Influence of Hardness and Particle Load on

Leakage of Hydraulic Seat Valves

Matthias Schmidt*, Hubertus Murrenhoff*, Henrik Lohrberg+, Franz-Josef Körber+
*
Institute for Fluid Power Drives and Controls, RWTH Aachen University,
Aachen, Germany, [email protected]

+
High Voltage Products, ABB AG, Hanau-Großauheim, Germany

Abstract As opposed to spool valves, seat valves are often as-

Seat valves constitute the biggest part of hydraulic sumed to be hermetically tight, but in practice they show
valves sold nowadays. Compared to spool valves, they a marginal amount of leakage. This is due to the fact that
provide the advantage of being able to isolate a hydraulic the surfaces of sealing elements exhibit a certain rough-
pressure nearly without leakage. However, in certain ap- ness. When closing the valve, the surface asperities come
plications any leakage whatsoever is prohibited. There- to rest on each other. Elastic and plastic deformations of
fore, efforts are carried out to increase the tightness of the asperities are induced by the sealing element’s contact
metallic sealing. pressure. However, there still remain microscopic gaps
A test rig was designed and built to investigate valves between the asperities, where micro flows can build up,
made from steel with different values of surface hardness. leading to leakage [2], [3], [4].
They can be tested with varying hydraulic and mechani- Publicly available literature includes only very few
cal parameters, such as particle load of the oil. To adjust experimental studies dealing with leakage of metal-
the contact force between piston and seat, an adjustable to-metal contacts. Mostly, the experiments were derived
pressure supply is being used. A second adjustable high from static gasket or similar seals. The fluid used for
pressure supply, which provides pressures of 350 bar, is measurements was often some kind of gas and no direct
used to create leakage. To exclude influences of other comparison to liquid leakage was proposed. In fact, to the
seals on the leakage, the flow is measured at the low authors’ knowledge, no publication deals with oil leakage
pressure side of the seat. in hydraulic seat valves directly. [5] gives a chronological
This paper presents the contact force depending leak- introduction of the experimental studies.
age of three valves with different hardness. It is shown As all seal contacts used in these experiments differ
that a harder valve material requires higher contact forces strongly from the design of hydraulic seat valves and the
to achieve acceptable tightness. Furthermore, the charac- fluid in use was gas or alcohol, the applicability of the
teristic of leakage over time and its change with increas- conclusions to common seat valves and particle contami-
ing particle load of the oil is discussed. nated oil is questionable. Furthermore, the time depend-
ency of leakage, observed in practice, could never be
Keywords
measured due to the different fluid in use.
leakage, seat valve, surface, tightness, oil contamina-
To abolish this drawback, a test bench has been de-
tion, particle
signed and set up, which enables simultaneous examina-
tion of several test valves. The design of the valves is
1 Introduction similar to common seat valves, hence results can be con-
ferred easily [5]. Furthermore, three test valves have been
A hydraulic seat valve is characterised by a sealing
built, with which hardness variations can be conducted.
element, which can separate or connect the hydraulic
Dependencies of different parameters on the leakage
ports by being attached or taken off a seal seat. These
of metallic seal seats have been determined experimen-
valves are distinguished – according to the type of sealing
tally. [5] presents results of measurements, showing the
element – into cone, ball, and disc seat valves [1].
dependency of leakage on time and geometry, whereas
[6] focuses on the description of the leakage, depending The test valves in use are made from 16MnCr5
on time and contact force. (1.7131), a case-hardened steel, with variable hardness.
This publication shows measurements of the leakage Table 1 shows the list of test valves. Valve I is not hard-
in test valves with varying hardness and particle load of ened at all, so its hardness is very low with 140 HV30.
the oil. Valve II and III are hardened by carbonisation to values
of 600 HV30 and 760 HV30 respectively.

2 Test rig and test items Table 1: Test valves in use

In Fig. 1 the measuring principle is depicted in a sim- test valve no. hardness

plified illustration. The piston of the test valve is pushed I 140 HV30
into the seat against the release spring by an adjustable II 600 HV30
pressure. A second adjustable high pressure supply, III 760 HV30
which provides pressures of up to 350 bar, is used to cre- Figure 2 shows the cross section of an assembled test
ate a differential pressure producing leakage. To exclude valve, consisting of the sealing element and the seal seat.
the influence of other seals on the leakage, the flow is Both parts are guided by a sleeve, which was manufac-
measured at the low pressure side of the seat. tured with very high accuracy to minimise influence of
tolerances on the leakage. Measurements revealed that the
pressure supply
for leakage creation concentricities of the guiding bores stay below 1 µm.
The piston of the sealing element separates the pres-
metallic seat
sure for generating the leakage (main pressure) from the
pressure supply for release spring pressure for the contact force. It also minimises the main
contact force
pressure’s influence on the contact force.

scale for
leakage measurement

Fig. 1: Simplified layout of the test rig

For measuring the leakage, precision scales were

chosen, offering the following advantages: Owing to the
high resolution of 1 mg, very small flow rates can be de-
termined. At the same time, it is also possible to reliably Fig. 2: Cross section of a test valve in the leakage test rig
detect conceivably larger flows. These are only restrained
by the overall weighing range of 0–160 g. Tests have
3 Measurement results
proven that the measuring range covers approx.
0.005–50 ml/min when measuring for a period of three For carrying out the measurements, firstly a certain
minutes. Measurement pipes conduct the leakage to beak- pressure for creating the contact force is applied to the
ers positioned on the scales. In order to prevent a stepwise piston of the test valve. Afterwards, the pressure for leak-
increase in leakage due to single drops, the pipes end be- age creation is raised to a predefined value and data ac-
low oil level. This ensures that a constant volume flow quisition is started. During the measurements, both pres-
rate can be measured by the scales. By differentiating the sures and temperature are recorded and kept constant, as
time-dependent signal and dividing it by the density of shown in [6]. The sampling of the scales provides the
the fluid, the flow can be calculated. actual test result, i. e. the accumulated leakage.
 The contact force FCF, taking effect in the sealing ference between the measurements was the particle load
contact, is calculated by the difference of the pressure of the oil.
generated force Fpressure and the force of the release spring The contamination level of measurement A was de-
Fspring, divided by the cosine of the angle α between the termined according to ISO Code 4µ/6µ/14µ to a value of
normal in the contact and the piston axis. To normalize 20/17/09. This is considerably high pollution for a hy-
the value regarding the size of the valve, the force is di- draulic system. In contrast, in measurement B the oil con-
vided by the length of the sealing contact with the diame- tamination was only 12/10/08, which means a very clean
ter dseal: fluid. The particle amount belonging to these values can
Fpressure − Fspring be seen in Tab. 2, according to ISO 4406/1999 [7]. Com-
FCF = (1)
π ⋅ d seal ⋅ cos α paring these values, it becomes clear that the oil used for
measurement A had approximately 250 times more parti-
The hydraulic force is composed of two parts. The
cles larger than 4 µm than the oil used for measure-
main pressure pLC has only minor influence on the contact
ment B.
force, as it acts on both the diameter of the seal dseal and
400 20
the piston dpiston, but in opposite directions. The main part

pressure [bar]

contact force [N/mm]

of the contact force is generated by pressure pCF, acting 350 17.5

on the piston diameter only, as in

300 15
π
Fpressure = ( pLC (d seal
2
− d piston
2
) + pCF ⋅ d piston
2
(2) 250 12.5
4
200 10
It was shown in [5] and [6] that leakage is a time-de-
pendent value. It was assumed that particles, being 150 7.5

washed into the small gaps, cause a clogging of the seal-

100 5
ing contact and thus gradually decrease the leakage.
pressure leakage creation, A
50 pressure leakage creation, B 2.5
12 contact force, A
leakage, accumulated [ml]

measurement A (20/17/09)
contact force, B
measurement B (12/10/08) 0 0
0 30 60 90 120 150 180
10 time [s]

Fig. 4: Main pressures and contact forces; measurements

8
with different particle loads

6
The curves in Fig. 3 prove that particle load has an in-
fluence on leakage at metallic seal seats. It can be seen
4
that, at the beginning of the measurement, both curves
2 have the same gradient. This means that the leakage flow
rate starts independently from contamination with the
0 same value, but decreases faster over time with more par-
0 30 60 90 120 150 180
time [s] ticles.

Fig. 3: Accumulated leakage over time; measurements with Table 2: Contamination of oil, according to [7]
different particle loads ISO-Code particles per 100 ml
(ISO 4406) from to
Figure 3 illustrates the leakage over time of two meas- 8 130 250
urements at the same test valve. The gradient of such a 9 250 500
curve represents the actual leakage flow rate. The main 10 500 1 000
pressure and the contact force have been constant and of 12 2 000 4 000
the same value, as demonstrated in Fig. 4. The only dif- 17 64 000 130 000
20 500 000 1 000 000
 In a second test series the effect of the valve material In a second test series three valves with varying hard-
hardness on the leakage was investigated. For this pur- ness were tested with identical hydraulic and mechanical
pose, the average leakage flow rates at different contact edge conditions, focussing on the leakage depending on
forces were measured and combined for each test valve. the contact force. The measurements reveal that a higher
Figure 5 shows the results for all three valves, plotted in a hardness of the valve material leads tonecessary higher
double-logarithmic diagram. contact forces, and thus support the theory of rough sur-
The curves reveal that leakage decreases with in- faces.
creasing contact force. This behaviour agrees with the
theory of rough surfaces, as described in the introduction.
Nomenclature
The gaps between the areas in contact become smaller
when the asperities are levelled due to higher contact α [-] Angle between the normal in the contact and
the axis of the piston
forces.
dpiston [m] Diameter of the piston at the sealing element
dseal [m] Diameter of the sealing contact
average leakage [ml/min]

10
FCF [N] Contact force in the sealing contact
Fpressure [N] Pressure-generated force on the sealing
element, acting in closing direction
Fspring [N] Force on the sealing element by the release
1
spring, acting in opening direction
pCF [Pa] Pressure for generation of the contact force
pLC [Pa] Main pressure for generation of the leakage
0.1

References
valve I (low hardness)
valve II (medium hardness) [1] Ebertshäuser, H., Helduser, S.: Fluidtechnik von A
valve III (high hardness)
0.01 bis Z. O+P Ölhydraulik und Pneumatik, Vere-
4 6 8 10 20
inigte Fachverlage Mainz, 1995.
contact force [N/mm]
[2] Marie, C.: Fuite monophasique au travers d’un con-
Fig. 5: Leakage over contact force; valves with different tact rugueux: contribution à l’étude de l’étanchéité
hardness statique. Thèse de doctorat, Université Bordeaux 1,
2002.
A higher hardness of the valve material leads to [3] Marie, C., Lasseux, D., Zahouani, H., Sainsot, P.: An
integrated Approach to Characterize Liquid Leakage
higher necessary contact forces to reach the same level-
Through Metal Contact Seal, Europ. J. Mech. &
ling. The curves in Fig 5 confirm this behaviour, as they Env. Eng., Vol 48(2), pp. 81–86, 2003.
show increasing leakage rates with increasing hardness. [4] Marie, C., Lasseux, D.: Experimental Leak-Rate
Measurement Through a Static Metal Seal, J. Fluids
Eng., Vol 129, pp. 799–805, 2007.
4 Conclusions
[5] Schmidt, M., Murrenhoff, H., Lohrberg, H., Körber,
F.-J.: Tightness investigation on hydraulic seat
The aim of this project is to investigate the influences
valves. Proc. ASME international Mechanical En-
on leakage in seat valves by an experimental approach. gineering Congress and Exposition IMECE2008,
For this purpose a test rig was built, which allows for the Boston, USA, 2008.
investigation of custom-made test valves with adjustable [6] Schmidt, M., Murrenhoff, H., Lohrberg, H., Körber,
hydraulic and mechanical edge conditions. F.-J.: Influencing parameters on tightness of hydrau-
lic seat valves. Proc. ASME Fluid Power and Mo-
Results of measurements carried out using oil with tion Control FPMC 2008, Bath, Great Britain,
different degrees of contamination have been presented. pp. 469–479, 2008.
These measurements support the theory, that particles in [7] ISO 4406/1999: Hydraulic fluid power – Fluids –
the oil are responsible for increasing the tightness of a Method for coding the level of contamination by
solid particles. International Organization for
hydraulic seat valve with time.
Standardization, 1999.