Testing valve wear in lean-burn large bore gas engines

A new test rig establishes a valuable research tool for the study of the wear mechanisms of valve closure and peak cylinder pressure in gas engines, write Oliver Lehmann and Alexander Renz

Large-bore gas-fuelled engines, with a cylinder diameter ranging from 20 to 50 cm, are often used as prime movers for cogeneration of heat and power. Next to that, such engines increasingly find applications as generator drives for stabilizing the electricity grid, especially in the case of a large fraction of volatile and intermittent renewable electricity generation based on wind and solar energy.

In order to be competitive in terms of investment costs, fuel efficiency and CO2 emissions, such engines are increasingly upgraded in terms of a higher output per cylinder volume. It will be clear that a higher power output from the same amount of steel can reduce the investment costs.

Further, a higher output per cylinder volume decreases the negative effect of friction losses and heat losses on the cylinder process. Two-stage turbocharging, based on two turbochargers in series with an intercooler in between, substantially helps to increase the power output per cylinder.

In addition, the emission of undesired species such as particulates is limited by increasingly stringent legislation. This requires a maximum reduction in lubricating oil consumption and also requires a combustion process which is as complete as possible. As a result, the interface between the valve and the valve seat becomes increasingly free of combustion residues consisting of, e.g., sulphur and particulates that used to form a protective layer on the valves and valve seats. The much higher peak cylinder pressure than in the past, with values up to 250 bar, increases the stress level that the valves and their seats experience. Improvements in engine performance have therefore resulted in unacceptable valve wear rates, especially for the exhaust valves.

Valve wear according to the state of the art for large bore diesel engines is about 0.5 mm after 24,000 operating hours, in other words a wear rate of about 21 nanometer per hour (nm/h). In contrast to diesel engines, approximately thirty times greater wear rates of about 600 nm/h were measured on several worn inlet and exhaust valves in lean-burn large bore gas engines by the authors, although state-of-the-art materials and proven valve designs were used.

à‚ 

Currently, acceptable lifetimes can only be achieved by a reduction of efficiency-relevant engine parameters. ‘Quick-fix’ measures are no longer able to cope with this technical challenge. Consequently, there is a need of a dedicated solution based on an in-depth understanding of the wear mechanisms.

Examples of actual wear rates

The authors have therefore designed a special test rig with which the effect on valve wear of the temperature, the valve impact velocity on its seat, the cylinder peak pressure and different atmospheres can be investigated. For the first time, this test rig allows us to investigate the effects of these parameters separately, which is not possible in an operating engine. In terms of the potential valve dimensions, this newly designed setup is the world`s largest valve wear tribometer.

However, in order to check if the test rig would give results representative for the actual valve wear, examinations on used valves from engine tests have been carried out to determine the wear rate of two different valve types in a real engine.

To quantify the valve recession, the wear scar width was measured using 3D-laser scan profilometry. In order to determine the high-resolution oxygen penetration, x-ray photoelectron spectroscopy (XPS) was employed. The oxygen penetration indicates to what an extent a valve has been subjected to wear.

The inlet and exhaust valves tested were bi-metal friction welded valve spindles (valve disc diameter of about 110 mm, total length about 550 mm) consisting of a martensitic and an austenitic stainless valve steel. The valves were taken from the same type of large bore gas engine. The first set was hardfaced with a Stellite 12 alloy and was removed from the engine after 250 hours of operation, while the second set was overlaid with a Tribaloy T400 and was investigated after 5000 hours of operation. Table 1 compares the composition of the two different protective layers used.

Figure 1. Wear profiles of inlet valves made of Stellite 12 (black line) and Tribaloy T400 (red line) after 250 and 5000 hours of operation, respectively
Figure 2. Wear profiles of exhaust valves made of Stellite 12 (black line) and Tribaloy T400 (red line) after 250 and 5000 hours of operation, respectively

The most prominent difference in wear rate between the valves is as shown as depth profiles in Figure 1 and Figure 2. While the Stellite 12 valves suffer from high wear (about 150 àŽ¼m for the inlet and 860 àŽ¼m for the exhaust) after only 250 hours, the Tribaloy valves show wear in the range of 20-30 àŽ¼m after 5000 hours of operation. The operational conditions that have prevailed during the 250 hours and 5000 hours of operation respectively allow for a fair comparison.

Investigations showed that no protective anti-wear tribofilm was formed on the sealing surfaces. Comparison of the microstructures below the worn valve seating surfaces gave hints to the reasons for different wear rates between the different protective layers. The Stellite 12 layer showed material transport as a result of plastic deformation. Plastic deformation means that a permanent change has taken place, in contract with elastic deformation. The Tribaloy T400 hardfaced valves had some visible cracks resulting from mechanical loading, but no plastic deformation of the material could be observed.

A typical valve for large-bore engines

The oxygen content of the Stellite 12 exhaust valve differed significantly from that of the inlet valve. The inlet valve showed a drop from 25% to < 5% over 1 àŽ¼m, whereas the exhaust valve showed high oxygen content above 35% that decreases slightly to 20% in 3 àŽ¼m depth. This indicates a high oxidation of the material which can be accounted for by the higher temperature of the exhaust valve compared to the inlet valve.

A closer investigation of the identified metal oxides within the Tribaloy T400 valve wear scars reveals an oxide layer on the inlet valve which consists of chromium oxide, silicon oxide and small amounts of molybdenum oxide. The latter was expected as a so-called tribochemical product that acts as a solid lubricant. In the wear scar of the exhaust valve, small amounts of zinc oxide are found together with molybdenum oxide. It can be suggested that the formation of such oxide films minimizes friction and hence lowers shear stresses. In effect, this means less valve wear.

The results presented thus far can be interpreted in several ways. Two different hardfacing alloys apparently lead to different wear responses. Thereby, it is difficult to explain the results directly when not all parameters during operation can be varied at liberty. Ultimately, there is the need to clarify whether the wear rate is material-related, design-related or operation-related.

Figure 3. Valve wear tribometer at Fraunhofer Institute for Mechanics of Materials, equipped for test mode I – valve closure

It is precisely for these reasons that a novel test rig has been developed to provide application-relevant testing conditions for large bore gas engines in combination with full control of the discrete parameters. The novel test rig is an empowering tool to answer the key questions on what the wear-relevant mechanisms are for valve wear in present lean-burn large bore engines running on natural gas.

Valve wear under different boundary conditions

The valve wear test rig was designed as a tribometer to simulate population-relevant large bore engine valves. A typical valve disc diameter is about 90 mm and the stem diameter is about 17 mm. The valve seat rings have to be matched with these valves. The possibility for a controlled variation of all test parameters offered the greatest challenge during the design phase. Besides the wear-related behaviour of the peak cylinder pressure and the valve closing velocity, the impact of the temperature as well as atmosphere are difficult to examine in detail in a real engine.

The inner atmosphere in the test chamber should resemble the conditions inside an engine. Therefore, a mixture of nitrogen, oxygen, carbon dioxide and water vapour could be created to image the working medium in the engine. To keep a constant temperature of the components, three cooling circuits are used: two water circuits and one oil circuit. On each component, three thermocouples are applied to monitor and control the temperature of both valve and seat ring in real time, whereas the signal cables of the valve spindle are passed by a hollow shaft. A force transducer is mounted between the seat ring holder and the test frame, to allow in-situ measurements of the impact forces.

Figure 4. Wear rates of valves tested in the new test rig

The valve lift can be varied by a hydraulic cylinder. The typical valve lift applied is 25 mm. The peak combustion pressure of up to 220 bar can be simulated at a valve disc diameter of about 90 mm and a frequency of about 8.3 Hz (equal to 1000 rpm).

Preliminary results

To investigate the correlation of wear behaviour and load parameters, the valve spindles were tested at various temperatures, atmospheres and valve closing velocities. In the first phase, inlet valves were examined.

The test temperatures range from room temperature up to 430à‚°C. In general, temperatures up to 900à‚°C can be reached by the heating system. The closing velocity is controlled by the movement of a hydraulic cylinder and was set to 0.2, 0.6 and 1.2 m/s.

Examples of results from test mode I (valve closure)

Lifetimes of valves in large bore engines are normally several thousand hours. As known, properly working tribological systems show a steady state of wear after running-in. With regard to efficient testing and to comply with large bore engine applications, a testing time of 100 hours was agreed.

Figure 4 shows that testing at a material temperature of ambient value showed almost no wear. The explanation is that the material has a high strength at such a temperature. At a material temperature of 380à‚°C, the wear rate was already a factor of 20 higher. Increasing the material temperature to 450à‚°C and doubling the valve impact speed resulted in a factor of roughly 60 higher wear rate than that at ambient conditions.

In conclusion

From the comparison of the engine tests with the lab-scale experiments, the measured linear wear rate of 2.2 àŽ¼m/h as found with the test rig is of the same magnitude as for the excessively worn valves of about 3.6 àŽ¼m/h from engine test. Hence the test rig has proven that it is able to produce wear rates equal to those found in engine tests.

Consequently, it establishes a valuable research tool for the study of the wear mechanisms of valve closure and peak cylinder pressure in detail in a laboratory environment. This opens up possibilities for a dedicated and fast development of intake and exhaust valves suitable for application in modern high performance gas engines.

This paper was presented and published on the occasion of the 28th CIMAC World Congress 2016 in Helsinki. The CIMAC Congress is held every three years, each time in a different member country.

The Congress programme centres around the presentation of technical papers on engine research and development, application engineering on the original equipment side and engine operation and maintenance on the end-user side.


Oliver Lehmann is Director of Research and Development at Màƒ¤rkisches Werk GmbH www.mwh.de

Alexander Renz is a post-doctoral researcher at the Fraunhofer Institute for Mechanics of Materials (IWM) www.en.iwm.fraunhofer.de This article is available on-line. Please visit www.decentralized-energy.com

No posts to display