Download PDF Version

Date: May 2007
Authors: Michele Grenier, XCG Consultants Ltd.
Jane Bonsteel, Region of Halton
George Lai, MOE
Linda Perry, XCG Consultants Ltd.
Presented at: OWWA 2007 Conference

INTRODUCTION
In the revised Procedure for Disinfection of Drinking Water in Ontario (June 2006), the turbidity performance criterion for filter effluent was decreased from ≤0.5 NTU in 95 % of measurements each month to ≤0.3 NTU in 95 % of measurements each month.

Performance evaluations were conducted at two surface water treatment plants based on the Guidance Manual for the Optimization of Ontario Water Treatment Plants Using the Composite Correction Program (CCP) Approach (Ontario Ministry of Environment, March 1998). The CCP approach was developed by the US Environmental Protection Agency and adapted for the MOE. The CCP approach consists of two main components, the Comprehensive Performance Evaluation (CPE) phase and the Comprehensive Technical Assistance (CTA) phase. The CPE is an evaluation approach that aims to estimate the capabilities of the existing facility. The objective of the CPE phase is to identify if significant improvements in the plant’s performance can be achieved without major capital improvements. The CTA is a performance improvement phase that addresses the issues identified during the CPE and implements the optimization of the existing facilities to achieve desired performance. This paper describes the CPE phase conducted at two Ontario water treatment plants.

FACILITY DESCRIPTIONS
Plant A

Plant A is a conventional filtration facility that obtains raw water from a surface water pond source, which is supplied by a river/lake system. The plant has a rated capacity of 92.75 L/s (8,014 m3/d) and serves a population of approximately 3,000. The plant is currently operating at about 30 percent of its rated capacity, with average daily flows of about 2,790 m3/d. The Permit to Take Water for the plant allows a maximum withdrawal of 5,616 m3/d.

A schematic of Plant A is shown in Figure 1. Powdered activated carbon (PAC) is added to the raw water year round for taste and odour control. Potassium permanganate (KMnO4) is added seasonally (January to March) to oxidize manganese. Liquid coagulant (Eaglebrook PASS-C [polyaluminum chloride] or PHAS [pre-hydroxylated aluminum sulphate]) is added and the water is conveyed to two (2) flocculation chambers, connected in series, where an anionic polymer (CIBA LT27AG) is introduced as a coagulant aid. The flocculated water flows to four (4) sedimentation basins that operate in parallel. Each of the basins is equipped with tube settlers and sludge hoppers for sludge removal. The water is then filtered through four (4) dual media filters. Filter effluent flows to either of two (2) clearwells where the water is chlorinated for primary disinfection. Backwash water from the filters is sent to a backwash tank equipped with a supernatant discharge pipe that drains to the sanitary sewer system. The sludge pumped out of the sedimentation basins is trucked to the municipal sewage lagoons for disposal.

Figure 1: Plant A Treatment Process Schematic

Figure 1: Plant A Treatment Process Schematic

Plant B
Plant B is a conventional surface drinking-water system that obtains its main raw water supply from a Great Lake. The raw water pumping station is located on the shore and supplies the treatment plant via a 10 km pipeline. Plant B has a rated capacity into the treatment trains (gross capacity) of 167.8 L/s or 14,500 m3/d according to its Certificate of Approval (C of A). The 2003 and 2004 treated water flow records indicate that the plant is operating at average daily flows of approximately 40 percent of the plant’s rated capacity.

A schematic of Plant B is shown in Figure 2. Pre-chlorination of the raw water is carried out for zebra mussel control at the intake crib. Raw water is delivered to a two (2) compartment concrete wet well, each side of which is equipped with a travelling screen. Pre-chlorination takes place in the wet wells during the winter months. Aluminum sulphate coagulant is added to the incoming raw water before it is directed to a flash mixer. After the flash mixer, the flow is split and conveyed to twin flocculation/sedimentation chambers that operate in parallel. Each flocculation chamber consists of six (6) cells, and the sedimentation basins are equipped with tube settlers.

The water is filtered through five (5) dual media sand/granular activated carbon (GAC) filters. The filter effluent is directed to an underground clearwell where it is chlorinated. The effluent from Filters 3, 4 and 5 first flows to a small clearwell that ultimately empties into the larger clearwell. Filters 1 and 2 effluent flows directly to the large clearwell.

Figure 2: Plant B Treatment Process Schematic

METHODOLOGY
The methodology that was used to conduct the CPE phase of the evaluation was divided into three main sections:
1. A performance assessment based on plant operating records provided for a full operating year;
2. An evaluation of the individual unit processes at the plant; and,
3. An identification of performance limiting factors. Jar testing and full scale stress testing were also conducted.

RESULTS – PLANT A
Performance Assessment

A full year of data was used for the evaluation of historical plant performance as recommended in the CCP. A timeframe from June 2004 to May 2005 was selected to conduct the performance assessment after plant staff reported that the plant’s former largest customer had shut down operations in March 2004. The selected timeframe between 2004 and 2005 was intended to accurately reflect the existing water demands at Plant A.

The characterization of the raw water quality according to the plant operating records and water samples taken in 2005 indicated that colour and total organic carbon (TOC) removal from the raw water, as well as turbidity removal and manganese control, were treatment challenges for Plant A. Colour and TOC removal are necessary, both for aesthetic reasons and to control disinfection by-products (DBPs).

The plant records indicated that variability in the quality of the raw water source affected the quality of finished water from June 2004 to May 2005. Individual filter performance data indicate that all four (4) filters experienced turbidity spikes higher than the previous acceptable limit of 0.5 NTU, generally from June through December; yet all filters were able to meet the previous performance criterion for filtered water turbidity of less than 0.5 NTU in at least 95 percent of the measurements each month. Filtered water turbidity data show that all four (4) filters had turbidity levels at times higher than the revised 0.3 NTU level.

Major Unit Process Evaluation
Following the process described in the CCP document, specific performance requirements were considered during the evaluation of each process. The criteria for major unit process evaluation presented in Table 2.1 of the CCP manual were used as the basis to rate each treatment process at Plant A.

A performance potential graph was prepared based on the calculated rated capacities obtained for each process. The graph shown in Figure 3 compares the calculated rated capacity of each process to the design flow capacity and the peak instantaneous operating flow.

Plant flow records show that maximum instantaneous raw water flows of approximately 0.069 m3/s (5,962 m3/d) occurred at the plant. However, operating staff indicated that these maximum instantaneous flows occurred for very short periods of time (seconds), and thus are not considered to be the peak instantaneous operating flow. For the purpose of this evaluation, a peak day flow of 4,895 m3/d, the maximum raw water flow at which the plant was operated between June 2004 and May 2005, was used.

Bars to the right of the dashed lines indicate that these unit processes have the capacity to treat the maximum flow or the rated capacity flow. Bars to the left of the dashed lines indicate that the size of the units is not sufficient to provide an adequate plant performance at maximum day and rated capacity flows; therefore, changes to these processes should be considered.

As shown in Figure 3, the limiting unit processes (bar to the left of the dashed lines) are flocculation, sedimentation and winter primary inactivation. The high coagulant dosages required at the plant are probably at least partially the result of the limitations of the flocculation and sedimentation processes.

Figure 3: Performance Potential – Plant A

A rating system was used to rate each unit process and the overall plant as either Type 1, 2 or 3. Type 1 plants are those where major unit processes are adequate and the problems are likely related to operation, maintenance or administration. Type 2 plants represent a situation where marginal capacity of unit processes could potentially limit the plant from achieving the desired performance level. Type 3 plants are those in which major unit processes are projected to be inadequate to provide the required capacity for existing water demands; therefore, major modifications are felt to be required to achieve the desired level of performance. A unit process is rated Type 1 when its calculated capacity exceeds the actual demand; Type 2 if its projected capacity is 90 to 100 percent of actual peak demand; or Type 3, if its projected capacity is less than 90 percent of actual peak demand. Table 1 below illustrates the results of the rating.

Upgrading the flocculation process would involve increasing the retention time. The mixing energy needed for optimal formation of floc in coloured water should also be evaluated. Although the sedimentation process has a Type 2 rating, the shallow design and the coagulant dosages necessary under current conditions create operational challenges such as the need to manually clean each of the tanks on a bi-weekly basis.

Table 1: Type Rating of Unit Processes – Plant A

Table 1: Type Rating of Unit Processes - Plant A

Full-Scale Testing
A summary of the results obtained for raw and filter effluent water samples collected during a plant trial on November 9, 2005 is shown in Table 2.

Table 2: Results of Sample Analysis

The combination of a new polymer (SNF 440) and PASS-C allowed a reduction in coagulant dosage by almost 25 percent. Stress-testing of Filter 4 during the polymer trial indicated that sedimentation basins and filters operated effectively in terms of turbidity removal during the limited testing period.

The raw water had high a TOC concentration of 16.4 mg/L. The TOC removal obtained during the plant testing averaged 75 percent, with an average filter effluent TOC concentration of 4.0 mg/L.

Filter effluent turbidity levels measured on-site with portable turbidimeters during the plant trial varied considerably from the results given by the on-line turbidimeters.

Performance Limiting Factors
Raw water quality characteristics such as high total organic carbon (TOC), colour and seasonal manganese concentrations have a detrimental effect on the performance of the overall treatment process.

Very high coagulant dosages are used at the plant (up to 360 mg/L of poly aluminum chloride (PASS-C) during summer months). The elevated natural organic matter in the raw water in combination with the non-ideal conditions provided during flocculation and sedimentation are major contributing factors to the current high chemical demand. The coagulant dosages have sometimes exceeded the NSF maximum use concentration for PASS-C of 250 mg/L (as supplied product). Although the residual aluminum concentration obtained for the treated water sample collected on the first visit to the plant was below the operational guideline of 0.1 mg/L, a reduction in the coagulant dosages is desirable in terms of compliance with the NSF maximum use concentration and to reduce costs.

The size of the flocculation tanks may be inadequate to provide sufficient detention time for proper floc formation for flows above the average daily flow. At design or maximum daily flows, the flocculation basins are capable of providing retention times of 11 and 18 minutes, respectively, which is not considered adequate to form dense settleable particles, especially with coloured water. Also, the low raw water temperatures (<5oC) experienced from mid November through the end of March further delay the formation of floc, resulting in a decrease of flocculation performance during the colder months.

The calculated mixing energy, “G”, for temperatures experienced at the plant (5oC to 20oC) is 126-154 s-1. This mixing energy is very high, since colour floc is typically light and fragile. Maximum G values for this type of water are generally less than 100 s-1. Tapered flocculation – reducing G values for each stage – is also recommended (MOE Design Guidelines, 1982). The current flocculation energy could be negatively affecting the floc formation process. Poor performance of the flocculation process can contribute to problems in downstream unit processes (sedimentation and filtration).

At the plant rated capacity, the SOR for the sedimentation basins is very high, given the light colour floc and the shallow basins and tube settlers. In each of the sedimentation tanks, sludge collects in the hoppers and is removed via a perforated pipe located at the bottom of each of the two hoppers. The sludge is not removed as quickly as it accumulates, however, and each of the tanks must be emptied and manually cleaned approximately every two weeks. The problem is probably due to the low density of the colour floc and the volume of sludge produced. If chemical dosages can be lowered and the density of the floc increased through changes to the coagulant and/or coagulant aid and flocculation process, the frequency of cleaning needed for the sedimentation tanks may be reduced. If the tanks still require frequent cleaning, changes to the sludge collection system should be investigated.

RESULTS – PLANT B
Performance Assessment

A full year of plant data from the year 2004 was used for the evaluation of historical plant performance as recommended in the CCP. Some additional data from the MOE’s Drinking Water Surveillance Program (DWSP) was also used to assess water quality.

The data provided in the DWSP report for parameters including pH, colour, alkalinity, dissolved organic carbon and trihalomethanes, indicate that the source water is of good quality and that their usual concentration levels should not compromise the chemical and physical treatment of the water at Plant B given the existing unit treatment processes.

Historical turbidity data trends for raw, filtered and treated water were taken from the plant process data reports for 2004. Raw water turbidity levels were generally below 5 NTU from January through May; raw water turbidities above 20 NTU were most frequently experienced from October through December, peaking to levels above 90 NTU on various occasions during September, October and December. The average annual raw water turbidity in 2004 was 15 NTU, indicating that particulate removal is an important treatment requirement for Plant B.

Individual filter effluent turbidity data for 2004 indicate that all five (5) filters provided an appropriate performance level in terms of turbidity as average levels for all filters were approximately 0.04 NTU, below the limit of 0.3 NTU. Filter effluent turbidity levels in the range of 0.3 NTU rarely occurred during the entire year.

Individual filter effluent turbidity data showed maximum filter effluent levels of 0.04 NTU during the period when higher combined treated water turbidity leaving the clearwell (>1.0 NTU) was reported; thus, the high levels in turbidity shown in the treated water were not the result of deficiencies in filter performance.

Major Unit Process Evaluation
Following the process described in the CCP document, specific performance requirements were considered during the evaluation of each process. The criteria for major unit process evaluation presented in Table 2.1 of the CCP manual were used as the basis to rate each treatment process at Plant B.

A performance potential graph (Figure 4) was prepared based on the calculated capacities obtained for each process and the peak instantaneous operating flow and rated plant capacity. For this evaluation, a peak flow rate of 11,941 m3/day, representing the maximum daily flow during 2004, was assumed as the peak instantaneous operating flow. As shown in Figure 4, the limiting unit process for Plant B is sedimentation. All other unit processes showed performance potential greater than the C of A rated capacity of the plant.

Figure 4: Performance Potential - Plant B

Figure 4: Performance Potential – Plant B

Although the evaluated capacity of the sedimentation unit is adequate to treat the current average flows, the capacity of this process may be limited in cold water conditions when the plant is operated at maximum flows of 11,940 m3/d. However, it is unlikely that peak flows would be experienced during the winter period. In addition, the stress test indicated that the sedimentation process is capable of performing at higher SORs in cold water conditions, at least for short periods of time, although this would need to be confirmed with additional testing.

As part of the evaluation, a rating system was used to rate each unit process and the overall plant as either Type 1, 2 or 3 as shown in Table 3.

Table 3: Type Rating of Unit Processes – Plant B

Table 3: Type Rating of Unit Processes - Plant B

As discussed previously, the capacity of the sedimentation process may be limited if the plant operates at maximum flows in cold water conditions. However, a higher rate of operation was successfully demonstrated during a short stress test period conducted with cold water conditions. A longer-term test at high flows should be conducted to ensure that, as solids accumulate in the basins, they are not carried into the launders.

Jar Testing and Full-Scale Trials
Jar testing was conducted in March and October, 2005. At both times the raw water temperature was approximately 5°C. Both alum and polyaluminum chloride (SternPac) were tested with no additional pH adjustment. Using both coagulants, filtered water turbidity levels were acceptable, but the aluminum residual concentrations were greater than 0.10 mg/L.

On December 13 and 14, 2005 stress testing was performed on Filters 1 and 2. Flow was incrementally increased on December 13th to the rated capacity for the two filters (52 L/s each or 104 L/s in total).

The stress-testing at the target flow rate was conducted for a period of approximately 2.5 hours on December 13th, at which point the plant was returned to normal operation. Flows were again increased in the morning of December 14th and the filters stressed for a further 5.5 hours. The raw flow was increased from the normal influent flow value of 35 L/sec to approximately 104 L/s (rated capacity flow) with the two filters in operation.

During the stress test the average raw water turbidity was 7.49 NTU. The plate settler effluent averaged 0.8 NTU, below the CCP goal of 2 NTU. The two filters performed well with effluent turbidity levels lower than 0.1 NTU at all times, as recorded by the on-line turbidimeters. The maximum effluent turbidity recorded was 0.026 NTU for Filter 1 and 0.025 NTU for Filter 2.

Stress testing of Filters 3, 4 and 5 was performed on January 24, and 25th, 2006. Flow to one side of the plant was increased to the C of A rated capacity for the three filters (31.6 L/s each or 95 L/s in total). Stress testing was performed for approximately 3.5 hours on the 24th and 5.5 hours on the 25th. The stress test could not be maintained for long periods on the 25th as the reservoir was filling up quickly; therefore the flows were turned up and down over the course of the 5.5 hours. Data were collected and recorded only during periods of higher flow.

During the stress test the average raw water turbidity was 37.6 NTU as measured by the on-line turbidimeter. The plate settler effluent averaged 2.03 NTU approximately equal to the CCP goal of 2.0 NTU. The filters performed well with effluent turbidity levels lower than 0.1 NTU at all times, as recorded by the on-line turbidimeters. The maximum effluent turbidity recorded was 0.033 NTU for Filter 3, 0.040 NTU for Filter 4 and 0.035 NTU for Filter 5.

The turbidity results obtained from on-site measurements using a portable turbidimeter differ significantly from those obtained from the on-line instrumentation during the monitoring. The filter effluent on-line turbidimeters consistently measured turbidities much lower than the portable turbidimeter.

Performance Limiting Factors
The aluminum residuals in the treated water were greater than the Operational Goal of 0.1 mg/L at times, and as such, pH adjustment may be needed to reduce the aluminum residual.

During the filter stress tests, flow through the plant was diverted through one plate settler tank. Surface overflow rates (SOR) calculated during the stress test period were 6.3 and 5.8 m/h respectively. These values are greater than the CCP recommended SOR of 4 m/h, but the settlers performed well. The plate settler maximum effluent turbidity during the Filter 1 stress tests was 1.33 NTU and 2.55 NTU during Filter 4 stress testing. A longer-term test at high rates should be conducted to confirm the rating of the sedimentation process.

The Recommended Standards for Water Works (“Ten State Standards”, 2003), states that the minimum backwash rate should be 37 m/h, but that a rate of 50 m/h or “a rate necessary to provide for a 50 percent expansion of the filter bed is recommended”. The MOE Guidelines for the Design of Water Treatment Works (1982) state that provision should be made for a backwash rate which “will expand the filter media 25 to 40 percent overall, typically 30 percent”. Because water viscosity changes with temperature, higher backwash rates are needed in the summer than in the winter and the backwash program should be adjusted accordingly. Each backwash pump at the Plant B provides a maximum rise rate of 51 m/h for Filters 1 and 2 and 56 m/h for Filters 3 to 5. This is acceptable, according to the Ten State Standards. However, in the current backwash sequences, the maximum valve opening is 45 percent. The media expansion should be measured to determine if the current backwash rates are sufficient to expand the media according to MOE recommendations.

CONCLUSIONS
Under current flow conditions, Plant A is producing filtered water that generally meets the MOE’s turbidity criteria. The filter stress testing indicated that if the filters are operated at design flows, the filter effluent turbidity would climb to levels above the 0.3 NTU criteria. The process evaluation also indicated that the performance potential of the prefiltration processes (flocculation and sedimentation) is less than the current rated capacity of the plant.

The current operational challenges for Plant A are the removal of TOC and control of disinfection by-products, and the very high coagulant dosages required.

Under current flow conditions, Plant B is producing filtered water that generally meets the MOE’s turbidity criteria. The filter stress testing indicated that if Filters 1 and 2 are operated at design flows for longer periods of time, the filter effluent turbidity would climb to levels above the 0.3 NTU criteria, limiting the filter run time. Filters 3 to 5 are rated and operated at filtration rates that are far lower than their actual capacity. The solids holding capacity of these filters allow them to capture the influent turbidity for long runs at their current filtration rate.

The process evaluation also indicated that the performance potential of the sedimentation process is less than the current rated capacity of the plant. However, short term stress testing indicated that the sedimentation process may be capable of satisfactory performance at higher rates.

A discrepancy between turbidity measured by different instruments was noted during on-site work at both plants, indicating that special attention should be given to the accuracy and precision of the turbidity measurements. In monitoring turbidity performance in the water treatment process, operators should be looking for changes in turbidity levels rather than focusing on absolute values.

ACKNOWLEDGEMENTS
The authors gratefully acknowledge the assistance provided by the operations staff at both water treatment plants, as well as the technical support received from MOE staff.

BIBLIOGRAPHY
Great Lakes – Upper Mississippi River Board of State and Provincial Public Health and Environmental Managers, 2003, Recommended Standards for Water Works (Ten State Standards)

Ministry of the Environment, 1982, Guidelines for the Design of Water Treatment Plants

Ministry of the Environment, March 1998, Guidance Manual for the Optimization of Ontario Water Treatment Plants Using the Composite Correction Program (CCP) Approach

Ministry of the Environment, June 2003 and revised draft March 21, 2005, Procedure for Disinfection of Drinking Water in Ontario

MWH, 2005, Water Treatment: Principles and Design, Second Edition, John Wiley & Sons, Inc.

USEPA, 1999, Microbial and Disinfection By-products Rules Simultaneous Compliance Guidance Manual (EPA/815/R-99/011)