Erosion corrosion is particularly prevalent in piping and tubing, where high flow rates across a metallic surface trigger erosion. The corrosion leads to pitting, which increases the turbulence of the liquid and can amplify the effects of erosion. This is the essence of accurate Thickness Measurement Location (TML) plan in the corrosion circuit.
This dataset and calculations are for educational purposes only.
Modeled and Analyzed by the author Roderick Paulino
Propylene glycol 80 is transported through a butt welded pipeline made of S275N material at an inlet pressure of (200 kpa - 2 MPa) and a temperature of 22°C. Analyze the fluid flow through a 6" schedule 80 pipe with a 90-degree elbow. Assume adiabatic condition. Design the pipe bend (radius and thickness), calculate the erosion rate and design the appropriate pipe support base on the static analysis of boundary conditions. Predict the life span of the manufacturer's elbow vs using your designed pipe bend (extrados/intrados location points). Stratify your solution using vertical and horizontal piping configurations on the downstream and upstream side with gravity effect using different pipe sizes. Perform sensitivity analysis to determine the feasible configuration(s) with respect to boundary conditions.
References:
https://www.youtube.com/watch?v=8wbbM3YwEZQ (Depiction of Erosion/Corrosion effect)
https://www.mdpi.com/1996-1073/16/6/2707 (Numerical Study of the Erosion Distribution of Sulfur-Containing Particulate Gas in 90-Degree Gathering Elbow)
ILLUSTRATED FLUID DYNAMICS FOR INITIAL SET-UP:
Assuming 30C relieving temperature and fluid compressibility factor of 1.0, inlet loss of 0 psig, and the system is capable of handling back pressure for relief. NOTE: Ideally, at this condition, the required flow rate is limited to 0.004 m3 per second (rated) only for line relief sizing with a back pressure of approximately 2960 psig with 10% over pressure and assuming 210 psig for additional pipe resistance downstream. (The maximum outlet pressure exceeds the pressure limit of 2160.00 psig for available relief valve group 2.2 material (Stainless Steel) and 900# flange rating (as per ASME B16.34).
Velocity in m/s
Pipe size analysis for the optimum flow based on 3.6 flow rate using Darcy-Weisbach equation assuming 100 m length of pipe. The pressure regulator can be designed from the table above. For this discussion, we will constrain our analysis to 1.5 meter downstream and 1.0 meter upstream of the elbow with no pressure regulator (No relief valve reaction forces) taking the full pressure at 2 Mpa.
Initially, we will select Carbon Steel Seamless ASTM A53-B pipe S80. (2-Mpa 290-psig, 72-Fahrenheit 22-Celcius)
Specifying the process conditions and mechanical properties of 6" pipe for initial boundary conditions. Assuming that the pipe is resting on a steel beam of a piperack with 2mm gap and 0.30 coefficient of friction using a guide on both horizontal and vertical segment while putting an anchor at node 35 and a free end (to allow movement) at node 5 for simplicity. In actual design, if the the piping segment will fail at this condition, appropriate support type will be needed to offset the cold and hot loads including the thermal expansion of the pipe material.
We will use the default static load cases to analyze our piping system.
Maximum DY and DX occuring at this piping segment.
Maximum stress occurring at the segment (yellow) showing 821x amplified displacement at operating load case. Note that the displacement is acceptable and slight lift off of supports. It is within the normal deflection assuming that the pipe was properly supported downstream where force 1411.62 N and moment of 296.03 N-m were mitigated.
Velocity contour on the elbow at 2 m/s and 200 kPa. Even at this condition, the intrados of elbow (90-deg long radius) is subjected to high velocity of the fluid. Note that during fluid surge, the velocity can amplify the velocity and can enhance the erosion rate of the pipe.
Using Calcium Carbonate as inert material. Calcium carbonate particles can cause erosive wear when they are carried by the flow of propylene glycol and impact the inner surfaces of pipes, especially in areas where the flow is turbulent, such as bends, elbows, and valves.
These particles are abrasive, and when they repeatedly strike the pipe walls, they can gradually wear away the material, leading to erosion. This is on top of stress static loads based on pipe stress analyses as calculated above which is not yet included in the equation. This will require a different approach called transient structural system using 2-way FSI (Fluid-Structure Interaction).
The erosion can be worsened if the system experiences high fluid velocity or if the calcium carbonate particles are of significant size (usually occur during fluid surge). Below are the erosion rates as predicted by different models. The views are directly looking at the extrados of the elbow.
Linear striations on the pipe bend along chord line as simulated the location of the flow using calcium silicate as inert material to check this kind of phenomena.
The presence of calcium carbonate particles in the propylene glycol solution may occur if the glycol comes into contact with hard water or if there are leaks in heat exchangers or systems where water is used for cooling.
DPM Erosion Rate
OKA Model
FINNIE Model
MCLaury Model
Calcium carbonate particles may also contribute to erosion-corrosion. This occurs when the particles erode the protective oxide layer on metal surfaces, exposing the bare metal to corrosion. This combination of mechanical wear and chemical attack can accelerate material degradation.
Propylene glycol is a stable organic compound and does not typically react with calcium carbonate. However, in real-world piping systems, corrosion reactions often involve water, dissolved gases, and metal surfaces in addition to propylene glycol and calcium carbonate. If water is present in the system, it can dissolve CO₂ to form carbonic acid (H₂CO₃), which can initiate corrosion processes.
Note that if there are metal surfaces (such as steel pipes) in contact with the propylene glycol and calcium carbonate solution, potential corrosion reactions can occur due to carbonic acid formation.
CO₂ Dissolution and Carbonic Acid Formation:
In the presence of dissolved CO₂ (either from the atmosphere or from the breakdown of carbonate/bicarbonate ions), water can form carbonic acid:
CO₂ + H₂O → H₂CO₃ (Carbonic Acid)
The carbonic acid can then dissociate into bicarbonate (HCO₃⁻) and hydrogen ions (H⁺), which can lower the pH and promote corrosion of metal surfaces.
Corrosion of Metal (Iron) Surfaces:
In the presence of carbonic acid (from dissolved CO₂ or water in the system), metal surfaces, particularly iron or steel, can undergo a corrosion reaction:
Fe + H₂CO₃ → FeCO₃ + H₂ (Ferrous Carbonate)
Ferrous carbonate (FeCO₃) is a corrosion product, and hydrogen gas (H₂) is released as a by-product.
FeCO₃ can precipitate and form a protective layer on the metal surface, but this layer may be disrupted by erosion, leading to continuous corrosion.
Bicarbonate and Calcium Carbonate Equilibrium:
If calcium carbonate (CaCO₃) is present, it can act as a buffer, but under high temperatures or lower pH conditions, it can dissolve to release calcium (Ca²⁺) and bicarbonate ions (HCO₃⁻):
CaCO₃ + H⁺ → Ca²⁺ + HCO₃⁻
The bicarbonate ions may further dissociate, contributing to the acidic environment that enhances the corrosion of metal components.
Overall Corrosion Reaction (for Metal in Contact with Propylene Glycol, Calcium Carbonate, and Water):
In the presence of water, assuming there is a contamination, and dissolved CO₂:
Fe + H₂CO₃ → FeCO₃ + H₂
Anodic Reaction (Oxidation):
At the anode, the metal (iron in this case) undergoes oxidation, losing electrons and forming metal ions. For iron (Fe), the anodic reaction is:
Fe → Fe²⁺ + 2e− (Normally with steel pipes)
Cathodic Reaction (Reduction):
At the cathode, a reduction reaction takes place. In this case, with carbonic acid (H₂CO₃) being present in the system due to dissolved CO₂ in water, the cathodic reaction involves the reduction of hydrogen ions (H⁺) or carbonic acid to form hydrogen gas (H₂):
2H⁺ + 2e− → Fe²⁺ + H₂ (Hydrogen Gas)
Alternatively, if carbonic acid is directly involved, the reduction reaction can be expressed as:
H₂CO₃ → HCO₃− + H₂
Below, is the convergence plot of species to form the mixture and products of reactants showing the 3 reactions.
Looking at the intrados of the elbow when reacting h2 to h+ (second cathodic reaction).
Corrosion prediction in mm / year occuring in the corrosion area where the x-axis is the varying curvature of pipe from straight pipe to 90 degrees short elbow. Note that the corrosion rate was purely based on pipe materials and design conditions of the pipe. Fatigue stress and thermal fatigue were not included in the calculations.
The combination of mechanical erosion and chemical corrosion can lead to material degradation and may result in crack formation. To simulate the thinning of pipe thickness, we created a geometry with uneven thickness for pipe elbow with 3mm notch inside to represent the initial damage due to erosion assuming that the pipe wall coating has been compromised.
The erosion results below show a reduced thickness of 7mm at both the intrados and extrados, down from the original 10.97mm (S80). Note the location of the pit, which measures 3mm in diameter and 3mm in depth. While the pit can still withstand operating loads, other parts of the elbow, particularly the eroded sections, are experiencing stresses beyond the material's allowable limits. Additionally, when analyzing fatigue stress using the Soderberg model (selected as the most conservative approach for engineering data), we observe that stress is beginning to accumulate around the pit.
As shown below, we compared the stress intensity to determine whether the fracture will occur first at the erosion-induced pit or at the weld seam due to fatigue stress. Based on the analysis, we found that the elbow section adjacent to the weld seams is more likely to fail first, as it has a shorter lifecycle, which aligns with the equivalent alternating stress results.
As predicted, random pitting inside the pipe may go unnoticed during operation, and potential failure might not be immediately visible. The most noticeable issue during a visual inspection would likely occur at the high-stress point of the elbow, where a hairline crack could develop along or adjacent to the weld seams at the outlet of the elbow (based on the vertical displacement from the stress analysis) in a vertical piping configuration.
From the probe measurements across each layer, the likely failure location is identified at the weld seam where the pipe and elbow are connected. In this scenario, our weld material is the same throughout, excluding the necessary root pass material typically used in TIG welding for joining the elbow to the pipe. However, the fusion area and mid-section have a longer lifespan compared to other segments.
If the crack initiates in the mid-layer of the pipe material thickness, shown in green to yellow colors, it can only be detected through non-destructive testing (NDT) using ultrasonic weld inspection equipment, not through visual inspection. In the illustration below, the most probable crack location is at the outer layer, along the periphery of the weld zones, indicated by the red color.
After predicting the probable location of the failure, we can now define the crack growth. In this case, we aim to analyze the behavior of the strongest part of the metal as the crack propagates based on the stress intensification factor. From the results below, where K1 is greater than or equal to K1C, we can conclude that the crack will grow even within the metal’s strongest layer. At this stage, there is already movement of slip bands (intrusion or extrusion) along the material’s slip plane.
To achieve more accurate results, we performed 50 iterations to gather additional data on the crack propagation rate. We applied a static crack growth model driven by the stress intensity factor (SIF) using fracture toughness. According to the theory, the displacement (in mm) can be derived from the strain integrals under defined boundary conditions.
It's important to note that crack propagation can be classified as either critical (where the crack will propagate until fracture) or non-critical (where the crack initiates but does not propagate).
STRESS INTENSIFICATION FACTOR MODE 1 (Opening - characterized by a tensile stress perpendicular to the crack surface. It causes the crack to open.) SHOWING 50 ITERATIONS.
STRESS INTENSIFICATION FACTOR MODE 2 (In-plane shear This mode involves a shear stress acting parallel to the crack surface and perpendicular to the crack front. It causes the crack to slide.) SHOWING 50 ITERATIONS.
STRESS INTENSIFICATION FACTOR MODE 3 (Out-of-plane shear - This mode involves a shear stress parallel to both the crack surface and the crack front. It causes the crack to tear.) SHOWING 50 ITERATIONS.
The chart above shows the crack growth over several time cycles using SMART (Separating Morphing Adapted Remeshing Technology) crack growth analysis. Initially, the crack measures 0.4mm in length and 0.1mm in width. It remains dormant for a period before starting to grow at 800 cycles, reaching 0.003345mm with an upward trend. In this analysis, the crack extends to 0.21468mm by the final iteration. To obtain a more detailed prediction of the crack's growth, we could adjust the iteration time to around 1000 cycles or more, depending on the desired timeline, though this would require more computational resources.
Below is a real-world sample of a corroded pipe from the field, where the crack formation (fracture path) aligns with the thinnest remaining wall, consistent with our observations. Q.E.D.
