GPF PROJECT:

AIR COMPRESOR PACKAGE VIBRATION CACULATION NOTE

VIBRATEC report ref. 001.026.RA.01.A for TECHNIP ABU DHABI

______________________________________________________________________________

1. INTRODUCTION

The aim of this study is to validate the mechanical behaviour of the air compressor
package (TAG, PU-2094 A/B of the TECHNIP ABU DHABI GPF PROJECT) from a dynamic
point of view, by avoiding coincidences between global skid resonances and rotating unit
excitation frequencies, while by respecting the vibration acceptance levels, as defined in
supplier specifications. The air compressor package was built by FLUIDES SERVICE.

For the dynamic behaviour, the modal basis was computed and analysed as follows:

The skid is considered safe from a dynamic point of view if no mode is computed around
the motor’s main excitation frequency [25 Hz]. Notice that the pump’s main excitation
frequencies are above 130 Hz, out of the critical frequency range.

In case of critical frequency close to [25Hz±30%] frequency range, structural modifications

should be proposed to shift the modes out of the frequency range. If it is not possible to
shift all modes, a response computation would be performed. Most worst-case rotating
unbalance scenarios were used.

For mechanical criteria, the computed velocity levels had to satisfy ISO 10816 criteria
(motor and compressor).
For human vibration exposure, the GS – SAF221 was also used for information, but this
specification is not contractual.

A conclusion summarizes the computed results.

______________________________________________________________________________
VIBRATEC report ref. 001.026.RA.01.B for TECHNIP ABU DHABI
5
2. MODEL DESCRIPTION

The air compressor package modelling and computation work was performed using
IDEAS V11 finite element model software.

The geometry was established using technical data (drawings) and the Vibration Data and
Studies furnished by FLUIDES SERVICE.

Drawing references: 5055789-P01: Detailed drawings of air and dryer frames

5055789-P03: Detailed drawings of the motor support
BRB220: Technical specifications for BRB 220 Anti Vibrating
Mountings (A.V.M)
Extract of Part List Atlas (ATLAS COPCO manual) :
Drawings of ATLAS COPCO frame ZT160
1.3 MLD35-0122: LOHER Motor dimension drawing

Figure 1 presents a global drawing of the air compressor package :

Figure 1 – Air Compressor Package upper view

The unit is composed of:

- a motor (red),
- 2 compressors with a gearbox (blue)
- a main skid (UPN beam) (provided by FLUIDES SERVICE, black)
- a secondary skid (thin shell) bolted to the main skid (yellow)
- 2 cooling units (green).

The global coordinate system is:

______________________________________________________________________________
VIBRATEC report ref. 001.026.RA.01.B for TECHNIP ABU DHABI
6
 - X: Longitudinal axis (parallel to the motor rotation axis, positive from motor to
pump),
- Y: Transversal axis (in the skid plane),
- Z: Vertical axis (perpendicular to the skid plane).

The air compressor package’s Finite Element Model (FEM) was made using thin shell
elements (primary and secondary skids) and lumped mass and inertia (see Figure 2 and
Figure 3). The enclosure and the 2 cooling units were taken into account as non-modelled
mass.

The units (motor, gearbox and pumps) were modelled as lumped masses and inertia. Their
operating characteristics are given in Table 1.

Mass Inertia (kg.m²)

Label
(Kg)
Ixx Iyy Izz
Motors 2350 149 718 718
Gearbox, Pumps
700 168 157 157
and auxiliary
Table 1: Operating mass and inertia characteristics

N.B.:

Components’ inertia were estimated by considering the motor as a cylinder and the whole
part of pumps, gearbox and auxiliary components as a cube. The liaison between electric
motor and pump was considered to be rigid as are the 2 main components. The working
hypothesis was that the components (motor, gearbox and pumps) are non-shrinking in the
0 – 35 Hz frequency range.
The liaison between the components (modelled as lumped mass) is modelled with rigid
elements; rotations remain free and the translations are considered to be very stiff.
Skid extensions were attached to the skid with a constraint element (mass repartition
without adding local stiffness).

______________________________________________________________________________
VIBRATEC report ref. 001.026.RA.01.B for TECHNIP ABU DHABI
7
The skid frame material is an isotropic steel, whose characteristics are:
- Mass density : 7,850 kg/m3,
- Modulus of elasticity : 2.11011 N/m²,
- Poisson ratio : 0.27,
- Elastic limit : 235 MPa.

N.B.: To take account of non-modelled masses, the skid beams’ mass density was increased.

FEM characteristics:

 Number of nodes: 3 803,

 Number of elements: 3 798,
 Total model mass: 10 500 kg.

The skid was welded on the GPF platform supporting deck (cellar deck) on a specific IPE
220 beam frame (see Figure 4).

Skid extension, including

cooler and enclosure

Gearbox & compressors

Electric Motor

Y X

Figure 2: Air Compressor Package Model - Top view

______________________________________________________________________________
VIBRATEC report ref. 001.026.RA.01.B for TECHNIP ABU DHABI
8
 Figure 3 - Air Compressor Package Model - Bottom view

109

109

Figure 4: Air Compressor package model – Top view with cellar deck frame

______________________________________________________________________________
VIBRATEC report ref. 001.026.RA.01.B for TECHNIP ABU DHABI
9
For dynamic computations, the supporting deck rigidity was modelled with part of the
cellar deck (beams) below the main skid (see Figure 4). These main beams are attached
with dynamic stiffnesses presented in Table 2.

Direction Boundary conditions

X 1.109 N/m
Y 1.109 N/m
Z 1.109 N/m
x Free
y Free
z Free
Table 2: Air compressor package –
Boundary conditions (dynamic stiffness) at fixation points

N.B.: 109 N/m is the estimated dynamic stiffness of the main piles under the cellar deck.

______________________________________________________________________________
VIBRATEC report ref. 001.026.RA.01.B for TECHNIP ABU DHABI
10
3. DYNAMIC ANALYSIS

3.1. MODAL BASIS

The modal basis was computed between 0 and 50 Hz.

The acceptance criterion was no mode around the motor and pump main excitation
frequencies (rotating frequency H1): 25 Hz  30 % [17 Hz – 33 Hz].

The main resonance frequencies are listed Table 3 and corresponding mode shapes are
presented in Appendix 1.

Comments:

Four resonances were computed inside the critical frequency range: 17.2, 20.1 and 21.4 Hz
involved the behaviour of motor/compressor component (see example Figure 5). The mode
at 30.5 Hz involved the deck’s behaviour. Regarding the modifications, these modes’
shapes must be shifted out of the frequency range [17 - 32 Hz], but as the component is
already packaged, it was decided to compute a dynamic response in order to compare
acceleration levels to acceptance criteria. A response calculation was performed around 25
Hz in order to validate the skid structure from a dynamic point of view.

Figure 5: Mode shape at 17.2 Hz

(In black: undeformed shape)

______________________________________________________________________________
VIBRATEC report ref. 001.026.RA.01.B for TECHNIP ABU DHABI
11
 Mode Frequency Deformed shape Figure
(Hz)

1 4.6 Translation mode (along Y axis) of the motor & compressors A1

2 6.1 Translation mode (along X axis) of the motor & compressors A2

3 8.6 Translation mode (along Y axis) of the compressors A3

4 10 Translation mode (along Z axis) of the compressors A4

Vertical motor translation in phase with a global flexion of

5 13.8 A5
the cellar deck frame.

Vertical motor translation in oppositional phase with a global

6 16.4 A6
flexion of the cellar deck frame.

Lateral motor translation in oppositional phase with a global

7 17.2 A7
flexion of the cellar deck frame.

Rocking of the motor around X and translation of the ATLAS

8 20.1 A8
skid (in phase)

Rocking of the motor around X and translation of the ATLAS

9 21.4 A9
skid (in oppositional phase)

11 30.5 2nd flexion of the cellar deck frame, including skid A10

35.8 and
12 Flexion of the cellar deck frame
higher

Table 3: Modal basis

Nota: X  longitudinal axis

Y  transversal axis
Z  vertical axis

______________________________________________________________________________
VIBRATEC report ref. 001.026.RA.01.B for TECHNIP ABU DHABI
12
 3.2. DYNAMIC COMPUTATION

For the dynamic response computations, a modal damping  = 4 % ( = 2C/Cc) according to

VIBRATEC’s database (measured modal damping on offshore/onshore structures for
structural modes) and a damping of  = 10 % for modes including the AVM were used.

The excitation is defined as rotating unbalance at 25 Hz (H1) on the motor. The excitation
of the both compressors was not taken into account (high frequency rotation speed). The
defined unbalance load case is the unbalance in normal operating conditions.

Estimated rotating unbalance (dynamic force) is summarized in Table 4. Dynamic forces

were computed using a balance quality factor of G = 2.5 mm/s for the electric motor, and
using a rotor mass of 600 kg (values issued by machine supplier).

Normal
rotating unbalance
(25 Hz)
Motor 236 N

Table 4: Rotating unbalance forces (plane YOZ)

3.2.1. Acceptance criteria

3.2.1.1. Component acceptance criteria

The acceptance criteria presented in Table 5 (according to ISO 10816-3) is applicable to

the motor. The ISO 10816-3 criteria places the velocity levels in the [10 Hz – 1,000 Hz]
frequency range. The proposed criteria for the computation was 80% of this level at the
main excitation frequency, as most of the excitation comes from the first harmonic
excitation frequency for this type of equipment.

VIBRATEC recommended limit Vibration acceptance criteria

at the main excitation frequency ISO 10816-3
H1 [Overall 10 – 1,000 Hz]

(mm/s RMS) (mm/s peak) (mm/s RMS)

A/B limt
Normal 3.6
5.1 4.5 (acceptable for
operation (80 % of 4.5)
long time running)

Table 5: Acceptance criterion for the motor

______________________________________________________________________________
VIBRATEC report ref. 001.026.RA.01.B for TECHNIP ABU DHABI
13
 3.2.1.2. Acceptance criterion for human exposure to vibrations:

According to SAF 221 (specification used for other projects), the horizontal and vertical
acceptance criteria for human exposure to vibrations, acceptable locally to equipment, are
indicated in Table 6 and Table 7. For GPF projects, there is no criteria regarding human
exposure to vibrations (nobody lives on the platform), therefore, these results are given for
information.

Limit L1 : restricted area,

Limit L2 : only acceptable locally to equipment,
Limit L3: upper limit for general work areas,
Limit L4: upper limit for offices/control rooms,
Limit L5: upper limit for living accommodations.

For the computation, the area was considered as an L3 area and this criteria is a single
frequency criteria.

Limit L1 Limit L2 Limit L3 Limit L4 Limit L5

Velocity RMS (mm/s) < 159.1 <11.94 <3.98 <1.59 <1
Velocity peak (mm/s) <225 <16.88 <5.63 <2.25 <1.41
Table 6: Velocity limits in mm/s – Horizontal axes (X,Y)

Limit L1 Limit L2 Limit L3 Limit L4 Limit L5

Velocity RMS (mm/s) < 55.7 <4.46 <1.47 <0.6 <0.36
Velocity peak (mm/s) <78.8 <6.3 <2.08 <0.84 <0.51
Table 7: Velocity limits in mm/s – Vertical axis (Z)

______________________________________________________________________________
VIBRATEC report ref. 001.026.RA.01.B for TECHNIP ABU DHABI
14
 3.2.2. Response computation results:

3.2.2.1. On components:

Response curves, resulting from compressor and motor dynamic rotating force actions were
computed at the vibrating components’ centres of gravity (compressor, motor). Response
curves are presented in Figure 6.

25 Hz ±30%

Limit 5.1 mm/s peak (motor)

MOTOR COMPRESSOR

Figure 6: Velocity response spectra – Magnitude: x ²  y ²  z ² - m/s peak

at the vibrating components’ centres of gravity – H1 frequency

The motor and compressor’s vibrating velocity levels respect the acceptance criteria. The
highest computed velocity levels are given Table 8.

Motor Compressor
Velocity response
(mm/s peak) 1.2 < 5.1 1.7 < 5.1

Table 8: Highest computed velocity responses (resonance frequency 18 Hz).

______________________________________________________________________________
VIBRATEC report ref. 001.026.RA.01.B for TECHNIP ABU DHABI
15
 3.2.2.2. Skid and cellar deck (human exposure):

The following response computations were performed on the skid structure. The highest
velocity level was computed on the ATLAS skid, vertical axis (Z): 1 mm/s (peak) at 17.5 Hz.

Figure 7 : Velocity response spectra – Magnitude: x ²  y ²  z ² - m/s peak

at skid locations,
2 points on the ATLAS skid (blue, red), 2 points on the main skid (orange, green)

The associated cartographies are presented in Appendix 2.

The vibrating velocity levels on the skid are far below the acceptance criterion. The
highest velocity levels computed are listed in Table 9.

Velocity response (mm/s

peak)
Level along vertical
axis (Z axis) 0.8 < 2.08

Level along horizontal

axis (Y axis) 0.7 < 5.63

Level along horizontal

axis (X axis) 1 < 5.63

Table 9: Highest velocity responses at 17.5 Hz on the skid

______________________________________________________________________________
VIBRATEC report ref. 001.026.RA.01.B for TECHNIP ABU DHABI
16
 3.2.3. Comments:

These levels are due to the modal appropriation of the 17.2 Hz mode by the H1 rotation
frequency. The main repartition and the position of the centre of gravity can significantly
modify this frequency. The masses used in the model were estimated, and from
VIBRATEC’s point of view, probably underestimated. In reality, the suspension frequency
will be lower and will probably be shifted out of the frequency range.

______________________________________________________________________________
VIBRATEC report ref. 001.026.RA.01.B for TECHNIP ABU DHABI
17
4. 4. CONCLUSION

This study’s aim was to validate the dynamic behaviour of a skid integrating a motor and
compressors.

The modal basis was computed taking into account the cellar deck’s rigidity.

Four resonances were identified in the [17 Hz ; 33 Hz] frequency range (rotating frequency
of the equipment ±30%):

 17.2, 20.1, 21.4 Hz motor and compressor equipment suspension modes (directly
impacted by the stiffness of the anti vibrate mounting),
 30.1 Hz global flexion of the cellar deck (This mode is due to the supporting floor
and not to the equipment).

As the modes involved in the frequency range were suspension ones and/or deck flexion,
responses due to rotating unbalance were computed according to an electric motor
unfavourable rotating unbalance case. The computations show that the vibration levels
computed on the components’ centres of gravity and on the skids are lower than
acceptance criteria in the critical frequency range.

The air compressor package is safe from a dynamic point of view.

______________________________________________________________________________
VIBRATEC report ref. 001.026.RA.01.B for TECHNIP ABU DHABI
18
 APPENDICES

______________________________________________________________________________
VIBRATEC report ref. 001.026.RA.01.B for TECHNIP ABU DHABI
Appendices - 1
 APPENDIX 1

Deformed modal shapes

______________________________________________________________________________
VIBRATEC report ref. 001.026.RA.01.B for TECHNIP ABU DHABI
Appendices - 2
(In black: undeformed shape)

Figure A1: Mode 1 – 4.6 Hz Figure A2: Mode 2 – 6.2 Hz

Figure A3: Mode 3 –8.6 Hz Figure A4: Mode 4 – 10.6 Hz

Figure A5: Mode 5 – 13.8 Hz Figure A6: Mode 6 – 16.4 Hz

______________________________________________________________________________
VIBRATEC report ref. 001.026.RA.01.B for TECHNIP ABU DHABI
Appendices - 3
 Figure A7: Mode 7 – 17.2 Hz Figure A8: Mode 8 – 20.1 Hz

Figure A9: Mode 9 – 21.4 Hz Figure A10: Mode 10 – 30.5 Hz

______________________________________________________________________________
VIBRATEC report ref. 001.026.RA.01.B for TECHNIP ABU DHABI
Appendices - 4
 APPENDIX 2

Velocity response cartographies on skid

______________________________________________________________________________
VIBRATEC report ref. 001.026.RA.01.B for TECHNIP ABU DHABI
Appendices - 5
 Figure A13: Computed velocity (mm/s peak) – on skid), 1.6E+01 Hz

Figure A14: Computed velocity (mm/s peak) – on skid, 2.54E+01 Hz

______________________________________________________________________________
VIBRATEC report ref. 001.026.RA.01.B for TECHNIP ABU DHABI
Appendices - 6