Research Progress on Corn Starch Wastewater
Treatment Process
ZHANG Chunyang1, ZHANG Congju1, LIU Jianguang2, HAN Lichao2
1. School of Thermal Energy Engineering; Shandong Jianzhu University; Jinan 250101 China
2. School of Municipal and Environmental Engineering, Shandong Jianzhu University, Jinan, 250101,China;
Abstract: The reclamation approaches and treatment
technology process of corn starch wastewater were re-
viewed. contamination Characteristic of corn starch
wastewater were concentration high and biodegradable,
and the reclamation potential was huge. Treatment tech-
nologies included biological treatment, physical chemistry
treatment, biological pond and photosynthetic bacteria
process. The optimized combine treatment process of re-
source reclamation from corn starch wastewater and treat-
ment technology was recommended.
Keywords: anaerobic, wastewater treatment, Corn Starch
Wastewater, reclamation
1 Introduction
Starch is a kind of important industrial raw materials and
widely used in the food, chemistry, textile, pharmaceutical
industry, etc. In the production and processing process of
starch, a large amount of high concentrated organic
wastewater is discharged, which mainly contains organic
matters such as starch, fiber, protein, etc. Once these or-
ganic matters are discharged into the water, it will lead to
severe water pollution and bring repercussion to the sur-
vival environment of human beings. Therefore, starch
wastewater should be processed to meet the standard be-
fore discharging. Further, the high concentration of organic
matter in the water has revealed significant resource poten-
tial.
2 Quality of the corn starch wastewater
The wastewater generated by the production of corn starch
can be divided into two parts, namely the pickling liquid
and process water. The former contains high content of
organic matters, mostly protein, but little water; and the
latter is generated by the whole production process of
broking corn, removing embryo and drying starch and
___________________________________
978-1-4244-9577-1/11/$26.00 ©2011 IEEE

Authorized licensed use limited to: Queensland University of Technology. Downloaded on August 15,2020 at 04:37:27 UTC from IEEE Xplore. Restrictions apply.
contains a lot of water which is about 4-5 times of the for-
mer. The pickling liquid is high concentrated organic
wastewater with the characteristic of four high and one low,
namely high CODCr (8,000 ~ 15000mg/L), high suspended
solids (SS) (1000 ~ 3000mg/L), high total nitrogen value
(240 ~ 540mg/L), high concentration of phosphate (in P
plan, about 15 ~ 130mg/L), and low pH (4.2 -5). The proc-
ess water is medium concentrated organic wastewater with
CODCr value in 2000 ~ 3500mg/L, not high ammonia ni-
trogen and phosphate concentration, respectively 20mg/L
and more than 14 ~ 32mg/L.
From the analysis of the above data, it can be concluded
that: (1) Corn starch wastewater is rich in carbohydrates
and nitrogen, phosphorus nutrients, belonging to better
biochemical high concentrated organic wastewater, which
is suitable for high concentration organic wastewater
treatment by biochemical process. (2) Wastewater sus-
pended solids and the colloid protein content is relatively
high, which will produce adverse impact to the develop-
ment of the anaerobic activated sludge system. (3) Corn
immersion process will produce a small number SO32 -. In
anaerobic treatment process, these sulfur-containing com-
pounds are deoxidized to hydrogen sulphide by microor-
ganisms, which is likely to produce certain inhibition on
the anaerobic system.
3 Corn starch wastewater treatment technology situa-
tion
3.1 Biological treatment method
Biological treatment method is to use the effect of the mi-
crobial metabolism, which degrades and converts the dis-
solution and colloid organic pollutants into harmless mate-
rial. The method of purifying wastewater can generally be
divided into anaerobic biological treatment and aerobic
biological treatment. Due to the characteristics of high or-
ganic matter in starch wastewater and difficulty in the
 
process, single biological treatment method is rarely used
and is usually combined to other kinds of biological treat-
ments. By this way, it can make its advantages and disad-
vantages to supplement each other and improve the effi-
ciency.
3.1.1 Anaerobic-aerobic combination
Anaerobic biological treatment can decompose a large
amount of high concentration organic compounds and
produce methane, while aerobic biological treatment is
used as the follow-up process to further remove the resid-
ual organic matters. The combination of these two methods
has a higher efficiency thereby applying in the practical
production process as a whole subject.
Wang Rongmin etc used the three-phase anaerobic-aerobic
one-piece baffle plate bioreactor and added inorganic
polymer abandoned packing rubber as the adherent grow-
ing packing for the aerobic microbes. According to the
result, in temperature between 25 to 35 ć, pH 5.0 ~ 850
and the three-phase anaerobic-aerobic one-piece baffle
plate bioreactor effluent 200mg/L COD concentration be-
low, COD total removal rate topped 98.7%;Effluent am-
monia nitrogen in 10mg/L or so, ammonia nitrogen re-
moval topped 82.3%, which makes the effluent water
standard.
Yang Kaiming and Yang Xiaolin etc introduced a UBF -
CASS combination process to handle starch wastewater in
the northwestern starch company. CODCr tanking water is
reduced by 12000mg/L to 125mg/L, NH3 - N is reduced
by the original 160mg/L to 20mg/L, BOD5 and SS re-
moval rates reached 99.4% and 87% respectively, which
makes the effluent water to achieve the secondary standard.
The system has high impact resistance for strong capability
of load, stable operation effect in biochemical treatment
unit. Moreover, the generated biogas and protein feed all
can have certain economic benefits.
Shi Hui etc used EGSB - A/O combination process to han-
dle high concentration starch wastewater in a starch com-
pany in Si Ping City. The result indicates that the process-
ing high concentration starch wastewater by methods is
completely feasible. When COD for wastewater is
8000~10000mg/L, the COD for processed wastewater is
less than 100mg/L; the reactor COD is stable in 20kg /

Authorized licensed use limited to: Queensland University of Technology. Downloaded on August 15,2020 at 04:37:27 UTC from IEEE Xplore. Restrictions apply.
(m3·d) or so; the combination process on COD, BOD5, SS
and NH3 - N can reach removal rates 99.1%, 99.6%,82%
and 96% respectively, which reaches the national sewage
comprehensive level 1 emission standard.
3.1.2 Anaerobic biological treatment method
The main processes in anaerobic biological treatment of
handling starch wastewater have ShengLiuShi anaerobic
sludge bed (UASB), anaerobic baffle plate reactor (ABR),
anaerobic fluidized bed (AFB), anaerobic contact method
(ACP), two-phase anaerobic digestion method (TPAD) and
anaerobic filter (AF), etc.
ShengLiuShi anaerobic sludge bed is a kind of anaerobic
reactor based on mature theory. It is developed success-
fully by Lettinga in Wageningen agricultural university in
the Netherlands in 1977. It has a lot of merits, such as high
volume load rate and high sludge load rate etc and it is
widely used in engineering. Li Yafeng etc introduced re-
actor UASB treatment effect used for starch wastewater
treatment in a Shandong company. This company adopted
2 UASB reactors whose diameter was 20m and height was
6.8m in starch wastewater anaerobic biological treatment.
Meanwhile, COD volume load was 8kg / (m3 · for d) with
stable treatment effect and effluent COD, BOD5, SS re-
moval rates were 83%, 90%, 62% respectively, which sat-
isfied the follow-up process requirements. Zhang Chunyan
analyzed the starch wastewater treatment process in Shan-
dong Qingyuan Food Company and proved that it was
economic and effective to use UASB to process high qual-
ity concentration corn starch wastewater, which operation
cost was low and the treatment effect was stable. It further
determined that the best temperature for the UASB treat-
ment was 35 ~ 40 ć; pH value of 6 ~ 7, CODCr reached
8kg / (d. m3) design load and the removal rate reached
above 90%.
Anaerobic baffle plate reactor used built-in vertical guide
plate to divide the reaction chambers into several relative
independent ShangLiuShi sludge bed systems. Several
small reaction chambers were installed in series, which
makes the processed wastewater flow up and down inter-
nally along its baffle plate. In the mixing effect of the flow
and the producing gases, organic matters in the wastewater
contacted with anaerobic sludge repeatedly and ware able
 
25h, COD removal rate could reach amount to 77% ~ 96%.
Both of them have obvious treatment effect.
3.2.2 Floating processing method
Floating processing uses high pressure to dissolve a large
amount of vapor gas as working liquid. After sudden de-
compression, it will release countless small bubbles, which
makes the flocculate in the wastewater stick to their surface.
Thus, the proportion of the flocculate is far less than the
actual proportion. Along with the rising of the bubbles, the
flocculate will float to the liquid surface, thereby achieving
the purpose of liquid-solid separation. Mu Jianbo etc took
the wastewater in a Hubei starch company and adopted an
air floating device to conduct an experimental research.
The wastewater infusing drug went through the pump into
the integration device. The micro bubble produced by the
dissolved gas water contacted with the counter-current
wastewater. Floc properties were clinging to the micro
bubble and ejected along with the rising, the disposed wa-
ter below the column went out through the liquid level
control device. According to the experiment and the analy-
sis of the flocculant, floating agent and the operation pa-
rameters on the treatment effect, it is concluded that the
best operating conditions can be achieved in feeding posi-
tion 70cm, air inflow 120L/h, feeding amount 100ml/min
and liquid surface height 127cm.
3.3 Other processing methods
3.3.1 Biological pond method
Biological pond technology uses the natural purification
capacity of water to handle wastewater. This technology
was developed rapidly after 1950s and mostly adopted in
sewage and organic industrial wastewater. According to
the characteristics of high organic content and rich nutri-
tion in starch wastewater, the combination of anaerobic
pond, facultative pond and aerobic pond is applied. Yang
Fengjiang etc designed a method to handle starch waste-
water with ShuiHuLu and XiLuPing based on the water
characteristic of Xinmin Starch Company in Liaoning
military logistic department. The precipitated wastewater
was ejected into the natural oxidation pond to have natural
fermentation. Then, it was ejected into the ShuiHuLu pond
to be purified by 7d and was ejected into the XILuPing to
be purified by 7d again. All these procedures made the ef-

Authorized licensed use limited to: Queensland University of Technology. Downloaded on August 15,2020 at 04:37:27 UTC from IEEE Xplore. Restrictions apply.
fluent water to achieve farm irrigation water quality. Prac-
tice has proved that through the metabolic process which
contained 4 degree of wastewater handling, 3 steps of
multi-function and multi-system, this method can achieve
all material energy conversion and recycle. Thus, it has
obvious popularized value.
3.3.2 Photosynthetic bacteria method
Photosynthetic bacteria (PSB) are mainly red pseudomonas
genera, which can conduct bad oxygen photosynthesis un-
der anaerobic condition. The organic matters are used as
the carbon sources and the hydrogen body for photosyn-
thesis, being decomposed and removed. Moreover, they are
able to withstand high concentration organic matters. Pho-
tosynthetic bacteria method is also good in removing con-
taminants, such as nitrogen and phosphorus which has
been widely applied to remove waste water in organic pol-
lutants. Wang Yuxin etc separated out a spherical red coli
L2 from the wastewater sludge in Shandong Wen Deng
Starch Company in order to handle the wastewater after
pretreatment. Under appropriate conditions, CODCr re-
moval rate could reach 95.7% which made the wastewater
to meet the effluent standard.
4 Corn starch wastewater recycling technology
There are a lot of researches at home and abroad using mi-
croorganism for starch wastewater resource disposal in-
cluding using starch wastewater to produce single cell pro-
tein, recycling protein for fodder, producing microbial
flocculants, producing polysaccharide, etc.
Li Suyu etc introduced research in a variety of microbial
purification corn starch wastewater synergy. The result
showed that per cubic meter of wastewater could be used
for producing feed additives SCP1.646 kg and could be
turn into resources of protein production. Meanwhile, mi-
crobial synergy still can make wastewater COD removal
rate more than 90% and the purified wastewater COD re-
moval under 300mg/L. Wang Yuanyuan etc used starch
wastewater to tame, and foster mixed microbial flocculants
to cultivate bacterium. Researching on the cultivated con-
dition, they found that in condition of COD concentration
was 4000mg/L; urea was the source of nitrogen; C: N: P =
100:5-2, training time was 42h, culture temperature was 30
ć, wave bed speed was 150r/min, the MBF - 17 floccula-
 
[23] LI Suyu, LI Yu, REN Jun, et al. Cooperative Effect of Varied Microor- ˖19̚23.
Wastewater[J]. China Water & Wastewater, 2007,23˄5˅
ganism in Purifying Maize Starch Wastewater[J]. Environmental Protec- ⥟ುುˈ⥟৥ϰˈ䰜Ꮰ.߽⫼⎔㉝ᑳ∈෍‫ݏ‬໡ড়ൟ㍂‫ࠖޱ‬ѻ⫳㦠ⷨお[J].
˄1˅
tion, 2003ˈ ˖22̚23. Ё೑㒭∈
ᴢ㋴⥝ˈᴢ⥝ˈӏ࿳ˈㄝ.໮⾡ᖂ⫳⠽‫㉝⎔㉇⥝࣪ޔ‬ᎹϮᑳ∈ⱘणৠᬜ
ᑨⷨお[J].⦃๗ֱᡸˈ2003ˈ ˄1˅ ˖22̚23.
[24] WANG Yunyun, WANG Xiangdong, CHEN Xi. Study on Culture of
Compound Microbial Flocculant-producing Bacteria Using Starch



Authorized licensed use limited to: Queensland University of Technology. Downloaded on August 15,2020 at 04:37:27 UTC from IEEE Xplore. Restrictions apply.
 1265 J. Fettig et al. | Treatment of tapioca starch wastewater Water Science & Technology | 68.6 | 2013

Convention on Climate Change (UNFCCC ). As a low- Table 1 | Composition of tapioca starch wastewater

cost solution, some middle-sized plants have recently started

This study (mean
to cover the first anaerobic pond with a synthetic canvas in Parameter values 2010–2012) Data from Mai (2006)
order to collect and utilize the biogas produced (Hoang
Conductivity 1,673 μS/cm –
). In addition, some small companies have constructed
pH 4.5a 4.0–4.2
anaerobic filters combined with aerobic post-treatment
TSS 1,700 mg/L 1,500–2,600 mg/L
tanks (Phuoc & Phuong ). Since there is no biogas collec-
tion, a major benefit of anaerobic processes is not utilized. COD (non-filtered) 11,800 mg/L 14,000–18,000 mg/L

The main objective of this study was to find out whether
a combination of technical and nature-based treatment pro-
cesses suited for small- and middle-sized companies can
meet the discharge requirements corresponding to 50 mg/L
biochemical oxygen demand (BOD5), 30 mg/L total nitro-
gen (total-N), 6 mg/L total phosphorus (total-P) and
0.1 mg/L total cyanide according to Vietnam Standard
TCVN 5945 class B (MONRE ), and whether it can
be operated reliably.
MATERIALS AND METHODS
In small companies the starch is separated by sedimentation,
while in larger plants centrifugal screen extractors are more
common. The latter separation process provides wastewater
with a higher fraction of dissolved organic substances and a
lower portion of particulate matter. In this study the pilot
plant was located at a company that applies centrifugation.
In Table 1 the composition of the wastewater investigated
is compared with data published by Mai ().
Accordingly, the wastewater undergoes acidification
caused by anaerobic micro-organisms in the pre-treatment
units. Although there is a certain amount of nutrients,
organic matter is the main component. A comparison of
non-filtered and filtered samples reveals that total suspended
solids (TSS) make up for about 25% of the chemical oxygen
demand (COD). Therefore, TSS removal prior to biological
treatment was considered an important element of the pro-
cess scheme.
The treatment concept developed includes physical pre-
treatment, anaerobic degradation of organic substances, and
aerobic post-treatment. The process scheme is shown in
Figure 1. As far as we know the specific combination of tech-
nical and nature-based processes is a novel approach for this
type of wastewater. A detailed description of the concept is
given elsewhere (Pick et al. ).
The pilot plant was designed to treat continuously up to
12 m3/d of wastewater. Before flowing into the plant, the
water passes through three buffer tanks with a total hydrau-
lic retention time of 3.5 h. In this stage pH decreases to 4.5
BOD5 6,900 mg/L 9,000–11,000 mg/L
Total phosphorus 71 mg/L –
Kjeldahl nitrogen 280 mg/L –
COD (filtered) 8,840 mg/L –
Total cyanide 22 mg/L 5.8–96 mg/L
a
After 3.5 h of microbial acidification. TSS: total suspended solids; COD: chemical oxygen
demand; BOD5: 5-day biochemical oxygen demand.
due to rapid microbial acidification. As a result, colloidal
organic matter flocculates without adding any chemicals
and the removal efficiency in the flotation stage is much
better than with fresh wastewater. After this effect had
been observed during the first months of operation, pH
was no longer adjusted in the neutralization unit.
Dissolved air flotation (DAF) was applied in order to
remove the major portion of TSS. An Aquatector® Micro-
float® unit (Enviplan Company, Germany) was operated at
a hydraulic surface load of 2.5–3.0 m/h. Since dosing of
polymeric flocculants prior to flotation improved TSS
removal only slightly, flocculants were not added during
regular operation.
The central treatment stage was an anaerobic process
(expanded granular sludge bed (EGSB) reactor, type
ANAFIT-AC, Hager þ Elsässer Company, Germany) which
converts organic matter into biogas. The performance of
the reactor largely depends on stable process conditions
and a low suspended solids loading. This was achieved by
an optimization of the upstream DAF process. The EGSB
W
reactor was operated at a temperature of 35 C and a hydrau-
lic load of 4–5 m/h. It was seeded with sludge from a
brewery wastewater treatment plant. pH was adjusted to
6.8 by adding sodium hydroxide. There was no need for
heating because raw wastewater temperatures were already
at the required level.
For post-treatment, a vertical flow constructed wetland
(VFCW) was designed. The hydraulic surface load of the
unit was about 30 L/(m2 · d) and the average organic surface
load corresponded to 72 g COD/(m2 · d). The effluent was
collected in a small basin which was the sampling point,
and discharged via a fluid tipper to a lagoon operated by
the tapioca starch company.
 1265 J. Fettig et al. | Treatment of tapioca starch wastewater Water Science & Technology | 68.6 | 2013

Convention on Climate Change (UNFCCC ). As a low- Table 1 | Composition of tapioca starch wastewater

cost solution, some middle-sized plants have recently started

This study (mean
to cover the first anaerobic pond with a synthetic canvas in Parameter values 2010–2012) Data from Mai (2006)
order to collect and utilize the biogas produced (Hoang
Conductivity 1,673 μS/cm –
). In addition, some small companies have constructed
pH 4.5a 4.0–4.2
anaerobic filters combined with aerobic post-treatment
TSS 1,700 mg/L 1,500–2,600 mg/L
tanks (Phuoc & Phuong ). Since there is no biogas collec-
tion, a major benefit of anaerobic processes is not utilized. COD (non-filtered) 11,800 mg/L 14,000–18,000 mg/L

The main objective of this study was to find out whether
a combination of technical and nature-based treatment pro-
cesses suited for small- and middle-sized companies can
meet the discharge requirements corresponding to 50 mg/L
biochemical oxygen demand (BOD5), 30 mg/L total nitro-
gen (total-N), 6 mg/L total phosphorus (total-P) and
0.1 mg/L total cyanide according to Vietnam Standard
TCVN 5945 class B (MONRE ), and whether it can
be operated reliably.
MATERIALS AND METHODS
In small companies the starch is separated by sedimentation,
while in larger plants centrifugal screen extractors are more
common. The latter separation process provides wastewater
with a higher fraction of dissolved organic substances and a
lower portion of particulate matter. In this study the pilot
plant was located at a company that applies centrifugation.
In Table 1 the composition of the wastewater investigated
is compared with data published by Mai ().
Accordingly, the wastewater undergoes acidification
caused by anaerobic micro-organisms in the pre-treatment
units. Although there is a certain amount of nutrients,
organic matter is the main component. A comparison of
non-filtered and filtered samples reveals that total suspended
solids (TSS) make up for about 25% of the chemical oxygen
demand (COD). Therefore, TSS removal prior to biological
treatment was considered an important element of the pro-
cess scheme.
The treatment concept developed includes physical pre-
treatment, anaerobic degradation of organic substances, and
aerobic post-treatment. The process scheme is shown in
Figure 1. As far as we know the specific combination of tech-
nical and nature-based processes is a novel approach for this
type of wastewater. A detailed description of the concept is
given elsewhere (Pick et al. ).
The pilot plant was designed to treat continuously up to
12 m3/d of wastewater. Before flowing into the plant, the
water passes through three buffer tanks with a total hydrau-
lic retention time of 3.5 h. In this stage pH decreases to 4.5
BOD5 6,900 mg/L 9,000–11,000 mg/L
Total phosphorus 71 mg/L –
Kjeldahl nitrogen 280 mg/L –
COD (filtered) 8,840 mg/L –
Total cyanide 22 mg/L 5.8–96 mg/L
a
After 3.5 h of microbial acidification. TSS: total suspended solids; COD: chemical oxygen
demand; BOD5: 5-day biochemical oxygen demand.
due to rapid microbial acidification. As a result, colloidal
organic matter flocculates without adding any chemicals
and the removal efficiency in the flotation stage is much
better than with fresh wastewater. After this effect had
been observed during the first months of operation, pH
was no longer adjusted in the neutralization unit.
Dissolved air flotation (DAF) was applied in order to
remove the major portion of TSS. An Aquatector® Micro-
float® unit (Enviplan Company, Germany) was operated at
a hydraulic surface load of 2.5–3.0 m/h. Since dosing of
polymeric flocculants prior to flotation improved TSS
removal only slightly, flocculants were not added during
regular operation.
The central treatment stage was an anaerobic process
(expanded granular sludge bed (EGSB) reactor, type
ANAFIT-AC, Hager þ Elsässer Company, Germany) which
converts organic matter into biogas. The performance of
the reactor largely depends on stable process conditions
and a low suspended solids loading. This was achieved by
an optimization of the upstream DAF process. The EGSB
W
reactor was operated at a temperature of 35 C and a hydrau-
lic load of 4–5 m/h. It was seeded with sludge from a
brewery wastewater treatment plant. pH was adjusted to
6.8 by adding sodium hydroxide. There was no need for
heating because raw wastewater temperatures were already
at the required level.
For post-treatment, a vertical flow constructed wetland
(VFCW) was designed. The hydraulic surface load of the
unit was about 30 L/(m2 · d) and the average organic surface
load corresponded to 72 g COD/(m2 · d). The effluent was
collected in a small basin which was the sampling point,
and discharged via a fluid tipper to a lagoon operated by
the tapioca starch company.

Pile design practice in the USA varies from state to state, and
some states have published amendments to Aashto (2007) to take
account of local conditions. In California for example, Caltrans
(2011), at Table 10.5.5.2.4-1 for bored (‘drilled’) pile construc-
tion, replaces the Aashto resistance factor on pile shaft resistance
of 0.45 with 0.7, and the Aashto resistance factor on pile base
resistance of 0.4 with 0.5. Caltrans (2011) cites engineering
judgement and past design practice as the primary reasons for
these amendments. Caltrans (2011) stresses the importance of
construction quality control, and states that the ‘full effectiveness
of the tip resistance should only be permitted when cleaning of
the bottom of the drilled shaft is specified and can be acceptably
completed before concrete is placed.’
Using the full Caltrans (2011) amendments, we calculate a value
of F of 2.58, which falls perhaps fortuitously within the range
2.43 to 2.66 calculated using Eurocode 7.
Readers may be somewhat surprised at Caltrans’ use of, in effect,
an equivalent Æ value of 0.7 for a stiff clay. This value can be
related to a reliability index, , that represents the number of
standard deviations between the mean safety margin and the
failure limit (Paikowsky, 2004). Using this approach, an Æ value
of 0.7 would imply a  value of about 2.5, rather than about 3.5
if using the lower, Aashto-specified resistance factors.
However, in California it is often the case that the governing load
case is from short-term, rapid loading from seismic action, in which
Authors’ reply
We welcome the clarification from the discussers regarding the
Caltrans modifications to the Aashto design approach for piles. It
seems that the USA may be no more ‘united’ than Europe in terms
of geotechnical standards. In the original paper (Vardanega et al.,
2012), a review was undertaken to compare the different design
approaches (DA1, DA2 and DA3) in Eurocode 7. A similar
comparative study could be undertaken with the different modifi-
cations or annexes to the Aashto design code from different
jurisdictions within the United States. It is interesting to note that
using the Caltrans modification of Aashto the pile design would be
fairly similar to that obtained using the Eurocode.
The discussers cite the work of Paikowsky (2004), which suggests
that the partial factors for pile design can be determined from
statistical calibration – again based on data. This implies that the
database used to calibrate any partial factor set is ‘representative’
of the designs that will be carried out under its auspices.
Eurocode 7 assigns partial factors based on the experience of
experts, who would have access to various databases, but it does
not claim to have a load-test database that represents pile designs
for the whole of Europe. The key is to have access to a database
that is representative of the designs that will be done under the
auspices of the code.
Other codes also facilitate risk analysis, but not using explicit
statistical procedures. Vardanega et al. (2012) reviewed the
approaches described in AS2159-2009 (Standards Australia,
2009) and Poulos (2004), where the code drafters set limits on
87
cleanings can be performed either as a preventive maintenance op-
eration to maintain permeate flux [through incorporating chemical
cleanings into regular physical cleaning processes, for example
chemically enhanced backwashes (CEBs)], or as a recovery oper-
ation to recover membrane flux from a severely fouled membrane
module (Ramos et al. 2014). Chemically enhanced backwashes
combine conventional physical backwashing with chemical clean-
ing in one process, which could be of high importance for the
design and operation of membrane bioreactor (MBR) systems
for wastewater treatment. Recent studies using CEBs have typically
been performed at long intervals (weekly or monthly cleaning proc-
esses) (Zhou et al. 2014). There have been few comparative studies
on different cleaning methods and conditions, and scant informa-
tion exists on CEBs, in general, in the literature (Ramos et al. 2014;
Zhou et al. 2014). Snowdon et al. (2018) observed that one 60 s
CEB every 24 h resulted in increased permeate flux, operational
permeability, and net flux, demonstrating CEBs’ effectiveness as
a preventive maintenance operation.
This research project compared the efficiency of two different
chemical cleaning processes applied to one membrane module
integrated within a continuously operated pilot-scale AnMBR sys-
tem. The two chemical cleaning processes were applied to the
membrane module only once it was severely fouled, acting as a
recovery operation. As a secondary objective, the project investi-
gated the impact of daily CEBs on the effectiveness of each
chemical cleaning process. The novel pilot-scale AnMBR system
consisted of a 1,000 L anaerobic continuously stirred tank reactor
(CSTR) integrated with an external nanofiltration tubular mem-
brane module. The reactor and membrane configuration provided
benefits over conventional cleaning processes in that the membrane
module did not need to be removed for chemical cleanings (since it
could be isolated from the CSTR through a recirculation loop). This
resulted in minimal downtime and maintenance work required for
chemical cleaning processes. With few direct comparative studies
on different chemical cleanings applied to the same membrane
module, and little information in the literature on CEBs, there
Fig. 1. Simplified PFD of experimental setup.
© ASCE 04019018-2
J. Environ. Eng., 2019, 145(5): 04019018
are, to the authors’ best knowledge, no studies investigating the
impact of daily CEBs on the effectiveness of chemical cleaning
processes.
Materials and Methods
Reactor Configuration
A 1,000 L stainless steel anaerobic CSTR was equipped with an
external nanofiltration tubular membrane module and operated in
cross-flow mode. A simplified process flow diagram (PFD) of the
configuration may be viewed in Fig. 1. The anaerobic CSTR was
equipped with a heat wrap, which maintained the reactor temper-
ature at an average value of 36.2°C (
0.5°C) throughout the dura-
tion of the study. The operational temperature of approximately
36°C was selected since it is the upper limit of mesophilic anaerobic
digestion (25°C–35°C) and maximized biogas generation, which
was desired (Metcalf & Eddy, Inc. 1991).
Reactor-mixed liquor was continuously recirculated from the
CSTR to the membrane module and back into the CSTR by a pro-
gressive cavity pump (Nemo, New Brunswick, Canada). This pump
maintained a mixed-liquor flow rate of 6,000 L=h, which resulted
in the targeted membrane surface cross-flow velocity of 2.5 m=s
and a complete recirculation of reactor contents every 10 min.
A cross-flow velocity of 2.5 m=s was selected since it is deemed
to be the minimum operational cross-flow velocity that must be
maintained for the membrane module based on manufacturer rec-
ommendations. It was desired to operate at the lower spectrum of
the operational cross-flow velocity to minimize energy demands.
The membrane module was cylindrical in shape, composed of poly-
vinylidene fluoride (PVDF), with an outer diameter of 50 mm and
a total length of 1.436 m. The module contained 13 tubular mem-
brane channels, each 8 mm in diameter, which resulted in a total
membrane area of 0.42 m2 . The membrane surfaces had a mean
pore size of 30 nm.
J. Environ. Eng.
 When mixed liquor flowing through the recirculation line
entered the membrane module, a pressure gradient generated from
a peristaltic pump (Masterflex, Ontario, Canada) instantaneously
and continuously filtered a portion of the mixed liquor through the
pores of the membrane generating filtrate, which will henceforth
be referred to as permeate. The permeate flow direction was con-
trolled by two solenoid valves. For forward filtration, the permeate
flowed through the permeate recycle line back into the CSTR. The
solenoid valves would flip every 15 min and a secondary peristaltic
pump (Masterflex, Ontario, Canada) would force a portion of the
permeate into a secondary holding tank, where it would exit as ef-
fluent or be recycled for backwashing processes. The backwashes
would occur every 30 min, again controlled by the solenoid valves
and secondary pump. For CEBs, a mixed solution of 2,600 ppm
sodium hydroxide and 2,600 ppm sodium hypochlorite was in-
jected into the permeate backwash just prior to entering the mem-
brane module (at a ratio of 80% permeate and 20% chemical
solution), creating an approximate concentration of 520 ppm so-
dium hydroxide and 520 ppm sodium hypochlorite entering the
membrane module. The chemical concentrations used for CEBs
were recommended by the membrane manufacturer based on past
performance data.
Waste anaerobic sludge was pumped out of the reactor on a
2-h cycle. Every 2 h a peristaltic pump (Masterflex, Ontario,
Canada) would pump mixed liquor from the base of the CSTR out
of the reactor as waste sludge. The specific flow rate of the waste
sludge was adjusted to maintain reactor-mixed liquor suspended
solids (MLSS) concentration at the desired level of approximately
19–21 g=L.
Biogas exited the CSTR through a gas trap at the top of the
reactor and passed into a plastic gas collection bag, where the flow
of biogas was regulated. The regulated outflow of biogas moved
through a mass-flow meter (AALBORG, Orangeburg, New York),
where its total mass was measured and logged, before exiting the
system. The mass-flow meter was calibrated with a headspace
gas having a composition of 70% methane gas and 30% carbon
dioxide gas.
Wastewater Characterization
The anaerobic CSTR was fed with a high-strength (average COD of
84.8
8.9 g=L) potato starch–based synthetic wastewater simulat-
ing typical food-processing industry wastewater. Snowdon et al.
(2018) provided a complete characterization of the synthetic waste-
water. The CSTR was equipped with a level gauge, and whenever
the mixed-liquor level fell below a preset threshold of 0.94 m,
an influent pump (Masterflex, Ontario, Canada) came online and
pumped wastewater into the system until the mixed-liquor level
returned to the threshold level. The daily influent flow rate was
equivalent to the sum of the effluent and waste flows, minus the
backwash flow.
Seed Sludge Characteristics
The reactor was initially seeded with granular sludge acquired
from an anaerobic digester treating a fruit-juice-processing plant
wastewater in Lassonde, Quebec, Canada. The seed sludge had
an initial total solid (TS) concentration of 23.2 g=L, total volatile
solid (TVS) concentration of 19.8 g=L, total suspended solid
(TSS) concentration of 21.4 g=L, volatile suspended solid (VSS)
concentration of 19.5 g=L, and a specific methanogenic activity
(SMA) of 0.18 g-CODCH4 =g-VSS=day. The membrane module
used throughout the duration of this study was first placed inline
171 days after initial seeding of the reactor. The consistency of the
© ASCE 04019018-3
J. Environ. Eng., 2019, 145(5): 04019018
sludge changed slightly over the course of operation (shifting from
a granular consistency to a more viscous consistency); however,
87 days after placing this membrane module inline the mixed liquor
had a TSS concentration of 21.3 g=L, a VSS concentration of
20.4 g=L, and a SMA of 0.17g-CODCH4 =g-VSS=day, making it
quite similar to the initial seed sludge in terms of sludge biomass
activity.
Analytical Methods
The COD, TSS, VSS, BOD, and total alkalinity were determined
following Standard Methods (APHA 2005). Total phosphorous
(TP), total kjeldahl nitrogen (TKN), phosphate-phosphorous, and
ammonia nitrogen were measured according to the HACH Water
Analysis Handbook (Hach Company 2002). The methane con-
tent of the biogas was measured using a Varian CP-3800 gas
chromatograph. The chromatograph was fitted with an Alltech
CTR 1 concentric packed column and a thermal conductivity de-
tector maintained at a constant temperature of 180°C. Helium was
used as a carrier gas operating at a flow rate of 30 mL=min. The
temperature of the injection port was set to 120°C. Biogas injec-
tions were done in triplicate, and SMA tests were conducted in
duplicate using batch studies operated at 36°C. The tests were con-
ducted in 500 mL serum bottles with a working volume of 450 mL.
The bottles were filled with reactor sludge, acetic acid as a sub-
strate, pH buffer, and mineral bases. Young and Cowan (2004)
outlined the procedures and mixed solutions used. All flow and
pressure readings were noted once daily at similar times using in-
line flow meters and pressure sensors.
Experimental Design
The experimental design was developed through applying the
pretest–posttest control group design methodology. This method-
ology demonstrated that an effect occurred only following the
change in a certain variable from control conditions (Leedy and
Ormrod 2005). Applying the pretest–posttest control group design
methodology an initial clean water flux, clean water operational
permeability (OP), mixed-liquor flux (MLF), and mixed-liquor op-
erational permeability (MLOP) were established for the membrane
module prior to its placement inline (for clean water values) and on
Day 2 of operation (for mixed-liquor values). Following a chemical
cleaning process, each of these parameters was measured and cal-
culated to identify how efficient each chemical cleaning process
was in recovering flux and OP.
A total of six chemical cleaning processes were applied to the
membrane module, and their respective cleaning conditions may be
viewed in Table 1. The two different cleaning conditions were
chosen because they offered the largest recoveries in clean water
flux and clean water OP during other membrane studies conducted
by the authors. Cleaning Processes 1 and 6 were dynamic cleanings
(with backwashing processes in place during the cleaning) and
consisted of three parts using three different chemical solutions.
Cleaning Processes 2–5 were static cleans (with no backwashing
processes in place) and consisted of one part with one chemical
solution. Through the application of each set of cleaning conditions
multiple times, repetitions were accomplished for comparison.
In Table 1, the recirculation time is the amount of time the mem-
brane module was operated in forward filtration mode during the
cleaning process. The soak time is the amount of time forward and
backward filtration was ceased, allowing the membrane module to
sit and soak in the chemical cleaning solution. The number of back-
washes and duration are the quantity and duration of backward fil-
tration periods, or backwashes, accomplished during a cycle of the
J. Environ. Eng.
Table 1. Chemical cleaning conditions
Cleaning Recirculation Soak time Number of backwashes and
process Cycle time (min) (min) duration (number and min) Chemical used
1 1 30 N/A 2 and 2 min 1% sodium hydroxide
2 N/A 20 N/A
3 30 N/A 2 and 2 min
4 30 N/A 2 and 2 min 1% citric acid
5 N/A 20 N/A
6 30 N/A 2 and 2 min
7 30 N/A 2 and 2 min 1% sodium hydroxide and sodium hypochlorite
8 N/A 20 N/A
9 30 N/A 2 and 2 min
2 1 30 N/A N/A 1% sodium hydroxide and sodium hypochlorite
2 N/A 60 N/A
3 30 N/A N/A
3 1 30 N/A N/A 1% sodium hydroxide and sodium hypochlorite
2 N/A 60 N/A
3 30 N/A N/A
4 1 30 N/A N/A 1% sodium hydroxide and sodium hypochlorite
2 N/A 60 N/A
3 30 N/A N/A
5 1 30 N/A N/A 1% sodium hydroxide and sodium hypochlorite
2 N/A 60 N/A
3 30 N/A N/A
6 1 30 N/A 2 and 2 min 1% sodium hydroxide
2 N/A 20 N/A
3 30 N/A 2 and 2 min
4 30 N/A 2 and 2 min 1% citric acid
5 N/A 20 N/A
6 30 N/A 2 and 2 min
7 30 N/A 2 and 2 min 1% sodium hydroxide and sodium hypochlorite
8 N/A 20 N/A
9 30 N/A 2 and 2 min

chemical cleaning process. Finally, the chemicals used outline the
exact concentration of chemical cleaning solutions applied to the
membrane module during each cleaning condition.
Clean water flux was measured by isolating the CSTR recircu-
lation line from the CSTR itself, allowing a 50-L tank to be filled
with clean water (maintained at around 30°C) and filtered through
the membrane module. Each clean water test recirculated the clean
water at five flow rates in increasing order (790, 1,580, 2,360,
3,150, and 3,940 L=h), and the resultant membrane fluxes and
TMPs were measured. Five different increasing flow rates were se-
lected to provide incremental increases in flux and TMP, reaching
the maximum flow rate of the recirculation loop at approximately
3,940 L=h. Operational permeability is defined as membrane flux
divided by instantaneous operational TMP, in other words pressure-
averaged flux, and may be defined by Eq. (1) (Hamden de Andrade
et al. 2013):
Ji
OPi ¼ ð1Þ
ΔPi
In Eq. (1), OPi is the instantaneous operational permeability
(L=m2 =h=bar), J i is the instantaneous membrane flux (L=m2 =h),
and ΔPi is the instantaneous transmembrane pressure (bar). Opera-
tional permeability was considered so that TMP could be factored
into the flux recovery analysis. Clean water recoveries were calcu-
lated by dividing the post–chemical cleaning clean water flux or
post–chemical cleaning clean water OP by the initial clean water
flux or initial OP values measured on Day 1 of operation, and then
multiplying by 100. Once the membrane module was placed back
inline for mixed-liquor recirculation, and after 24 h of recirculation
at a flow rate of 6,000 L=h, the flux and TMP of the membrane
module were also measured. These values allowed the MLF
and MLOP to be calculated. The mixed-liquor recoveries were
calculated by dividing the post–chemical cleaning MLF or post–
chemical cleaning MLOP values by the initial MLF or initial
MLOP values measured on Day 2 of operation, and then multiply-
ing by 100. Equations used to calculate clean water and mixed-
liquor recoveries are displayed in Eqs. (2) and (3), respectively:
J; OPPC
CW J;OP ð%Þ ¼ × 100 ð2Þ
J; OPo
J; OPPC
MLJ;OP ð%Þ ¼ × 100 ð3Þ
J; OPo
In Eq. (2), CW J;OP is the clean water flux or operational per-
meability recovery (%), J; OPPC is the post–membrane cleaning
clean water membrane flux or clean water operational permeability
(L=m2 =h, L=m2 =h=bar), and J, OPo is the initial clean water mem-
brane flux or clean water operational permeability (L=m2 =h,
L=m2 =h=bar). In Eq. (3), MLJ;OP is the mixed-liquor flux or opera-
tional permeability recovery (%), J; OPPC is the post–membrane
cleaning mixed-liquor membrane flux or mixed-liquor operational
permeability (L=m2 =h, L=m2 =h=bar), and J; OPo is the initial
mixed-liquor membrane flux or mixed-liquor operational per-
meability (L=m2 =h, L=m2 =h=bar).
Membrane resistance coefficients were calculated using the
resistance-in-series model [Eq. (4)]. This equation was adopted and
developed from other studies (Ruigomez et al. 2017):

© ASCE 04019018-4 J. Environ. Eng.

J. Environ. Eng., 2019, 145(5): 04019018

 ΔP
J¼
μ × ðK M þ K IR þ K R Þ
In Eq. (4), J is the permeate flux (m3 =m2 =s), ΔP is the trans-
membrane pressure (Pa), μ is the dynamic viscosity of the permeate
(Pa=s), K M is the membrane resistance coefficient (m−1 ), K IR is
the irreversible fouling resistance coefficient (m−1 ), and K R is the
reversible fouling resistance coefficient (m−1 ). The membrane re-
sistance coefficients are important to calculate for a membrane
module because they provide indicators of the contributions from
various types of membrane fouling resistances to the total mem-
brane fouling resistance. The membrane resistance coefficient, K M ,
is the resistance of the membrane module prior to exposure to
mixed liquor. This value is found by filtering clean water of a
known temperature through the membrane module at monitored
TMP levels and measuring the resultant membrane flux. The
reversible fouling resistance coefficient, K R , is resistance to mem-
brane filtration contributed from foulants that may be removed
through regular physical and chemical cleaning processes. Gener-
ally, over the course of membrane filtration operation, K R gradually
increases as regular physical cleaning processes are used to reduce
its rate of accumulation. Eventually, these physical cleaning proc-
esses alone cannot remove a sufficient quantity of the reversible
membrane fouling, and a chemical cleaning is performed to im-
prove the amount of reversible membrane fouling removed. Thus,
K R may be calculated by comparing the total membrane resis-
tance prior to a chemical cleaning to the total membrane resistance
following a chemical cleaning, where the quantity of membrane
resistance removed from the cleaning process was contributed from
reversible membrane fouling. The irreversible fouling resistance
coefficient, K IR , is resistance to membrane filtration contributed
from foulants that, generally, may not be removed through the
regular physical and chemical cleaning processes. Over the course
of membrane filtration, K IR gradually increases and irreversible
foulants cannot be removed unless the regular cleaning processes
already set in place are improved.
Following chemical cleaning processes, there is typically a
sharp increase in TMP as foulants removed by the cleaning process
4
Clean No.
0 10 20 30
Mixed-Liquor OP Recovery
Clean Water OP Recovery
Fig. 2. Flux and operational permeability recoveries for both mixed liquor and clean water following each chemical cleaning.
© ASCE
J. Environ. Eng., 2019, 145(5): 04019018
rapidly reaccumulate on the membrane surface or in its pores. After
ð4Þ the rapid increase in TMP, a slower, linear increase in TMP is often
noted, and the rate of membrane foulant accumulation may be cal-
culated during this linear rise. However, in this experiment’s setup,
since both TMP and flux were dependent on the permeate pump
flow rate and, therefore, simultaneously adjusted over the duration
of data collection, TMP increase alone would not be entirely rep-
resentative of membrane foulant accumulation. Therefore, follow-
ing each chemical cleaning process, the rate of membrane foulant
accumulation was quantified by calculating the rate of decline
in operational permeability since operational permeability factors
both flux and TMP into its value. Eq. (5) was developed through
the application of this concept, where the rate of fouling is quanti-
fied through the loss of OP between membrane module chemical
cleaning processes:
OPi ¼ OPPC þ rOP t ð5Þ
In Eq. (5), OPi is the instantaneous operational permeability
(L=m2 =h=bar), OPPC is the operational permeability measured
after membrane cleaning (L=m2 =h=bar), rOP is the rate of decline
in operational permeability (L=m2 =h=bar=d), and t is the time in-
crement of measurements (days).
Results and Discussion
Flux Recoveries and Membrane Resistance
Coefficients
The clean water flux, clean water OP, MLF, and MLOP recoveries
following each chemical cleaning process may be viewed in Fig. 2.
No chemical cleaning processes were able to recover more than
53% of the original membrane clean water flux. Cleaning Processes
1, 5, and 4 offered the largest clean water flux recoveries with val-
ues of 53%, 34%, and 30%, respectively. One reason that Cleaning
Process 1 offered 56% greater performance in clean water flux re-
covery over the next highest recovery offered by a cleaning process
may be due to its occurring 5 days after the membrane module was
40 50 60 70 80 90 100
Recoveries (%)
Mixed-Liquor Flux Recovery
Clean Water Flux Recovery
04019018-5 J. Environ. Eng.
Table 2. Membrane resistance coefficients calculated following each
chemical membrane cleaning process (unit: 10−12 × m−1 )
Cleaning process KM KR K IR
1 0.62 4.66 0.73
2 9.33 2.00
3 17.3 2.60
4 18.7 1.45
5 6.27 1.32
6 6.34 3.72
first placed inline. Reversible membrane fouling occurred rapidly
when the membrane module was first placed inline, resulting in a
quick drop in membrane flux (a loss of 35% in MLF over 5 days),
which merited the first chemical cleaning. The membrane resis-
tance coefficients were calculated (Table 2) to further support this
observation. Following Cleaning Process 1, K R was found to be
much greater than K IR (values of 4.66 × 10−12 m−1 and 0.73 ×
10−12 m−1 , respectively), indicating that the majority of foulant
removed was reversible membrane fouling. Since this was the first
chemical cleaning applied to the membrane module, K IR was also
at its lowest value throughout the study. As the majority of mem-
brane fouling was contributed from reversible foulants, which are
removed effectively through chemical cleaning processes, and there
was little contribution from irreversible membrane foulants relative
to other cleaning methods, Cleaning Process 1 appeared much more
effective. With this consideration in mind, it is difficult to distin-
guish whether Cleaning Process 1 was superior in terms of clean
water flux recovery over all other cleaning processes due to its
cleaning conditions or whether its enhanced performance was sim-
ply due to occurring so soon after the membrane was first placed
inline. Wang et al. (2017) did find that the maximum amount
of organics in their chemical cleanings was desorbed through the
use of an alkaline and acidic cleaning process when treating MBR
effluent. Cleaning Process 1 results agree with this observation, as
the cleaning conditions provided both alkaline and acidic chemicals
and it had the largest clean water flux recovery, indicating more
foulant removal.
Cleaning Processes 4 and 5 outperformed Cleaning Processes 2
and 3 in terms of clean water flux recovery. This result was not
initially anticipated because, due to the accumulation of irreversible
membrane fouling over the course of membrane operation, succes-
sively applying the same cleaning conditions typically results in
a decrease in flux and OP recoveries with each successive clean
(Sun et al. 2016). Since all four of these cleaning processes were
identical in conditions, it was expected that irreversible membrane
foulants would accumulate, making each successive membrane
cleaning process less efficient than the one preceding it. A differ-
ence that was present leading into Cleaning Processes 4 and 5 was
that there were CEBs present 60% (73 days) and 100% (77 days) of
the time, respectively, leading into each of these chemical cleaning
processes. Cleaning Process 5 had more CEBs prior to it and out-
performed Cleaning Process 4. Similarly, Cleaning Process 4 had
more CEBs prior to it compared to Cleaning Processes 2 and 3 (as
there were no CEBs in place leading into Cleans 2 and 3) and out-
performed them. From this observation, it would appear that regu-
lar CEBs synergize with chemical cleaning processes and enhance
their performance in terms of clean water flux recovery. Regular
contact between a low-strength caustic chemical cleaning solution
known for its efficiency in removing organic membrane foulants
and a wastewater/mixed liquor composed almost entirely of organic
matter as a source of membrane foulants (with average volatilities
for the wastewater and mixed liquor of 99% and 94%, respectively)
© ASCE 04019018-6
J. Environ. Eng., 2019, 145(5): 04019018
would allow the CEBs to be quite efficient in both removing
built-up membrane foulants and preventing their rate of accu-
mulation while operating. Zhang et al. (2007) found that, when
operating a tubular polyethersulfone side stream AnMBR for the
treatment of swine manure, long-term operation without chemical
cleaning processes resulted in residual membrane fouling that was
difficult to remove with the chemicals used in their study (nitric
acid and ethylenediaminetetraacetic acid). By using a constant
chemical cleaning process, the build-up of residual membrane
foulants might have been prevented. In this study, the prevention
of the build-up of difficult-to-remove membrane foulants was ob-
served, and a constant source of chemical cleaning was supplied
through CEBs, resulting in more effective fouling mitigation. It was
found that K R increased by 85% over 106 days between Cleaning
Process 2 and Cleaning Process 3, whereas K R increased by only
8% over 121 days between Cleaning Process 3 and Cleaning
Process 4, indicating a decrease in the rate of K R accumulation.
With the cleaning conditions of Cleaning Processes 2 and 3 being
identical, and operating conditions being fairly consistent in the
periods following these cleaning processes, the primary reason for
the decline in reversible foulant accumulation may be attributed to
CEBs being in place 60% of the days following Cleaning Process 3.
K R was 6.27 × 10−12 m−1 following Cleaning Process 5. A de-
crease in this value from Cleaning Process 4 to 5 indicates that the
rate of reversible foulant build-up was greatly reduced during
this period of operation, which may again be attributed to regular
CEBs being in place.
The irreversible fouling resistance coefficient typically increases
over the duration of membrane filtration and would not be antici-
pated to decrease unless operational or cleaning conditions could
be enhanced significantly. Considering that K IR decreased from
Cleaning Process 3 to 4, and again from Cleaning Process 4 to 5,
while all of these cleaning processes operated under the same
conditions, indicates that the presence of CEBs might also en-
hance regular cleaning protocols in such a way that irreversible
foulants (that could not be removed without CEBs in place) could
then be removed. Zhang et al. (2007) concluded that weekly mem-
brane cleanings provided no significant benefits in membrane per-
formance compared to monthly membrane cleanings, attributing
the lack of benefits to the fact that irreversible fouling accumulated
more rapidly than the frequency of either cleaning. Perhaps the
daily CEBs in this study occurred frequently enough that irre-
versible foulants were unable to accumulate effectively between
membrane cleanings. From Cleaning Process 5 to 6 the K R value
remained quite steady between cleans, whereas K IR increased by a
factor of 2.8 with CEBs regularly in place for 26 days of operation.
This indicates that Cleaning Process 5 may have been superior to
Cleaning Process 6 since K IR increased once again between cleans.
Similar to the results for clean water flux recovery, Cleaning
Processes 1, 5, and 4 offered the largest clean water OP recoveries
with values of 49%, 31%, and 30%, respectively. Since the recov-
eries in clean water operational permeability are near identical to
the recoveries in clean water flux, this indicates that the chemical
cleanings had similar recoveries in TMP. One reason why Clean 1
outperformed all of the others by a large margin may again be
attributed to its early application and lack of irreversible foulant
accumulation.
In terms of MLF and MLOP recoveries, all cleaning processes
exhibited good performance, with the lowest MLF recovery being
Cleaning Process 3 (with a value of 67%) and the lowest MLOP
recovery being Cleaning Process 2 (62%). Compared to the largest
clean water flux and OP recoveries of 53% and 49%, respectively,
this indicates that the chemical cleanings were much more effective
at recovering inline mixed-liquor flux rather than clean water flux.
J. Environ. Eng.
Geotechnical Engineering
Volume 167 Issue GE1
the geotechnical reduction factors (similar to partial factors), and
a risk analysis matrix is used to search for a value of the
reduction factor applicable to the specific design project being
undertaken. This gives the design engineer the flexibility to
reduce the partial factors if good ground investigation data and
load testing are carried out, or increase them if designing in an
unfamiliar soil deposit. The code itself gives advice on the
elements of geotechnical risk that the designer must consider; in
other words, they recognise that a ‘one size fits all’ approach is
perhaps limited. Eurocode 7 aims to achieve a similar result by
encouraging the designer to vary the characteristic values of soil
properties in relation to the quality of the supporting data, while
keeping the values of partial factors constant.
The discussion of the adhesion factor Æ for bored pile design is
interesting. The parameter Æ relates to the soil deposit being
studied; the value Æ ¼ 0.5 is commonly used in London Clay.
Increasing Æ to account for rate effects in clay might be
warranted, but only if load-test data were also available (e.g.
Burland et al., 1966). For example, previous use of ‘constant rate
of penetration’ tests generally showed slightly higher values of Æ
for London Clay (Patel, 1992). For static loading, Æ ¼ 0.7 used in
the Caltrans amendment, as noted by the discussers, does seem
surprisingly high for a stiff clay (API, 1984); the overall
reliability of the design will depend not only on this and the
applied strength reduction factor, but also on the factors applied
to loads. We understand that, in practice, the use of this value is
often replaced by results from load testing.
REFERENCES
Aashto (2007) LRFD bridge design specifications. Customary US
Units, 4th edn. American Association of State Highway and
Transportation Officials, Washington, DC, USA.
API (1984) Planning, Designing and Constructing Fixed Offshore
88
Discussion
Vardanega, Kolody, Pennington et al.
Platforms. American Petroleum Institute, Washington, DC,
USA.
Burland JB, Butler FG and Dunican P (1966) The behaviour and
design of large diameter bored piles in stiff clay. Proceedings
of the ICE Symposium on Large Bored Piles, London, UK,
pp. 51–71.
Caltrans (2011) California Amendments to Aashto LRFD Bridge
Design Specifications, 4th edn. California Department of
Transportation, Sacramento, CA.
FHWA (Federal Highway Administration) (2010) Drilled Shafts:
Construction Procedures and LRFD Design Methods, Report
No FHWA NHI-10-016. National Highway Institute,
Washington, DC, USA.
Guha S (1995) Dynamic Characteristics of Old Bay Clay
Deposits in the East San Francisco Bay Area. PhD thesis,
Purdue University, Ann Arbor, MI, USA.
Paikowsky S (2004) Load and Resistance Factors for Deep
Foundations, NCHRP Report 507. Transportation Research
Board, Washington, DC, USA.
Patel D (1992) Interpretation of results of pile tests in London
Clay. In Piling: European Practice and Worldwide Trends
(Sands MJ (ed.)). Thomas Telford, London, UK, pp. 100–
110.
Poulos HG (2004) An approach for assessing geotechnical
reduction factors for pile design. Proceedings of the 9th
Australia New Zealand Conference on Geomechanics,
Auckland, New Zealand, vol. 1, pp. 109–115.
Standards Australia (2009) AS2159-2009: Piling: design
and installation. Standards Australia, Sydney, NSW,
Australia.
Vardanega PJ, Kolody E, Pennington SH, Morrison PRJ and
Simpson B (2012) Bored pile design in stiff clay I: codes of
practice. Proceedings of the Institution of Civil Engineers –
Geotechnical Engineering 165(4): 213–232.