[modula id=”4206″]

Water Filtration

In the later 1970s, and even before, many design changes to air conditioning systems for high rise office buildings resulted in the installation of smaller diameter condenser water pipe.  Rather than large 18 in. pipe runs straight to and from the roof to the central chiller plant, smaller package A/C units and heat pumps were installed on each floor using 4 in. or smaller diameter pipe – and lots of it.

• Same Issue – Greater Impact

While rust product and airborne captured particulates have always been a problem for open condenser water systems, their impact to individual A/C units has been far greater due to smaller diameter schedule 40 pipe of lesser wall thickness, threaded assembly in many cases, and extended horizontal runs.

Moderate and even low corrosion activity significantly impacted the smaller pipe. Rust product generated anywhere throughout the system then had a tendency to settle out in low flow areas, and especially at the lower floor horizontal lines.  Package A/C units and heat pumps designed to shut down flow during off hours were especially vulnerable.  Green variable flow pumps made the issue even worse.

For many buildings, indication to a corrosion problem was discovered when examples of pipe were removed, or when maintenance was performed to the cooling towers to refrigeration condensers.  Rust debris and scale collected in buckets spelled more and more trouble as their volume increased and the rust filled pails multiplied in number.

• Early Solutions

In order to reduce the higher level of rust and particulate deposits, the very first solution to the problem was what had worked so well for swimming pools – the sand filter.

Building engineers and managers installing pool filters marveled at how quickly they turned their cloudy and rusty colored water back to clear.

For many reasons, most residential and even some commercial pool filters failed to hold up to the workload – resulting in the appearance of more rugged, advanced units.  While offering many design benefits, greater controls, monitoring, safety, different  mediums, and varying particle retention efficiency, all still remained essentially sand filters.

• Multiple Limitations 

The limitation of sand filters were immediately obvious, or should have been.

First was their low flow capacity in comparison to the much greater size and GPM circulation rate of a typical condenser water system.  This eliminated any possibility of filtering the entire loop, and in most cases filtered only 5% of the flow rate or less – thereby gaining the term, “side stream filtration.”

Given the low flow rate and filtering capacity of sand filters, the mistaken assumption was quickly adopted that whatever rust and other particulates were not filtered out on the first pass would simply continue to circulate around and around the loop forever until they were eventually captured.  Of course, there was no possibility that such an assumption would hold true for the overwhelming volume of rust and debris common to any condenser water system.

First of all, most filter units were installed in response to a visual observation of hard rust deposits, not colloidal and essentially invisible fine particulates.  Generally, expensive sand filtration units were installed only in response to a clearly visible rust threat, and often panic, rather than as a precaution to a problem that may not even exist.  Such larger rust deposits were there for a very good reason – gravity.  After years or decades, they were quite unlikely to start circulating back throughout the piping system on their own.  Likewise, newly created or loosened rust product would likely settle out just as quickly as they had in producing the original observation.

In either event, heavier settled rust product was never going to again move unless some outside force were applied.  The fact that strainers, cooling tower pans and condenser heads still filled up with scale and rust deposits requiring regular maintenance was alone proof to the fact that the sand filter installed to control the problem was not very effective.

This soon led to modifications whereby suction headers were installed inside the lower cooling tower basin in order to capture the already settled debris.   Relocating the filter inlet from the side of the pipe to the tower sump placed it where higher concentrations of debris existed.  Filtered return water to the outer perimeter header at the cell pan then helped push settled debris toward the intake.

Yet again, the low GPM flow of the sand filter, divided by potentially 100 or more perforations at the inlet header, provided insufficient suction power to function as claimed.  Settled rust and other materials still clogged strainers and cooling tower nozzles, reduced heat transfer efficiency across the condenser tubes, and built up in the cooling tower pans.  Rust and scale deposits piled up right next to water jets intended to benefit the system.

Another limitation to a sand filter was its very limited loading capacity, and the fact that the influx of any larger rust particulates would quickly throw it into a backwash cycle.  Once backwash efficiency declined, an expensive re-bedding with new media was required. Small strainers installed prior to the sand bed itself proved that manufacturers knew it incapable of handling any larger debris.

Furthermore, the microbiological activity always present in any open condenser water system would eventually clog the pores between the sand or media particles in a process termed “Bioclogging.”   This in turn reduced filter flow rate even further; requiring both new media and sterilization to correct.  Where chemical treatment and biological monitoring was not maintained, the development of Legionella on the media added another threat.

In some cases, “channeling” would occur, with a pathway created through the media in effect providing no particle retention.  For sand type filters, a worst condition existed where the sand or filtration media was lost into the piping system itself – creating tremendous new problems to resolve.  This was especially damaging when occurring to closed piping systems.

• Apparent Success

What the sand filter produced, however, was very clear water.  Cloudy or rusty looking cooling water inside the pan, in a sample container, or observed flowing through the corrosion coupon sight glass, quickly cleared up once the filter was in operation.

Competition between filter manufacturers focused toward lowest retention rating rather than removal capacity.  Units were selected exclusively based upon their ability to remove sub-micron size particles for a higher degree of filtration than provided to most drinking water.  Sand filtration systems installed in response to known large volumes of rust product, buckets of rust in many cases, were credited with cleaning the entire piping system remarkably fast.

Problem solved, a pat on the back, job well done – success!

Yet, in reality, the filtration unit had only removed the smallest suspended particulates causing turbidity, instead of cleaning the piping system of the rust for which it was installed.   Just an illusion of success.

• An Unequal Threat

In fact two separate problems and consequences exist for all piping systems depending upon contaminant weight and size.

The first relates to all heavier settled out deposits which are the primary cause of under deposit corrosion common to by-pass lines, crossovers, futures, and long horizontal pipe runs.  These are generally large and unmovable deposits that produce the most severe damage to a condenser water system by greatly reducing the effectiveness of the chemical treatment program.  Larger settlements of rust also facilitate micro-biologically influenced corrosion (MIC), and establish very aggressive corrosion cells capable of 25 MPY and greater localized wall losses.

Yet such heavier deposits are virtually untouched by the addition of not only sand filtration, but by any form of filtration unless some added chemical or physical force is applied to return them back into circulation for capture.  Nevertheless, sand filtration is recommended as the solution even though no possibility exists for those deposits to even reach the filter in the first place.

In contrast are the finest rust and other particulates that impart cloudiness, discoloration, or turbidity to the water, but present no real threat in terms of an aggressive corrosion loss or pipe constriction.  These smaller particles of size 25 micron and less typically remain in motion for at least a few passes before coagulating together in sufficient size to settle out.

• The Most Important Factor

One of the most critical aspects to any water filter installation is the selection of the location where water is to be drawn.  But for virtually all of the ultrasonic investigations of heavily corroded condenser water systems we become involved, that decision has been poorly arrived.  Or has it?

Given the critical importance of water filtration to virtually all HVAC applications, we typically address the subject in our ultrasonic based piping assessment reports.

• • • Is water filtration provided?
    • Has the appropriate type and sized filter unit been installed?
    • Is the filter operational, or has it been removed from service?
    • Has the filter been effective?
    • How much rust and other debris is removed per cycle?
    • Is it installed correctly?

• Actually, No

Quite often, the answer to the last question answers all others, and in fact, we find most filtration units installed incorrectly.   Not only are most water filters installed in the worst possible location, but they have been installed incorrectly under the strict direction of either the filter manufacturer, sales representative, or system design engineer.

At left, the take-off to the sand filter at a 24 in. condenser water return main is at the 9 o’clock side  position, or about 1 ft. above the 2 in. of known rust sediment on the bottom of the pipe which prompted the filter purchase.  Multiple better locations were available including at the bottom elbow of a vertical drop although the filter representative identified this as the ideal location.

In another condenser water example below left, water moving to the right will force any rust through inertia and centrifugal force against the outside pipe wall.  Yet the sand filter inlet has been firmly specified to the inside wall.  At center, this hot water heating system flows to the left which again forces any particulates through centrifugal force to the outside wall.  But again the take-off to the filter is installed at the inside wall.  At right, the take-off from an 18 in. chill water main is at the 11 o’clock position.  While not as poorly planned as it would at the top, the draw location still makes this $100,000 sand filtration unit not worth the electricity it costs to run.

In the following example at left, three individual sand filtration units provided to the condenser water, chill water, hot water heating systems at the same property were all installed with their take-off inlet at the 12 o’clock top dead center position where no heavier rust particulates travel.

At a cost of over $325,000, not one of these sand filters will remove any of the rust deposits for which they were intended.  Property management will believe their problems have been solved up to the time a corrosion related piping failure occurs.

At far left, the take-off to a centrifugal separator offers a stunning example of multiple design errors.

First, the installation of a centrifugal separator to a closed heat pump system having a minor fine rust problem is worthless since it can only remove large particulates of around 75 micron and greater.  The photo at right shows this smaller diameter run-out pipe to a heat pump unit in nearly pristine condition.

Second, the same exact principal behind the operation of the centrifugal separator is used against it.   Separators function by using velocity and centrifugal force to spin out heavier debris for capture and removal.  Yet that same inertial force will now propel any particulates capable of being captured by such a filter straight up and right past its inlet. Indeed, even rust particles of lightest mass are not going to stop dead in order to make this 90° left hand turn into a filter fundamentally incapable of effecting their removal.

And finally, a valve throttled down to 40% reduces any possible benefit from a unit which requires highest possible  water velocity.  Obviously, this expensive filter installation was worthless from day one!

• Proof

Always overlooked, definitive proof to the effectiveness of every aspect of any water filter installation is provided by what it has removed.  For those units backwashing into a settling tank, the capture of a quart of silt after 2 years of operation quickly answers the question.  A backwash counter showing 33 cycles after a year of operation similarly proves that either it is unsuited for the system, is malfunctioning, or most likely – that it has not been installed optimally.

For most installations, discharging the filter directly into a floor drain obscures filter effectiveness entirely.  It also proves that had that large volume of rust deposits in the condenser water system been suddenly transferred over to the waste system, entirely new pipe clogging issues and emergencies would now exist.

• Supporting The Error

Remarkably, such poor installations have been heavily defended by those in the filtration and water treatment industries through a variety of arguments ranging from questionable to utterly absurd.

Clients questioning the placement of a filter take-off at the top or side of the pipe  have been advised it better not to second guess experts in the field.  Others have been advised that scientific studies have proven the top of the pipe as the absolute most effective location for the filter draw.

Our recommendations to either relocate the existing filter take-off or install a new unit at its most effective location are inevitably changed back to certain failure.

• But Why?

Sand filters are spectacular at keeping a swimming pool crystal clear due to the fact that they have minimal fine particulates to remove.  No one shovels buckets of rust, dirt, scale and other debris into a pool on a daily basis.

Conversely, a 30 story office building with 24 in. risers and a 5 MPY corrosion rate will  lose about 934 lbs. of steel every single year!   Oxidized, that wall loss translates into approximately 37 cubic feet of rust product annually, which is a monumental task for a sand filter or any form of water filtration unit to control.

For building management, unfortunately, clear water still remains the sole indicator of success up until a piping failure occurs due to the problem which was supposedly solved.

Meanwhile, sand filters will continue to be specified for installation where they provide the illusion of a benefit, as opposed to be installed correctly – which would then prove them as totally unsuitable and inappropriate for the task.

You can view and download our two page handout on this subject below.

© Copyright 2023 – William P. Duncan, CorrView International, LLC