water Utilities (GPWU) in
C Monitor and reduce process variability. the past. The GPWUs provided their supply from

Corresponding Author: Abubakar Usman, Department of Mathematics/Computer Science, Federal University of Technology, Minna,
Nigeria
387
Pak. J. Nutr., 9 (4): 387-391, 2010

conventional water treatments plants that uses water

from impounded reservoirs, flowing perennial streams,
lakes and deep boreholes. As the country population
grows and industries increase, the supply of water by
the GPWUs becomes inadequate in quality and quantity.
This led to the emergence of some Privately Owned
Water Enterprises (POWE) that operated side by side
with the GPWUs within the water sector (Onemano and
Otun, 2003).
One of the most popular POWE in Nigeria is the sachet
water sold in polythene sachet otherwise called ‘Pure
Water’. The POWEs mainly collect their water as the end
product of initially treated water supplied by the GPWUs
and do little treatment such as the removal of the
suspended solids to make the GPWUs water more Fig. 1: A typical control chart
potable. They also do some minor treatment on water
from natural springs, open wells and deep boreholes. x-bar. Let w have a true population mean (µ) and
Some also collect water directly from the GPWUs kiosks standard deviation (F).
and later resells them at a higher price (Onemano and The centre line, upper control limit and lower control limit
Otun, 2003). have the general form:
The production, marketing and consumption of sachet
water have increased tremendously. There are now UCL = µ + kF
several brands of these type of packaged water CL = µ
marketed in Nigeria and other developing nations LCL = µ - kF
(Ogan, 1992 and Kassenga, 2007). This so called Pure
Water in sachets is readily available, easy to serve and The upper and lower control limits are chosen to be a
the price is affordable and finds patronage from the distance of K standard deviation (F) for the mean (µ). It is
middle class and members of low socio-economic customary to use k = 3, Which gives 3-sigma control
classes, but there are concerns about its purity. The limits that are standard deviation away from the centre.
integrity of the hygienic environment and the conditions Research as show that if the distribution is normal,
where majority of the water in sachets are produced has approximately 99.73% of the values production will fall
also been questioned (CAMON, 2007). between the 3-sigma control limits and 0.27% of the
We use the Statistical process control chart to monitor values of the production will fall outside the control limits
the production process of pure water with the giving a false alarm approximately only 27 out of 10,000
assessment of some basic chemicals used in the times.
production and process of pure water.
A control chart includes three horizontal parallel lines: a Objectives of the study: The research work was aimed
centre line, an upper control limit above it and a lower at ensuring that the production process is monitored to
control limit below it. The centre (CL) on a control chart reduce variability and maintain the process target. To
points are expected to cluster in the absence of an determine when a process needs “tweaking” (adjusting)
assignable cause. The centre line is usually set at and when it does not. To establish process stability and
average, the median, the mode or the target value of the detect process changes so that corrective action can be
points being plotted. The Upper Control Limit (UCL) line taken and finally, to improve quality and productivity by
and the Lower Control Limit (LCL) defined a region improving the process, which reduces product
where most observations are expected to fall. The upper inspection, scrap and rework at the end of the line.
and lower control limits refer to as statistical control
limits or Statistical Process Control (SPC) limits reflect MATERIALS AND METHODS
the natural variability of the process and are constructed A total of hundred sachets (100) from ten different
in such a way that when the process is in control most of brands of the Pure water were randomly selected in
the points will falls inside the control limits in random Minna, Niger state which was the study area. Ten
fashion. If a point on the control charts falls above the samples of each were collected direct from hawkers and
upper limits or below the lower control limits the process wholesalers at different locations of the town which
is said to be out of control, and assignable causes need includes; motor parks, markets and other busy areas of
to be searched for eliminated. the town. These were then transferred to the laboratory
Many control charts have a common form. Let w be an for analysis. For physical and laboratory analysis, the
observed process characteristics that is determined physical visual examination of the water samples; odour,
from a sample, such as the sample mean (average) and appearance such as colour, turbidity and presence

388
Pak. J. Nutr., 9 (4): 387-391, 2010

of floating particles were all noted. While in the The upper control limit is 8.5 while the lower control
laboratory, each sample was subjected to both physical limit is 6.5 from the statistical process chart it seen
and chemical tests in accordance with standard that three of the points are out of statistical process
methods as obtained from the American Public Health control, most of the points are at the lower control
Association (APHA), 2005. The chemical tests limit, which according to Veronica and Patrick (1997)
conducted include chemical analysis of Lead (Pb), when eight points fall at one side the production is
chloride (Cl), Iron (Fe) and Aluminum (Al). The physical out of process control. This might lead to increase
tests carried out were those of pH and conductivity. The of dissolution of certain metals whose
summary of the Raw data is as in appendix1 (Table2). concentration above normal could have detrimental
health effect.
Data analysis and interpretation: Data obtained were
analyzed using means, standard deviation, standard
error of mean and control charts with the use of Instat
and Minitab 14 statistical software.
All the tests were conducted between January-June,
2009. The choices of the periods were as a result of the
realization of the large production and consumptions of
pure water during these periods. The analysis is as
follows:

Table 1: Mean, standard deviation, standard error of mean of the

chemicals tested in January-June, 2009
Groups Mean SD SEM
pH 7.1600 0.712 0.225
Conductivity (µS/cm) 0.3610 0.1760 0.0556
Iron (Fe+2) 0.0870 0.1114 0.0352 Fig. 3: Plot showing the control chart for conductivity
Lead (Pb) 0.0340 0.0836 0.0264
Aluminum (Al) 0.0310 0.02558 0.00809
Chloride (Cl) 8.4000 3.3000 1.0400 From the Plot above, it concord with the assertion
SD = Standard deviation made by Veronica and Patrick (1997), that an indicator
SEM = Standard error of mean for out of control is that if eight consecutive points are on
one side of the centre line or two out of three consecutive
From Table 1, the mean values Iron (Fe+2), Aluminum points are outside the 2-sigma warning limits on one
(Al), Conductivity (µS/cm); Chloride (Cl) are less that side of the centre line, then the productions are out of
maximum permitted by Standard organization of control. Thus, conductivity could be classified has been
Nigerian, while, mean of the Lead (Pb) is greater than out-of-control.
the maximum permitted by Standard Organization of
Nigerian (SON) standard for portable drinking water as
in appendix 1 (Table 3). This shows statistical out of
process which invariably has health impact on the
consumers.

Fig. 4: Plot showing the control chart of Iron

Figure 4 shows that two of the values are out of control

and looking at the plot most of the points are on one
side of the center, that is , the lower limit. This could lead
to water having a metallic taste and could cause stains
Fig. 2: The control chart for the pH to laundry.

389
Pak. J. Nutr., 9 (4): 387-391, 2010

Table 2: Raw Mean score of parameters analyzed for the duration of

Cancer, interference with vitamin D metabolism, affects mental development in infants, toxic to the central and peripheral
work
Cond. Fe Pb Al Cl
Sam. pH (µS/cm) (mg/L) (mg/L) (mg/L) (mg/L)
1 6.80 0.38 0.02 0.27 0.02 11.20
2 6.40 0.24 0.01 0.00 0.06 4.33
3 7.00 0.62 0.15 0.02 0.03 4.75
4 7.20 0.13 0.03 0.00 0.03 7.96
5 8.80 0.18 0.00 0.01 0.04 15.74
6 7.20 0.38 0.24 0.00 0.04 7.34
7 6.20 0.51 0.04 0.00 0.01 8.71
8 7.10 0.61 0.03 0.00 0.08 6.80
9 7.30 0.36 0.03 0.03 0.00 9.638
10 7.60 0.20 0.32 0.01 0.00 7.53
Sam. = Samples; Cond. = Conductivity

Fluorosis, skeletal tissue (bones and teeth) morbidity

Very toxic to the thyroid and the nervous system
Potential Neuro-degenerative disorders

Fig. 5: Plot of the control chart for Aluminum

Source: Nigerian Industrial Standard (NIS) 554: 2007, Nigerian standard for drinking water quality
The range of Aluminum is between 0.00-0.0.155 mg/L
which is within the acceptable limit of 0.2 mg/L.
Gastrointestinal disorder

Consumer acceptability
Neurological disorder
Toxic to the kidney

nervous systems
Health impact

Hypertension

Cancer
Cancer

None

None

None
None
None
Maximum permitted
Table 3: Chemical parameters-inorganic constituents

0.003

1000
0.01

0.05

0.01

0.05

0.01

0.20
250

150
0.2

0.7

1.5

0.3

0.2
1

Fig. 6: Plot of control chart showing the responses for

µS/cm

Lead (Pb)
Mg/L
Mg/L
Mg/L
Mg/L
Mg/L
Mg/L

Mg/L
Mg/L
Mg/L
Mg/L
Mg/L
Mg/L
Mg/L

Mg/L
Mg/L
Unit

From the figure above it was observed that the element

Hydrogen sulphide ( H2S)

Lead (Pb) is out of control. Given the upper control limit

Hardness (as CaCO3)

to be 0.01 mg/L, showing most of the points to be out of

Manganese (Mn+2)
Magnesium (Mg+2)

controlled, this could have deleterious health effect

Chromium (Cr6+)
Cadmium (Cd)
Aluminum (Al)

Cyanide (CN-)
Copper (Cu+2)

as lead is associated with cancer, interference with

Chloride (Cl)
Arsenic (As)

Conductivity

Fluoride (F)
Parameter

mental development and vitamin D metabolism and is

Iron (Fe+2)
Lead (Pb)
Barium

also implicated in central and peripheral nervous

system toxicity.

390
Pak. J. Nutr., 9 (4): 387-391, 2010

conform to standards led out for quality drinking water,

the chemical constituents are out of process control; as
such this could pose health risk to the consumers. The
result also confirms other reports on the health risk of
pure water.
We therefore recommend that a statistical process
control chart should be plotted periodically
productivity by improving the process, which the
process production, scrap and rework at the end of
the days production.

REFERENCES
Adekunle, L.V., M.K.C. Sridhar, A.A. Ajayi, P.A. Oluwade
Fig. 7: Plot of control chart showing the responses for and J.I. Olawuyi, 2004: An assessment of the health
chloride and social economic implications of sachet water
in Ibadan Nigeria: A public health challenge. Afr. J.
Chloride concentration in the samples has a mean Biomed. Res., 7: 5-8.
range of 4.33 and 15.74 which is within the American Public Health Association (APHA), 2005.
accepted limit of 250 mg/L. The range is 15.74-4.33 Standard Methods for the Examination of Water
= 11.31. and Wastewater. 21st Edn., Washington,
DC.
RESULTS AND DISCUSSION Consumer Affairs Movement of Nigeria (CAMON), 2004.
The results were compared with Nigerian Standard for NAFDAC to Ban Pure Water-97% Samples
Drinking Water Quality as outlined by Standard Contaminated. Consumer Link 1:1.
Organization of Nigeria (SON), 2007. As outlined in the Douglas Montgomery, C., 1985. Introduction to Statistical
table the mean pH ranges from 6.2-8.8, which indicates Quality control. John wiley and sons Inc., USA.
that some of samples are falling out of range of the Kassenga, G.R., 2007. The health-related
normal pH of drinking water (6.5-8.5). Though, this microbiological quality of bottled water sold in Dar
might not have any direct effect but could lead to es Salam, Tanzania. In Dada, A.C. (2009), Sachet
increase of dissolution of certain metals whose water phenomenon in nigeria: assessment of the
concentration above normal could have detrimental potential health impacts. Afr. J. Microbiol. Res., 3:
health effect. 15-21.
The conductivity has a mean range of 0.13-0.62 µS/cm, Ogan, M.T., 1992. Microbiological quality of bottled sold
while Iron has a mean range of 0.00-0.32 mg/L in retail outlets in Nigerian, In Dada, A.C. (2009),
indication a deviation of the accepted level of 0.3 mg/L, Sachet water phenomenon in Nigeria: Assessment
this could lead to the water having a metallic taste and of the potential health impacts. Afr. J. Microbiol.
could cause stains to laundry. Lead has a mean range Res., 3: 15-21.
of 0.00-0.03 mg/L with the upper limit well above the Onemano, J.I. and J.A. Otun, 2003. Problems on Water
accepted limit of 0.01 mg/L this could have deleterious Quality Standards and Monitoring in Nigeria:
have effect as lead is associated with cancer, Towards the Millennium Development Goals. 29th
WEDC International Conference, Abuja, Nigeria.
interference with mental development and vitamin D
Standards Organization of Nigeria (SON), 2007.
metabolism and is also implicated in central and
Nigerian Standards for Drinking Water Quality.
peripheral nervous system toxicity. Aluminum has a
Nigerian Industrial Standards NIS 554.
mean range of 0.00-0.08 mg/L which is within the
Third World Forum on Water, 2003. Blockade, Myth,
accepted limit of 0.2 mg/L. Chloride concentration in the
Illusions in Development and Cooperation.
samples has a mean range of 4.33-15.74 which also
January, 2003; 30 (1).
within the accepted limit of 250 mg/L.
UN Department of Economics and Social Affairs
(UNDESA), 2004. Urban Agglomerations in Dada,
Conclusion and recommendations: Production process A.C. (2009), Sachet Water Phenomenon in Nigeria:
control chart can monitor and reduce process variability, Assessment of the Potential Health Impacts. Afr. J.
determine when a process needs adjusting and when Microbiol. Res., 3: 15-21.
it does not, establish process stability and detect Veronica Czitron and D. Patrick Spagon, 1997.
process changes. Statistical case studies for industrial process
As evidence from the results some of the chemicals improvement. America Statistical Association
used in the production of packaged water at times do not Alexandria, Virginia.

391