As a result of our evaluation of the pit type defects observed on all of the Cameron Pinion Shaft teeth, DST offers the following proposed mechanism of failure.
Failure Mechanism
a. The semi-circular rough, somewhat shiny surface pit type defect present on each Pinion tooth was the result of cavitation erosion. Cavitation is defined as the formation and rapid collapse within a liquid of bubbles that contain vapor or gas or both. It occurs where the local hydrostatic pressure becomes lower than the liquid’s vapor pressure. Vapor pressure is defined as that pressure exerted by the gaseous state of a liquid that is in equilibrium with its liquid phase.
b. The most common explanation of the phenomenon is the destruction of material by violent highly localized shock waves originated from the implosion of bubbles, when passing through a pressure differential zone. Data in the literature denotes that the bubbles are microscopic but countless and that the erosion observed is caused by the effect of a ‘zillion’ pulse loads on the surface. The process is not instantaneous but occurs over a period of time. Cavitation caused by severe turbulent flow often results in the degradation of the surface and can result in the loss of material, surface deformation, and changes in properties and appearance. In regards to damage of a gear tooth surface, the mechanism is the result of the nucleation and implosion of bubbles in the lubricating fluid. From a microscopic standpoint, the craters that form are deep, rough, and clean and have a honeycomb appearance. Damage rate and erosion severity is considered dependent on the liquid properties and the time period over which it occurs.
c. Characteristics of the defects observed on the Pinion helical teeth showed evidence of erosion of the coast side tooth tip and flank. It was interesting to note that the erosion of each tooth was at the same location and about the same size on the tooth. Deep cavities formed during service life in essentially the center area of the eroded zone. Macroscopic examination of each cavity examined disclosed a relatively clean, rough pit, which had honeycomb type features. These features were confirmed by microscopic analysis of cross sections through the deep pits.
d. Close examination of the defect features revealed that erosion of the surface first initiated, most likely due to the impingement and implosion of bubbles in the cavitating fluid on the gear tooth surface. Each defect was located within 0.5 inches of one end of the helical pinion tooth. Erosion was observed on the coast side of the tooth tip and extended down into the addendum portion of the coast side of the tooth. Continued operation of the system resulted in the formation of deep cavities. Small, incipient cracks were observed to extend from the base of these cavities. These cracks, with further service time, eventually became part of the original cavity and cause a deepening of it. It is also noted that erosion features were observed in the roots between several teeth. The erosion characteristics in these areas were not as severe as that on the tooth tip and flank; however two locations were observed to be eroded. One area was relatively close and in line with the large defect while the other area was located at the end of the root. This latter eroded area could be seen by looking at the end of the teeth with a microscope. Continued operation would have most likely resulted in the cavity extending completely through the center portion of the tooth. The deepest cavity observed in this study was equal to about one third, 1/3 of the Pinion tooth height.
e. Cavitation erosion failures observed on the Pinion teeth is believed to be related to sudden changes in the liquid (oil) pressure due to changes in the liquid (oil) velocity. Assisting the mechanism could be some form of contamination present in the oil system. This could be steel particles left from the machining operation, the presence of other manufacturing debris in the oil, water in the system and/or the surface hardness of the nitrided gear being somewhat lower than necessary.
f. No evidence of rehardening of the surface in the eroded/pitted areas was observed. This indicated that the defects were not related to electrical current damage. The lack of rehardening and/or any evidence of molten/frozen material in the defect areas helped to confirm the failure mode, i.e., cavitation erosion as opposed to a form of electrical discharge.
Comments
DST offers the following comments in regards to this failure:
a. Based on the fractographic examination of the failed component, it is apparent that the all of the defects initiated due to cavitation erosion of the gear tooth surface. This mechanism is not normally associated with gear tooth problems; however, DST has observed this mechanism on a few other parts. Data in the literature indicate that gear teeth have been affected by this mechanism and that the use of hard materials, such as tool steels, e.g. stellite, have been used to mitigate the problem. In contrast, it is noted that electrical discharge problems are not typically observed on gear teeth but rather are most prevalent in bearing applications. In these latter situations, the metal surface appears etched, while in the former, the metal surface exhibits erosion of the surface and deep rough pits.
b. In order to minimize cavitation erosion problems, the major options are related to the lubricant system utilized. These include (1) increasing the supply oil pressure, (2) changing to a higher viscosity oil, (3) re-evaluate the oil feeding configuration, and (4) design changes to blend edges or contours areas to promote streamline flow. Another option is to change to a harder more resistant material. In the present case, this could involve attaining a higher surface hardness on the 4340 steel than presently has been attained or the use of a carburized and hardened component.
c. In order to minimize the erosion characteristics of the lubricant, the oil system should be reviewed and analyzed for contamination. Contamination of the system could be from inadequate removal of machining particles prior to installation of the part and/or inadequate flushing of the gearbox after complete assembly. Periodic examination of the oil during service life for metal contamination could also be considered. Water in the lubricant should also be monitored and controlled. In this regard, a review and analysis of the oil used in the system in which the LS Pinion operated should be accomplished. Similarly, the oil presently being used should be similarly evaluated.
d. It is probable that there may have been other parts in this system that may also show evidence of cavitation erosion. For example, review of the HS Pinion used in the gearbox also appears to show possible similar damage, albeit to a much lesser degree. This component, shown in Figure 1d of this DST report shows what may be evidence of the failure phenomena beginning on this part. Arrows point out the location of the defect area, which seems to be similar to the LS Pinion evaluated herein. This component should be visually examined for evidence of cavitation erosion. Close examination of other parts in this gearbox should also be reviewed.
e. In regards to the heat treat processing of this part, AISI 4340 steel is an acceptable nitriding material. However, as observed in this report, the surface hardness of the teeth was at the minimum requirement of R/C 47. When considering the nature of cavitation erosion and as stated above, a harder material is less affected by this failure mode than a softer material. It is commonly known that the core strength of 4340 steel can be controlled by both carbon level of the material and the tempering temperature used. Based on our microhardness studies, it is apparent that the hardness of the material was at a minimum. DST would recommend that the heat treatment of this material, prior to nitriding should be optimized to produce core hardness values in the R/C 36 – 40 range. This is important because the nitride surface hardness is a function of the core hardness of the material. The higher the core hardness, the higher the nitrided layer surface hardness. Typical surface hardness range for nitrided 4340 steel should be in the R/C 55 – R/C 58 range. Surface hardness values in this range could have a positive effect in preventing cavitation erosion failures.
f. The type and heat treat level of the material utilized to make this pinion component was not given. In addition the engineering drawing for the LS Pinion was not available for DST review. We recommended that the engineering drawing(s) for all critical components that are being utilized in critical gearboxes be obtained but this is often difficult to accomplish because of proprietary data restrictions. In addition, metallurgical data such as the material quality level, surface and core hardness requirements and the case depth details if the part is surface hardened should be obtained from the OEM but for the reasons noted above this is sometimes not practical. In addition, the method to determine the required case depth (particularly for nitriding – since various requirements are possible) along with certain microstructural requirements should be given and shown to have been met. In regards to nitriding, whether or not the white layer is permitted or not should be detailed.
g. As indicated above, using a material with a harder surface can have a significant improvement to cavitation erosion issues. For the LS Pinion application, consideration should be given to the use of a carburized and hardened material. Carburizing a steel alloy such as 17crNiMo6 will insure that the surface hardness will be in the R/C 58 – 62 range, as compared to the hardness requirement for nitrided through hardened steel of R/C 47 – 55.
h. The axial velocity of this gear mesh is 82,731 FPM. At this velocity, the movement of the air/oil mixture trapped in the tooth mash will cause the gear tooth temperature to be non-uniform across the face width. As the axial meshing velocity (axial velocity of the contact lines and thus the trapped air/oil mixture flow) 100,000 FPM differential tooth surface heating will occur across the face width. The temperature at the mesh exit side of the face width will be greater than that on the mesh entering side of the face width. The example shown in Figure 22 (drawn from DST files, for illustrative purposed only), shows this effect clearly for a somewhat similar gear set. In such cases, the design of the lubrication system including jet location, pressure, flow volume, oil type and temperature, is very critical. While we see no indication of an extreme temperature condition as the primary causative agent in our evaluation of the subject pinion, higher temperatures tend to foster the erosion type failures we have identified on the subject pinion. The mechanism by which this happens is this. The vapor pressure of a lubricating oil (or any other liquid) increases exponentially with increasing lubricant temperature. The atmospheric pressure boiling point of lubricating oil is the temperature at which the vapor pressure equals the ambient atmospheric pressure (within the gearbox near the gear teeth. As the lubricant temperature continues to increase it may reach a critical point at which the vapor pressure becomes sufficient to overcome the local atmospheric pressure allowing vapor bubbles to form inside the bulk of the substance. Cavitation occurs when a fluid’s operational pressure drops below its vapor pressure causing gas pockets and bubbles to form and collapse, as detailed above. This can occur in what can be a rather explosive and dramatic fashion.
In view of this connection, a thorough review of the lubrication system used in this gearbox should be implemented to ascertain what improvements can be made, especially in terms of lubricant delivery, to minimize the deleterious effects of high axial velocities. Additionally, since increasing the helix angle reduces the axial velocity, a study should be implemented to optimize the overall gear design without prior restrictions.
i. This optimization should include an evaluation of the use of double helical gears with higher helix angles and a smaller face width on each helix. Double helical gears will also provide a theoretically zero thrust load and an apex gap that also allows the “hot” axial flow constituents to exit the tooth space earlier in the mesh cycle. In addition, the use of thrust collars creates an effective “dam” right up against the ends of the helical teeth (shown in Figure 23) further limiting both the exit of the high axial velocity oil/air mixture. This factor can also be addressed through the use of double helical gears.
While not related to the subject failure in any way, as we examined the pinion teeth carefully we noted another condition worthy of note. As Figure 24 shows there appears to be a slight step at the root of the pinion tooth where the involute tooth profile joins the trochoidal fillet curve. This is condition occurs when the teeth are flank only ground after initial tooth cutting. While not contributory to the damage which is the subject of this report, this condition significantly reduces the bending strength of the pinion teeth. Requirements for future gear procurements should address this issue by adding a requirement to indicate that a slight undercut is permitted (Figure 25a) but a step at the root (Figure 25b and C) is not permitted when gears are finished by flank only grinding. This condition does not generally occur when teeth are finished by full form grinding but a similar notation is always good insurance.
Conclusions
Based on our evaluation of the failed LS Pinion, we reached the following specific conclusions:
a. Metallurgical evaluation of the singular defects on each tooth of the Pinion revealed the surface damage initiated due to a cavitation erosion mechanism. Erosion was caused by bubble implosion in the cavitating liquid (i.e., lubrication fluid) impinging on the coast flank, top land and root areas of the teeth. Continued operation resulted in deep cavernous pits which extended down into the tooth cross section. Cavitation erosion was characterized by the erosion of the system being somewhat sponge like and the deep rough pits that exhibited a honeycomb characteristic appearance.
b. The location of the cavitation erosion pits being on the coast side tooth tip and addendum and in the root area contiguous with the deep pitting suggests a somewhat concentrated lubrication impingement system on the Pinion teeth coast side tip corners. This should be reviewed and altered appropriately. This concentration of lubricant impingement on the coast side of the teeth most likely contributed to the cavitation erosion characteristics observed on each tooth.
c. The low surface hardness of the Pinion teeth obtained by nitriding the 4340 component was found to be R/C 47. Low surface hardness values most likely contributed to the cavitation erosion mechanism observed on this part. Higher hardness materials more effectively resist erosion.
d. Evaluation of the LS Pinion for certain metallurgical characteristics revealed the following:
1. The surface and core hardness values obtained were considered typical for nitride 4340 steel. However, it was noted that the surface hardness requirement was at the minimum acceptable range.
2. Evaluation of the nitride case depth was found to be within the requirements given to DST.
3. Chemical analysis indicated the material to be typical of vacuum degassed quality 4340 steel.
4. Chemical etching of test samples for microstructural constituents indicated the carburized case to consist primarily of tempered martensite and some nitrided nitrates. The core microstructure consisted primarily of tempered martensite, which is typical of heat treated 4340 steel.
5. No evidence of white layer was observed on any of the tooth profile surfaces. Nitride white layer was observed in the root areas of the teeth. The depth of this constituent was observed to be less than 0.001 inch.
e. The high axial velocity of this gear mesh combined with the damming effect of the thrust collars contribute to conditions that foster the type of cavitation erosion observed here.
f. Based only on our evaluation of the damage observed, there is no indication that the inherent load capacity of the gear set is in question. However, we have not accomplished an independent load capacity rating of the gear set and the operating life thus far is relatively short thus any presumption that the gears are adequately rated for this application would be premature. In addition, the presence of a step at the fillet/profile junction raises some question concerning the actual reduction in bending fatigue that may accrue due to this issue.
Recommendations
As a result of our evaluation and brief literature study, we developed the following recommendations:
a. Review of the compressor lubrication system in which the LS Pinion operates should be accomplished. Specifically the following should be considered:
1. Avoid restrictions and obstructions to lubricant flow
2. Reduce vibration, flow velocities and pressures
3. Minimize sudden changes in liquid pressure due to changes in liquid velocity or to shape or motion of parts
4. Avoid low vapor pressure, aerated wet oils
5. Use non-corrosive oils
6. Consider the use of additives, cavitation inhibitors, which are specifically related to minimizing cavitation erosion issues.
b. In addition to the above, the following should also be considered in order to minimize cavitation erosion problems:
1. Increasing the supply oil pressure
2. Changing to a higher viscosity oil
3. Re-evaluate the oil feeding configuration
4. Design changes to blend edges or contours areas to promote streamline flow.
c. In order to minimize the erosion characteristics of the lubricant, the oil system should be reviewed and monitored for metal and other forms of contamination. Debris of the system could be related to inadequate cleaning of machining particles on the part prior to installation in the gearbox and/or inadequate flushing of the gearbox after complete assembly. Both size and type of contaminant should be determined. Periodic examination of the oil could also be considered. Water in the lubricant should be monitored and controlled.
d. It is also recommended to review and obtain a contamination analysis of the oil in the compressor system presently being used with the replacement LS Pinion. Particle type and size, if present, should be determined.
e. Review of the heat treat processing of the 4340 steel using in the manufacture the LS Pinion. This should be done to assure that core hardness properties of the material are optimized such that maximum case hardness values after nitriding are obtained. This is because the hardness of the nitrided surface/case is modified appreciably by core hardness. For nitriding purposes, 4340 steel is usually provided with maximum core hardness by tempering at the minimum allowable tempering temperature.
f. Since surface hardness of the component improves its resistance to cavitation erosion type failures, we recommend that consideration should be given to the use of a carburized and hardened steel component in this application. Both surface hardness and case depth can be significantly increased by using a material such as 17crNiMo6 in the LS Pinion application.
g. As indicated herein, other parts in the gearbox compressor system may show beginning evidence of cavitation erosion. One such part is the HS Pinion shown in Figure 1d (seen in previous column). We recommended that this component be visually examined in the near future for evidence of cavitation erosion. Close examination of other parts in this gearbox should also be reviewed.
h. Serious consideration should be given to reviewing the basic gear system design in an effort to develop a new approach that minimizes conditions that tend to produce cavitation erosion. This review should be undertaken without prejudicial preconditions so that all possible configurations can be adequately considered.
i. Although the failure observed is not related to the basic inherent load capacity of the gear set, it would be prudent to conduct a detailed, independent load capacity rating1) of this gear set, including the effect of the step at the fillet/profile junction, in order to ascertain its suitability for long life in this application.
Brief Description of the Cavitation Mechanism
Liquid, in this case the lubricating oil is the medium that causes cavitation wear. Unlike adhesive wear, which involves the interaction of two surfaces (typically the mating pinion and gear tooth flanks for a gear set) of three body wear which involves the interaction of these same two gear tooth surfaces and an abrasive contaminant particle which is trapped in the mesh, cavitation wear occurs on a single surface and does involve a second surface. It requires only that high relative motion exists between the surface and the fluid. Such motion reduces the local pressure in the fluid. When the liquid reaches its boiling point and ebullition occurs, vapor bubbles form, which produce cavitation, as shown schematically in Figure 26a. Each vapor cavity lasts a short time because almost any increase in pressure causes the vapor in the bubble to condense instantaneously and the bubble to collapse and produce a shock wave, as shown in Figure 26b. This shock wave then impinges on adjacent metal surfaces and destroys the material bonds (Figure 26b and Figure 26c show this process). The shock wave first produces a compressive stress on the solid surface, and then when it is reflected, produces a tensile stress that is normal to the surface. This stress reversal process results in the characteristic “spongy” appearance of a cavitation failure, as shown enlarged in Figure 26d.
Entrained air and dust particles in the fluid serve as nucleation sites for the formation of vapor cavities. These nuclei can be small gas-filled pockets in the crevices of the container or simply gas pockets on contaminant particles moving freely in the flow stream. Therefore, all confined fluids may contain sufficient impurities to produce cavitation. In a gear system lubricating oil system entrained air from excessive churning of the lubricant, small amounts of water contamination, among many others can precipitate this process. While the fluid flow velocity in a typical gearbox lubrication system is generally low (by fluid dynamic standards), the combination of a tightly constrained flow region (the small space between the coast flanks of the pinion and gear created by the backlash in the system), the high axial velocity of the air/oil mixture in this space and the damming effect of the thrust rings on the end of the pinion face create a condition which can foster this cavitation mechanism.
Small voids near the surface or flow field, where minimum pressure exists, indicate that cavitation has begun. The action of centrifugal force and the damming effect of the thrust rings cause the flow to turn at right angles at the exit end of the face width. Because of this cavitation damage is less in the root area, as indicated by the right most red arrow in Figure 6C, for example, than it is at the tip of the tooth where velocity the cavitation effect is greatest. Once initiated, bubbles continue to grow as long as they remain in low-pressure regions. As the bubbles travel into high-pressure regions, they collapse, producing intense pressures and eroding any solid surfaces in the vicinity. This is readily apparent in Figure 6C in the region indicated by the blue arrow. During the collapse, particles of liquid surrounding the bubble quickly move to its center. Kinetic energy from these particles creates local water hammers of high intensity (shock), which grow as the front progresses toward the center of the bubble. This is the mechanism that produces the damage observed.
Closing Comment
It should be apparent that the cavitation erosion failure mode identified here is far from “common.” In fact, while we have seen this mode several times before, even in our combined more than 100 years of gear system experience we have seen this mode just a few times. For this reason, we believe that it is important to provide as complete and detailed description of the phenomena as possible so that it can be recognized more easily by other investigators.
** Printed with permission of the copyright holder, the American Gear Manufacturers Association, 1001 N. Fairfax Street, Suite 500, Alexandria, Virginia 22314. Statements presented in this paper are those of the authors and may not represent the position or opinion of the AMERICAN GEAR MANUFACTURERS ASSOCIATION.
