Amine Solvent Change and Sample Report Interpretation
An amine solvent change, nine years of lean solution samples, and an inspection decision made on the plant’s own data.
The Problem
A US onshore gas plant treating high CO2 and H2S gas had replaced a major amine vessel after roughly five years of service. Around the same time the plant changed solvents, moving from DGA to a formulated MDEA blend. Four years later, an advanced NDE scope on several insulated vessels came due, with insulation removal as the largest cost driver, and the owner asked a fair question: the chemistry changed, so does this scope still have to happen in the current budget year?
The data that could answer it already existed. Routine amine sample reports were going to operations and the chemical supplier, not to the integrity engineer, and the lab flags on those reports grade against solvent performance targets, not against corrosion.
Approach & Solution
Integrity EPC read nine years of routine lean solution samples, 35 in total, as integrity data anchored to API RP 571 Section 3.2, Amine Corrosion. Iron and manganese are corrosion tracers: both are constituents of carbon steel and dissolve into the solution during a corrosion reaction. API 571 makes the connection directly. Rising iron coincides with rising amine degradation products (3.2.7), and the alkanolamines rank from most to least aggressive with DGA second from the top and MDEA at the bottom (3.2.3). Degradation markers, CO2 loading, and sample appearance carried the rest of the read.
Execution
Both solvent eras were trended, 23 DGA samples against 12 MDEA blend samples, and the final DGA sample was compared to the latest MDEA sample flag by flag.
Heat stable salts moved the other way, rising from 0.03 to 0.17 percent of amine while every corrosion indicator improved. Both values sit far below the roughly 2 percent level where API 571 expects a significant rate increase in hot lean amine. An integrity review that read HSAS alone would have reached the wrong answer at this unit. The sample trends were then set against advanced NDE results already in hand, which showed the system in good condition outside of localized low areas at the amine filters.
Results
The data supported a change in timing, not in scope. The insulated vessel advanced NDE moved one budget year. The reboiler, the still, and the hot lean and hot rich circuits stayed at the top of the plan, where API 571 puts them regardless of solvent. No cracking mechanism was retired: amine stress corrosion cracking applies to MDEA service too, weld inspection coverage is set by the service and by PWHT status per API RP 945, and an H2S non-detect on a lean sample is not a clearance for wet H2S damage.
Six years after the solvent change, an internal inspection of the unit’s reboiler found adherent scale and localized attack that could be measured and trended, on equipment that has been in the same hot amine duty the whole time. The same read has since been applied at two more gas plants run by another operator. There, the plant with the rising iron trend was also the plant with a corrosion related valve failure on record. The sample data and the failure history agreed.
Takeaways
1. Amine sample reports are integrity data. Iron, degradation products, and appearance belong on the integrity engineer’s distribution.
2. Read the lab flags, then read past them. A value flagged Low for treating performance can be favorable for corrosion.
3. Solvent type sets the expectation. Operation sets the result.
4. A solvent change is a management of change event for the integrity program and should trigger a damage mechanism review.
5. Correlate sample trends with NDE results before crediting either one.
The full technical write-up, including the complete API 571 clause alignment and the limitations of the comparison, is available on request.