1. INTRODUCTION

The purpose of this investigation is to estimate, through a finite element analysis, the torque and leakages of a double and engineered sealing unit, to then validate them with those obtained by testing the same Unit on an instrumented bench.

Moreover, we wish to underline how the values of the leakages are affected by a number of operating factors, such as the pressures and temperatures of the various fluids, therefore it is not possible to speak of unique values for them, but rather highlight variation ranges.

For the tests we used one of the benches, equipped with data logger, installed in the FLUITEN test rooms and a picture of which is shown.

PICTURE OF THE TEST BENCH – FIG.1

As Seal Unit, the GTP-16141-0000 was used, consisting of two opposing and pressurised seals, intended for use on a multiphase pump in the field of oil & gas.

On the plant the GTP is pressurised through a modified and engineered API PLAN 53B type system.

Real operating conditions.

• Pump shaft diameters: fitted with step with two diameters 120 and 140 mm
• Number of revs: 993 RPM

Process fluid

• Pressure : 3.8 barG
• Type: Desalted Stabilised – Crude Oil

Barrier fluid:

• Maximum pressure: 10.8 barG
• Barrier liquid: ISO VG 32

Seal rings materials:

• Process side: silicon carbide/Silicon carbide (presence of abrasive solids in the pumped liquids)
• Atmosphere side: silicon carbide/Graphite.

The two seals, process and atmosphere side are similar to each other and the variations are only dimensional, as shown in FIG. 2.

2. DESCRIPTION OF THE TESTS

Test Conditions

The need to balance the axial thrusts on the bench makes it necessary to assemble two opposed cartridges joined by a specially constructed stuffing box. Liquid circulation is assured by two separate units, each equipped with heat exchanger and circulation pump, so as to remove the power dissipated by friction from the seals and limit heating of the liquids.

Circulation diagram

The barrier liquid circulates in series between the two cartridges and the outlet of the first coincides with the inlet of the second, as shown in FIG. 2

ISO VG32 type oil is used as process and barrier fluid, whose pressures are as follows:

• Process pressure: 3.8 barg
• Barrier pressure: 6.6 barg

With regard to the number of revolutions, it is assumed to be equal to the operating one, i.e. equal to: 998 RPM

Temperature reading.

The data logger makes it possible to record the fluid temperatures and this is essential for defining the thermal boundary conditions of the seal rings, in the FEA simulation.

Barrier fluid

The temperature on first cartridge inlet and second outlet is recorded and the total increase is divided, exactly, between the two.

LBI Inlet temperature : 50 °C
LBO Outlet temperature : 88 °C
Average delta per cartridge : DT=(88-50)/2=19°C

We can, therefore, assume an average temperature of the barrier between the two cartridges, as shown in TAB.1.

Fluid simulating the product

The temperature on stuffing box inlet and outlet is monitored, as shown in TAB.2.

3. FINITE ELEMENT ANALYSIS

The SEALHDNL software is used, solely intended to analyse mechanical seals and the formation of liquid film between the sliding faces.

The steps followed are listed briefly below, which are the typical ones followed in all FEA modelling.

• Geometric design of the seal rings and their subdivisions in mesh.
• Definition of the various concentrated loads, such as springs and hydraulic thrusts, and their application points.
• Definition of the pressure distribution.
• Definition of the temperature distribution.
• Definition of the fluid distribution.

All significant quantities are summarised in table FIG. 4 and TAB.3

ATMOSPHERE SIDE AND PROCESS SIDE SEAL. FIG.4

ATMOSPHERE SIDE AND PROCESS SIDE SEAL. FIG.4

Distribution and values of the pressures, temperatures and fluids used in the FEA schematisation.

4. THEORETICAL RESULTS OBTAINED

DRAG TORQUE

This analytical value can be easily compared with the real value obtained on the bench on which a torque meter is placed. The reading and recording are continuous during the tests, its average value is around 115 N*m for two cartridges

TAB.4 compares the theoretical and measured values and, from this, it can be seen that the values match by order or magnitude.

It is immediately noted that, in terms of order of magnitude, the measured data match the theoretical data obtained from the analysis.

TEMPERATURE OF THE RINGS

Another interesting set of analytical data is the distribution of temperatures on the faces but, unfortunately, difficult to measure and, therefore, one must restrict oneself, with all due doubts, to the theoretical one.

The two slides show that the atmosphere side seal is affected by a greater thermal increase, as it is subjected to the total pressure and, moreover, on the inner part it is lapped by the air which, being gaseous, has low thermal inertia and a poor ability to remove heat. As a positive aspect one notes its even distribution of temperature over the entire height of the face which, due to the “compliance” of graphite, signifies regular contact of the faces.

With regard to the process side, it is affected only by a pressure differential but, having two hard materials, a slight edge contact is observed on the inner edge where the maximum temperature reached is highlighted.

LEAKAGES

For leakages, we generally think of a unique value or, at most, variable with pressure. In actual fact the change of each operating parameter involves a small or significant change of the leakages, whose visibility is linked to their size. These pages do not expect to identify exact values or parameters but simply highlight how “ranges” are present in which the leakage values are contained as some operating parameters change. For operational simplicity, only the temperature and pressure of the barrier fluid are changed.

Two pressure limits are assumed: P= 7.6 bar a and 5.6 bar a, and two temperature limits: T= 50° and 90° C.

ATMOSPHERE SIDE SEAL

T1= barrier oil temperature
T2= ambient temperature
P2=atmospheric pressure
S= 918 rpm

It is observed that the two lines do not have equal slope and that the one at a pressure of 7.6 bar a is greater.
Tab. 5 shows the two values with their percentages.

Let us also examine the graphs relating to the power dissipated by friction and the height of the liquid meatus between the faces.           

As with the leakages, their percentage changes are taken into account, as shown in TAB. 6a.

PROCESS SIDE SEAL

The situation on the process side seal is more complex than that on the atmosphere side because, in addition to the temperature and pressure changes of the barrier, one should also assume those of the liquid simulating the process and, obviously, all that would involve a considerable complication in case management.

For operational simplicity, it is assumed that the process does not undergo changes in magnitude.

5. CONCLUSIONS

The different behaviour of the two seals is clear. The atmosphere side seal, despite being subjected to the total pressure, has a lower power dissipated by friction and a better formation of the liquid film and this, obviously, leads to greater leakages than the process side seal. All  that is due to the presence of a graphite seal ring, a yielding material with excellent tribological features, which assures perfect coupling of the sliding faces. Whereas, the process side seal pays for the presence of two hard, rigid and little deformable materials, whose deformations, induced by pressure as well as temperature, cause edge contacts that locally tend to break the liquid film. Obviously, the choice of the two hard materials is dictated by the fact that, in these applications, the process products are dirty and contain solid and abrasive particles. With regard to the leakage values, they vary, depending on the pressure and temperature, between a minimum of 1% and a maximum of about 20%.

Table 11 shows the totals for a cartridge referred to grams per hour.

The real leakages, measured during the tests and after a running-in period, were variable and included between a minimum of 3 g/h and a maximum of 6 g/h. The graphs of FIG. 9 and 10 show how the leakages vary according to the temperature and pressure. In the former the leakages are expressed in g/h, while in the latter they are in mg/h

FIG.9 ATMOSPHERE SIDE SEAL

FIG.10 PROCESS SIDE SEAL