ABSTRACT
The Greater Vancouver Sewerage & Drainage District (GVS&DD) is a major regional wastewater agency serving Vancouver, British Columbia and surrounding municipalities. In the late 1980’s it was mandated to upgrade its largest treatment plant from primary treatment to secondary treatment. From 1990 until the present the GVS&DD has engaged in Predesign, Design, Construction and Commissioning of the Annacis Island WWTP Secondary Upgrade. The size of the facility and the fact that no secondary treatment components were previously in place combined to make the upgrade a very large project valued at nearly $470 million CDN. This paper describes the major project components of the liquid stream process improvements. In addition to the major biological treatment components, other aspects of the project are described including, odour management, influent pumping, and primary treatment upgrades.

KEYWORDS
Wastewater treatment; design; secondary treatment; trickling filter/solids contact process; odour control; biofilter

INTRODUCTION
The major portion of the secondary treatment upgrade involved adding facilities to accomplish biological secondary treatment for flow exiting the existing preaeration and primary sedimentation processes. Originally, the project was conceived as a single step addition of an entire secondary treatment process train. Funding constraints required that an alternative implementation strategy be adopted - a portion of the facilities was to be commissioned at an earlier date to provide secondary treatment for a portion of the primary effluent flows. This change in approach led to the Phase I Secondary Treatment upgrade, which is discussed later in this paper and which is planned to operate for a little more than one year before the full secondary treatment process is commissioned.

TRICKLING FILTER/SOLIDS CONTACT (TF/SC) PROCESS SELECTION
A principal activity of the predesign effort was to select the biological treatment process (Parker et.al., 1993). There were few constraints imposed on the selection process since no form of secondary treatment was already in place.

The major design requirements for the process were the quality of effluent and the range of flows to be treated. Within Canada, effluent criteria are set by the provinces rather than the federal government. The British Columbia Ministry of Environment, Lands and Parks set fairly stringent criteria for the Annacis Island plant. Composites for any single day cannot exceed BOD5 values of 30 mg/L and TSS values of 40 mg/L. These daily criteria are considerably more stringent than requirements for secondary treatment in the US, where the USEPA criteria require BOD5 and TSS values of less than 30 mg/L for monthly averages rather than for daily values.

The requirement to meet effluent criteria on a daily basis created the need for the treatment process to perform optimally at rated peak wet weather flows (PWWF) which are two times the rated average dry weather flows (ADWF) of 5.6 m3/s. PWWF can be maintained for intervals of 12 hours or more. Above the rated PWWF, the criteria allow for blending of primary effluent flows that are greater than the PWWF receiving complete secondary treatment.

A total of 17 secondary treatment processes were screened for consideration. Initial screening was done on the basis of noneconomic criteria. Preestablished screening criteria included process characteristics, operation/maintenance considerations, planning criteria, and environmental factors. The screening criteria were weighted using nominal group technique involving participation of GVS&DD stakeholders. After screening, three processes remained; the air activated sludge process (AAS), the oxygen activated sludge process (OAS), and the trickling filter / solids contact process (TF/SC).

Present value cost analysis yielded the cost comparison summary shown in Figure 1. While OAS was a slightly lower cost process to build, TF/SC showed operation and maintenance savings that resulted in an estimate of an overall lower total present value cost for TF/SC.

In addition to the economic analysis, the three process alternatives were subjected to a technical evaluation using non-economic factors developed in the screening evaluation. A consensus procedure was used by GVS&DD staff and consultants in two workshops to establish numerical scores for each factor. In the first workshop, issues, concerns, or unknowns about each process were listed. Information on each identified technical concern was then gathered. The evaluation group visited four OAS plants, two AAS plants, and two TF/SC plants. The final workshop resulted in establishment of ratings for the various categories of concern. Each process ranked higher in some categories and lower in others than the other two alternatives. Overall, however, the OAS and TF/SC processes were found to have essentially equal ratings. The AAS process was found to be 15 percent lower in its non-economic score than the other two processes.

The TF/SC process was ultimately selected as the biological secondary treatment process. It was competitive with the next best alternative (OAS) in terms of capital costs, and had significantly lower O&M costs because of lower labour and energy requirements. Three key non-economic characteristics favoured the TF/SC process: robustness or the ability of a process to maintain high effluent quality under stressed conditions, ease of operation, and ease of maintenance.

TF/SC PROCESS DESIGN

Figure 2 illustrates the TF/SC process. The TF/SC process consists of high rate attached growth reactor (trickling filter) followed by a short solids residence time (SRT) aerobic suspended growth reactor (solids contact tank) which improves the settling characteristics of biomass wasted from trickling filters as well as provide additional oxidation of carbonaceous organic materials. The solids contact tanks are coupled with secondary clarifiers which return biological solids to the front end of the solids contact tanks. Low SRTs are maintained in the solids contact portion to prevent nitrification thereby limiting aeration capacity requirements.

Trickling Filters Design Features. Table 1 describes the basic design data for the trickling filters. The trickling filters remove soluble carbonaceous BOD by means of biofilm attached to the media. Each of the trickling filters is covered for odour control. Covering of the filters necessitated forced air ventilation of the filters to assure supply of oxygen for treatment and for conveyance of foul air to odour treatment.

Table 1 — Trickling Filters Design Data Number of Units 4 Media Type Plastic Crossflow Diameter, m 52 Media Depth, m 6 Total Media Volume, m3 52,000 Total Organic Loading, kg BOD5/m3 - day at ADWF 1.40 at Maximum Month Flow 1.63 Ventilation Rate, m3/min, each Supply 255 Recycle 2,600 Exhaust 338 Hydraulic Application Rate, m3/m2 - day at Minimum Diurnal Flow 18 at ADWF 57 at PWWF (4 filters) 118 at PWWF (3 filters) 158

After pumping, wastewater is applied to the top of the plastic media by means of six arm, mast type rotary distributors. The distributor arms rotate by reactionary force of the discharging primary effluent. Accommodations have been made for reversing discharge nozzles as a means of controlling arm rotation rate in order to assure adequate washing of the media. The magnitude and low frequency of occurrence of minimum diurnal flows allowed the trickling filters to be designed without a recirculation system.\

An expanded section of the trickling filter effluent channel was designed for the purpose of removing many of the snails which discharge from the trickling filters. The expanded section is aerated much like a grit removal tank and a submersible solids handling pump is located to remove the snails to be disposed with primary grit.

Solids Contact Tank Design Features. Table 2 describes the basic design criteria for the four solids contact tanks. These tanks were configured to operate in reaeration mode. Return secondary sludge will be returned to one or two of the tanks where it will be aerated prior to blending with the trickling filter effluent. This configuration allows some of the tankage to be maintained at higher MLSS levels, while allowing the design SRT to be maintained and lowering the solids load to the secondary clarifiers. The major control parameter will be the solids contact tank MLSS rather than the SRT. Clarifier design was based on an MLSS concentration of 2000 mg/L. Wasting rates and return rates will be manipulated to ensure that this basic parameter is satisfied.

Tank geometry was finalized after considering the space and geotechnical constraints of the site and optimal depth to achieve low energy aeration. The selected six metre depth provided the lowest overall cost and minimized the space required for the tankage.

Special features were included in the tankage specifically due to the nature of the trickling filter effluent entering the tanks. Snails can grow in trickling filters and eventually are flushed into the trickling filter effluent. Other TF/SC plants have experienced maintenance difficulties due to substantial accumulations of snail shells in the contact tanks. To preclude this problem at the Annacis Island plant, the floors were sloped at three percent to a drainage trench along one wall. In turn, this trench slopes to a central large drain. This configuration is intended to minimize the problems associated with cleaning the snails from the tank floor when access for maintenance is required.

The solids contact tank aeration system was designed to meet the oxygen demands through a wide range of conceivable operating scenarios. Extensive modeling was conducted to estimate the oxygen demand during average, peak and minimum load conditions, in different process configurations (different combinations of reaeration and contact) and through the various seasons of the year.

Fine bubble aeration was chosen for the solids contact tanks. The contract documents were prepared allowing for the tender of flexible membrane or ceramic diffusers, both meeting basic performance and design criteria. Flexible membrane systems with 225 mm diffusers proved the most cost effective and were incorporated in the work. The standard oxygen transfer efficiency (SOTE) limit specified for the reaeration tanks was 33 percent and for the contact tanks, 35 percent. Clean water performance testing conducted after installation proved that the system exceeded these stipulations.

An analysis of the air supply options led to the selection of single stage blowers for the Annacis Island plant. The blowers were designed to provide the necessary aeration air supply plus that required for agitating deep, mixed liquor channels. Three 800 hp blowers, each capable of 6.5 m3/s, were provided. Separate, smaller multi-stage blowers also have been incorporated to provide shallow channel agitation air.

Table 3—Secondary Clarifiers Design Data Number of Units 12 Diameter, m 40 Sidewater Depth, m 5.7 Surface Overflow Rate m3/m2 - day ADWF 32 Max Month Flow 51 PWWF 64.1 Solids Loading Rate kg/m2 - day ADWF 80 Max Month Flow 121 PWWF 192 Design Settled Volume Index, ml/g 135

Secondary Clarifiers Design Features. Table 3 describes the basic design criteria for the secondary clarifiers. The clarifiers are circular flocculating well type with inboard launders. Performance data at high loading rates were reviewed from other treatment plants having similar clarifier designs in order to build confidence that stringent effluent criteria would be reliably achieved at the high loading rates experienced with Number of Units PWWF. Solids loading rates are sustainable so long as Settled Volume Index (SVI) remains below the design value. Typically, TF/SC plants produce biological solids with SVIs significantly lower than traditional suspended growth processes.

Settled solids are removed from the clarifier floor where they fall by means of two arm rotating suction headers directly connected to return secondary sludge pumps. Scum removal is achieved with a novel traveling beach and trough mechanism that removes scum from the surface where it appears rather than pushing the scum to the perimeter of the launders. With this system, most scum removal takes place within the area contained by the flocculating well. A vertical line shaft non-clog centrifugal pump is located at the center area of the clarifiers to pump secondary scum off to the plant headworks.

PHASE I SECONDARY TREATMENT

Project funding mandated the phased implementation of the Annacis Island Secondary Treatment Upgrade. To achieve the earliest possible improvement in plant effluent quality, the components of the plant constructed under this strategy included the solids contact tanks, the secondary clarifiers, DAF thickening, and the digesters. This strategy allowed a portion of the plant flow to receive secondary treatment as soon as possible, thereby reducing the loads discharged to the Fraser River at the earliest possible date.

Concept. The intent of Phase 1 was to operate the secondary components of the plant as an activated sludge facility, without the trickling filters. However, the aeration system is limited in the amount of oxygen it can supply and would not be able to satisfy the requirements when the activated sludge configuration was used to treat the entire primary effluent flow. Accordingly, a relatively constant primary effluent flow has been diverted to the secondary treatment area of the plant while the remainder is directed to the plant effluent.

The activated sludge process design envisioned operation in a relatively high rate mode, with SRTs between 2 and 2.5 days. At this SRT, oxygen demands are low, as the potential for nitrification is limited; although higher residual dissolved oxygen concentrations are needed to inhibit bulking. Seasonal analyses were conducted to determine a reasonable expectation for secondary treatment capacity through the year. It was determined that the secondary plant could reliably treat the following flows:

Summer—140 ML/d

Spring-Fall—160 ML/d

Winter—260 ML/d

The seasonal differences reflect the affect upon the aeration system of the differing oxygen transfer characteristics, bacterial respiration rates, and environmental conditions through the course of a year.

Interim Process Features. Secondary treatment flows were controlled by modulating the primary effluent bypass gates to maintain a constant head in the primary effluent channels. The flow is directed through the trickling filter pump station area to a series of bypass gates that connect to the mixed liquor channel feeding the solids contact tanks. By only partially opening one of these gates, it acts as an orifice and at a relatively constant upstream head, admits a constant flow into the secondary treatment area.

The possibility of sludge bulking was addressed by installing a temporary hypochlorite system. Typical SVIs through the first several months of operation have been in the range of 170 mL/g. As there is significant excess secondary clarifier capacity at the present flows and the SVIs have not proven excessive, the temporary hypochlorite system has not been used.

Sludge is wasted at a constant rate from the reaeration tank to maintain the target SRT range - 2.0 to 2.5 days. As the mixed liquor concentration tends to vary, the rate of wasting is moderated while maintaining the SRT within the desired range.

Process Performance Data. Startup and commissioning of the Phase 1 Secondary plant commenced during mid March 1997. The system was seeded with 90 m3 of thickened waste activated sludge obtained from another nearby plant. Within a few weeks, the process had stabilized and was generally achieving the effluent quality expected. Figure 3 illustrates the BOD and Figure 4 shows the TSS for the period starting April 1 and ending June 20 1997.

           

During this period, the flows averaged approximately 180 ML/d. The average TSS was 8.2 mg/L and the average BOD, 10.1 mg/L. A slight upset during the second week of April coincident with the startup of secondary sludge thickening was quickly rectified and the system has been stable since that time. Only once during the upset period did the BOD exceed the limit of 30 mg/L while the TSS has not exceeded the 40 mg/L standard. During this period, the TSS load to the Fraser River has been decreased by about 5,000 kg/d and the BOD load has been decreased by about 6,500 kg/d.

ODOUR TREATMENT
The Annacis Island WWTP is located within a populated light and heavy industrial zoned area. Additionally, the plant is located immediately adjacent to a major arterial overhead traffic bridge that passes thousands of cars daily. Odour treatment systems were provided for both liquid stream and solids handling facilities. The liquid stream processes are served by two major odour treatment systems: the primary treatment odour control system which serves influent sewer ventilation, preaeration, and gravity thickening; and the secondary treatment odour control system which serves trickling filters, solids contact tanks, and mixed liquor conveyance channels. Each system has its own foul air collection network and treatment system. Both systems use the same process design for odour treatment.

Process Concept Development. Figure 5 illustrates the odour treatment process flowsheet used for both systems. The GVS&DD made an early decision to embrace the use of biofilter technology for odour removal. Biofilters demonstrated excellent removal performance for a wide range of organic and inorganic odourous compounds, including sulphides. There was information which showed that applying too much sulphide to a biofilter could result in too much sulphuric acid production which would ultimately destroy the filter beds ability to host odour removing microorganisms. The raw sewage contained high sulphide levels and it was probable that hydrogen sulphide concentrations in foul air could peak in the 20 - 50 ppm range. With these considerations, a process flowsheet was developed with scrubbing towers for peak sulphide removal followed by biofilters for removal of all other odorous compounds.

For the scrubbing tower portion of the flowsheet, a traditional caustic/hypochlorite scrubbing system was compared against a caustic only scrubbing system. Caustic/hypochlorite scrubbing systems are more typical because the hypochlorite oxidizes the absorbed sulphides as well as a range of other absorbed odourous compounds. A concern with caustic only scrubbers is that absorbed odorous compounds are not oxidized and can be re-released when spent scrubbing solution is discharged back into the treatment plant liquid stream. For this application, two factors mitigated the concerns related to a caustic only system; the scrubbers main purpose was to protect the downstream biofilter from excessive sulphides as the biofilter was expected to remove most other odourous compounds, and spent scrubber solution could be discharged to the highly oxidizing conditions of the solids contact tank to promote oxidation of absorbed sulphides. The operational savings associated with not utilizing hypochlorite in the scrubbing solution were estimated at over $500,000 per year. For the cost savings and the mitigation of potential concerns, the decision was made to not utilize hypochlorite in the scrubber system.

Caustic Scrubber/Biofilter Design Features. Table 4 describes the design data for the packed bed scrubber systems. Packed bed scrubbers reduce the concentration of odourous compounds (primarily hydrogen sulphide) with a dilute caustic soda solution. Caustic soda creates an alkaline solution that has a high affinity for acid, such as hydrogen sulphide, and therefore absorbs acid gases from the air. Caustic soda, at 25 percent concentration, is supplied by truck and stored in a bulk storage tank. A portion of the scrubbing solution is pumped to the solids contact tanks for disposal. The aerobic biomass within the reaeration tanks oxidizes the sulphide ions into soluble sulphate ions and elemental sulphur.

Foul air enters the bottom of a scrubber, flows upward through random packing, through a demister, and exits the top of the scrubber. Two variable speed blowers, located between the scrubbers and the biofilters, force the partially treated foul air from the scrubbers to the biofilters. If bypassing the biofilters is desired, the scrubbers can discharge directly to the atmosphere. During periods of low sulphide production (winter) the packed bed scrubbers can be bypassed to send foul air directly to the biofilters.

Table 5 describes the design data for the biofilters. The biofilters are either the second treatment stage or the only treatment stage of foul air. Biofilters reduce the concentration of odorous compounds by absorbing them onto the filter media, where microbial oxidation occurs.

To compensate for varying pressures, variable frequency drives regulate the speed of the biofilter blowers. As the pressure in a biofilter changes, the flow controller maintains a set flow rate and provides a constant flow of foul air to the media. For uniform distribution, the foul air is diffused beneath the media by perforated piping.

The media layers include a base liner, a graded gravel underdrain, a pea gravel layer for air dispersion, a bark layer, and the main media consisting of a combination of bark mulch, peat moss, topsoil, and oyster shells. An automatic sprinkler system provides water to keep the media moist during dry weather periods.

OTHER FEATURES
Influent Pumping. For the secondary treatment project the firm influent pumping capacity was required to be upgraded. The original pumping station incorporated bar screens with 25mm spacing. It was determined that with the installation of trickling filters in the secondary process, screens with 12mm spacing would be required. After an evaluation of alternatives, it was concluded that construction of a new influent pumping station, with new climber-type bar screens, would be more economical than expanding the existing station and replacing the existing screens. The existing station is being converted to a maintenance facility.

Three 6.3 m3/s influent pumps lift screened raw sewage from the wet well to the influent channel of the preaeration tanks. Pump speed and the number of pumps operating is varied based on the water level in the wet well. This allows the wet well to be relatively small. There are three influent pumps which are vertical mixed flow dry pit units, each driven by an 1100 hp, 275 rpm electric motor, with variable frequency drive.

The new influent pumping station includes screenings handling facilities as well as mechanically-raked bar screens. Screenings are raked from the screen bars into shaftless screw conveyors and directed to pneumatic ejector vessels. At preset intervals, the screenings are pneumatically transferred from the ejector vessels to screenings compactors in the screenings handling area. The compactors dewater and compact the screenings that are then discharged to screenings hoppers located above truck bays.

Primary Treatment Improvements. As the project was basically an upgrading of a primary treatment plant to secondary treatment, modifications to the primary treatment area were limited to those necessary to incorporate secondary treatment. These included reconstruction of the influent channel which feeds the preaeration (grit) tanks, covering of the preaeration tanks to allow foul air collection and replacement of the sedimentation tank v-notch weirs with submerged launders. The submerged launders reduce odour production and allow sedimentation tank level to be controlled by varying the trickling filter pump speed. Hydraulically operated sluice gates were installed at the end of the primary effluent channel to divert primary effluent to the new secondary treatment facilities while still allowing emergency bypass of primary effluent to the outfall.

SUMMARY
Over the 1990’s decade, the GVS&DD has engaged in the process of bringing about a comprehensive upgrade to secondary treatment at the Annacis Island WWTP which serves a population equivalent of nearly one million. A comprehensive selection protocol resulted in choosing the TF/SC process for providing secondary treatment. New large trickling filters, aerated solids contact tanks, secondary clarifiers and an array of other related facilities and improvements resulted in a project valued at nearly $470 million CDN. An interim Phase 1 Secondary Treatment process was successfully commissioned which used the aerated solids contact tanks and secondary clarifiers to provide full activated sludge secondary treatment to a portion of the plant flow for about a one year period prior to completion of the trickling filters. The trickling filters are expected to be commissioned in late 1998, thereby allowing secondary treatment of design flows. The plant's proximity to nearby businesses and a major commuter bridge required that odour treatment facilities be built for both primary treatment and secondary treatment facilities. Odour treatment was achieved with use of caustic only scrubbers for sulphide removal followed by biofilters for foul air polishing. The entire Annacis Island Secondary Treatment Upgrade project is anticipated to be complete in the year 2000.

REFERENCES
Parker D., Krugel S., McConnell H., Littleford D., Palsenbarg R., and Esping D. (1993). Selecting the TF/SC process for secondary treatment for Vancouver, Canada. Proceedings of the 66th Annual Conference of the Water Environment Federation.