(47aa) Identification of Fatigue Risk in Mixing Tees Using an Operating History Based Analysis Method

Identification of Fatigue
Risk in Mixing Tees Using an Operating History Based Analysis Method

Michael
F.P. Bifano, Ph.D.

Seetha Ramudu Kummari, Ph.D,

Robert
G. Brown, P.E.

The Equity
Engineering Group, Inc.

Shaker Heights, Ohio
USA

ABSTRACT

The mixing of hot and cold fluids at industrial
pipe tee junctions is highly turbulent and can result in thermal fatigue
cracking even when the incoming stream temperatures and velocities remain
constant.  Thermal fatigue cracks occur when
the mixed fluid streams create hot and cold thermal shocks on the unprotected
pipe walls.  Global unsteady changes in
flow rate, temperature, and pressure are an additional fatigue concern.  The combination of both mechanisms may result
in through-wall fatigue cracks and thus leakage, causing a structural integrity
and process safety concern.  Standard transient
Computational Fluid Dynamics (CFD) models average-out local turbulence patterns
(Reynolds Averaged Navier Stokes Models) and
therefore cannot model or predict fatigue damage due to turbulent thermal
mixing.  Instead, computationally expensive
Large Eddy Simulations (LES) are required to model the turbulence and resulting
thermal transients at the pipe walls.  Unfortunately,
LES simulations can be impractical for industrial applications since they are
time intensive and require significant computing resources.   In lieu of computational modeling, some
industry practices mitigate thermal fatigue damage at mix points with the
instillation of mixing quills or thermal sleeves when stream temperature
differences exceed 300°F.  This
temperature limit has not been validated, and may not be a conservative
practice.  Fatigue damage at mix points have
been found for stream temperature differences of approximately 300°F; however,
global fluctuations in temperature, stream velocities, and pressure may be a significant
contributor.  The 300°F limit is not
regularly implemented in the petrochemical and refining industries.  To date, without costly CFD modeling, a standard
process hazard analysis or practice does not exist to determine which mix
points are at-risk for fatigue damage.   Instead, in many cases, costly mixing quills
are conservatively designed and installed.  The standardized process hazard analysis presented
can also be used to prioritize inspection by comparing multiple mix points to determine
those most susceptible to cracking.  

The presented methodology offers a novel
approach that determines the severity of mixing by assessing past operating
data such as process stream temperatures, velocities, and fluid properties
which are readily available or can be derived from known information.  The severity of mixing is determined from
previous LES CFD modeling and supplemented with industry data.  The LES CFD modeling has been previously
benchmarked against observed in-service failure data. With the addition of further
industry failure data, the presented model is capable of continuous improvement,
and eventual industry wide standardization.  Currently, the presented model can be
effectively used on a plant wide level to compare multiple mix points to
determine those most susceptible to fatigue damage.  Global cycle counting of past operating
history is also performed using well known rain-flow cycle counting algorithms,
since the total fatigue damage is a combination of both turbulent thermal
mixing and global fluctuations in process variables.

Lastly, a case study is presented for two
mixing tees that operate below the conventional 300° F temperature limit. Through-wall
fatigue cracks were found in one of the two mix points in the case study.  The applied method identifies these mix points
as high risk and good candidates for mixing quills.  

Keywords:  Turbulent thermal mixing, Thermal Striping, Fatigue
Crack Growth, Computational Fluid Dynamics (CFD), Process Hazard Analysis