Clear Choice Emerges For Small-System Iron And Manganese Removal

By Kevin Westerling,
@KevinOnWater

Pennsylvania American Water (PAW) was receiving complaints about the unsightly color of their water, which contained manganese well above U.S. EPA guidelines. The quest began for a cost-conscious and easy-to-operate solution (appropriate to a small, stand-alone system) that would restore water quality by greatly reducing both manganese and iron.

It’s tough to hide excess iron and manganese in the drinking water supply (not that a public water supplier would be so inclined) when customers show up to the utility’s front doors with stained clothes and faucet filters as evidence. That potential scenario was a rising concern for the staff at Mid-Monroe Water System, a small community water system serving 600 customers (0.110 MGD) in Middle Smithfield Township, Monroe County, PA.

More textbook indicators of a problem are the U.S. EPA’s secondary maximum contaminant levels (SMCLs) — benchmarks for maintaining proper taste, color, and odor that are non-mandatory on a federal level, but are enforceable by the Pennsylvania Department of Environmental Protection. The SMCLs for iron and manganese are 0.3 milligrams per liter (mg/L) and 0.05 mg/L, respectively, and Mid-Monroe was egregiously over the guideline for the latter. With peaks of about 1.0 mg/L, manganese was the (not-so-clear) culprit for the majority of discolored-water complaints. Iron also contributed to the problem with measurements as high as 0.45 mg/L.

The chart below shows the typical (average) raw water quality concentrations at Mid-Monroe. The system is comprised of four groundwater wells, two treatment facilities, and two storage tanks, privately owned and operated by PAW since 2002. Wells 1 and 2 combine for treatment into the south entry point (EP 102), while wells 3 and 5 combine for treatment at the north entry point (EP 105).

Raw water quality at the Mid-Monroe Water System

With EP 102 identified as the trouble spot, PAW focused on a solution that would reduce iron to less than 0.1 mg/L and manganese to less than 0.03 mg/L at that location, and do so in a cost-effective way. As a stand-alone, mainly remotely-operated system, capital expenditure, operational expenses, and operations and maintenance (O&M) requirements — or lack thereof — were all important considerations.

The previous approach to controlling iron and manganese (Fe/Mn) consisted of sequestering via blended polyphosphate chemical addition at the entry points, system-wide flushing twice per year, and simply maximizing the use of EP 105’s better water quality to lessen reliance on EP 102. However, due to the ineffectiveness of sequestering at these concentrations and a declining yield from EP 105, it was time to bring EP 102 up to standards that would meet customer expectations on its own. This meant a new treatment technique for secondary contaminants, particularly manganese, would need to be implemented.

“We were basically shutting off EP 102 manually and only turning it on as needed because the water quality was so bad,” said PAW Project Manager Daniel Rickard, P.E. “We were running EP 105 about 23 hours a day, and we were dewatering the wells.”

Multiple Approaches Considered

A number of treatment techniques were considered, including membrane filtration, ion exchange resins, manganese greensand filtration, and proprietary media filtration. Here’s how each fared in the evaluation.

• Membrane filtration: The microfiltration process would require a strong oxidizer to effectively precipitate 100% of the iron and manganese prior to filtration, and perhaps additional operational steps, such as removing the oxidant before it gets to the filters, in order to protect the membranes. The filters would need to be backwashed regularly, and would also likely require a periodic clean-in-place procedure utilizing a strong acid to remove stubborn foulants from the membranes that would cause high pressure loss across the filters. That special attention, in addition to the task of setting and monitoring enough oxidant contact time, would be operationally infeasible for the limited staff available, as well as cost-prohibitive.


• Ion exchange resins: Often used for water softening, demineralizing, and nitrate removal, ion exchange resins would be a better fit if the Fe/Mn in the source water was completely dissolved, but at Mid-Monroe it is 95% dissolved. The non-dissolved 5% would likely clog the resins and cause operational problems. In addition, ion exchange is not recommended for applications where the combined concentration of Fe/Mn exceeds 0.3 mg/L, which was the case at Mid-Monroe. These operational caveats meant that ion exchange resins were less than ideal for this application.


• Manganese greensand filtration: Greensand, a naturally occurring mineral mined out of the United States’ eastern seaboard, offers natural ion exchange and good adsorption properties — proven over 100 years of water treatment use. However, the oxidation/reduction process required for Fe/Mn filtration requires pretreatment with a potassium permanganate chemical feed, as the manganese greensand needs to be continuously, or at least intermittently, regenerated.  In addition, the natural greensand media is relatively limited in hydraulic loading and contaminant removal capacity, as well pressure differential tolerance, compared to some other available filter-media types.  Once again, the operational drawbacks were too onerous for the site in question.


• Oxidation-filtration with propriety media: The solution chosen by PAW uses a portion of the applied disinfectant — in this case, liquid sodium hypochlorite — to reduce the insoluble Fe/Mn (oxidation) by the presence of a pre-applied manganese dioxide coating on the engineered media (Omni-Sorb by Severn Trent), which then adsorbs the oxidized metals (filtration). According to Rickard, the benefits include:
  
  
  • Higher hydraulic loading rates (7 to 10 GPM/SF [gallons per minute per square foot]) and capacity (700 to 1,200 grains/SF) than greensand (2 to 5 GPM/SF and 500 to 700 grains/SF)
  • Small footprint (two 6-ft. diameter pressure vessels, which fit within a 1,500 sq. ft. building)
  • Does not require long contact time (reaction is instantaneous)
  • Does not require preconditioning (permanganate feed, coagulants, etc.) or regeneration
  • No brine or salt required
  • Lower head loss across filters
  • No clearwell storage and/or high-service booster pumping required with pressure filters
  • Fully automated
  • Low operating cost (~$2,900/year, including oxidant, media replacement, power, etc.)

Dual pressure filter system

Design Elements And Considerations

In addition to the new media pressure filtration system put in place for Fe/Mn removal, the complete treatment system features a pre- and post-filter chlorine feed, a zinc orthophosphate corrosion inhibitor feed (post-filter), 4-log disinfection of viruses (achieved via 12.5% sodium hypochlorite and 60 feet of a 36-inch contact main), and a 22,000-gallon backwash/filter-to-waste tank.

The facility is fully monitored and controlled through remote terminal units (RTUs) and programmable logic controllers (PLCs) facilitated by touch-screens and cellular data/voice signals to the regional office. The system allows for automated control of the chemical feed systems (chlorine residual, etc.) the filter system (pressure differential, turbidity, backwash, etc.) and security features (intrusion, fire, etc.), as well as the level, flow, and pressure of both the wells and the backwash tank (recycle and solids pumps).

The backwash aspect is significant because the treatment process was estimated to generate an average of 11,380 gallons of wastewater per day, which would be cost prohibitive to discharge into the public sewer system. The township sewer authority charges $6,500 for the purchase of a single EDU (equivalent dwelling unit), resulting in a total charge of $370,500 for the 57 EDUs that would be required. Clearly, an alternative was needed to handle the system backwash.

PAW equipped the wastewater tank with dual submersible recycling pumps, controlled by VFDs (variable frequency drives) to operate at speeds of 15 GPD (~10% of raw water influent). The pumps return the clarified supernatant from the tank back to the headworks of the plant for a recycle capacity of 94% of the total backwash volume.  Dual submersible residual waste pumps (15 GPD) discharging to the municipal-owned public sewer system were also added, along with an air gap, to dispose of the settled sludge after the oxidized iron and manganese is settled out — a process that requires approximately 8 hours.

A schematic of Entry Point 102 (well 1 and well 2)

Putting The System Together, On A Budget

At a cost to American Water of $1.25 million (see table below for breakdown), EP 102 came online in December of 2011 after roughly a year of construction. The two wells were able to remain in operation for the bulk of that time, with the exception of about 30 days — during which time pump, piping, and electrical maintenance took place — when treatment was assumed entirely by the Mid-Monroe’s two source wells at EP 105. However, the increased burden that had fallen on EP 105 — during and even prior to the new construction — was lifted for good upon EP 102’s completion. In fact, EP 105 can lean on the new facility, if need be, rather than vice versa. Both facilities, however, are equipped with emergency generators and are built in accordance with Pennsylvania flood plain management regulations.

Construction cost summary

Additionally, Rickard noted the following capital cost-saving measures that were implemented by PAW:

• The project was competitively bid for construction.
• The high-cost, long lead-time equipment was pre-procured to avoid schedule delays and contractor mark-up.
• Engineering for the project was completed in-house, so an outside engineering consultant was not required.

A Clear Winner

Manganese levels in the drinking water are now less than 0.01 mg/L, which is a 98% reduction over the previous average concentrations of 0.54 and well below the SMCL of 0.05. Though iron was less of a problem, previously at 0.18 mg/L (SMCL = 0.3), it was reduced 94% to less than 0.01 mg/L. Furthermore, the water is crystal clear, rated at 0 platinum-cobalt (Pt-Co) units compared to 5 Pt-Co prior to the site rehabilitation.

Having all the wells producing quality water also provides much-needed efficiency and redundancy of sources, according to Rickard. “We optimized our reliability in the system by getting Entry Point 102 back in a state of regular functionality, so we're able to give the other wells a break as needed,” he said. “In the long run, it's much less wear and tear on the well pumps and the wells themselves.”

While the project’s return on investment is hard to quantify in dollars — the capital and operations costs were kept at a minimum, but are expenditures nonetheless — the significantly improved quality and reliability of the water supply has appeased regulators and made for much happier customers.

These days, if Middle Smithfield Township’s whites aren’t gleaming white on laundry day, they can't blame the water.

Are customers vocal about the water quality at your municipality? Absent a strict mandate, are customer complaints enough to compel action (and funding)? Please share your thoughts below.