Hydrocarbon Processing - January 2021 - 71

Maintenance and Reliability
FIG. 1 shows the distribution curves of
chlorides concentrations measured in a
system during three separate time periods, using three different measurement
scenarios. In all three scenarios, caustic is
injected downstream of the desalter and
adjusted to limit the chloride content in
the overhead accumulator vessel. The median of all three data sets are essentially
identical, at 6 ppm-7 ppm, which is considered appropriate. However, the standard deviation of the data in the scenarios
is quite different, depending on the way
the data is obtained and how the caustic
injection rate is controlled.
The purple curve in FIG. 1 shows manual sampling and analysis in a refinery onsite laboratory, while the blue curve uses
an onstream analyzer. Manual adjustment
of the caustic rate was used in both scenarios. While a slight improvement to the
standard deviation is seen when sampling
and analysis moves from manual spot
testing to automatic onstream analysis,
manual adjustments to the caustic injection rate limit much of the potential
for improvement. In contrast, the green
curve uses the same onstream detector
for measurement, but now the caustic injection rate is controlled using automatic
closed-loop control, and the resulting
variation of chloride concentration is significantly smaller than the measurements
obtained using the other two methods.
The advantage of this is dramatically
lower overall salting potential, the ability
to easily change chloride setpoint on an
as-needed basis, and opening the opportunity to both expand and take advantage
of a dynamically changing, safe operating
window to allow more naphtha or kerosene/diesel to be produced as the market
and margins dictate.
In parallel, more frequent and more rigorous amine speciation analysis with rapid
generation of results brings a much larger
amount of ongoing information about
crude contaminant changes and their subsequent impacts on the corrosion behavior of the overhead system. The diversity
of organic amine compounds used for a
variety of chemical treatments and unit
operations, both throughout the refinery
and in the upstream/midstream segment,
has increased over time in attempts to
solve specific problems. As a result, the
number of amine compounds detected in
overhead systems has increased. In a modern refinery environment, it is not uncom-

mon for six or more amine compounds
to be detected in an overhead boot water
sample within a short timeframe. Not all
compounds react to form problematic
salts, but the common practices of infrequent rigorous measurement or frequent
measurement for a limited set of compounds leave significant gaps that often
lead to undetected corrosion events.
FIG. 2 evaluates salting above the dewpoint temperature across five refineries
that have experienced transient corrosion
episodes. Each dataset was taken over
a 4-mos period. Rigorous amine analysis was done 1-3 times per week in each
system. The color scale indicates the frequency for each amine chloride salt that
resulted in salting above the dewpoint
for each system. Uncolored cells indicate
that the amine was not detected in any
analysis over the evaluation period. Clear
variation is seen from system to system
in both the amines detected and the ones
that produce detrimental salts. Even when
an amine is detected across most systems,
the frequency of detrimental salt is not
the same. While not specifically shown
in FIG. 2, it is also notable that the amine
species that produce detrimental salts in
a given system often change significantly
over time. Therefore, to provide adequate
information about salting corrosion in a
system, many amines should be tested on
a frequent basis to accurately assemble
a picture of salting and how the salting
changes as time and conditions fluctuate.
Moving the rigorous analysis of multiple amines from a remote laboratory to
the refinery site enables dynamic salting
potential and safe operating window calculations in a near-continuous fashion. This
is important, as many aggressive corrosion
events are of a short-term nature and are

missed otherwise. Coupling this with improved measurement of the salting state for
the system allows more rapid identification
of out-of-control events and the ability to
make quantitative decisions for mitigation,
further accelerating the ability to detect
and respond to event-driven corrosion. Ultimately, lower corrosion, greater flexibility
and extended equipment life result.
FIG. 3 shows three separate, 60-d periods where corrosion exceeded target on
the same system. The blue region represents continuous corrosion measurement with manual control; the yellow
region shows the value of closed-loop
control of the chemical treatment and
automated alarming of corrosion rates;
and the gray region shows further improvement from dynamic system salting
calculations, using local onsite amine
analysis and subsequent automated salt

FIG. 2. Salting frequency in multiple systems.

FIG. 3. Corrosion excursion improvement.
Hydrocarbon Processing | JANUARY 2021 71

Hydrocarbon Processing - January 2021

Table of Contents for the Digital Edition of Hydrocarbon Processing - January 2021