IMPROVEMENTS IN ORRELATING TO TURBOCHARGERS. AND A TURBOCHARGERTHRUST BEARING

The invention relates to a method for limiting damage to a turbocharger in the event of an interruption in the supply of lubricant to a thrust bearing of the turbocharger, and to a turbocharger thrust bearing.

Turbochargers are often employed in industrial diesel engines, since they can provide a significant increase in the power output and efficiency of the engine. A cross-section of a typical turbocharger, as used in an engine, is shown in Fig. 1. The turbocharger comprises a turbine section 10 and a compressor section 12. The turbine section 10 includes a set of turbine blades 14 mounted to one end of a shaft 16, while the other end of the shaft has mounted to it an impeller 18, which is part of the compressor and includes a set of blades 19. In use, exhaust gas from the engine is directed into the turbine section from the right-hand side (see arrows 20), where it impacts, and thereby turns, the turbine blades 14. Since both the turbine blades and the impeller 18 are firmly fixed to the shaft 16, the impeller blades 19 turn as well. Air from outside the engine is directed into the compressor from the left-hand side (see arrows 22), and this air impacts on the impeller blades 19, which impart a centrifugal component of motion to the air, forcing it through a volute 24 and out of the turbocharger. The exiting compressed air is fed to the engine, which uses it to develop power.

The shaft 16 is mounted in the turbocharger via journal bearings. In Fig. 1 the bearings lie at the two ends of the shaft 16, and comprise a first journal bearing 26 at the turbine end and a second journal bearing 28 at the compressor end. Thus, in the example shown, both bearings lie "in-board" of the turbine and impeller. However, other designs are possible, in which at least one of the bearings lies "out-board" of the turbine/impeller, i.e. axially outside one or both of these components. Also at the compressor end is an additional bearing component, which is a thrust bearing 30.

The thrust bearing is shown in greater detail in Fig. 2. It is usually a hydrodynamic bearing and comprises a thrust collar 40 (see also Fig. 1) and an annular thrust pad 42. The

thrust collar 40 is mounted to the shaft 16 as an interference fit and lies in contact with the thrust pad 42, which in turn lies against part of the compressor-end journal bearing 44. In use, the shaft is forced in a direction 46 by the impact of the incoming exhaust gas at the turbine end and by the effect of a low-pressure region 48 (see Fig. 1) caused by the action of the impeller 18. The resultant combined force is a so-called thrust force, which urges the thrust collar hard against the thrust pad 42. The thrust pad 42 may be either a separate component from the journal bearing or it may be integral with the journal bearing. When it is separate, it may either be fixed to the journal bearing or float with respect to it. In the latter case there will normally be two thrust-pad components: a fixed component, which is fixed to the journal bearing, and an additional floating thrust plate, which lies between the fixed component and the thrust collar and rotates (notionally) at half the rotor speed.

In order to absorb the thrust force in the direction of the arrow 46, those surfaces of both the thrust collar 40 and the thrust pad 42, which face each other, are provided with a wear-resistant coating. In the event that the thrust pad is movable relative to the journal bearing, the facing surfaces of the fixed thrust-pad component and the floating thrust plate, and the facing surfaces of the floating thrust plate and the thrust collar 40 are provided with the wear-resistant coating. The thrust collar and pad, which act as substrates to the coating, are composed of a suitable metal. Conventionally the metal is an aluminium-tin alloy, while the coating is applied to the metal by an anodizing process.

To prevent wear due to the thrust force, it is necessary to ensure that the facing surfaces of the thrust collar and thrust pad (and of the thrust pad and journal bearing, where applicable) are well lubricated. To achieve this, it is known to provide the thrust pad with a series of circumferentially spaced apart grooves, as shown in plan and side views, Figs. 3(a) and 3(b), respectively. In this configuration a number of platform portions 50 are provided around the circumference of the thrust pad 42 between grooves 52. The platform portions 50 are not level over their circumferential extent, but consist of an inclined portion 54, which rises from an adjacent groove and leads into a plateau portion 56, the plateau portion 56 in turn leading into the next adjacent groove. Thus the platform portions, which are often referred to as "pads" (not to be confused with the "thrust pad", which is the whole component), are wedge-like in shape. It should be appreciated that the angle of the inclined

portions is shown greatly exaggerated for clarity. In use, the thrust pad is supplied with lubricating oil, which enters the grooves 52 and rises up the inclined portions 54 onto the plateau portions 56. It is the plateau portions which are in direct contact - via the resultant oil film - with the thrust collar 40 in normal use of the turbocharger.

The arrangement described ensures a reasonable life of the thrust bearing, provided that the oil supply does not fail or become severely reduced. If it does, the thrust forces and high speed of rotation arising from normal use of the turbocharger will bring about complete destruction of the bearing. In practice, this is likely to result in considerable damage to other components of the turbocharger and even to parts of the engine itself. This is because the thrust collar erodes away the above-mentioned platform portions, thereby allowing the shaft to move axially in the direction of the compressor. Consequently the impeller 18 collides with other, neighbouring components in the compressor area, and the damage which is caused to these other components can then create problems to parts of the engine adjacent the turbocharger. Thus, failure of the lubrication supply can prove very expensive to remedy, and gives rise to an undesirable downtime of the engine, while damaged parts are being replaced.

In an attempt to avoid these problems, much effort has been directed toward making sure that the oil supply to the bearing is not interrupted. Thus, although electrical oil pumps have been used in this context, mechanical pumps are often preferred, since they are not susceptible to electrical breakdown. Consequently, as long as the engine is running, there should be a supply of oil to the bearing. Furthermore, in some applications a gravity-fed reservoir of oil is supplied to the turbocharger, in order to ensure that, if the lubrication pump fails, some oil is supplied to the turbocharger during the time it takes for the engine to run down. However, despite these precautions, the interruption of the oil supply to the turbocharger thrust bearing is still the prime cause of turbocharger failure.

It is therefore desirable to be able to limit the damage caused to a turbocharger, and to its associated engine, in the event of a failure in the lubricant supply to the turbocharger thrust bearing.

In accordance with a first aspect of the present invention there is provided a method for limiting damage incurred by a rotating machine in the event of an interruption in the supply of lubricant to a thrust bearing of said rotating machine, the method comprising the steps of: providing as a thrust bearing a thrust-bearing substrate comprising a material of high melting temperature, the substrate being provided with a coating of high hardness and low coefficient of friction; providing a sensing means for sensing said interruption in the supply of lubricant, and providing a monitoring means connected to the sensing means for providing an indication of said interruption in the supply of lubricant; the method further comprising the steps of: indicating said interruption in the supply of lubricant before significant damage has been done to said turbocharger, and stopping said turbocharger following said indication.

The method may comprise the further steps of: removing said thrust bearing, and replacing said thrust bearing with an undamaged thrust bearing.

The substrate material is preferably a metal having a melting temperature of > 1000 ⁰ C and the coating preferably has a hardness of > 1000, HVO.05, and a coefficient of friction of < 0.2. More preferably, the thrust bearing comprises a steel thrust collar and a steel thrust pad, the mutually facing surfaces of which are coated with a ceramic coating. Most preferably the ceramic coating comprises carbon-enriched tungsten carbide.

In a second aspect of the invention, a thrust bearing for a rotating machine comprises: a thrust collar and a thrust pad comprising a material of high melting temperature, the mutually facing surfaces of the thrust collar and thrust pad being coated with a coating of high hardness and low coefficient of friction.

Preferably the thrust-collar and thrust-pad material is a metal having a melting temperature of > 1000 ⁰ C and the coating has a hardness of > 1000, HVO.05, and a coefficient of friction of < 0.2. More preferably the thrust collar and thrust pad are composed of steel and the coating is a ceramic coating. Most preferably, the ceramic coating comprises carbon-enriched tungsten carbide.

The thrust bearing may further comprise a means for sensing an interruption in the supply of lubricant to the thrust bearing.

In both aspects of the invention the sensing means may be a temperature sensor embedded in the thrust pad. Alternatively, the sensing means may be a resistance sensor for detecting changes in the electrical resistance of said ceramic coating.

The resistance sensor may comprise an array of interconnected resistors embedded in the coating, the array being disposed so that erosion of the coating results in destruction of one or more of the resistors. Alternatively, the resistance sensor may comprise a single resistive element embedded in the coating, the resistive element being disposed so that erosion of the coating results in a reduction in a cross-sectional area of the resistive element. Yet a further alternative is for resistance sensor to comprise a pair of low-resistance conductors embedded in a substrate of the thrust pad in spaced-apart manner, at least an end- face of the conductors being in contact with the coating.

The rotating machine may be a turbocharger.

An embodiment of the invention will now be described, by way of example only, with reference to the drawings, of which:

Fig. 1 is a cross-sectional view of a typical turbocharger;

Fig. 2 is an enlarged view of a portion of the turbocharger shown in Fig. 1;

Figs 3 (a) and 3(b) are a plan view and a side view, respectively, of a thrust pad employed in the known turbocharger of Figs. 1 and 2;

Fig. 4 corresponds to Fig. 2, but with the provision of a temperature-sensing device in the thrust bearing, the temperature-sensing device being employed in a method of limiting damage to a turbocharger in accordance with a first embodiment thereof;

Figs. 5(a) and 5(b) are two variants of a resistance-measuring scheme that may be employed in a method of limiting damage to a turbocharger in accordance with a second embodiment thereof;

Fig. 6 illustrates a further variant of the resistance-measuring scheme of Figs. 5(a) and 5(b), and

Fig. 7 is a block diagram of a system for limiting damage incurred by a turbocharger in accordance with the present invention.

A turbocharger thrust bearing in accordance with the present invention is composed of different materials than the materials conventionally employed. More specifically, the soft aluminium alloy, which normally forms the substrate of the thrust collar and pad, is replaced with a material of a higher melting temperature, and the anodized coating is replaced with a hard coating having a lower coefficient of friction than the anodized coating. Possible materials for the substrate and/or coating are "hard" metals, including high- temperature alloys, e.g. nickel-based materials, and/or materials that can sustain high pressures, such as ceramics. It is not anticipated that organic materials or polymer-based composites, such as Kevlar or carbon fibre, will be particularly suitable, despite their good energy-absorbing quality, due to contact with lubricant oil, which would degrade such materials. In a preferred embodiment of the invention, the substrate is a hard metal having a melting temperature of > 1000 ⁰ C, so that, in the event of a serious drop in oil pressure to the bearing, the bearing will survive for long enough for an alarm to be triggered and for the turbocharger to subsequently run down. (This aspect of the invention is described in detail later on.) Such a metal is phosphor bronze, for example, or preferably steel. The preferred embodiment also has a coating with a hardness > 1000 (HVO.05), which refers to Vickers hardness evaluated using a force of 0.05kgf. The coating also has a low coefficient of friction of maximum value 0.2. This is to ensure that, in the event of an oil failure, the risk of scuffing and cold-welding of the facing bearing-surfaces is minimized, till the turbocharger can be brought to a standstill. Thus the coating may be a ceramic coating, for example, and preferably a carbon-enriched tungsten carbide coating. This provision of the high-melting-temperature substrate and tougher, lower-friction coating enables the thrust bearing to survive a failure in the lubricant supply for a longer period than the conventional bearing. More specifically, the use of a high-melting-temperature substrate ensures that the wedge-shaped bearing pads do not lose their shape in the event of a lubricant failure.

To test this principle, a turbocharger with a thrust bearing of the above-mentioned preferred constitution - namely, a steel substrate and a carbon-enriched tungsten carbide coating - was mounted in a test rig, in which a small burner was situated in the turbine entry

duct (see location 20 in Fig. 1). This burner produced the hot exhaust gases necessary to set the turbine rotating, and therefore in that capacity acted as a substitute for the reciprocating engine that would normally be present. No auxiliary lubricant-oil pumps were employed. During the test the turbocharger was allowed to undergo a crash stop, in which the burner was turned off and/or a mechanical failure was simulated, for example through a broken crank. The turbocharger was then left to freewheel until it stopped. In an alternative set-up, the test may be done with the turbocharger mounted to a reciprocating engine. In this case, to initiate the test, the fuel to the engine is turned off, as is also the lubrication pump. Again, the turbocharger is allowed to freewheel to a stop. With the conventional (i.e. anodized aluminium) bearing this test would have caused catastrophic failure of the turbocharger, but with the ceramic-coated steel bearing the turbocharger operated normally after the test. During the test, the temperature of the bearing increased from around 110 ⁰ C to 190 ⁰ C in a few seconds. On inspection, the coating on the pads was found to be damaged. During a subsequent test using the same damaged bearing, the oil pressure was reduced while the turbocharger was operating. This test resulted in failure of the bearing, although the damage to the turbocharger was limited due to the harder bearing-substrate material. Bearing temperatures in excess of 700 ⁰ C were measured during this subsequent test.

While the use of the steel-substrate/ceramic-coating bearing material by itself helps to limit damage to a turbocharger when the oil supply fails, there is a second measure which the present invention takes to further ensure that no extensive damage occurs to the turbocharger and its associated engine. This measure provides the bearing with a means for detecting a serious reduction in the pressure of the oil supply, and provides an indication of such failure before significant damage has been caused to the turbocharger.

In a preferred embodiment of the invention a serious reduction in oil pressure is detected indirectly by monitoring a parameter which is affected by such a pressure reduction. A first such parameter is the temperature of the bearing, and in particular the temperature of the ceramic coating, which is immediately affected by an oil-pressure reduction. A number of known sensors are available to achieve this, including thermocouples, thermistors and semiconductor devices, e.g. diodes. The sensor element is inserted into a bore provided in the thrust-pad substrate and wires are run from the sensor element out of the thrust pad to a

monitoring device, which in practice may be part of the existing engine control system. Ideally, the sensor element should be in contact with the ceramic coating, but since steel is a good conductor of heat, it is sufficient if the sensor is located in the substrate only, but as close to the ceramic-coated surface of the substrate as possible. An example of the use of a thermocouple to sense temperature in this manner is shown in Fig. 4. In Fig. 4 the thermocouple is item 58 received in a bore 60, and the wires emerging from the thermocouple are shown as item 62. The bore may be disposed parallel to the longitudinal axis of the turbocharger or perpendicular thereto, i.e. radially oriented, or indeed at any convenient orientation.

A second parameter, which may be used as an indirect indication of oil-pressure reduction, is electrical resistance. In this case, the resistance of the thrust pad is sensed between two points thereof by running two wires to the thrust pad from the monitoring device and passing a small current through them. Either the current can be supplied from a constant-current source in the monitoring device, or it can result from the application of a constant voltage across the wires. In the former case the magnitude of the voltage across the wires is a measure of the resistance between the two points, while in the latter case the current through the wires is a measure of the resistance between the two points. In either case it is assumed that the resistance of the wires is either known or is negligible compared with the resistance of the thrust pad.

An example of such a resistance-measuring scheme is illustrated in Fig. 5(a). In Fig. 5(a) an array 70 of fine resistors is embedded in the coating 72, which is shown disposed on the substrate 74 of the thrust pad 42. This array may be employed with one or more of the platform portions 50 (see Fig. 3(a)) of the thrust pad. When the thrust collar 40 comes into contact with the coating 72, it shaves off one or more of the resistors, which changes the overall resistance of the array. Note that, although the array is shown as a parallel connection of resistors, it could equally well be a series-connected array. Furthermore, the array could be voltage-driven, rather than current-driven, as shown. The monitoring circuitry, which senses the resistance, is arranged to trigger an alarm when a given number of resistors have been eroded away. The larger the number of resistors, which are required

to be eroded in order to give the alarm, the greater the wear allowed to the thrust pad and the higher the temperature suffered by it.

A variant of the arrangement just described is shown in Fig. 5(b). In this variant a substantially axial bore 75 is provided in the thrust pad, the bore penetrating the whole of the depth of the coating 72 and at least part of the substrate 74. The bore 75, which may be repeated at predetermined points around the circumference of the thrust pad, receives a plug 76, which consists of an insulating material, in which is embedded a conducting plate 77, to which are connected a pair of current-carrying wires, in the manner of the Fig 5(a) variant. When the bore 75 does not extend all the way through the substrate 74, a pair of narrower bores will be provided to receive the wires. These narrower bores will likewise contain an insulator to prevent shorting between the wires and the substrate. The insulating material is selected so as to wear away approximately as fast as the coating 72. Again, as in the Fig. 5 (a) variant, the resistance of the plate 77 changes as it is worn away by frictional contact with the thrust collar, the change being due to a reduced cross-sectional area presented to the current, which flows through the plate. As with the previous variant, the monitoring circuit will trigger the alarm once the cross-sectional area has been reduced by a predetermined amount.

Although Fig. 5(b) shows the use of a laminar plate 77, it may alternatively be a three-dimensional body. The same principle of reduction in cross-sectional area applies here also, leading to a change in resistance of the 3-D body. Similarly, in the Fig. 5(a) case the individual resistors may be laminar or three-dimensional. In the latter case, the cross- sectional area offered to the sensing current will be greater and the resistance of each resistor proportionally less. (Of course, a laminar body in the real world has finite thickness and is therefore to that extent three-dimensional. Hence what is meant by "three dimensional" in this explanation is a resistor (Fig. 5(a)) or a conducting plate (Fig. 5(b)) which has appreciable thickness - for example, in the case of the resistor array, a resistor thickness approximately equal to, or even greater than, the resistor height, i.e. the dimension of the resistor extending through the thickness of the coating. The latter scenario is quite feasible, given that the extent of the coating 72 into the page in Figs. 5 (a) and 5(b) is much greater than its extent in the plane of the page. At all events, the choice of resistor dimensions, all

three of them, will be readily selectable on the basis of the resistive material to be used, which sets the conductivity value, and the value of resistance desired, given the sensing circuitry connected to the resistors or conducting plate, and in particular the planned sensing current or voltage, the voltage dynamic range of any amplifiers used, and so on.)

In another scheme, it is the resistance of the coating itself, which is directly sensed, rather than the resistance of a resistive element embedded in the coating. This may involve sensing purely the coating resistance, or the series resistance of the coating and part of the substrate. It is assumed here that the resistance of the substrate will be significantly lower than that of the coating. Where this is not the case, sensing of the coating only is to be preferred.

An example of this scheme is shown in Fig. 6. In one or more of the platform portions 50 of the thrust pad (see Fig. 3(a)) a pair of bores 71 are made through at least part of the depth of the substrate 74 and slightly into the coating 72. Each bore receives a low- resistance conductor 73 surrounded by an insulator 78. The insulator extends to the end of the bore and the conductor protrudes beyond the end of the bore, so that an end-portion of the conductor contacts the coating 72. In a variant thereof, the conductor terminates at the end of the bore 71, so that only the end-face of the conductor contacts the coating. Wires 79 are connected to the other ends of the conductors 73 for the passing of a current therethrough. In this scheme, the current senses primarily the resistance of the coating itself. Note that the same comments above regarding the laminar or 3-D nature of the resistive body apply here also.

The invention further includes a situation in which, for each of the pads in which the resistor array 70 is provided, further such arrays are provided one behind the other extending into the page (see Fig. 5(a)). The resulting stack of linear arrays will be either monitored individually or together by interconnecting the arrays in a suitable manner. Thus, for example, the connecting wires shown feeding the array 70 in Fig. 5(a) could constitute common connecting wiring for all of the arrays, so that the arrays are all connected in parallel. Similar considerations apply to the other variants of the resistive-sensing principle.

Thus, there could be two or more conducting plates 77 one behind the other going into the page in Fig. 5(b), and similarly for the Fig. 6 variant.

When resistance or temperature sensors are employed with a number of pads in the bearing, they will be monitored either individually or together. In the latter case the sensors of one pad may be simply connected in series or parallel with the sensors of the other pad or pads and be fed with a common sensing voltage or current.

When the thrust bearing is a floating bearing, as mentioned earlier, the sensing measures described above are employed in the fixed part of the bearing, rather than in the floating part.

A block diagram of the monitoring system is shown in Fig. 7. Here it is assumed that the sensor element in the thrust pad is the thermocouple 58 first shown in Fig. 4. The thrust pad with its thermocouple 58 is part of the turbocharger 80. The wires 82, 84 from the thermocouple 58 are taken to a temperature monitor 86, which amplifies the signal from the thermocouple in a differential amplifier 88 and compares the amplified signal with a reference voltage, V _REF , in a comparator 90. The output of the comparator 90 is fed to an alarm indicator 92, which may be an audible and/or visual alarm, such as a siren, flashing light, etc. In addition to, or instead of, such an alarm, a line 94 may be taken to the engine control system to provide for automatic run-down of the engine and turbocharger in the event that the comparator 90 signals an interruption in oil supply. Items 86-92 may be separate from the engine control system, or may form part of it.

Reference voltage V _REF is set so that the comparator provides an indication of lubrication failure before the thrust pad has become so severely damaged that the shaft moves forward axially and causes further damage to the turbocharger and possibly the engine, as already described earlier. In view of the use of a thrust bearing with a hard steel substrate and a hard ceramic coating, there is sufficient time for the comparator to respond before this point of excessive damage is reached. In practice, V _REF may be set so that it corresponds to a rise in temperature of between 50 ⁰ C and 100 ⁰ C. A preferred level is roughly midway between these limits, namely 75°C. IfVREF is set too low, it is possible that

an alarm could be triggered by the normal temperature excursions experienced by the thrust bearing, whereas if it is set too high, the alarm might be triggered too late, since damage may have started to be done to the engine at that stage.

Although it has been assumed that the interruption of the oil supply will be monitored by detecting changes in a parameter such as temperature and resistance, the invention also envisages the use of pressure sensors for directly sensing a significant change in oil pressure. Such pressure sensors, however, will not normally be disposed in the thrust bearing itself, unlike the two sensor examples described above. Other sensing methods are also envisaged. For example, a system such as described in European Patent Application 0999113, published on 10th May 2005, may be employed. This system employs a plug material, which plugs one end of a passage provided in a bearing component. The other end of the passage is connected to one end of a conduit, the other end of which is connected to a source of pressurized gas. When the temperature of the bearing component rises above a given value, the plug melts and the gas pressure in the conduit falls. This fall in gas pressure is detected and provides an indication of over-temperature in the bearing. This arrangement could be incorporated into the thrust pad of the present invention. The choice of plug material would be important, since it would be necessary to choose a material which melted at a temperature, at which the thrust pad was not severely damaged. Such a material would be, for example, a tin or zinc alloy. In the case of the present invention, the removal of the plug would lead to a stream of lubricant oil entering a new cavity, where the leakage would be detected.

In summary, the present invention provides a turbocharger thrust bearing, which is capable of surviving at least one interruption of the lubricating oil supply to the bearing with minimal damage. This is due to the composition of the bearing components: the thrust collar and pad are made of a high-melting-temperature (e.g. > 1000 ⁰ C), preferably hard, material such as steel, and the coating is a low-friction (e.g. coefficient of friction < 0.2), hard material (e.g. hardness > 1000 (HVO.05)) such as a ceramic, e.g. a carbon-enriched tungsten carbide material. Thus, while the bearing will sustain some damage, it will not be sufficient for engine damage to occur, provided the bearing can be replaced soon after the event. To help achieve this, the interruption of the oil supply is signalled to maintenance personnel as

soon as it occurs. These personnel then take the turbocharger out of service and replace the bearing with an undamaged one. Automatic shut-down of the engine may also take place when the oil-supply failure has been detected, in order to ensure that the turbocharger does not continue to run during the time it takes for the maintenance personnel to attend to the fault.

As regards the replacement of the bearing mentioned in the preceding paragraph, whether it will be necessary to replace the thrust collar or the thrust pad will depend on which of these components is the softer. If they are of the same hardness, it may even be necessary to replace both. In practice, it is envisaged that it will be the pad that is more likely to need replacement. This is because, although the tendency is to flatten both surfaces, the thrust-collar surface is already flat to begin with, and it is the flattening of the wedges on the thrust pad that destroys the ability of the bearing to carry a load.

Whereas it has been assumed that the invention will be applied to turbochargers, it may also find application in other environments, which rely on the integrity of a thrust bearing under heavy load. Examples are turbo-expanders, compressors, pumps and wind turbines.