Komati Power Station has installed a membrane plant consisting of ultrafiltration, double
pass reverse osmosis and continuous electro-deionisation to treat cooling tower
blowdowns in order to produce demineralised water and to conduct sidestream chemistry
control of the cooling water circuit. This plant has replaced the existing ion-exchange
plant that was used for the production of demineralised water and thus serves to reduce
the loading of mobile salts in the ash dam (90% reduction) by eliminating regeneration effluent from the ion-exchange plant.
Due to oil contamination in the cooling water circuit (when oil from oil coolers leaks into
the cooling water), the membrane plant was also designed to operate on raw water from
either the Nooigdedacht or the Vygeboom Dam or a blend of both dams. This is
considered to be an emergency intervention under abnormal conditions to prevent
possible irreversible fouling of the membranes due to oil in the cooling water. The
Nooigtedach Dam water contains high concentrations of organic matter and is also
enriched with nutrients due to raw sewage influent into the Dam water. This poses a
challenge with regard to treatment of the high fouling feed water on the membrane plant.
Natural organic matter in water has the ability to foul reverse osmosis membranes. This
adversely affects the operation of the reverse osmosis process. However, very little
information is available regarding the fouling characteristics of natural organic material in
the raw and cooling water at Komati Power Station for the reverse osmosis membranes.
Therefore, a pilot study was undertaken to determine the influence of natural organic
matter on membrane fouling, to optimise the process for the removal of natural organic
matter and to assess the ability of two different reverse osmosis membranes to
effectively treat the high fouling feed water at Komati Power Station. The ability of a polyethersulphone hollow-fibre ultrafiltration membrane system was first
evaluated to remove natural organic matter in the feedwater, by conducting pilot tests,
initially without coagulation of the raw water and thereafter with in-line coagulation for
organics removal. Jar tests were conducted in the laboratory to determine the most
suitable coagulant and dosage for turbidity and natural organic matter removal. Various
coagulants were tested and, based on the results of the jar tests, a coagulant (U3000)
was identified based on optimal removal of both total organic carbon and turbidity at a
dosing level of 20 mg/L. During the operation of the ultrafiltration pilot plant, permeate flow; feed pressure and
feed temperature were monitored. Performance of the ultrafiltration membrane was
monitored in terms of flux versus time for operation with and without a coagulation
process. The results indicated that there was very little total organic carbon removal
(maximum removal of 4%) without coagulation and a slight decrease in flux. The flux
declined as a result of fouling but could be recovered by performing hydraulic
backwashes and CEB procedures. Permeate flux, however, could be maintained at
about 90 Lmh (from 642 hours of operation). Since most of the organics passed through
the ultrafiltration membrane, it was concluded that the loss in flux was due to colloidal
fouling of the membrane. This was observed when the operation was carried out using
raw water as feed as well as when cooling water was used.
The total organic carbon removal increased to 30% when the plant was operated with inline
coagulation. The flux remained relatively stable during the first 600 hours of
operation and only decreased significantly during the last 200 hours of operation as a
result of fouling. The reduction in flux prior to cleaning was less than the 15% (maximum
flux decline of 9.9% during the test period) which is acceptable according to the industry
norm of 15%. It appeared that flux could be maintained at around 90 Lmh which was
about the same as when no coagulant was applied. The 30% total organic carbon
reduction that was obtained was not sufficient to reduce the organics to the level of 6mg/L dissolved organic carbon that was specified by the membrane manufacturer for the
standard brackish water reverse osmosis membrane.
Two reverse osmosis membranes – the standard brackish water reverse osmosis
membrane (BW30-2540) and the extra-low-fouling membrane (BW30XFR-2540) – were
assessed in terms of their ability to remove dissolved organic carbon, ease of cleaning of
the membrane and the ability to recover flux after cleaning. This was done to establish
which membrane is more suited to Komati’s high-fouling feedwater.
The evaluation of the performance of the two reverse osmosis membranes was
conducted using pre-treated water (filtered water after in-line coagulation, anti-scalant
and biocide dosing) as well as using water that was not pre-treated. During operation
(under both conditions), the normalised permeate flux, conductivity, dissolved organic
carbon and organics absorbing at UV254 were monitored.
It was established that in terms of flux decline that the extra low-fouling membrane gave
slightly superior performance to that of the standard membrane, achieving longer
production runs (up to 5 days compared with 3 days achieved by the standard brackish
water membrane) without requiring chemical cleaning. The low fouling membrane
achieved better CWF recovery after the cleaning cycles (81.26% Lmh of the virgin
membrane on the occasions when there was flux loss) compared to the standard
membrane (restored to 77.35% of CWF of the virgin membrane) when using untreated
feed water. This performance improved when pre-treated feed water was used and the
low fouling membrane’s CWF regained after the CIP was 95.89% which was within the
industry norm of a flux recovery of 95%, indicating that the CIP had been effective. It was
determined that the TOC rejection of the low-fouling membrane was higher (average
TOC rejection of 97%, maximum TOC rejection of 99%) than that of the standard
membrane (average TOC rejection of 95.3%, maximum TOC rejection of 97%).
Preliminary efforts to optimize the pre-treatment for organics removal in order to reduce
organic loading for the RO membranes confirmed that the use of granular activated
carbon and use of an organic scavenger resin might not be economically feasible due to
the relatively quick TOC breakthrough (8910BV, approximately 18000BV and less than
18000BV for the Filtrasorb 300, Filtrasorb 400 and organic scavenger resin,
respectively). Although further investigations should still be conducted, the preliminary
results indicate that it would be beneficial to also identify other options that can be further
investigated for optimization of organics removal at Komati Power Station.
Decline in the normalised flux as well as the evidence of biofouling were witnessed
during the pilot operation suggesting that the membranes were fouled. Autopsies were
performed on both membranes to identify foulants responsible for the decline in flux that
was observed during the pilot study. The results did not indicate an organic foulant on the
membrane surface. Biofouling should however, be monitored in the main plant as this
was suspected to have resulted in the flux decline during the pilot study.
The low fouling membrane demonstrated a better capability to treat the Komati raw and
cooling water and would be expected to achieve lower operating costs for the plant (CIP
costs and membrane replacement costs) while achieving better organics removal and it
is therefore recommended that the low-fouling membranes be used at Komati Power
Station as they are superior to the standard membrane and the cost of the low-fouling
membranes is comparable to that of the standard membrane. While this would provide
somewhat better performance than that obtained with the standard brackish water
membranes, it is proposed that further investigation into pre-treatment optimization for
organics removal as well as more efficient cleaning solutions be investigated to improve
the performance and economics of the main water treatment plant at Komati power
Station. / Dissertation (MSc)--University of Pretoria, 2013. / gm2014 / Chemical Engineering / unrestricted
| Identifer | oai:union.ndltd.org:netd.ac.za/oai:union.ndltd.org:up/oai:repository.up.ac.za:2263/33319 |
| Date | January 2013 |
| Creators | Dladla, Zanele |
| Contributors | Schoeman, J, zanele.dladla@eskom.co.za, Gericke. G |
| Source Sets | South African National ETD Portal |
| Language | English |
| Detected Language | English |
| Type | Dissertation |
| Rights | © 2013 University of Pretoria. All rights reserved. The copyright in this work vests in the University of Pretoria. No part of this work may be reproduced or transmitted in any form or by any means, without the prior written permission of the University of Pretoria. |
Page generated in 0.0048 seconds