The Kenosha Water Utility's (KWU) 22-mgd microfiltration membrane plant broke ground a decade ago as the world's largest low-pressure membrane system producing potable water, the first in Wisconsin and the first direct-membrane filtration process on Lake Michigan. Since the O. Fred Nelson Water Treatment Plant went on line, hundreds of people from countries such as Canada, France, Germany, Australia, New Zealand, China, Japan, Korea, and England have toured it to examine the leading-edge technology.

KWU is well satisfied with the decision to build a membrane plant. However, since the December 1998 start-up, the utility's demand for water has evolved. Operations and maintenance methods have improved, as have membrane quality and replacement options. Lessons learned since then include the importance of regular membrane integrity testing, regularly scheduled chemical cleaning, preventive maintenance, staff training, financial planning, and a positive relationship with regulatory agencies.

KWU built the membrane plant to update a 40-mgd conventional treatment facility, which consisted of two plants. The original 20-mgd “west plant” started as a steam-powered pumping station in 1895 and was upgraded to a conventional filtration plant in 1917. A 20-mgd “east plant” was added in 1964. After evaluating several options, and in the wake of the 1993 cryptosporidiosis epidemic in nearby Milwaukee, KWU decided to upgrade and continue operating the east plant and replace the west plant with a new membrane filtration system.

While the installed cost of the membrane plant turned out to be less than the projected cost of a conventional process, the primary motivation behind KWU's decision to become a membrane pioneer was the commitment to provide the highest possible water quality for its customers. KWU's vision has been confirmed every day since the membrane plant's start-up by producing high-quality potable water that meets all U.S. EPA and Wisconsin Department of Natural Resources (DNR) drinking water requirements.


In a direct filtration process of Lake Michigan water, Wisconsin DNR requires testing membrane integrity three times per day, at 8-hour intervals, along with continuous on-line turbidity measurements. Results from a series of fiber cut tests conducted during the plant's start-up convinced the DNR that a pressure decay test (PDT) would provide more sensitive measurement of unit integrity and reflect actual performance better than that obtained from turbidity or particle count measurements alone. Membrane-filtered water turbidity measurements are at the lowest end of the on-line turbidity instruments' range, always <0.02 nephelometric turbidity units (NTU), regardless of raw water turbidity. Turbidity can exceed 100 NTU, and in some years has exceeded 200 NTU, from storms on Lake Michigan.

The PDT used at KWU is a fully automated sequence conducted with a challenge test pressure of 15 psi on each membrane skid, and can detect a breach in membrane integrity as small as 3.0 microns. Wisconsin DNR requires operating staff to respond to the smallest change in PDT value by finding the individual membrane module(s) that triggered the change, isolating the leak, and scheduling a repair.

With more than 10 million fibers in each skid, looking for a leak may seem like trying to find a needle in a haystack. However, using an air hold test when responding to a change in PDT allows air bubbles to be heard, pinpointing a specific membrane module leaking air. This enables plant staff to identify even the smallest breach in integrity, isolate the leak, and fix it as part of scheduled maintenance.

The Wisconsin DNR's “zero tolerance” to changes in PDT created more fiber repair work than KWU initially anticipated, but never required additional staff. A set of replacement membranes installed in 2004 has proven to be more durable than the original supply, reducing the fiber repair work.


Maintaining membrane integrity is essential to providing the best possible treatment. Regular, effective membrane cleaning is the most important way to extend membrane life and ensure that the filtration capacity will consistently meet needs. Unlike high-pressure membranes that can weaken with repeated chemical cleaning, low-pressure membranes will decline in capacity if the membranes are not washed or are cleaned ineffectively. Low-pressure membranes are washed with chemicals using temperatures that do not degrade performance.

Transmembrane pressure (TMP) is the driving force that moves water through the membranes to the filtrate side of the system. TMP increases as membranes get dirty (foul) when filtering the same amount of water. Chemical cleaning removes foulant from the membrane surface and lowers TMP. New membranes can be operated for 2,000 to 3,000 hours before the TMP rises to the point where a chemical clean-in-place (CIP) is required.

The dirtier the membranes become, the harder they are to clean. Eventually, no matter how often the membranes are washed, some foulant remains. At this point, the membranes must be replaced to restore lost capacity. KWU has implemented a 500-hour run time between CIP for all membranes regardless of age, and expects to get five years of life between membrane changes. Also, following high-turbidity events, “deep cleans” are performed to further maintain filtration capacity.

Initially, KWU used proprietary chemicals supplied by the membrane manufacturer for the cleaning process. Process engineers from the manufacturer worked with plant staff to develop an effective cleaning regimen using locally purchased bulk chemicals, saving KWU a considerable annual cost. Commodity chemicals also cost less than replacement membranes. It is significantly less expensive to chemically clean the membranes regularly and often, thus extending plant capacity, than it is to forgo cleaning and replace membranes more often.


Before the Kenosha plant went on line KWU staff visited the Marquette, Mich., plant. In addition, a membrane “buddy system” allows plant personnel at new membrane installations to visit plants already in operation. Kenosha has hosted personnel from Carmichael, Calif., and they meet regularly with neighboring plants like Manitowoc, Wis., to share information.

KWU is a lean operation with a single plant operator on duty per shift. An extensive and modern supervisory control and data acquisition (SCADA) system enables the operator to run both the membrane and conventional plant. KWU's six operators are cross-trained in operations and maintenance duties. Two teams of three work a 10-day rotation, during which operators work three maintenance days and seven operating days. Repairs are made on maintenance days, while membrane monitoring occurs on all days.

The membrane system is a highly automated process with advanced programming to help operators do their job. A two-level alarm system is built into the controls to protect the membranes and equipment from extended operation when something is amiss. The lower level sends a signal with a printout describing the problem in detail while the system continues to run. The higher level can shut down a skid if an emergency condition is detected. By paying close attention to all alarms and making timely corrections, operators can avoid an emergency shutdown condition.


KWU has a long-term agreement with the manufacturer that calls for membranes to be replaced on an as-needed basis, regularly scheduled service visits, and an annual process audit. The agreement helps KWU to budget its costs up front.

In the 10 years following the membrane plant's start-up, the manufacturer has developed an improved process using a different, chlorine-resistant membrane material. While the cost of replacement membranes is included in the long-term agreement, the cost to convert the existing system to operate under the new process is not included. Any investment in remodeling must be justified by operating cost savings, including longer membrane life, lower power consumption, fewer chemicals needed for cleaning, and less backwash wastewater generated for disposal.

Another development since the plant's start-up is that several of Kenosha's largest industrial customers have left the area, significantly reducing the demand for water. KWU has responded by operating the plant at a higher rate through the night, taking advantage of reduced electrical load charges while replenishing storage tanks within the distribution system, and then coasting during the day.

Today, KWU is at a fork in the road: to reload with another set of membranes and continue as before, or invest additional money to convert to the lower-cost operating option. If the membrane plant were producing near its rated capacity, the cost of converting would be more easily justified. The manufacturer, who has a partnership interest in the long-term success of the plant whichever path is chosen, is maintaining the data required to support Wisconsin DNR's permitting of the new process, should this be selected.

— Davis is technical sales manager for Memcor Products at Siemens Water Technologies. Lewis is superintendent of the O. Fred Nelson Water Treatment Plant in Kenosha, Wis.