|
Oil Analysis Without Particle Counting?
Mark Smith, Technical Administrator, Analysts,
Inc.
First, I would make sure the lab manager was aware of the foundation
for why the particle counts were essential:
(a) System engineering: Turbines, hydraulics and compressors
have precise tolerance clearances approaching 5 microns. This
close dimensional tolerance and the long service interval of these
systems make them susceptible to both rapid and progressive damage.
From a reliability standpoint it is essential to monitor conditions
contributing to both.
(b) Testing limitations: Internally generated particles from
degrading non-metallic parts cannot be monitored through spectrographic
metals analysis. Even when the particles are metallic wear metals,
the spectrographic analysis is typically limited to particles
in the 3 to 10 micron size range, and particles outside this range
may only be partially quantified. Particle counting is the only
reliable means of monitoring and trending the presence of non-metallic
particulates--while at the same time, also offering an indicator
of larger metallic particles being present. Ferrographic trending
could be performed, but there are limitations regarding non-ferrous
materials, and ferrography is more costly than particle counting.
The lab manger is only seeking to drop a test, not offer an effective
alternative.
(c) Uncontrollable factors: The system performance, servicing
fluid quality and good maintenance practices which led to the
1 percent test exception percentage being used as evidence by
the laboratory manager is a bonus, but it cannot be assumed that
this favorable combination will continue indefinitely. New oils
are notorious for randomly high particulate levels--not necessarily
due to negligence by the blender or vendor--but simply as a consequence
of the manufacturing process. Major oil additions often exceed
the system filters ability to compensate for this condition. Breather
caps are left off, filters burst in service, bypass indicators
fail, and life rapidly becomes difficult without the proactive
means to detect these conditions.
Secondly, after the proper research I would present examples
of the significant cost savings achieved as a direct result of
the particle counting, and secondary examples of the cost of predictive
analysis failure. For example, "On our hydraulic systems alone,
we have changed our choice of elements and extended the filter
service intervals by 50% which has resulted in an annual savings
of $---- which more than justifies the time and cost of the analysis."
Or if he remains stubborn, "If we have a major unit problem due
to failure of the main feedwater pumps to trip in sequence with
the main turbine because of suspended particles in the main turbine
oil system and subsequent flow blockages in the control oil system
dump valves, our site will lose your entire lab's operating budget's
worth of revenue before you can finish your third cup of coffee."
(A failure of this type being a matter of NRC record, in case
he checks).
Third, I would ask about any outside pressures on the lab manager.
Is he under pressure to cut internal staff or expenditures? If
so, are these decisions being made with an overall look at system
reliability and safety or just the bottom line in one area of
company operations?
Finally, I'd make it personal: "Can you guarantee that the remaining
test results can determine that the oil is clean? Are you willing
to take responsibility for failures in these systems whose root
cause is determined to be undetected particulate contamination?"
The lab manager may be one step removed from system reliability
issues, but you aren't. Fight to keep a full toolkit...jour job
may be at stake.
See other responses
to this Readers Challenge.
|
