A Review of the Critical Design Parameters for

Labyrinth Type Separation Seals on Dry-gas-seals

Dian J. Hanekom

Dian is a Rotating Equipment Engineer with 30 yearsexperience in the Petrochemical and Nuclearindustry, currently working for TASNEE Co. in SaudiArabia. His main expertise in rotating machinery isdesign, reverse engineering, vibration analysis,reliability engineering, best practices and projectspecifications. His specialties are dry-gas-seals,lube-oil-systems, centrifugal compressors andsteam turbines.

Biography: Dian J. Hanekom

An ethylene compressor experienced a dry-gas-seal(DGS) hang-up condition and a RCA concluded thata possible root- or contributing cause is lube oilpassing the labyrinth type separation seal into thesecondary DGS cavity, causing the dynamic O-ring tomalfunction. Detailed modeling and design review ofthe separation seal system enabled the investigatorsto postulate several scenarios to explain how lube oilcould migrate into the DGS. The presentation presentsa procedure how to calculate the exit velocities oflabyrinth type separation seals and the effects ofchanges in various design parameters.

Abstract

Contents Introduction

Calculation Theory

CFD Analysis

Calculation Procedure

Design Criteria

Results Outcome

Sensitivity Analysis

Conclusions

Introduction

A DGS hang-up situation developed after a boiler trip in an

Olefin’s Plant Ethylene Compressor, HP-DE side;

The seal almost returned to its normal operating position in

following-up months, but during upcoming TAM inspection,

the secondary seal was found failed;

Turn-around-maintenance inspection also revealed that oil

had entered into the secondary DGS-cavity;

A detailed investigation was launched into the root cause of

oil passing and seal hang-up;

Introduction (Cont.)

A – Filtered seal gas supply to primary

B – Primary seal vent to flare header

C – Secondary seal gas supply

D – Secondary seal vent to atmosphere

E – Seal gas supply to separation seal

Bearing

Vent

Process

Gas

Side Secondary

Seal

Primary

Seal

Separation

Seal

Introduction (Cont.)

V = 15 m/s

Requirement: –

Design velocity

5 m/s at double

max. clearance

and 15 m/s at

max. clearance

Introduction (Cont.)

Calculations Theory

The calculation of exit velocities were based on the theory

of gas flow in straight-through labyrinth seals;

The flowrates and pressure distributions were calculated by

using the Neumann Modified Method (“Moody’s Friction-

Factor Model”),Y. Dereli & D. Eser, ;

The model incorporate the orifice plate and the labyrinth

together with a supply pressure of 3.6 kg/cm2;

Flow through the orifice plate was calculated using the

computational procedure detailed in ISO 5167-2.

CFD analysis

ORIFICE PIPE WALL

P = 3.6 kg/cm2

P upstream is known

P downstream is unknownM mass flow unknown

Labyrinth teeth

V, P = atmV, P = atm

Bearing2nd DGS

N2

Calculations Procedure

Starting with a known N2-supply pressure (3.6 kg/cm2) and an

assumed mass flow rate, pressure drop through the orifice plate

and labyrinth seal is calculated.

The mass flow rate is then varied in an iterative process until the

discharge pressure equals atmospheric pressure.

For each iteration, the gas properties are re-calculated as the pressure varies.

In this manner the correct mass flow rate is established and the

flow velocities in the different stages of the labyrinth can be calculated.

Design criteria

5 m/s at double the maximum design clearance.

15 m/s at maximum design clearance.

Best practice design values varied between 12 – 18

m/s, based on two compressor manufacturers, one

seal vendor and an international consultant.

Calculations and Analysis ResultsDesign velocities ≥ 15 m/s & ≥ 5 m/s at 2x max Cl)

Dereli/Eser

Model

CFD

Model

3rd Party OEM P&ID

Volume

flowrateNm3/h 24.82 24.85 26.1 26.1

Orifice

massflowkg/s 0.0086 0.0086 - -

Laby set 1 mass

flowkg/s 0.0039 0.0039 - -

Laby set 2 mass

flowkg/s 0.0048 0.0047 - -

Laby up-stream

PressureBar(g) 0.0012 0.0018 0.0011 0.0039

Calculations and Analysis ResultsLabyrinth tooth velocities

Dereli/Eser

Model

CFD

Model

3Rd Party OEM P&ID

Set 1, Tooth 1 m/s 8.77 8.55

9.08

15.1

Set 1, Tooth 2 m/s 8.77 8.55

Set 1, Tooth 3

(bearing side)m/s 8.77

(-42 %)

8.55

Set 2, Tooth 1 m/s 10.74 10.21

11.12Set 2, Tooth 2

(DGS side)m/s 10.74 10.21

Seal supply pressure sensitivity

3.9

5.7

7.6

8.89.5

6.6

9.9

13.1

15.0

16.3

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

1.0 1.5 2.0 2.5 3.0 3.5 4.0

Ve

loc

ity [

m/s

]

Supply Pressure [kg/cm2]

Variable Inlet Pressure vs Seal Exit Velocities - 260 mm Ø seals

(OEM Design for 15 m/s @ max seal Cl)

Set 1 - Vel @ Max Cl Set 2 - Vel @ Max Cl Set 1 - Vel @ 15 m/s design

Original

Design

Proposed Design Point

15 m/s @ 3.6 kg/cm2 Ø4.7

mm orifice

Seal back pressure sensitivity

8.78

6.79

4.72

2.53

10.75

12.73

14.80

16.98

0.00125

0.00175

0.00236

0.00310

0.0000 0.0005 0.0010 0.0015 0.0020 0.0025 0.0030 0.0035

0.00000

0.00050

0.00100

0.00150

0.00200

0.00250

0.00300

0.00350

0.00

2.00

4.00

6.00

8.00

10.00

12.00

14.00

16.00

18.00

0.0000 0.0005 0.0010 0.0015 0.0020 0.0025 0.0030 0.0035

Lab

y U

pst

rea

m P

ress

ure

[kg

/cm

2]

Se

pa

ratio

n S

ea

l Exit V

elo

citie

s [m

/s]

Back Pressure [kg/cm2]

Effect of Back Pressure on Seperation Seal

[Orifice size 3.6 mm]

Set 1 - Bearing Side Set 2 - DGS side Upstream pressure

Design Point

BP=200 Pa

15 m/s

Separation seal clearance sensitivity

11.674

8.782

6.881

5.5464.571

3.8303.2493.123 2.784 2.405

0.000

2.000

4.000

6.000

8.000

10.000

12.000

14.000

16.000

18.000

20.000

0.400 0.500 0.600 0.700 0.800 0.900 1.000 1.100 1.200 1.300

Se

al E

xit

Ve

loc

ity [

mm

/s]

Seal Clearence [mm]

Effect of Seal Clearence Enlargement on Seal Exit Velocities

Laby Set 1 - Bearing side Laby Set 2 - 2nd DGS side Laby Set 1 - Meet API design

Design @ max Cl = 0.525 mm

2 x max Cl = 1.05 mm,

API specify 5 m/s, V achieved = 3.1 m/s

API @ 0.525 Cl = 14.1 m/s

5 m/s line

Design Deficiencies

O-ring missing on the

inside of hold down bolts.

Oil drain in labyrinth

towards bearing side

installed upside down.

Drain should be at bottom of seal with alignment pin for correct

orientation

No O-ring installed. Oil

leak path

N2

Conclusions

The OEM design velocities are 42% (8.8 m/s) lower than

what was specified by the OEM (15 m/s);

The design velocity specification of 15 m/s at max

clearance is on par with other design practices;

Back pressure on the bearing housing was a likely

contributing cause of low seal gas exit velocities;

Oil leaked through the hold-down-bolts of the seal with

no O-ring installed on the inside of the bolts.

Conclusions (Cont.)

The labyrinth seal oil drain was installed upside down

due to no alignment pin.

Other important factors like a) Nitrogen “on” before

lube-oil start-up, b) cleanliness of supply orifices, c)

orifices installed sizes, were all correct.

Suggested improvements: -

Have a shutdown interlock and timer between LO- and seal-

gas shutdown. Seal gas “ON” for 3 hours after LO-stopped;

Have a back pressure measurement;

Have an online DGS monitoring system.

Thank You

References: - “Flow Calculations in Straight-Through

Labyrinth Seals by using Moody’s Friction-Factor Model”,

Yilmaz Dereli and Dursun Eser.

Recognitions: - My thanks to Mr. Gysbert van Zyl for his

support for software debugging, calculation review and for

reviewing this presentation.