AUSBRID PLUS AND MEGA-TONNE NEW BEARINGS FOR IN- CREASED RELIABILITY AND CAPACITY IN IRON-ORE OPERATIONS

Transcription

1 AUSBRID PLUS AND MEGA-TONNE NEW BEARINGS FOR IN- CREASED RELIABILITY AND CAPACITY IN IRON-ORE OPERATIONS Mark Fetty Principal Product Development Engineer ; Virginia, USA SUMMARY The iron ore railroads in the Pilbara region of Australia run the heaviest axle loads in the world. Today, it is common for short-g bearings to carry axle loads of 40-tonne with peak loads approaching 42 tonnes. These are well above the Association of American Railroad (AAR) load limit of 35.7 tonnes per axle for the standard class G wheelset. To ensure reliability at these loads, bearing usage cycles are limited and in-service bearing performance is closely monitored. With economic goals demanding greater productivity, Pilbara region customers are asking for increased load capacity without sacrificing reliability. The current short-g cannot meet these greater expectations without sacrificing longevity. With this in mind, it is clear that new bearings are required. This paper introduces two solutions offered by Amsted Rail Brenco and compares the performance of each to other international freight applications. It also introduces a new seal design that is standard on these new bearings, a seal that offers significant economic benefits. INTRODUCTION During the 1990s, BHP Billiton (BHP), an iron ore operator in the Pilbara region, was steadily increasing the size of wagon loads as a means to reduce transportation costs i. At this time, they conducted a Failure Mode and Effects Analysis (FMEA) that rated bearing performance as one of the major impediments to increasing axle loads ii. In 1999, this barrier was overcome by the invention of the BRENCO Ausbrid (hybrid class G) bearing. This first short-g size bearing became the solution that allowed BHP to increase loads to 38-tonne per axle (40-tonne peak) while avoiding the excessive fretting that was compromising their bearing clamp. This solution was also economical because it used the same class G raceway components. This allowed the use of the standard class G adapter and eased the logistics of reconditioning iii. It is well understood that as axle loads increase, that bearing fatigue life is reduced exponentially. With very heavy axle loads, Pilbara region railroads often limit bearing usage intervals to a period of less than the wheel life as a means to maintain bearing reliability. Additional increases in axle loading without any changes to the bearings could further reduce the bearing usage interval with an accompanying reduction in the system economics. CURRENT BEARING LIMITATIONS Bearing Fatigue Life, General Bearing fatigue life is predicted using an equation that combines the bearing load rating with the applied load and other factors. The result is a life period that, statistically, 90% of bearings are expected to exceed. This L 10 (or B 10 ) life can be plotted over a range of loads to see the exponential nature of the load/life relationship. Figure 1 shows the generic example of this result for tapered roller bearings using relative loads and relative life. Figure 1: Bearing Life, Generic L 10 Curve Perth, Australia

2 The predicted life of different bearing sizes can be compared by plotting their varying L 10 curves on the same graph. Individual applications can be represented by plotting a point on the corresponding L 10 curve aligned with the applied load. Figure 2 shows the L 10 curves of many of the tapered roller bearings commonly used in international freight wagon applications. Included are the AAR bearing sizes, and the common metric sizes, 130x230 and 150x250. The GG locomotive bearing was included as a reference. All curves represent bearings running fully loaded at all times on appropriate wheel diameters. Note that the class F, class K, and 150x250 bearings all share the same raceway geometry. Therefore, they share the same load capacity, predicted fatigue life L 10 results, and L 10 curve in the graph. Figure 2: Bearing L 10 Life for Various Current Freight Car Bearings In Figure 2, the L 10 curve for each progressively larger bearing is above and/or to the right of the next smaller one. This makes it easy to see the that a larger bearing will have longer life for the same load and/or the higher load capacity for the same expected life. Bearing Fatigue Life, Application Examples AAR wheelset (axle) load limits The blue dots in Figure 2 represent the various AAR wheelset (axle) load limits. More recently, the AAR established bearing load limits based on a 1M mile life (class E = 750k miles). Note that most coal wagons equipped with class-k bearings run loads very close to the AAR limit. However, most class-e and G applications run loads far less than their respective AAR limits. This is because they run in applications where the bearings are sized for rare peak loads. High Reliability Applications o Short G at 40 tonnes per axle. The green dots in Figure 2 represent international freight applications where greater L 10 life is speci- (2) Perth, Australia fied for a resulting higher reliability. This is of benefit in applications where track conditions may be poor, bearing fault detection is less accurate or less frequent, or where system capacity cannot tolerate stoppages. o The 130 x 230 bearing is used in Russia at 23.5 tonnes per axle. o The 150 x 250 bearing is used in Russia and China at 25 tonnes per axle. o A metric equivalent Class-M bearing is being pursued in China to run 30 tonne axle loads. Pilbara Iron Ore Railroad Applications Lastly, the red dots in Figure 2 represent load applications generally used by the Pilbara railroads for the last decade. o Class-K at 35 tonnes per axle o Short G at 38 tonnes per axle.

3 Bearing Limitations in the Pilbara Region It can be concluded from Figure 2 that the Pilbara applications have loads that far exceed the AAR limits for each of the bearing sizes. The result is that their L 10 lives are less than the 1M mile AAR minimum. This indicates the potential for a less reliable application. To offset this limitation, Pilbara region railroads have adopted practices that help to maximize bearing reliability. Several railroads have hot box detector (HBD) systems that can monitor a bearing s temperature station-to-station. Being able to see a bearing s temperature trend over time can provide an early warning of impending failure as compared to a single absolute temperature limit. Several use acoustic detection systems to detect damage in bearings even before heat is detected. Reconditioning standards are typically strict in order to remove any bearings that may be questionable. Lastly, bearings are removed from service based on established time/mileage intervals as to preclude unexpected removals. These time periods are often prior to wheel wear-out. This requires the use of another bearing to complete the wheel life. This bearing and its installation both increase the bearing cost per wheel for the railroad. Bearing manufacturers are always looking for ways to improve bearing life. There is always potential to tweak a parameter or two here or there, and new technologies may also be able to offer some gains, but the results of these evolutionary changes will be incremental, not drastic improvements. Only an upgrade to a larger bearing can move the system to the next level of reliability. For the current Pilbara region class-k application, the investment may be substantial, but there is the opportunity to adopt larger class-m or short-g bearings. For the current short-g application, there are no larger AAR freight car bearings available. These applications are currently limited to the short class-g bearing regardless of any willingness to invest. NEW DEMANDS, NEW OPTIONS A New Push for Increased Axle Loads Since the 90s, new iron ore suppliers have entered the market. Fortescue Metals Group (FMG) began hauling iron ore to port by rail in April of 2008 iv and the first shipments from a new Roy Hill rail operation are expected sometime this year (in 2015 v ). Along with the increased competition, recent economic conditions in the iron ore markets have created a renewed interest in increasing railroad productivity. Since increased axle loads were used to successfully increase productivity in the 90s, this solution is again being explored. FMG started feasibility planning in 2011 to increase axle loads from 40, to 42 tonnes per axle vi. Others are evaluating increases as well. Bearing Options for Increased Axle Loads The demands of higher axle loads were given to the wagon builders who, in turn, inquired the same of the bearing manufacturers. Brenco product engineers looked at many options. Of note: Adopt the Class-GG locomotive bearing? Brenco product engineers first considered adopting the class-gg, their heavy locomotive bearing, to a wagon application. This bearing already had a rated load capacity well above that of the class- G. Several issues however, indicated that this was a less than optimal option. This bearing, designed for the exceptional demands of a locomotive, has a contact angle greater than that of typical freight wagon bearings. This results in a radial /axial capacity ratio, or K- Factor, of 1.69 as compared to the 2.21 of most other freight wagon bearings. This relative shift toward greater axial capacity increases the internal friction inherent in a bearing. This was considered less desireable. The capacity of the GG bearing is great as shown in Figure 2, but with an outside diameter (OD) almost one-inch larger than the class-g, this bearing is large and heavy. The wide gap between the G and GG L 10 curves indicates that other sizes, more efficient and less costly, could be developed. The GG bearing is typically marketed with a 6-7/8- inch inner diameter (ID). This is smaller than the 7 ID of the class-g bearing. The use of a smaller journal was immediately dismissed, as it was unacceptable to gain L 10 life at the expense of increased fretting and a new axle. Increasing the GG to a 7 ID was considered as well, but with the existing raceway geometry, the section thickness of the inner ring, or cone, would become too thin and had the potential to through-harden. This condition, while acceptable for industrial applications, is not considered acceptable for railroad bearings. New Raceway Geometry for the Short-G? From a cost and acceptance perspective, being able to utilize the existing short G journal and adapter was seen as a distinct advantage. Since many railroad bearings adopted their raceway geometry from existing all-purpose bearings, there was the potential for a clean-sheet approach to yield improvements. Different roller diameters, contact angles, and raceway lengths were investigated. In-depth calculations indicated that up to a 5% gain in rated load could be gained with new raceway geometry in the current short-g envelope. New raceway geometry would take a significant investment in time, tooling, development, and test- (3) Perth, Australia

4 ing. Also, it would result in same-size bearing that was not interchangeable with the existing design. While there were real gains to be had, this option was seen as too little gain for the required level of investment. An investment in all new components has the potential for greater gains. Additional Load Capacity for the Ausbrid While studying raceway geometry possibilities, it was seen that the load capacity of the current Ausbrid could be increased relatively easily. One important bearing design parameter is the number of rolling elements (rollers) that can be used in a given raceway volume. Consider an axial section of a bearing through the mid-point of one row of rollers. One can consider the orbital path made by the center of each roller as the bearing pitch circle. See Figure 3. Considering the rollers and their spacing, most of the pitch circle is occupied by the rollers. The remaining space is reserved for use by a cage, also called a roller separator or roller retainer. The ratio of the roller portion, to the total pitch circle circumference is a measurement used to quantify roller density and is estimated as: 1. Adding an additional roller to a bearing raceway reduces the space between rollers and results in less room available for cage bars. With a fixedthickness steel cage, this would result in a weaker cage. In railroad applications, where high impact loads are common, a weaker cage would lead to an increased risk of cage fracture and bearing lockup. However, with a polyamide cage, the molding process allows a thickness change to be made as easily as a width change. Since the Ausbrid uses a polyamide cage as standard, this became a viable option for maintaining cage strength and thus, for increasing bearing capacity. A New Bearing Clean-Sheet Approach With new short-g geometry being too little, and GG geometry being too big, work began on designing a new bearing in between these two contrasting options. The design process began by targeting an increase in rated load of between 20 and 25%. The raceway angle was to maintain a K-Factor close to the 2.21 level. To maintain economic viability, the new bearing would need to fit the existing short-g journal. Finally, the bearing OD would be allowed to grow, but minimally. With these factors in place, new raceway geometry was successfully developed. At 290mm, the resulting cup OD is just over ½ larger than the short-g. It is believed that this can be fit to most pedestals, some with just some minor modifications. TWO OPTIONS SELECTED MEGA-TONNE A Game-Changing Solution With a firm dedication to customers in the Pilbara region, and encouraged by their early feedback, Amsted Rail made the decision to pursue the new, clean-sheet bearing. Dubbed the BRENCO MEGA-TONNE (MT), the resulting bearing represents an increase in capacity that is much more than a small evolutionary step. Figure 4 shows rollers from a class-k, a class-g, the new MT, and a GG bearing. This comparison clearly shows the increased roller size of the MT over the class-g and thus provides an indication of the increase in rated load. Figure 3: Bearing Pitch Circle, Roller Spacing While studying raceway geometry possibilities, it was seen that the current class G cone assembly had one of the lowest roller densities. Adding an additional roller would give it one of the highest. In the past, this would not have been allowed, but today it is a possibility. Figure 4: Rollers for Class-K, G, MT, and GG (4) Perth, Australia

5 Work started on the new MT bearing, but with this ambitious project taking place in a hi-demand market, development of manufacturing capacity trailed the desired scheduling. It became clear that an interim solution was needed; one that required less manufacturing development and that could fit the existing bogie without modification or customer investment. Ausbrid Plus A Short-Term Solution As previously stated, load capacity could be added to the Ausbrid relatively easily and quickly. The addition of a 24 th roller to each raceway represents a 4.3% increase in rollers and thus provides an incremental increase in rated load, a true Plus. While this provided a smaller increase than new raceway geometry would have, it is enough to allow for a small, 5%, increase in load without a significant reduction in reliability. In addition to increased load capacity, this solution has other benefits as well. It utilizes existing raceway components already proven in decades of service. This precludes risk. This also means that the cups and cone assemblies are completely interchangeable with the current Ausbrid. This eases reconditioning practices and precludes the potential for mistakenly mismatched parts. Comparing Options Figure 5 is similar to Figure 2 but it now shows the two new Brenco bearing offerings. It clearly shows the incremental improvement of the Ausbrid Plus and the prominent improvement of the MEGA- TONNE. Also added to Figure 5 are zones that indicate different levels of reliability. These zones are based on L 10 life and bearing usage in various international heavy-haul applications. The green zone represents the high reliability of many metric freight applications. The yellow zone represents high utilization, loading to the typical AAR maximum load limits. The pink zone represents extreme utilization by applications that exceed AAR limits but are successfully being used (i.e., the Pilbara iron-ore applications). The red zone represents overload conditions that often lead to early failure. This leaves a white zone believed to be the best economic balance between utilization and reliability. Figure 5: Bearing L 10 Life for Various Current and Future Freight Car Bearings For 40-tonne axle loads, the Ausbrid Plus moves the application into the Yellow Zone and to a level equivalent to that of the class-k in North America. At 42-tonne axle loads, the Ausbrid Plus is solidly in the Pink Zone and away from the red. The MEGA-TONNE, being the balanced long-term solution, moves both of these load levels completely into the white zone, approximately doubling the L 10 (5) Perth, Australia

6 life. This provides a game-changing solution where not only can the loads be increased to the desired levels, but bearing service intervals can likely be extended to those matching the wheel life. This can potentially provide an additional economic advantage. A NEW SEAL ANOTHER PLUS Over the years, in addition to increasing bearing life and reliability, Brenco product engineers have developed new sealing technologies that lower bearing operating torques and temperatures. These seals have been benefiting customers in the form of reduced set-outs and fuel savings. In fact, The Ausbrid bearing came standard with a BRENCO ST-212 seal, our first low-torque seal without a garter spring. More recently, a new sealing technology has been developed that reduces seal torque to essentially zero while improving moisture exclusion. The new seal, dubbed the BRENCO Tru-Guard, has been successfully running in the Russian market in metric bearings since This seal utilizes the successful labyrinth insert and rotor from the BRENCO Efficiency Plus seal, and replaces the dust guard with a slinger / shield that forms a labyrinth with the seal case. The result, shown in Figure 6, is a seal that retains grease and excludes contaminants without molded rubber lips and the friction they generate. Friction is only produced by the shearing of a small quantity of grease located between the walls of the labyrinth. In addition to running in the Russian market, a class-k version of the Tru-Guard has been conditionally approved by the AAR and is running in North American service. Figure 7 shows a chart of data collected over a 1.5 day period during a seal life test. With above ambient cup temperatures of about 15 F [8 C] and unloaded bearing torques averaging 10 to 11 lb. in., the test result were similar to those of bearings without seals. As a performance benefit, the Tru-Guard seal has been made standard on both the Ausbrid Plus and MEGA-TONNE bearings. Figure 7: The new BRENCO Tru-Guard Figure 6: The new BRENCO Tru-Guard ( Zero-Torque ) Seal The first field testing of a Tru-Guard seal took place on the JSC Railway Research Institute (VNIIZhT) test loop in Shcherbinka, Russia. The bearings were run under load at various speeds. During periodic stops for bearing inspection, the bearings were found to be cool enough to make full prolonged hand contact. BRENCO BEARINGS CURRENT STATUS Ausbrid Plus Status: The Ausbrid Plus bearing, has successfully passed all of the necessary AAR related laboratory tests. In addition to these, the new cage successfully passed tests outlined in sections of European Norm, EN Annex D. A significant number of Ausbrid Plus bearings are currently operating in service. Bearings were mounted in mid-march, 2013 and the wagons entered service in late May. No operational issues have been reported. Wayside data from late December was received that included three wagons running Ausbrid bearings and three wagons running Ausbrid Plus bearings. Since bearing temperature is dependent primarily upon load and speed, the data was parsed to show temperature values for loaded ver- (6) Perth, Australia

7 sus empty conditions (indicated by direction) and temperatures at each individual speed. Figure 8 shows a graph with the results of averaging all values (all eight positions x three wagon sets) for each bearing (seal) type, for each speed in each direction. It is clear that the new bearing and seal provide a reduction in temperature across all speeds both loaded (positive speed) and empty (negative speed). The 3 C average reduction represents 20% of the above-ambient temperature. CONCLUSION Increases to the already high axle loads on Pilbara region iron ore railroads will shorten the overall fatigue life of existing bearings and ultimately result in reduced reliability. The BRENCO Ausbrid Plus, a bearing Amsted Rail is currently supplying to users in the region, has an additional level of capacity that can help to offset the increased loads. This, the ability to interchange all bearing components with existing Ausbrid bearings, and the inclusion of a new, essentially zero-torque, Tru-Guard seal, all combine to make the Ausbrid Plus a great short-term solution. In the past, without the availability of a larger freight wagon bearing, these extreme-utilization applications did not have an option to increase load capacity to any large degree. To provide for this, Amsted Rail has created a completely new, clean-sheet, BRENCO MEGA-TONNE. This bearing has the capacity to essentially double the predicted L 10 fatigue life for a given load. This is done in a bearing that fits the existing short-g axle, and with a cup OD just slightly larger than the short G. When combined with the, zero-torque Tru-Guard seal, this bearing becomes a great long-term solution. REFERENCES Figure 8: Results of Wayside Data - Brenco Ausbrid with ST-212 seals vs. the new Ausbrid Plus with Tru-Guard seals. As of this writing, a teardown and internal bearing inspection is scheduled to take place in February of MEGA TONNE Status: Product and manufacturing development for the MEGA TONNE has taken place and a number of prototype bearings have been produced. The bearing has successfully passed all of the necessary AAR related laboratory tests. As of this writing, an initial sampling of bearings are scheduled to be fitted to wagons for initial field testing. (7) Perth, Australia i Moynan M, Cowin A, Offereins G, Tew G, Axle Loads The BHP Iron Ore Experience. Railway Gazette International, June ii Tew, G., Darby, M., Donnelly, R., Offereins, G., Epp, K., The role of technology in the operations of the BHP Iron Ore railroad, Proceedings of 7th International Heavy Haul Association Conference, Brisbane, June 2001, June, Brisbane Qld Australia. iii Fetty, M., Evolutionary and Revolutionary Changes of the Journal Roller Bearing for Today's Heavy Haul Railroad Markets, Proceedings of 7th International Heavy Haul Association Conference, Brisbane, June 2001, June, Brisbane Qld Australia. iv Wikipedia contributors. Cloudbreak mine [Internet]. Wikipedia, The Free Encyclopedia; 2014 Jun 4, 13:57 UTC [cited 2015 Jan 26]. Available from: ak_mine&oldid= v Wikipedia contributors. Roy Hill [Internet]. Wikipedia, The Free Encyclopedia; 2015 Jan 14, 14:22 UTC [cited 2015 Jan 26]. Available from: oldid= vi Ker, P., Fortescue takes another step in plan to boost Pilbara railway loads, The Sydney Morning Herald, October 24, 2011, Stdney, AUD