When a bare metal surface is exposed to the environment, a surface layer of oxide typically forms relatively quickly. This layer then acts as a physical barrier, greatly reducing the rate of further corrosion. However, water flowing rapidly over the surface can quickly dissolve the oxide layer. When this happens, a fresh surface is exposed and the oxide layer forms again, leading to a cycle of accelerated corrosion.
Flow-accelerated corrosion is a particular risk in high-pressure piping, such as in power stations. It may also occur where sub-sea structures are in contact with flows of seawater, or on high-speed vessels. The most high-profile failure due to flow-accelerated corrosion was the rupture of a pipe elbow at the Surry Nuclear Power Station in 1986 (not pictured). This caused four fatalities and tens of millions of dollars in repair and downtime costs. Due to the significance of this accident within the highly regulated nuclear power industry, it resulted in a comprehensive international response to understand the parameters which affect flow-accelerated corrosion, as well as methods of detecting and controlling it. Much of this work was carried out by the Electric Power Research Institute and utilities EDF and Kraftwerk Union.
Flow-accelerated corrosion continues to cause sudden failures of high-pressure and high-temperature feedwater pipes. These dangerous accidents can result in deaths as well as very high downtime costs. Below are listed the tools and techniques that can be used to identify likely locations and prevent this problem.
CAUSES
As described above, flow-accelerated corrosion (FAC) is caused by the protective oxide layer being continuously dissolved by water flowing rapidly over the surface. This is similar to erosion corrosion, in which particles, bubbles or cavitation mechanically remove protective oxide. Because FAC depends on the oxide being dissolved, it requires specific electrochemical circumstances to occur, and it increases with flow velocity. Stainless steel pipes do not experience FAC; it is typically seen in carbon steel pipes carrying wet steam or ultra-pure deoxygenated water. The presence of dissolved oxygen or increased pH greatly reduce FAC.
In practice, erosion corrosion and FAC often work together to damage pipework. When considering FAC, both flow through the pipes and chemistry must be considered.
Due to a physical characteristic called the ‘no slip condition,’ a fluid (gas or liquid) that is in direct contact with a surface cannot flow across the surface. The fluid can be thought of as composed of molecules that act like little rubber balls. The molecules can bounce off the surface, but they have perfect grip and cannot slide along the surface when they are in contact with it. Although the fluid directly touching the surface can’t slide along it, a layer of fluid just a tiny distance away can slide past the layer below. The resistance of the fluid to this shearing action is known as its viscosity.
The actual cause of viscosity is the exchange of kinetic energy between the molecules in the layers, which are constantly exchanging molecules due to Brownian motion. Flow of a fluid over a surface therefore always has zero velocity at the surface, and as each layer slides over each other, the velocity increases with distance away from the surface. The region of increasing flow velocity is known as the boundary layer. In a fully-developed laminar flow through a pipe, the boundary layers merge into a velocity profile which has maximum velocity at the centre of the pipe and a parabolic velocity distribution. However, when the fluid first enters the pipe, the boundary layer may be very thin, with approximately constant flow velocity over most of the cross sectional area of the pipe. If flow becomes turbulent, the velocity profile will be flatter and velocity close to the surface will be higher.
Anything which increases the flow velocity close to the surface of a pipe will increase the risk of flow-accelerated corrosion. This may be caused by acceleration of the flow, diversion of the peak velocity towards a surface, or the flow becoming turbulent. Physical causes include restrictions, bends and irregularities.
Flows tend to become turbulent at high Reynolds numbers. The Reynolds number is therefore an important concept to understanding flow, it is given in Figure 1 below. For a given flow rate, the diameter of the pipe also affects the velocity. For pipe flows, the Reynolds number is often given in a different form (Figure 2). For a circular pipe, the hydraulic diameter is simply the diameter. Since the area is a function of the diameter, it is possible to substitute for the area and simplify to give Figure 3. So it can be seen that, for a the same flow rate and fluid, reducing the pipe diameter will actually increase turbulence.
_0759a00f.jpg)
Where pipes bend, velocity at the elbow is increased dramatically and flow separation may occur. This means that bends are important sites with enhanced flow-accelerated corrosion and erosion corrosion. Similar effects are seen in manifolds and interconnections. Valves and throttles are also areas of enhanced risk for these forms of corrosion.
“Whenever a fluid flows through a bend, there is a radial pressure gradient developed by the centrifugal force acting on the fluid. Because of this, a double spiral flow field and a pair of counter-rotating vortices can also be observed inside the bend,” says Prasun Dutta of the department of aerospace engineering and applied mechanics at the Indian Institute of Engineering Science and Technology.
FEEDWATER CHEMISTRY
pH and oxygenation are critical parameters for corrosion. Corrosion in iron pipes increases dramatically when pH moves in either direction outside the optimal range. A typical safe operating zone for condensate and feedwater pipes is between pH 9.0-9.6. Ammonia and organic amines are typically used to control the pH in feedwater systems.
There are two stages to identifying FAC sites. Firstly, analysis of pipe flows can be used to predict where FAC is likely to occur, and secondly, measurements of pipe wall thickness can be used to measure the actual effects.
At its simplest, a basic understanding of flow, as described above, can be used to consider where FAC might be worst. Once potential sites have been identified, non-intrusive wall thickness measurement can be performed, using ultrasound or eddy current methods, to measure the actual