Journal of Food Engineering 222 (2018) 276e283

Contents lists available at ScienceDirect

Journal of Food Engineering

journal homepage: www.elsevier.com/locate/jfoodeng

The impact of clean-in-place parameters on rinse water effectiveness

and efficiency
Mengyuan Fan, David M. Phinney, Dennis R. Heldman*
Dale A. Seiberling Food Engineering Laboratory, The Ohio State University, 2015 Fyffe Road, Columbus, OH 43210, USA

a r t i c l e i n f o a b s t r a c t

Article history: Although Cleaning-in-Place (CIP) is abundantly used throughout the food industry, it is recognized that
Received 21 April 2017 CIP operations use significant amounts of water and energy. The overall objective of this investigation
Received in revised form was to evaluate parameters needed to improve the effectiveness of water use during CIP pre-rinse. A
20 November 2017
pilot-scale CIP system was operated over a range of Reynolds number (Re) from 16,000 to 260,000, while
Accepted 21 November 2017
evaluating the effectiveness of rinse water to remove a reconstituted skim milk residue film from
Available online 22 November 2017
stainless steel pipe surfaces. Rinse water effectiveness was quantified by comparing the protein con-
centration on the pipe surface after pre-rinse to the initial level. As the Re increased, the effectiveness of
Keywords:
Clean-in-place
rinse step increased, but not in linear proportionality. The efficiency of the rinse water decreased
Rinsing significantly as the volume of rinse water increased. The results of this investigation provide the basis for
Optimization reducing water and energy requirements during CIP operations.
Water conservation © 2017 Elsevier Ltd. All rights reserved.
Milk deposit

1. Introduction fouling have evolved, including modification of surfaces, pulsed

Cleaning is one of the most critical stages in quality control of
food processing operations (Kulkarni et al., 1975). Failure to prop-
erly clean processing equipment not only increases the processing
cost due to the decreased heat transfer coefficient, but also
threatens plant operations with microbial contamination (Fryer
et al., 2006; Gillham et al., 1999; Kulkarni et al., 1975; Mattila
et al., 1990). Cleaning-In-Place (CIP) is a widely-used method of
automated cleaning without the need for disassembling equip-
ment, while reducing labor and time for cleaning (Seiberling, 1968).
However, CIP systems require significant quantities of water,
detergent, and energy. In addition, in-place cleaning generates
large quantities of waste water, with the additional economic
burden on the industry, and an environmental burden on the
community (Lyndgaard et al., 2014). Optimization of CIP operations
to reduce water and energy requirements, while maintaining hy-
giene of the CIP process is a worthwhile goal.
Over the past two decades, the basic understanding of the
fouling and cleaning processes in dairy plants has increased
considerably (Fryer and Robbins, 2005; Wilson, 2005; Xin et al.,
2002). Novel approaches to enhance cleaning and reducing
* Corresponding author.
E-mail address: heldman.20@osu.edu (D.R. Heldman).
flow, ozonated water rinse, electrolyzed oxidizing rinse water, and
ultrasonic cleaning (Augustin et al., 2010; Boxler et al., 2013;
Christian and Fryer, 2006; Dev et al., 2014; Gillham et al., 2000;
Guzel-Seydim et al., 2000; Muthukumaran et al., 2004; Rosma-
ninho et al., 2007). All these methods require additional devices,
capital investment, and in some cases; energy inputs. When
compared to the aforementioned solutions, the adjustment of CIP
control parameters (flow characteristics, water temperature, and
contact time) to achieve improved cleaning efficiency while
reducing water consumption would seem more practical for the
industry.
Pre-rinse, as the first step of a CIP process, is to rinse fresh water
through the processing system to drain until the discharge is clear
(Seiberling, 1968). In addition to providing the initial soil removal,
the pre-rinse may also assist the following alkaline cleaning step by
wetting the severe deposits (Khaldi et al., 2015). The wetted deposit
is consequently more vulnerable to the detergent (Goode et al.,
2013). It is known that the direct physical force provided by the
fluid would help the deposit to pass an energy barrier and detach
(Fryer et al., 2006; Grant et al., 1997; Weidemann et al., 2014). Thus,
any parameters influencing the physical force of the fluid would
affect pre-rinse step. Reynolds number (Re) is commonly used to
determine the flow characteristics (Singh and Heldman, 2013).
Besides Re, some researchers believed other parameters to also be
important in developing the prediction models for CIP operation.

https://doi.org/10.1016/j.jfoodeng.2017.11.029
0260-8774/© 2017 Elsevier Ltd. All rights reserved.
 M. Fan et al. / Journal of Food Engineering 222 (2018) 276e283
For example, Timperley and Smeulders (1987) concluded that flow
velocity had a larger impact on the cleaning rate than Re. The
correlation between wall shear stress (tu ) and the removal of the
foulant from the surface has been demonstrated by many re-
searchers (Sharma et al., 1992; Visser, 1970). Lelie
vre et al. (2002)
demonstrated that both the mean wall shear stress and the asso-
ciated time-dependent fluctuation rate would impact the detach-
ment of the foulant under turbulent flow conditions.
It is important to note that the papers mentioned above were
based on the alkaline cleaning step which follows the pre-rinse step
in CIP operation. Therefore, results from these studies would serve
as good benchmark for investigating the mechanism of pre-rinse
step, which has not been studied much by comparison. Further-
more, to our knowledge, there are no prediction models in litera-
ture to describe the impact of Re and its interactions with other CIP
control parameters on the effectiveness of pre-rinse. Once built,
these models can be used to optimize the CIP pre-rinse step,
ensuring rinse effectiveness while reducing water/energy outlay by
locating the optimum CIP parameter sets.
The objectives of this investigation were: (1) to establish the
relationship between rinse water effectiveness and Reynolds
Number at various temperatures and contact times, (2) to develop a
relationship between the efficiency of water use and rinse water
requirements, and (3) to develop recommendations on CIP pa-
rameters for maximum water effectiveness while reducing water
and energy demands.
2. Materials and methods
2.1. Creating the deposit film
Rinsing experiments were conducted using pipe sections with
known quantities of a deposit film. The deposit film was created
using an established protocol (Fan et al., 2015). The deposit film was
applied to pipe sections with three different diameters: 0.025 m (1
inch), 0.038 m (1.5 inch) and 0.050 m (2 inch). Seven pipe sections
of the same diameter were combined to create a test line. The
middle and two end sections of each test line were tested after each
experiment. The deposit film contained nonfat-dry-milk (NFDM)
(U.S. Foods, Inc., Rosemont, Ill., U.S.A.) reconstituted to 20% w/w in
deionized water at ambient temperature. The pipe sections were
filled with the reconstituted milk and held for 5 min at room
temperature. The reconstituted milk was then removed and the
pipe sections were allowed to drain at room temperature for
30 min. Next, each pipe section was placed in a 75
C oven for
30 min to dry the deposit film. Three pipe sections were selected to
be used in the CIP pre-rinse experiment. The remaining pipe sec-
tions were used to measure the protein content in the initial de-
posit film. The protein content of the deposit film was measured by
washing the pipe interior using a sodium hydroxide solution. The
protein assay was completed with QuantiPro™ Bicinchoninic Acid
(BCA) (Sigma-Aldrich Co.).
2.2. Cleaning-in-place (CIP) pre-rinse
The cleaning-in-place (CIP) pre-rinse experiments were per-
formed by mounting the fouled pipe sections on a test manifold,
which is a part of a pilot-scale CIP system. The CIP system and test
manifold were described in Fan et al. (2015). Tap water e pH from
7.2 to 7.7 (Model HI, 2020; Hanna Instruments, Inc., Woonsocket,
RI., U.S.A.) e was used in this study. Tap water was recirculated
through the CIP system, except for pipe sections with deposit films,
until the flow rates (0.0016, 0.0028 or 0.0039 m3 s1) and tem-
peratures (22, 45 or 67
C) stabilized. Next, the flow of rinse water
was directed to pipe sections with deposit films. Pre-rinse was
277
programmed to be a single-pass step in this experiment, which
means the rinse water was pumped to rinse the fouled pipe sec-
tions then drained directly without reusing. The contact time be-
tween the rinse water and the pipe sections with deposit films was
controlled to 20 s or 60 s during this study. Due to the flow sepa-
ration and gravity, nine velocities levels were achieved in three
diameter pipes (0.025, 0.035 and 0.050 m) at three system set flow
rates. At the same system set flow rate, the velocity of the flow in
the 0.035 m (1.5 inch) pipe section is highest due to the gravity,
while the velocity of the 0.025 m (1 inch) pipe section is the lowest.
The pre-rinse experiment was performed three times from which
three subsamples were obtained under each condition.
2.3. Data treatment
After each CIP pre-rinse, the test sections were disassembled
and the residual deposit film was removed from each pipe section
using a sodium hydroxide solution and the protein was quantified
by Quantipro assay. The effectiveness of the CIP pre-rinse was
determined from the percent residual film, which was calculated by
comparing the protein content after pre-rinse to the protein con-
tent of the deposit film prior to rinsing (Equation (1) and Equation
(2))
Percent residual film (RF%) ¼ 100 [C/C0] (1)
Rinse effectiveness ¼ 100 e RF% (2)
Where C (mg cm2) is the protein concentration after pre-rinse and
C0 (mg cm2) is the initial protein concentration.
The efficiency of water consumed during the pre-rinse step was
evaluated from the rinse effectiveness data. Water efficiency (WE)
was defined as the amount of deposit protein (kg) removed per unit
volume (m3) of water, as expressed in Equation (3)
WE ¼ PR / WV (3)
Where PR is the mass of protein removed per surface area (kg m2),
which equals to C0 minus C. WV is the volume of rinse water spent
per surface area (m3 m2).
As suggested in Equation (4), the volume of water consumed
(WV) was computed from the volumetric flow rate (V) of the rinse
water and the contact time (t) between the rinse water and the pipe
section, which was either 20 or 60 s in this study.
WV ¼ (V)(t) (4)
In order to minimize the impact of the system size on the
models, both WE and WV were normalized to per fouled surface
area. The fouled surface areas were calculated based on the inner
diameter and length of the pipe sections.
2.4. Statistical analysis
SPSS.21 (IBM, Inc,. USA) statistics software was used for analysis
of all data. The data have been reported as the means ± standard
error. Analysis of covariance (ANCOVA) with 95% confidence in-
terval was used to evaluate the statistical difference of the intercept
and slope between the linear regressions. A one-way ANOVA with a
95% confidence interval was used to evaluate the statistical differ-
ence between the means.
3. Results and discussion
The influence of three CIP parameters (temperature, velocity,
278 M. Fan et al. / Journal of Food Engineering 222 (2018) 276e283

contact time) on rinse effectiveness was evaluated. Three rinse 30 a

water temperature levels (22
C, 45
C, and 67
C) were selected to 20 Seconds
represent the range of temperatures used during CIP operations; 25 b b 60 Seconds
the ambient water temperature (22
C), 67
C - a recommended

Residual Film (RF%)

bc
rinse water temperature (between 60
C and 80
C) by researchers 20 c
(Li et al., 2013; Liu et al., 2006; Xin et al., 2002), and an intermediate
d
temperature between the two extremes (45
C). Nine rinse water 15
velocities between 0.73 and 2.48 m s1, including the recom- e e e
mended CIP operation flow velocity of 1.52 m s1, were applied in 10 f f f
f
this research. j j j j
The magnitude of the residual film on the interior surface of the 5
pipe sections after pre-rinse is expressed as the residual film per-
centage (RF%) to describe the rinse effectiveness. A lower magni- 0
tude of residual film percentage implies a better rinsing 16080 26190 62300 77160 87380 108150 176140 211060 261230
effectiveness. Reynolds Number (Re)

Fig. 2. Improvements in removal of residual film (RF%) deposits as a function of

3.1. Effect of Reynolds Number on rinse effectiveness Reynolds Number (Re) for 20 and 60-s rinse water residence times.

The analysis was conducted over a range of Reynolds Numbers

between 20,000 and 300,000. In Fig. 1, the residual film percentage
(RF%) after a 20-s rinse with water at 22
C is presented as a
function of Reynolds Number. A linear correlation is found between
natural logarithm of residual film percentage (RF%) and Reynolds
Number (Equation (5)).
Ln[RF%] ¼ 3.23 - 1.03  105 [Re]
Based on the linear correlation expression, the (deposit) residual
film removal is 81.4% at Re ¼ 20,000, and increased to 91.6% at
Re ¼ 100,000. Although the rinse effectiveness improved as the
Reynolds Number increased (as expected), the improvement in
rinse effectiveness is not proportional to the increase in Reynolds
Number. For example, when Re increased by five times from 20,000
to 100,000; the deposit film removal only increased from 81.4% to
91.6%. The five times higher change in Reynolds Number is achieved
by elevating temperature and increasing flow velocity by increasing
pump speed, both of which require a certain amount of energy
input.
Similar to the results from this investigation, Hankinson and
Carver (1968) also found that rinsing effectiveness improved
significantly as the pumping rate increased (Reynolds number
ranges from 85,000 to 600,000). This could be attributed to the
larger mechanical force and a better mixing effect provided by the
high level of turbulence (higher Re). It has been previously reported
that generally, a high Re implies instability of flow and significant
mixing within the region of flow (Meyer, 1971). These flow condi-
(5) tions are expected to provide larger mechanical forces and
enhanced turbulent diffusion (Timperley and Smeulders, 1987).
Both of these mechanisms should have a positive impact on the
rinse effectiveness. Larger direct mechanical forces can improve
rinsing by breaking down and removing the deposits during the
initial phase of rinsing (Blel et al., 2009). Flow characteristic can
impact the mass transfer rate of the detached deposit particles into
the cleaning fluid bulk (Gallot-Lavallee et al., 1984; Jensen et al.,
2005). A high mass transfer rate enhanced rinsing due to turbu-
lent mixing and the exchange of deposit materials between surface
and bulk (Gillham et al., 1999; Weidemann et al., 2013; Xin et al.,
2002).
Fig. 2 shows the influence of Reynolds number on the RF% after
various rinsing times. Similar to the results presented in Fig. 1,
higher percentages of residual film removal occurred in

4
3.5
ln (Residual Film%)

3
2.5
2
1.5
Ln[RF%]= 3.23 - 1.03 x 10-5 [Re]
1
R² = 0.77
0.5
0
0 20000 40000 60000 80000 100000 120000
Reynolds Number (Re)

Fig. 1. Influence of Reynolds Number (Re) on residual film deposit (RF%) with rinse water temperature of 22
C, and rinse water residence time of 20 s.
 M. Fan et al. / Journal of Food Engineering 222 (2018) 276e283
Table 1
ANCOVA results on Regressions of ln (residual film%) over Reynolds Number.
Temperature (
C) Contact time (s) Slope Intercept
22 20 1.03E-05a 3.23
22 60 1.01E-05a 2.99
45 20 6.53E-06b 3.10
45 60 6.04E-06b 2.83
67 20 4.23E-06b 2.68
67 60 4.58E-06b 2.73
Note- Values with the same superscript letter are not significantly different
(a ¼ 0.05).
experiments conducted at higher Reynolds numbers. It needs to be
noted that extending the contact time from 20 s to 60 s under a high
level of turbulence (high Re range) did not significantly influence
the rinse effectiveness. In another words, the effect of the Re level
on rinse effectiveness seems to be the most dominant factor in the
system. If that is the case, it would suggest that longer rinse contact
time is not needed at higher Re level.
A one-way ANOVA with a 95% confidence interval was used to
analyze the difference in the means of percent residual film after 20
and 60 s rinsing over the range of Reynolds numbers. Similar to the
results presented in Fig. 1, higher percentages of residual film were
removed in experiments conducted at higher Reynolds numbers. It
was observed that when the Reynolds Number was equal to or
below 87,380, the effectiveness of rinsing was always significantly
improved at longer contact times. Extending the contact time (from
20 s to 60 s) at higher Re did not significantly influence the rinse
effectiveness. This suggests Reynolds number, rather than contact
time, was the dominant factor impacting rinse effectiveness.
Therefore it follows that longer rinse contact times are not needed
at higher Re levels.
This result is consistent with literature. It has been previously
suggested that higher turbulence plays a key role in affecting the
mass transfer during the initial seconds of contact (Weidemann
et al., 2013). According to Goode et al. (2013), there is a point
during rinse where no further deposit can be removed by water
alone and the remaining deposit can only be removed after
chemical agents are introduced. It was further stated that all water-
removable deposit is removed within a short contact time in a high
Re system due to the strong physical force introduced by the tur-
bulence. The remaining deposit is not removable by rinse water,
which explains the observation from this study that extended
contact times do not improve rinse effectiveness. Thus under a high
level of rinse water turbulence, usually measured by a high level of
Reynolds Number, extending the rinse water contact time is un-
necessary and cannot improve the rinse effectiveness significantly.
3.2. Combined effect of temperature and contact time on rinse
effectiveness
Under experimental conditions stated in materials and methods
section, the coefficients of natural log of residual film percentage as
a function of Reynolds Number were determined at various contact
times and temperatures, similar to the regression presented in
Fig. 1. These slope and intercept coefficients are reported in Table 1.
Higher slope and/or lower intercept of these regressions generally
implies a better rinse effectiveness. The influence of contact time
and rinse water temperature on rinse effectiveness was evaluated
by conducting “analysis of the correlation coefficients” (ANCOVA).
The analysis of covariance (ANCOVA) was used to compare the
slopes and intercepts of each regression (a ¼ 0.05) and the results
were presented in Table 1.
The results in Table 1 indicated that the differences among
279
slopes of different contact times at the same rinse water temper-
ature were not significant. Since the slopes were not significantly
different, the intercepts of the regression expressions were iden-
± 0.14a
tified as an indicator of the rinse effectiveness. Lower values of the
± 0.21b intercept suggests the presence of less residual film on the pipe
± 0.18b interior surface after rinsing, which meant achievement of better
± 0.30c rinse effectiveness.
± 0.25c
The impact of contact time on the rinsing effectiveness was
± 0.21c
observed in Table 1. The intercept coefficient after 60 s rinsing at
22
C was significantly lower than that after 20 s of rinsing at the
same temperature. This indicates that a 60-s rinse at 22
C is
significantly more effective than the 20-s rinse at the same tem-
perature. A similar trend was observed on comparison of the in-
tercepts after 20 and 60 s of rinsing at 45
C. However, when the
rinse water temperature was increased to 67
C, the difference of
intercepts between the 20 and 60 s contact times was not signifi-
cant. These results suggest that rinse effectiveness can be improved
by rinsing for longer periods at 22 and 45
C, but no significant
improvement is achieved when the water temperature is 67
C. It is
possible that 20 s of contact was sufficient to remove all the water-
soluble deposit due to the higher diffusivity at elevated tempera-
tures (67
C). Therefore, when the contact time was increased from
20 to 60 s at 67
C, the rinse effectiveness was not improved. The
portion of the deposit that was easily soluble in water was removed
after 20 s of contact with rinse water and the portion of the deposit
attached more tightly to the wall could not be removed even after
60 s of rinsing. As previously suggested, the deposited film cannot
be removed by rinse water only, and must be removed by the
cleaning cycle of the CIP operation (Goode et al., 2013).
The impact of rinse water temperature on the rinsing effec-
tiveness is also revealed in Table 1. The slopes of regressions from
experiments when rinse water was at 22
C as compared to rinse
water at 45
C were significantly different for both 20 s pre-rinse
and 60 s pre-rinse. These observations suggest that the rinse
effectiveness was improved by increasing the rinse water temper-
ature from 22 to 45
C. Otto et al. (2014) reported a similar result
with an approximately 37% increase in cleaning rate of whey pro-
tein deposit when the temperature of the cleaning fluid was
increased from 5 to 40
C. These observations indicate that tem-
perature of the rinse water impacts the diffusivity of water within
the deposited film and consequently, influences the rinse water
effectiveness. This is consistent with common observations about
water diffusion into porous media (Rastogi et al., 2002). Higher
temperatures are expected to enhance the diffusivity of water into a
porous media which in this case is the dairy deposit film. Other
researchers have also reported the increase in mass transfer at the
interface between rinse water and deposit film at higher temper-
atures (Mercade -Prieto and Chen, 2006; Xin et al., 2002). These
studies concluded that high temperature water enhances rinsing
through better mixing and more rapid exchange between film
surface and water flowing over the surface (Gillham et al., 1999).
Interestingly, compared to the significant benefits of increasing
the rinse water temperature from room temperature to an inter-
mediate temperature of 45
C on deposit removal, increasing the
water temperature to the recommended CIP temperature of 67
C
did not significantly improve the slope of the regression (Table 1).
In other words, the difference between regression slopes for rinse
water at 45
C and 67
C was not statistically significant. However,
the intercept for a 20-s rinse at 45
C was found to be significantly
higher than the following-rinsing longer at the same temperature,
20 s at 67
C, and 60 s at 67
C. As mentioned previously, lower
values of the intercept indicate lower RF% or better rinse effec-
tiveness. This suggests that rinse effectiveness was improved by
longer contact times at 45
C. Furthermore, rinse effectiveness was
not influenced by an increase in temperature from 45 to 67
C when
280 M. Fan et al. / Journal of Food Engineering 222 (2018) 276e283

3.5

3
Ln (Residual Film%)
2.5

1.5
Ln[RF%] = 3.28-0.99 [WV]
1 R² = 0.77
0.5

0
0 0.5 1 1.5
3 -2
Volume of Consumed Water/Fouled Surface Area, WV (m m )

Fig. 3. Improvements in removal of residual film deposits (RF%) with increased water use at rinse water temperature of 22
C and after 20-s rinse water residence time.

the contact time was 60 s. A possible explanation could be that 60 s
of rinsing is long enough to remove all the water-removable de-
posits regardless of the mass transfer rate which is impacted by the
rinse water temperature.
3.3. Effect of the used water volume on rinse effectiveness
In this section, the correlation between percent residual film
and the volume of water used per fouled surface area was evaluated
at three different temperatures (22
C, 45
C and 67
C) and two
different contact times. The range of water volumes used per fouled
surface area was from 0.18 to 3.54 m3 water per m2 fouled surface.
The percent residual film (RF%) after the 20 s of pre-rinse at
22
C as a function of the volume of water used per fouled area is
presented in Fig. 3. A linear correlation was found between the
natural logarithm of the residual film (RF%) and the water volume
(WV) per fouled surface area as given by Eq. (6).
Ln [RF%] ¼ 3.28e0.99 [WV]
Based on this regression, removal of 79.6% of the film residue
required only 0.27 m3 of water per m2 fouled surface. However,
removal of an additional 12.7% of the film residue would require
1.24 m3 of water per m2 fouled surface. These results indicate that
the effectiveness of rinse water decreased as the level of RF%
became smaller. It must be emphasized that relatively small vol-
umes of rinse water (0.18 m3 of water per m2 fouled surface) were
required to remove significant amounts (79.6%) of the residue film.
The decrease in effectiveness of the rinse water as larger volumes
are used suggests that reducing the duration of the rinse cycle
would provide an opportunity for significant reduction in water
effectiveness rinsing due to higher mechanical forces, along with
better mixing and transfer at the interface between the residue film
and the rinse water.
Similar to the linear correlation obtained for 22
C and a 20-s
rinse in Fig. 3, a linear correlation was found between the natural
logarithm of RF% and volume of water (WV) per fouled surface at
67
C after a 60-s rinse (Eq. (7))
Ln [RF%] ¼ 2.66e0.3 [WV] (7)
Based on this regression, only 0.38 m3 of water per m2 fouled
surface was needed to remove 87.3% of the residue film. An addi-
tional 3.7 m3 of water per m2 fouled surface was required to remove
and additional 8.5% of residue film. This indicates that the water
cost increases by nearly 10 times for an improvement in residual
removal from 87.3% to 95.8%. When compared to the film deposit
removed after 20 s of rinse with water at 22
C, improved rinse
effectiveness was achieved with higher water temperatures and
(6) longer contact times.
Similar regressions were developed for all combinations of
contact times and rinse water temperatures. Additional analysis
was accomplished by dividing the regression coefficients into two
groups based on contact times. An analysis of covariance (ANCOVA)
was used to compare the slopes and intercepts from each regres-
sion (a ¼ 0.05) and the results are presented in Table 2. The results
of the analysis indicated that the slopes of the linear regressions
Table 2
ANCOVA results for coefficient for regressions of ln (residual film%) over volume of
consumed water per unit fouled surface after 20 and 60 s of pre-rinse.

use. Contact time (s) Temperature (
C) Slope Intercept

For a fixed CIP system, the volume of consumed water is a
20 22 0.99 3.28 ± 0.12a
dependent on the flow velocity and the contact time between the 45 0.99a 3.01 ± 0.10b
rinse water and the residue film. Larger amounts of water are used 67 0.89a 2.65 ± 0.13c
when the water velocities are higher. The water consumption 60 22 0.31a 3.09 ± 0.19a
values presented on the horizontal axis were achieved by the 45 0.32a 2.85 ± 0.14b
different flow velocities as illustrated in Fig. 3. Previous researchers 67 0.30a 2.66 ± 0.12b
(Gillham et al., 1999; Weidemann et al., 2013; Xin et al., 2002) have Note- Values with the same superscript letter are not significantly different
illustrated that higher velocities of rinse water can improve (a ¼ 0.05).
 M. Fan et al. / Journal of Food Engineering 222 (2018) 276e283 281

0.7

Area, WE (kg protein removed in m-3

0.6

Water Efficiency per Fouled Surface

0.5

0.4

rinse water)
[WE] = 0.0495 [WV] -1.41
0.3 R² = 0.83

0.2

0.1

0
0 1 2 3 4
Volume of Consumed Water/Fouled Surface Area, WV (m3 m-2)

Fig. 4. The influence of rinse water volume on the efficiency of residual film removal.

were not significantly different for both contact times and tem-
peratures. For the 20-s contact time, the intercept of the linear
regression decreased significantly when the rinse water tempera-
ture was increased from 22 to 45
C and from 45 to 67
C. These
results suggest that an increase in rinse water temperature from 22
to 67
C provided an improvement in the effectiveness of rinse
water at the shorter contact time of 20 s. The same improvement
was not found for the 60 s pre-rinse. It was observed that there was
no significant difference between intercepts of the linear re-
gressions when the temperature was increased from 45 to 67
C.
These results suggest that higher temperatures (67
C) do not
improve rinse effectiveness when the contact times are relatively
long (60 s).
Based on the results and discussions in this section, it was
concluded that a pre-rinse of 60 s at 67
C is excessive. At these
conditions, no significant improvement occurred in deposit
removal. Moreover, these conditions required high energy and time
inputs. Pre-rinse for 20 s at 45
C would be an appropriate balance
between rinse effectiveness and energy and time input. A more
specific consideration about water cost and rinse effectiveness is
discussed in next section.
3.4. Changes of rinse water efficiency during the pre-rinse step
In this study, water efficiency (WE) has been defined as the mass
of protein film removed per fouled surface area over volume of
rinsing water per fouled surface area (Eq. (3)). Previous research
has indicated that the residue removal rate during the rinse step is
not constant (Xin et al., 2002; Weidemann et al., 2013), and the
water efficiency (WE) is not expected to be constant. Based on re-
sults from this study, as well as observations from other researchers
(Arnold and Maxcy, 1970; Harper, 1972), longer contact times may
not improve rinse effectiveness when the water temperatures are
high. Optimization of the water rinse step may be achieved by
knowing the relationship between water efficiency and volume of
used water.
The results in Fig. 4 illustrate the relationship between water
efficiency (WE) and water volume (WV) at 22
C, both WE and WV
mentioned here are normalized to per unit of fouled surface area.
When the volume of water used was lower, the water efficiency
was very high. As the water consumption increased, the efficiency
decreased rapidly. One interpretation of this relationship is that
during the initial phases of rinsing, the amount of residue removed
per unit of water used is largest. Obviously, the initial contact be-
tween water and deposit film is when the efficiency of the rinse
water is highest. When the quantities of water used are greater
than 1 m3 m2, the water efficiency decreased until it reached a
relatively low and constant level. Even though the magnitudes of
WE continued to be positive, the small WE values may not justify
the large quantities of water required (WV) as we move to the re-
gion in the right of this curve (Fig. 4). In this region, each increment
of deposit removed required a significantly greater amount of
water.
These results indicate that the rinse water will continue to
remove film deposits from the pipe surfaces after the water effi-
ciency reach a constant level. The relationship describing the water
efficiency as a function of water volume used, as given in Eq. (8),
must be considered when attempting to optimize the entire
cleaning operation.
1.41
[WE] ¼ 0.0495 [WV] (8)
In order to evaluate the impact of temperature on water effi-
ciency, the relationships between water efficiency and water vol-
ume at various temperatures were investigated. In order to change
Eq. (8) into a linear relationship, the natural logarithm was applied
to both water efficiency and volume of water. As reported in Table 3,
after the conversion, the natural logarithm of water efficiency
decreased in a linear fashion (R2 ¼ 0.83) as the nature logarithm of
water volume increased. The results indicate that the water effi-
ciency decreased rapidly with an increase in water volume at all
temperatures investigated.
By comparing the slopes and intercepts of each linear regression
Table 3
ANCOVA results on regressions between ln (water efficiency) and ln (water volume/
fouled surface area).
Temperature (
C) Slope Intercept R2
22 1.41a 3.01 ± 0.012a 0.83
45 1.45b 2.94 ± 0.013b 0.83
67 1.45b 2.97 ± 0.014b 0.84
Note- Values with the same superscript letter are not significantly different
(a ¼ 0.05).
282 M. Fan et al. / Journal of Food Engineering 222 (2018) 276e283
(p ¼ 0.05) using ANCOVA, the impact of water temperature on
water efficiency was analyzed (Table 3). The regression parameters
for 22
C rinse water were significantly lower when compared to
parameters for the linear regression at 45
C and 67
C. These re-
sults suggest that at a lower temperature (22
C), the water effi-
ciency is low. When the temperature was increased from 45 to
67
C, the differences in regression coefficients were not significant.
This suggests that an increase in rinse water temperature from 45
to 67
C does not improve water efficiency. The observation tends to
confirm earlier results about the effect of rinse water temperature
on the effectiveness of rinsing.
3.5. Overall observations
Rinse effectiveness is not the only factor required during opti-
mization of water rinse step. Water efficiency may also be consid-
ered since it has a huge impact on the conservation of water for the
entire cleaning process. According to results of the current study,
the water efficiency decreases rapidly as more water is used
regardless of the change in temperature.
A balance between rinse effectiveness and water efficiency is
required for CIP optimization. A CIP rinse water regime targeting
higher rinse effectiveness regardless of the lower water efficiency is
not recommended. There are limited benefits from using additional
amounts of water during pre-rinse to remove strongly attached
deposits, as any residual deposit can be removed by the detergent
and acid wash steps to follow.
4. Conclusions
The results of this investigation confirm the feasibility of opti-
mizing water usage during CIP operations through the selection of
control parameters. Effectiveness of rinse water in deposit removal
was 81.4% at Reynolds Numbers of 20,000, and improved to only
91.6% at a Reynolds number of 100,000; an improvement of only
10% in effectiveness.
Rinse water effectiveness was not significantly improved when
the rinse water temperature was increased from 45 to 67
C. At a
rinse water temperature of 67
C, an increase in contact time from
20 to 60 s did not improve the rinse water effectiveness signifi-
cantly. At 67
C, only 1.01 m3 of water per m2 fouled surface were
required to remove 89% of the residue film from the surface. The
improvement of residual removal from 89 to 95% required 3.64 m3
water per m2 of fouled surface. It seems evident that significant
reduction in rinse water use (72%) can be achieved by reducing the
amount of residue removed during the initial rinse water by only a
few percent. When the efficiency of water use is expressed as the
quantity of residue removed as a function of water volume per unit
area of fouled surface, the water efficiency decreased from
0.017 kg m3 at limited amounts of water to approximately
0.001 kg m3 when the amounts of rinse water were very high. The
efficiency of rinse water use decreased rapidly with rinse time to
about one cubic meter of water per square meter of fouled surface.
A rinse water step at 45
C for a residence time of 20 s was more
efficient than 65
C for 60 s. In general, the efficiency in deposit
removal is not improved significantly at volumes of water used
above 1 m3 m2 of fouled surface.
Acknowledgements
The authors wish to acknowledge the generous donations from
Dale A. Seiberling to the Food Engineering Research Laboratory at
The Ohio State University. This work was also supported in part by
the USDA National Institute of Food and Agriculture, Hatch project
232768.
References
Arnold, R., Maxcy, R., 1970. Evaluating circulation cleaning by analysis of soil
depletion from surfaces. J. Dairy Sci. 53 (11), 1540e1544.
Augustin, W., Fuchs, T., Fo €ste, H., Scho €ler, M., Majschak, J.P., Scholl, S., 2010. Pulsed
flow for enhanced cleaning in food processing. Food Bioprod. Process. 88 (4),
384e391.
Blel, W., Legentilhomme, P., Benezech, T., Legrand, J., Le Gentil-Lelievre, C., 2009.
Application of turbulent pulsating flows to the bacterial removal during a
cleaning in place procedure. Part 2: effects on cleaning efficiency. J. Food Eng. 90
(4), 433e440.
Boxler, C., Augustin, W., Scholl, S., 2013. Cleaning of whey protein and milk salts
soiled on DLC coated surfaces at high-temperature. J. Food Eng. 114 (1), 29e38.
Christian, G.K., Fryer, P.J., 2006. The effect of pulsing cleaning chemicals on the
cleaning of whey protein deposits. Food Bioprod. Process. 84 (4), 320e328.
Dev, S.R., Demirci, A., Graves, R.E., Puri, V.M., 2014. Optimization and modeling of an
electrolyzed oxidizing water based Clean-In-Place technique for farm milking
systems using a pilot-scale milking system. J. Food Eng. 135, 1e10.
Fan, M., Phinney, D.M., Heldman, D.R., 2015. Effectiveness of rinse water during in-
place cleaning of stainless steel pipe lines. J. Food Sci. 80 (7), E1490eE1497.
https://doi.org/10.1111/1750-3841.12914.
Fryer, P.J., Robbins, P.T., 2005. Heat transfer in food processing: ensuring product
quality and safety. Appl. Therm. Eng. 25 (16), 2499e2510.
Fryer, P.J., Christian, G.K., Liu, W., 2006. How hygiene happens: physics and chem-
istry of cleaning. Int. J. Dairy Technol. 59 (2), 76e84.
Gallot-Lavallee, T.H.I.E.R.R.Y., Lalande, M., Corrieu, G., 1984. Cleaning kinetics
modeling of holding tubes fouled during milk pasteurization. J. Food Process
Eng. 7 (2), 123e142. https://doi.org/10.1111/j.1745-4530.1984.tb00642.x.
Gillham, C.R., Fryer, P.J., Hasting, A.P.M., Wilson, D.I., 1999. Cleaning-in-place of
whey protein fouling deposits: mechanisms controlling cleaning. Food Bioprod.
Process. 77 (2), 127e136.
Gillham, C.R., Fryer, P.J., Hasting, A.P.M., Wilson, D.I., 2000. Enhanced cleaning of
whey protein soils using pulsed flows. J. Food Eng. 46 (3), 199e209.
Goode, K.R., Asteriadou, K., Robbins, P.T., Fryer, P.J., 2013. Fouling and cleaning
studies in the food and beverage industry classified by cleaning type. Compr.
Rev. Food Sci. food Saf. 12 (2), 121e143. https://doi.org/10.1111/1541-
4337.12000.
Grant, C.S., Webb, G.E., Jeon, Y.W., 1997. A noninvasive study of milk cleaning pro-
cesses: calcium phosphate removal. J. food process Eng. 20 (3), 197e230.
Guzel-Seydim, Z.B., Wyffels, J.T., Greene, A.K., Bodine, A.B., 2000. Removal of dairy
soil from heated stainless steel surfaces: use of ozonated water as a Prerinse1.
J. Dairy Sci. 83 (8), 1887e1891.
Hankinson, D.J., Carver, C.E., 1968. Fluid dynamic relationships involved in circu-
lation Cleaning1. J. dairy Sci. 51 (11), 1761e1767. https://doi.org/10.3168/
jds.S0022-0302(68)87273-X.
Harper, W.J., 1972. Sanitation in dairy food plants. In: Gothie, R.K. (Ed.), Food Plant
Sanitation. Avi Pub. Co., Westport, USA, pp. 130e160.
ne
Jensen, B.B.B., Friis, A., Be zech, T., Legentilhomme, P., Lelievre, C., 2005. Local wall
shear stress variations predicted by computational fluid dynamics for hygienic
design. Food Bioprod. Process. 83 (1), 53e60. https://doi.org/10.1205/fbp.04021.
Khaldi, M., Blanpain-Avet, P., Gue rin, R., Ronse, G., Bouvier, L., Andre , C.,
Delaplace, G., 2015. Effect of calcium content and flow regime on whey protein
fouling and cleaning in a plate heat exchanger. J. Food Eng. 147, 68e78.
Kulkarni, S.M., Maxcy, R.B., Arnold, R.G., 1975. Evaluation of soil deposition and
removal processes: an interpretive Review1. J. Dairy Sci. 58 (12), 1922e1936.

vre, C., Legentilhomme, P., Gaucher, C., Legrand, J., Faille, C., Be
Lelie nezech, T., 2002.
Cleaning in place: effect of local wall shear stress variation on bacterial removal
from stainless steel equipment. Chem. Eng. Sci. 57 (8), 1287e1297. https://
doi.org/10.1016/S0009-2509(02)00019-2.
Li, L., Che, L., Liang, Q., Mercade -Prieto, R., Wu, X., Chen, X.D., 2013. Study on the
dissolution of heat-induced ovalbumin gel in alkaline solutions relevant to the
removal of fouling deposits. J. Food Eng. 114 (4), 550e557.
Liu, W., Zhang, Z., Fryer, P.J., 2006. Identification and modelling of different removal
modes in the cleaning of a model food deposit. Chem. Eng. Sci. 61 (22),
7528e7534.
Lyndgaard, C.B., Rasmussen, M.A., Engelsen, S.B., Thaysen, D., van den Berg, F., 2014.
Moving from recipe-driven to measurement-based cleaning procedures:
monitoring the Cleaning-In-Place process of whey filtration units by ultraviolet
spectroscopy and chemometrics. J. Food Eng. 126, 82e88. https://doi.org/
10.1016/j.jfoodeng.2013.10.037.
Mattila, T., Manninen, M., Kyla €siurola, A.L., 1990. Effect of cleaning-in-place disin-
fectants on wild bacterial strains isolated from a milking line. J. dairy Res. 57
(01), 33e39. https://doi.org/10.1017/S0022029900026583.
Mercade -Prieto, R., Chen, X.D., 2006. Dissolution of whey protein concentrate gels
in alkali. AIChE J. 52 (2), 792e803. https://doi.org/10.1002/aic.10639.
Meyer, R.E., 1971. Introduction to Mathematical Fluid Dynamics, vol. 24. Courier
Corporation, North Chelmsford, 289e205.
Muthukumaran, S., Yang, K., Seuren, A., Kentish, S., Ashokkumar, M., Stevens, G.W.,
Grieser, F., 2004. The use of ultrasonic cleaning for ultrafiltration membranes in
the dairy industry. Sep. Purif. Technol. 39 (1), 99e107. https://doi.org/10.1016/
j.seppur.2003.12.013.
Otto, C., Zahn, S., Plenker, J., Rohm, H., 2014. Application of a flow cell for the
comparative investigation of the cleaning behavior of starch and protein. J. Food
Eng. 131, 1e6.
 M. Fan et al. / Journal of Food Engineering 222 (2018) 276e283
Rastogi, N.K., Raghavarao, K.S.M.S., Niranjan, K., Knorr, D., 2002. Recent de-
velopments in osmotic dehydration: methods to enhance mass transfer. Trends
Food Sci. Technol. 13 (2), 48e59.
Rosmaninho, R., Santos, O., Nylander, T., Paulsson, M., Beuf, M., Benezech, T., Müller-
Steinhagen, H., 2007. Modified stainless steel surfaces targeted to reduce fou-
lingeevaluation of fouling by milk components. J. Food Eng. 80 (4), 1176e1187.
Seiberling, D.A., 1968. Equipment and process design as related to mechanical/
chemical cleaning procedures. In: Chem. Eng. Prog., Symp. Ser, vol. 64, p. 86.
Sharma, M.M., Chamoun, H., Sarma, D.S.R., Schechter, R.S., 1992. Factors controlling
the hydrodynamic detachment of particles from surfaces. J. Colloid Interface Sci.
149 (1), 121e134. https://doi.org/10.1016/0021-9797(92)90398-6.
Singh, R.P., Heldman, D.R., 2013. Introduction to Food Engineering, fifth ed. Gulf
Professional Publishing, pp. 80e95.
Timperley, D.A., Smeulders, C.N.M., 1987. Cleaning of dairy HTST plate heat ex-
changers: comparison of single-and two-stage procedures. Int. J. Dairy Technol.
283
40 (1), 4e7.
Visser, J., 1970. Measurement of the force of adhesion between submicron carbon-
black particles and a cellulose film in aqueous solution. J. Colloid Interface Sci.
34 (1), 26e31. https://doi.org/10.1016/0021-9797(70)90254-7.
Weidemann, C., Stahl, S., Nirschl, H., 2013. Development of a qualitative test method
for the cleanability of polymer woven filter media. Food Bioprod. Process. 91
(4), 515e524. https://doi.org/10.1016/j.fbp.2013.05.005.
Weidemann, C., Vogt, S., Nirschl, H., 2014. Cleaning of filter media by pulsed
floweEstablishment of dimensionless operation numbers describing the
cleaning result. J. Food Eng. 132, 29e38.
Wilson, D.I., 2005. Challenges in cleaning: recent developments and future pros-
pects. Heat. Transf. Eng. 26 (1), 51e59. https://doi.org/10.1080/
01457630590890175.
€
Xin, H., Chen, X.D., Ozkan, N., 2002. Cleaning rate in the uniform cleaning stage for
whey protein gel deposits. Food Bioprod. Process. 80 (4), 240e246.