Patrick W. Given, M.Sc., P.Eng.
Associated Engineering
Edmonton, Alberta
Contact: givenp@ae.ca

Dennis Mutti, P.Eng., M.A.Sc.
Acres & Associated Environmental Ltd.
Etobicoke, Ontario

Doug Scott, P.Eng.
City of Thunder Bay, Ontario

Darrell Matson
City of Thunder Bay
Thunder Bay, Ontario

ABSTRACT

The Ontario Ministry of Health placed a boil water advisory on the City of Thunder Bay following detection of a Giardia cyst in the water supply. The advisory impacted only the portion of the water system served from Loch Lomond, a pristine lake in a protected watershed.

Two microfiltration (MF) pilot plants were tested on the Loch Lomond water supply with and without chemical coagulation, and on the City’s second source – Lake Superior – without chemical coagulation. Particle challenge testing showed that both test units could reliably achieve 4-log (99.99%) removal of 2-15 m size particles. With chemical coagulation, significant colour removal could also be achieved.

The successful pilot test program led to the fast-track design and construction of a temporary MF facility which enabled lifting of the boil water advisory. Performance of the facility has exceeded expectations. The facility not only is removing cyst-size particles, but also some colour constituents to the extent that disinfection byproducts from chlorination of the filtered water appear to have been reduced. The equipment supplier now classifies the membrane treatment facility as an ultrafiltration plant.

Successful experience with the pilot MF and temporary ultrafiltration facilities has provided confidence leading to the City’s decision to incorporate the ultrafiltration equipment in a permanent facility.

INTRODUCTION

The City of Thunder Bay has two water supply sources – Loch Lomond and Lake Superior. Loch Lomond is located in an elevated, protected watershed. Water quality is generally excellent in terms of turbidity (generally about 0.4 NTU); however colour, organics and corrosivity are concerns. Up to November of 1998, treatment of this water was limited to disinfection with chlorine and conditioning with sodium silicate. The minimally treated water from Loch Lomond was distributed by gravity principally to the south portion of the city. The other source, Lake Superior, normally has higher turbidity levels (about 1 NTU but with occasional upward excersions due to storms). This water is treated at the Bare Point direct filtration water treatment plant and distributed by pumping principally to the north portion of the city.

In October, 1997, a Boil Water Advisory was issued by the Medical Officer of Health in response to detection of a Giardia Lamblia cyst in the south distribution system. The south distribution system serves about 40,000 persons, or about 40% of the City’s total population.

By coincidence microfiltration (MF) pilot plant trials, under the direction of Acres & Associated (A&A) Environmental Ltd., were already in progress when the boil water advisory was issued. MF pilot testing of Loch Lomond water was conducted from August to November 1997. Successful performance of the MF trials led to the consideration of this technology as one of several alternative means of expediting removal of the Boil Water Advisory. MF was selected as the preferred alternative in early 1998 leading to the tendering and commissioning of temporary MF facilities by November of the same year. The decision to implement temporary microfiltration of the Loch Lomond source was based largely on the recognized capabilities of microfiltration to remove cyst-size particles, as well as the relative speed at which temporary microfiltration facilities could be commissioned.

In this paper, we will briefly review the concepts of MF technology, MF pilot plant test results, temporary MF facility performance, and plans for transferring the temporary MF facilities at Loch Lomond to a permanent location at the Bare Point Water Treatment Plant.

MICROFILTRATION TECHNOLOGY

Microfiltration is one of several membrane filtration processes. Membranes are thin films of synthetic organic or inorganic materials which can selectively separate constituents from a fluid. Fig. 1 illustrates four common membrane processes and their effective treatment capabilities.

Transmembrane pressures range from 1 to 3 atmospheres for microfiltration up to 30 to 150 atmospheres for reverse osmosis. Also with respect to microfiltration, one manufacturer utilizes vacuum instead of pressure as the driving force.

MICROFILTRATION PILOT TESTS

Test Equipment

Memcor (Memtec Group) and Zenon Environmental Inc. provided pilot plant facilities. The Memcor plant, rated at 85.5 L/min, was utilized throughout the test program. Zenon supplied a 26.5 L/min pilot plant which was tested on the Loch Lomond water supply, and a 38 L/min pilot plant which was tested on the Lake Superior water supply and in particle challenge testing.

Aside from proprietary differences in the membrane materials, other noted differences between the two systems were as follows:

Memcor Zenon

Transmembrane pressure/vacuum Pressure to 150 kPa Vacuum to 60 kPa

Filtration direction Inside to outside Outside to inside

Membrane housing Pressure housing Aerated tank

Regular cleaning Back pulse air Back pulse chlorinated water

Off-line cleaning (in place) Caustic Chlorine or citric acid

Test Program

Pilot plant performance test results presented in this paper refer to Unit 1 (Memcor) and Unit 2 (Zenon).

Microfiltration pilot testing began at the Loch Lomond site near the beginning of August 1997 and ended mid-November 1997. Testing at the Bare point site started mid-November 1997 and ended mid-February 1998. A&A prepared the test work plan, the plant suppliers provided initial hands-on training, and the City provided daily plant operation. Plant operation included record keeping, adjustments and collection of water samples for analysis.

Raw and treated water samples were analyzed for particle counts, turbidity, colour, and other parameters. Only the listed parameters are reviewed in any detail in this paper.

Pilot Plant Performance

Table 1 illustrates the particle removal efficiencies of the two MF pilot plant units during short-term particle challenge testing. Backwash water from the Bare Point Water Treatment Plant on Lake Superior was utilized as the feedwater. Raw water particle counts in the 2-15 micron size range totaled 1.54x10⁶ per mL. The associated raw water turbidity exceeded 2000 NTU. Both test units demonstrated the capability to remove at least 4-log (99.99%) of the particles 90% of the time. Apparent differences between the two units in the particle log-removal results are not necessarily significant since testing was short term. Filtered water turbidities were less than 0.1 NTU.

Table 1. Particle challenge test results

Table 2 summarizes the turbidity and colour performance data for the two MF pilot plants operating under three test conditions -- Loch Lomond source with no chemical pretreatment, Loch Lomond source with chemical pretreatment, and Lake Superior source (Bare Point) with no chemical pretreatment. Chemical treatment trials at Loch Lomond utilized polyaluminum chloride for Unit 1 and alum for Unit 2.

Table 2. Microfiltration pilot test summary for turbidity and colour

* 10^th % and 90^th % refer to 10^th percentile and 90^th percentiles, respectively.

** Polyaluminum chloride with Unit 1 and alum with Unit 2.

Both raw water sources exhibited low turbidity levels: median less than 0.4 NTU for Loch Lomond and 1.0 NTU for Bare Point. Treated water turbidity was generally excellent. The 90^th percentile filtered water turbidities for both units were between 0.1 and 0.2 NTU when treating Loch Lomond water without chemical pretreatment. Under the other two test conditions, 90^th percentile filtered water turbidities were less than 0.1 NTU.

With respect to raw water colour, the Loch Lomond source showed moderate levels with a median colour around 15 TCU. The Lake Superior source was significantly lower at 4 TCU. Therefore, chemical pretreatment trials to enhance colour removal were not undertaken for the latter source. Colour and related dissolved organic carbon (DOC) are parameters not expected to be significantly removed using microfiltration since the contributing constituents have minute particle sizes, less than the MF 0.2 m nominal membrane pore size. Not only are moderate to high levels of colour a possible aesthetic concern, but they could also be a health concern if subsequent chlorination processes lead to excessive formation of chlorinated organic compounds which are suspected carcinogens. Tests without chemical pretreatment confirmed that the MF units did not remove colour and DOC effectively without chemical pretreatment. However, with such pretreatment at Loch Lomond, significant removal of colour and DOC was achieved by both test units.

Chemical cleanings were required for both units at about four week intervals to restore transmembrane pressures/vacuums to their original values.

TEMPORARY MICROFILTRATION

Pretendering of microfiltration equipment was undertaken in February 1998 with tender award to Zenon in April of that year. The tender was based on the supply of packaged microfiltration equipment providing a capacity of 30 ML/d at <2 ^oC. The tender documents included a buy-back clause in the event that the City of Thunder Bay chose not to incorporate the temporary facilities in a permanent structure. Zenon’s proposal incorporated the use of seaway containers for the tanks with no building except to house blowers, pumps and controls. Engineering work on the microfiltration installation contract commenced in May 1998. The completed facilities were operational and running at maximum design capacity on November 8, 1998. The total installed cost of the MF facilities was approximately $6 million.

The City of Thunder Bay has collected various performance data for the temporary MF facilities including the main parameters of focus in this paper: particle counts, turbidity and colour . On-line particle counters appear to be showing extremely low and consistent filtered water particle counts. The apparent excellent results will be undergoing checking in the near future and therefore will not be reported at this time. Checking of the particle counters is underway. In addition, the current turbidimeters consistently register "0" as they are incapable of measuring turbidities less than 0.1 NTU. Therefore, our discussion of performance data will be limited to observations of colour removal.

Figure 2 plots colour measurements taken by Thunder Bay staff from August to October 1998 (prior to microfiltration), and from November 1998 to June 1999 (following microfiltration). The results show significant improvement in filtered water colour during the first month of MF operation. Subsequently, colour removal reduced somewhat but generally appeared to remain slightly better than observed during the pre-MF period. It should be noted that some of the observed colour removal likely is due to chlorine oxidation of residual organics in the water. Based on pilot plant test results, improved colour removal was not expected. The plant supplier subsequently re-classified the facility as an ultrafiltration plant.

In addition to the colour data reported by the City, some laboratory test results from the Ontario Ministry of Environment & Energy (OMEE) were available for analysis. Table 3 summarizes data provided by the City of Thunder Bay for the pre-MF and post-MF periods. Summarized results for raw and treated Loch Lomond water include turbidity, colour, dissolved organic carbon, total trihalomethanes (THMs), and total haloacetic acids (HAAs). The data show excellent filtered water quality in terms of turbidity was provided by the MF plant on the sampling dates shown in Table 3. Also, the data appear to show improved colour removal is provided by the membrane filtration plant, helping to support Zenon’s statement that the plant should be classified as an ultrafiltration facility. Although limited sample tests show dissolved organic carbon removals appeared minor, the levels of THMs and HAAs appear improved (reduced from >100 g/L to <100 g/L). Additional testing will be required to determine if these apparent improvements continue in future.

The Loch Lomond membrane treatment plant is manned 8 hours per day and is monitored via PLC from the Bare Point plant during the remaining periods. During normal plant operation, the membranes are automatically back pulsed every 20 minutes for a 30 second duration. Off-line membrane cleaning has been required approximately at six week intervals during winter operation and up to five months during summer operation. Sodium hypochlorite solution is used for cleaning the membranes by soaking them up to eight hours, followed by dechlorination and a rinse cycle before returning the units to service. To date, citric acid has not been required for cleaning.

Table 3. OMEE test results before and after "microfiltration"

* The raw water colour on May 25^th requires confirmation. Independent testing by Thunder Bay operators showed colour readings in the 10 to 20 range during that time period.

FUTURE PLANS

The City of Thunder Bay is proceeding with plans to convert the City's water supply to a single source at the Bare Point Water Treatment Plant on Lake Superior. The plant will also be upgraded to achieve 4-log removal and/or inactivation of Cryptosporidium and Giardia cysts. Evaluated major technologies to meet this treatment goal included ultrafiltration membranes and direct filtration with ultraviolet (UV) or ozone disinfection. Since the City already has temporary ultrafiltration (UF) facilities in place at the Loch Lomond location, the application of UF technology at the permanent location became more economically feasible. The current recommendation calls for:

The capacity of the microfiltration facilities was originally rated at 30 ML/d but is being re-rated and guaranteed by the supplier at 38 ML/d. The expanded and upgraded facilities will have and an overall capacity of 122 ML/d. Of course, secondary disinfection with chlorine will remain.

CONCLUSIONS

The main conclusions in this paper are summarized as follows: