Circulation induced pipe corrosion

By Timothy Allinson, P.E.,
Murray Co., Long Beach, Calif.

Back in April, I wrote about copper pipe corrosion accelerated by water chemistry, mainly thought to be chloramines (and their byproducts) that the water utilities are using as a water disinfectant with growing popularity. The article noted that copper pipe corrosion can also be accelerated by design issues, one being water velocity and pipe sizing. This design issue is of particular concern when it comes to hot water circulation systems.

The most common type of copper tube failure is cold water pitting, also referred to as Type 1 pitting. The cause of Type 1 pits is the subject of much debate, but it is clearly a complex combination of water chemistry and is often attributed in part to fluxes used in the installation process. I am of the personal belief that it is more frequently the result of chloramines than fluxes, but much study needs to be done before that can be proven.

The second most common type of copper tube failure is erosion-corrosion caused by rapid turbulent pipe flow and aggravated by aggressive water quality. This type of impingement failure is more common in circulated hot water systems than in cold water piping.

Water velocity in hot and cold water distribution is fairly easy to control. Most codes are written around maxima of 8 fps for cold water and 5 fps for hot water. The reason the hot water velocity is less than the cold is because hot water is more aggressive than cold water. That is why it is better for cleaning. Since it is more aggressive, velocity-induced erosion-corrosion is more likely to occur.

When copper pipe systems are newly installed they begin to form a protective film of initial corrosion, which is sometimes referred to as a bio film. This film protects the tubing from further corrosion. When water velocities are elevated, the water can erode the bio film and make the pipe more vulnerable to further corrosion.

It is quite uncommon for erosion-corrosion to occur in straight lengths of pipe where flow tends to be laminar at the relatively low velocities of domestic plumbing. More often, impingement occurs at changes in direction, where flow becomes turbulent, or in areas where irregularities in the pipe’s inner surface cause turbulence. The turbulence erodes the bio film, allowing impingement damage to propagate.

Figure 1 is a typical example of erosion-corrosion in a 90-degree elbow in a hot water circulation system. The location of the impingement indicates that the water was flowing counterclockwise through this fitting. The turbulence occurs where the water exits the bend rather than where it enters the bend.

There are several reasons that erosion-corrosion is most common in hot water circulation piping. First, the water in hot water circulation piping flows constantly, 24/7, in most systems. In contrast, flow velocity in cold water piping is constantly changing from as little as zero fps to a theoretical peak of about 8 fps. This variance in velocity acts to protect the piping from experiencing continual turbulence in the exact same area.

Second, large circulated hot water piping systems are inherently difficult to balance. Installation conditions always differ from design conditions, and calculated circulation rates give way to rule-of-thumb practical minima. A system with 100 hot water risers might have a calculated circulation rate of 25 gpm, or about 0.25 gpm per riser. But in the field it is difficult, if not impossible, to balance a system and have it stay in balance with as little as 0.25 gpm flowing through each riser. A more practical flow rate per riser would likely be 0.5 gpm to achieve balance.

Third, the actual pump head required for the circulation pump of any large system is as difficult to calculate accurately as the system flow rate — and the two go hand-in-hand. Most engineers will add a factor of safety to the pump head. If in practice the pump requires less pressure than it is capable of delivering, then the pump will ride out on its curve and produce excess flow and, with it, excess velocity.

Fourth, the circulation piping is often sized for the same design conditions as the hot water distribution piping, i.e. 5 fps. This velocity is fine as the theoretical peak of the distribution piping, where flow increases and decreases as a function of fixture demand, but is excessively fast for the circulation mains where flow is constant. For circulation mains, a velocity of 2 – 3 fps would be more appropriate.

Fifth, since design flow is often underestimated and pump head is frequently overestimated, piping winds up undersized for the actual field flow conditions. A pipe sized for 5 fps might have an actual flow of double that when all is said and done. Couple that velocity with the increasingly aggressive nature of America’s water supply, and it won’t take long for erosion-corrosion to cause leaks in the copper tubing.

The designer should practice sound judgment and not use the factors mentioned in this article as cause to dramatically oversize hot water circulation systems; for example, many engineers shy away from using ½-inch water piping. But the example above of a system with 100 circulated hot water risers, each with a practical 0.5 gpm minimum, would be well served by ½-inch piping for each of its circulated risers. At 0.5 gpm, the velocity in ½-inch tubing is still well below the 2 – 3 fps maximum suggested. The key is to identify where the cumulative effect of flow, pump head, water temperature and water quality might give reason for caution when sizing the hot water circulation piping.

Timothy Allinson is a senior professional engineer with Murray Co., Mechanical Contractors, in Long Beach, Calif. He holds a bsme from Tufts University and an mba from New York University. He is a professional engineer licensed in both mechanical and fire protection engineering in various states, and is a leed accredited professional. Allinson is a past-president of aspe, both the New York and Orange County Chapters. He can be reached at laguna_tim@yahoo.com.