Vibration Testing and Troubleshooting of Coke DrumsTroubleshooting of Coke Drums
Mahmod Samman, Ph.D., P.E.Senior Associate
Stress Engineering Services, [email protected]
Coking.com Coke Drum WorkshopRio de Janeiro, Brazil
August 2009
1
OVERVIEWOVERVIEW
• Vibrations BasicsVibrations Basics• Drum Dynamics
Wh t i Aff t d• What is Affected• Potential Consequences• Case study
2
VIBRATIONS BASICSVIBRATIONS BASICS Response of a flexible structure to excitation
(fluid flow pressure pulsations sloshing etc )(fluid flow, pressure pulsations, sloshing, etc..) Components
• Mass (inertia)• Spring stiffness (Force–Displacement)• Dashpot viscous damping (Force–Velocity)
Basic characteristics: Basic characteristics:• Natural frequency (or frequencies) • Mode shapes
Input – output• Force• Displacement
3
• Displacement
DRUM DYNAMICSDRUM DYNAMICS
4
WHAT IS AFFECTEDWHAT IS AFFECTED
Pipes and pipe supports Base-plate bolts and grout Non-structural (stairs, lights, guardrails,..)( , g , g , ) Machinery (elevator, pumps, ..) Superstructure (concrete and steel)Superstructure (concrete and steel) Foundation system (substructure and soil)
Humans!
5
POTENTIAL CONSEQUENCESPOTENTIAL CONSEQUENCES
• Operator discomfort or fatigue• Operator discomfort or fatigue • Poor performance• A l ti f i d• Acceleration of corrosion damage• Interruption of operations during repairs
( / f )• Fatigue cracks (leaks/ fires)• Bodily injuries
6
WHAT TO DO?WHAT TO DO?
• Measure• Vibrations• Strain• Process variables• Process variables
• Analyze• Stress / fatigueg• Human tolerance
• Mitigate• Process• Structure
7
CASE STUDY - 1T d it• Two-drum unit
• “Significant” vibrations in structure• Conflicting views on When they vibrate the most Which one vibrates more Which one vibrates more
• Failures Piping supportsp g pp Anchor bolts Base plate grout
• I th it f ?• Is the unit safe?• Can the process be optimized to
minimize vibrations?8
minimize vibrations?
DRUM DESIGN AND OPERATION1996 API Coke Drum Survey, 2003
16 f i di l i l ll f h id• 16 feet in diameter - relatively small for the mid 80’s
• 76 f t i h i ht t t ll• 76 feet in height - average to tall• Slender drums
V 1 ¼ C 1/2M i l• Very common 1 ¼ Cr -1/2Mo material• 17 hour fill cycle - average to relatively slow
tioperation
9
OBJECTIVES
• Monitor vibrations in the drums, piping, and structure.
• Obtain synchronized temperature and strain measurements.
• Determine the timing and characteristics of maximum vibrations.
• Determine severity of dynamic stresses and potential for fatigue damage in the structure.
• Conduct sensitivity analysis of vibrations versus process variables.
10
PROCEDURE• Installed 33 sensors and two data acquisition
systems on the drums piping and structuresystems on the drums, piping, and structure.• Monitored the unit for a period of 20 days (14
cycles)cycles) • Processed and analyzed the data in time and
frequency domain.eque cy do a• Analyzed the correlation between key process
variables with vibration, strain, and temperature , , pmeasurements.
11
INSTRUMENTATIONVibration data acquisition system
16 channel unit16 channel unit16 seismic accelerometers @ 102 samples per second2 strain gages (low-speed DC channels)2 strain gages (low speed DC channels)
Temperature data acquisition system StrainDAQ unit16 thermocouples 1 sample per two seconds
Data collection was continuous without interruption during the entire monitoring period.
12
Sensor Layout
13
Example Quench TransientTEMPERATURE
Quench
TIMETIME
ACCELERATION
14
Integrated Displacement Data
0.4
0.5Drum B - West Dir. (in)Drum B - North Dir. (in)Drum B - Down Dir. (in)
0 1
0.2
0.3
NT
(IN)
Drum B Down Dir. (in)
-0.1
0
0.1
PLA
CEM
EN
-0.3
-0.2DIS
P
Large Transient Displacement
-0.5
-0.4
11035 11037 11039 11041 11043 11045 11047 11049 11051 11053 11055
g p
15
TIME FROM START OF CYCLE 9 (SEC)
Vibration Versus Flow Rate0.12 250
DRUM B PIPEDRUM A PIPEDRUM A
0.08
0.1
es R
MS 200
(GPM
)
DRUM ADRUM BSTRUCTUREDRUM B ONLINE
0.06
on in
Inch
e
100
150
um V
alue
s
0 02
0.04
Vibr
atio
50
100
Max
imu
0
0.02
1 3 5 7 9 11 130
16
1 3 5 7 9 11 13Cycle
Strain Gauge DataStrain Gauge Data
Low dynamic strains
17
CASE-1 SUMMARY• Drum vibration magnitude was maximum during the
quench part of the cycle in the East-West direction.• The two drums vibrated in a comparable manner both
from magnitude and frequency standpoints. • The maximum recorded peak displacements were 0 58The maximum recorded peak displacements were 0.58,
0.35, and 0.23 inches for the drums, the piping, and the structure, respectively.
• Measured dynamic strains in the structure were below• Measured dynamic strains in the structure were below the fatigue-inducing levels
• The correlation between vibration levels and recorded bli h dprocess parameters was established
18
CASE STUDY -2
• Four-drum unit.• Blow-down line.• Cracks and leaks.• Angled-tee joint.Angled tee joint.• Thermal cycles.• 3D loads3D loads.• Vibrations.• Why?• Why?• How to fix it?
19
ACTION PLAN1. Instrumentation
ACTION PLAN
Strain gages Thermocouples
2 E t ti f l di diti2. Extraction of loading conditions3. Finite element analysis4 F ti A t4. Fatigue Assessment
20
INSTRUMENTATION
Intrinsically Safe Instrumentation System
21
FIELD MONITORING
22
TEMPERATURE MEASUREMENTS
23
THERMAL GRADIENTSEvery forth cycle
24
CLOSEUP
25
PAD THERMAL GRADIENTS
26
Extracted load casesFive thermal profiles during the coking cycle
27
Thermal Load CaseThermal Load Case
28
Von Mises Stress in the pipep p
outside inside
29
Von Mises Stress inside the padVon Mises Stress inside the pad
outside inside
30
Fatigue AssessmentFatigue Assessment
• Maximum thermal stress range is 92 ksiMaximum thermal stress range is 92 ksi.
• Alternate stress is 49.8 ksi.
• Correction for E at 4000F
• Fatigue life is 4,441 cycles.
31
CASE-2 SUMMARYCASE-2 SUMMARY• Vibrations and pressure are not the main problem.
• The failure is caused by severe thermal transients that generate 92 ksi stress range in the pipe at the locationgenerate 92 ksi stress range in the pipe at the location of cracks.
R d i i i i d i• Recommendations to minimize stresses and increase fatigue life:
• Redesigned integral fitting.
• Fatigue-resistant welds.
32
g