RTDF ROOT CAUSE OF UNDESIRED HOSE SEPARATIONS AND A SOLUTION

Transcription

1 Proceeding of the ASME 29 Rail Tranportation Diviion all Conference RTD29 ctober 19-21, 29, ort Worth, Texa, USA RTD RT CAUSE UNDESIRED HSE SEPARATINS AND A SLUTIN Edgardo Jimenez Product Engineer Strato, Inc. 1 New England Avenue Picataway, New Jerey 8854 Tel: ; ax: ejimenez@tratoinc.com Lin Hua Chief Product Engineer Strato, Inc. 1 New England Avenue Picataway, New Jerey 8854 Tel: ; ax: lhua@tratoinc.com Michelle Monear Manager Mechanical BNS Railway ort Worth, Texa 7616 Tel: michelle.monear@bnf.com Kevin Nilon Sr. Implementation Leader BNS Railway ort Worth, Texa 7616 Tel: Kevin.nilon@bnf.com ABSTRACT Hoe eparation remain a ignificant problem for the cla 1 railroad. They average about ix thouand undeired hoe eparation a year. There are a lot of fact that the railroad are aware of, but lack good explanation. Why do more eparation happen in the winter than in the ummer? Why are old end hoe more prone to eparation? Why doe replacing the gaket more frequently lead to a lower eparation rate? Subtantial reearch ha been done to figure out the root caue of the undeired eparation problem and to anwer the above quetion. The data analyzed from the cla 1 ha hown that the trolley arrangement account for 52% of the undeired eparation. The force to move the end hoe in a cuhioned car trolley arrangement i ometime greater than the force required to pull apart the gladhand. It i probable that mialignment of the gladhand, coupled with an end of car (EC) arrangement that hinder mooth movement of the end hoe, will lead to an undeired hoe eparation. Reearch point to the fact that if there i a way that the force required to pull apart the end hoe connection i kept greater than the force required to move the hackle on a trolley, then thi will lower the frequency of undeired hoe eparation. By changing variou characteritic of the gladhand gaket, it can be enured that the pull-apart force i alway around the 6lb range. Thee change have been incorporated into the deign of the double wide lip gaket and are currently in field teting. By reducing the occurrence of undeired end hoe eparation (UDE), the amount of time a train i topped can be greatly reduced. INTRDUCTIN The main function of the end hoe on a railcar i to provide a method for connecting the brake ytem on two adjacent car, enabling the brake ignal to travel from the locomotive through the entire length of the train. The end hoe ue a pecial fitting called a gladhand to maintain the connection between two adjacent train. The deign of the gladhand and the gaket within it allow the two end hoe to be coupled together eaily, to be able to hold the trainline air preure without leaking, and to allow uncoupling by pulling apart the endhoe with ufficient force. The brake ytem on a railcar i deigned a a fail-afe ytem. Any udden lo of air preure in the main trainline, including a eparation of the train, will caue the brake to be applied, bringing the train to a top. Unfortunately, not all of the eparation of the end hoe are planned. There are intance when the end hoe between car become uncoupled for unknown reaon; thee are known a UDE. Data compiled by the cla 1 railroad how that the trolley arrangement i the mot problematic, accounting for 52% of the undeired hoe eparation. 1 Copyright 29 by ASME

2 Undeired hoe eparation are a major problem for the railroad indutry. Every time a UDE occur, it caue a train to top moving. Thee unplanned toppage reult in ubtantial lo in time from having to locate and re-couple the eparation and wait while rebuilding brake preure. Thi alo reult in additional train delay behind the topped train, potential damage to railcar due to the emergency braking, wheel wear, aociated repair, and potential damage to the commodity being carried. The total etimated cot of thee eparation to the railroad indutry i 15 million dollar a year. Undeired hoe eparation erve to drive down the average velocity of a train. When a UDE hoe eparation occur, it cot the indutry both time and money. BACKGRUND n cuhioned car, the EC arrangement i deigned to let the end hoe move along the centerline of the car, depending on whether the car i in buff or in draft. When the arrangement doe not allow the end hoe to move freely, it reult in a greater pulling force being applied to the end hoe. The force to move the end hoe i ometime greater than the force required to pull apart and eparate the gladhand [1]. Thi i epecially true when the trolley arrangement ha ome defect or damage, whether it i weld platter on the rod from repair, for example, or damage from ervice. It ha been believed that the gladhand gaket alo contribute to UDE. The force required for pull-apart uing tandard gaket in gladhand that are correctly connected range from 25 lb 4 lb (ig. 1) L 3 B S 25 Standard Gaket Cycle Teting (Pull Apart) IGURE 2: RCE VERLAP A tudy wa conducted on the eparation force reulting from mialigned gladhand, and it wa found that they can be le than the force required to move the hackle on a trolley. In turn, thi led to the real poibility that mialigned gladhand can be a caue for UDE. New deign of the gladhand were tudied to find one wherein the gaket doe not roll or twit when coupled. They were alo tudied to enure that the pull-apart force for the end hoe aemblie alway tay well above the maximum force needed to move a trolley arrangement. It wa found that if thee two condition are met, the chance of a UDE i ubtantially reduced. A new gaket wa deigned that prevent mialignment and alo eliminated rolling and twiting in the gaket groove. The new deign i meant to enure that the pull-apart force i alway around the 6 lb. range. Thi new gaket i now undergoing field teting via in-ervice freight car. MISALIGNMENT INDINGS It wa oberved at everal railroad facilitie [1], a well a demontrated in the tet lab, that a gladhand gaket can both roll and twit when two gladhand are coupled while till maintaining preure (ig. 3, 4) 2 15 R C 1 E ITERATIN Serie1 Trendline IGURE 1: STANDARD GASKET CYCLE TESTING However, if the gladhand gaket were mialigned, it wa theorized that the eparation force could be much lower. Gladhand gaket were believed to be able to roll and twit while holding air preure and not maintaining their pull-apart trength. If a trolley arrangement hould be coupled with a gladhand gaket connection that ha not been properly aligned, thi could lead to a ituation where the required force to pull the gladhand apart and to move the trolley overlap, thu reulting in a UDE (ig. 2). IGURES 3 & 4: MISALIGNED GLADHAND GASKETS Lab tet were performed to determine the pull-apart force neceary for gaket that have been mialigned. To determine the force required, a imulation ytem wa ued where one end of a car wa imulated by a fixed frame that followed the Aociation of American Railroad (AAR) S-421 Specification [2]. The et-up had a load cell mounted to the end hoe. The output wa connected to a digital reader that logged the highet force meaured. The adjacent car wa imulated by a moving mechanim that imulate the tructure of the AAR S- 427 trolley pecification [3]. At the bottom part of the ytem two aluminum channel were poitioned to imulate the rail. The height of all the component related to top-of-rail were 2 Copyright 29 by ASME

3 meaured from the top urface of the channel. All dimenion and part were deigned and intalled to follow AAR requirement S-427 [3]. A hydraulic cylinder wa applied to drive the moving mechanim. The eparation peed of the two hoe wa et at about 1 mile/hour. The tet wa performed with a peron that wa inexperienced with coupling together end hoe in order to repeat the chance of a mialigned gladhand in the real cenario. The end hoe were coupled together and eparated a total of 1 time. Each end hoe conited of a gladhand, gaket, male pipe fitting, two ferrule and a hoe blank. The hoe blank met the AAR M- 61 pecification [4]. The end hoe aemblie were from tandard inventory. Each time the gladhand were eparated, the maximum force required to pull apart the gladhand wa recorded. It i evident from the tet reult (ig.5) that there are many intance where the force required to eparate the gladhand wa ubtantially lower than the force that i commonly aumed to be needed to eparate end hoe. L B S R C E Standard Gaket Cycle Teting (Pull Apart, Inexperienced Carman) ITERATIN IGURE 5: MISALIGNED GASKET CYCLE TESTING The data gathered during thi experiment confirm that the force required to pull apart end hoe i ignificantly lowered when the gaket are mialigned. Serie1 Trendline IGURE 6: STANDARD GASKET VS. DUBLE WIDE GASKET The double wide gaket ha a mating idewall thickne of.312, compared with the current idewall thickne of.156. The bigger idewall give the gaket added trength which prevent it from rolling and twiting in the gaket groove. It incorporate much more material than the current deign, thu making it harder to roll and deform in the gaket groove. The double wide gaket ha a urface area of 1.47 in 2, compared to the.659 in 2 urface area of the current deign. The increae in urface area prevent the gaket from being mialigned, ince the contact area i ignificantly greater than the amount the gaket can deform in the gaket groove. The greater urface area, coupled with a lightly lower durometer rubber, alo reult in greater adheive friction between mating gaket, helping them tay together. The inner diameter of the double wide gaket ha not been changed; it i till 1 3/16 a noted in AAR requirement M-62 [5] to allow for normal air flow in the brake ytem. All of thee factor hould enure that the pullapart force for an end hoe equipped with the double wide gaket i alway around the 6lb range and never drop ubtantially, even with an inexperienced peron coupling them. Data that wa gathered in a recent tudy [1] point to the fact that if there wa a way that the force required to pull apart the end hoe wa kept ubtantially greater than the force required to move the hackle on a trolley, thi would lower the frequency of UDE. By enuring that the pull-apart force alway tay around the 6 lb. range, it i believed that thi will ignificantly diminih the amount of UDE. PRPSED SLUTIN Gladhand gaket can become mialigned eaily with current dimenion. Thi i primarily due to the geometry of the gaket [5]. The current gaket can eaily be moved and hifted inide the gaket groove. The propoed olution i to double the width of the gaket idewall thickne (ig.6) thu increaing it reitance to deformation when coupling two gladhand. Thi will alo prevent it from being mialigned due to the larger urface area. LABRATRY TESTING All of the teting wa performed uing the ame equipment a denoted previouly in Mialignment inding. The teting wa et to run with block of 1 iteration each and everal tet performed. The firt tet ued two hoe aemblie with the double wide gaket, being coupled by an inexperienced peron. Thi peron wa not the ame peron that coupled the tandard gaket. When comparing the reult of thi tet with the tandard gaket reult (ig.7), it i evident that the new deign maintain the force needed to uncouple the hoe at a afe level. 3 Copyright 29 by ASME

4 STANDARD GASKET VENDR 2 DUBLE WID E DUBLE WID E VS STANDARD L 7 b 6 f 5 4 o r 3 c e 2 Gladhand Gaket Comparion, Inexperienced Carman STANDARD GASKET DUBLE WIDE GASKET Inexperienced Carman Inexperienced Carman tet wa that when the tandard gaket wa mialigned with the double wide gaket, it would elf-center and correct itelf when the hoe were preurized. Thi combination did not reult in any mialigned gaket. And the eparation force wa conitently around the 5 lb. range. The next erie of tet were a comparative of tandard gaket from two different vendor, the double wide & tandard combination, and two double wide aemblie (ig.9). All of the tet were repeated to verify their conitency Iteration IGURE 7: STANDARD VS. DUBLE WIDE, PULL APART RCE Gladhand Gaket Comparion, Pull Apart Even with an inexperienced peron, there were no intance where the hoe were coupled incorrectly with the double wide deign. Even when the gaket were lightly off center, there wa till enough material making contact to keep the gaket properly aligned, and to provide an adequate eal. There wa no added difficulty in coupling the double wide gaket when compared to the tandard gaket. The elimination of the mialigned gaket, coupled with the improved geometry of the new gaket, reulted in a eparation force that wa conitently around the 6-7 lb. range. The next tet conited of the double wide gaket remaining on the moving mechanim, and the fixed hoe wa replaced with a tandard gaket aembly. Thi wa to imulate the behavior of thee two diimilar hoe being coupled together in the field (ig.8). 7 6 L B 5 S 4 3 R 2 C E 1 DUBLE WIDE & STANDARD GASKET CYCLE TESTING (PULL APART) ITERATIN IGURE 8: DUBLE WIDE & STANDARD, PULL APART RCE When a tandard gaket i coupled with a double wide gaket, the force i conitently around the 5 lb. range. Even though the pull-apart force wa lower than two double wide gaket, it wa till ubtantially higher than the force required to moving a trolley arrangement. The behavior noted from the Serie1 L b f o r c e Iteration IGURE 9: GASKET CMPARATIVE, PULL APART RCE The gaket comparative excluded any mialigned gladhand. It wa done trictly to compare the pull-apart force with good connection. Thi demontrate the difference between the type of gaket connection. The higher value are directly linked to the double wide gaket. The tandard gaket, even from different vendor, exhibited force threhold that were lower than what wa previouly thought. It ha been hown that when gaket are mialigned, the force required to eparate the end hoe i ubtantially reduced. Introducing the double wide gaket, even when coupled with a tandard gaket aembly, greatly increae the pull-apart force required. It alo eliminate the rik of a mialigned gaket. In all of the teting that wa performed, the double wide gaket ha never created a mialignment. When two double wide gaket are coupled together, the gaket do not mialign. The force to eparate them i ubtantially higher than a regular gaket and thi i reflected in the tet data. Whenever the double wide gaket wa ued in teting, the pull off force required to eparate the hoe never dropped below 4 lb. The idewall of the double wide gaket i much tiffer than a tandard gaket and thi help it reit the rolling and bending motion that caue a regular gaket to mialign. The double wide gaket alo allow for increaed frictional force between the gaket urface when eparated, due to the increae in urface area. Normal frictional 4 Copyright 29 by ASME

5 calculation ignore the role that urface area play. Thi i due to the fact that the area of coverage of a liding block i actually much larger than the area of contact. In rubber, however, it low elatic modulu reult in a much greater area of contact at it interface [6], thu urface area play a very big role. The increaed urface area of the gaket will reult in greater frictional force between the gaket when two end hoe are coupled together. The greater adheive friction help raie the force required to pull apart the hoe. Raiing the force required to eparate the end hoe i very important, thi will prevent the force required from going into the range of force required to move the trolley. Data that wa gathered in field teting how that the force to move the hackle on a trolley arrangement varie between < 1lb. to 63 lb., with an average of 19 lb [1]. Thi data wa taken and then 1, intance were randomly generated that matched the logitic ditribution of the field data. They alo had the ame parameter a the field data. The ame random generation wa performed uing the data from tandard gaket eparation, a well a double width gaket eparation. By analyzing the trolley movement data together with the gaket eparation data, there i a clear overlap where the force required for both are the ame. It i believed by the author that thi overlap i a caue of UDE. Uing the double wide gaket raie the pull-apart force well beyond the range that i neceary to move the trolley arrangement (ig.1). I n t a n c e Intance Standard Gaket, Pull Apart orce Intance Double Width Gaket, Pull Apart orce Intance Trolley Pull orce orce Range, Lb. IGURE 1: INSTANCES SEPARATIN RCE REQUIRED The data repreented in figure 1 i the count of how many time the force required to perform the pecified action fell in a given range. The reaon that the trolley pull data i piked i becaue the force required to move the trolley fall in a very narrow range, from 7 lb, while the force to eparate end hoe ha a much broader range. The data for the tandard end hoe form a double bell curve becaue there are two ditribution aociated with it. The lower force range i aociated with intance of mialigned gladhand. The econd bell curve reult from the force ditribution aociated with correctly connected gladhand. The data for the double wide gaket end hoe doe not form a double bell curve, becaue there are no intance of mialigned gladhand, or any other condition that would weaken the gladhand connection, occurring. The force required to eparate the tandard gaket overlap with the force to move the trolley in the 3 7 lb range. In the random imulation, there were a total of 1 intance occurring during the imulation for an average of 1 occurrence per 1 intance. Thi data matche up with the percentage of intance that are happening in the indutry [1]. Thi data confirm that the problem lie within the gaket connection. The overlap between the tandard gaket and the trolley movement force i very problematic. Raiing the eparation force beyond the range of the force required to move the trolley arrangement will greatly reduce the amount of undeired hoe eparation, a upported by teting. IELD TESTING The double wide lip gaket ha been well received by the indutry. The repreentative of TTCI, a few cla 1 railroad, and Strato met in Strato facility in June and dicued the undeired hoe eparation iue and poible olution. After the demontration of the mialigned gladhand connection and ome tet, the group believed that thi wa the main reaon behind the hoe eparation problem. Many of the fact that the railroad are aware of, but lack explanation, can be anwered by thi finding. 1. Why do trolley arrangement have a much higher eparation rate than fixed bracket? Becaue the end hoe on a trolley arrangement i puhed or pulled to follow the cuhion unit travel, but the end hoe attached to a fixed bracket ha almot no force applied to it. 2. Why do more eparation happen in the winter than in the ummer? Becaue the end hoe and intermediate hoe are tiffer under low temperature o a higher force i applied to move them, which increae the chance of a eparation. 3. Why i it eaier to have eparation on old endhoe? Becaue the more wear the gaket and gladhand have, the eaier it i to mialign the connection between to end hoe. Thi i alo why replacing the gaket more often can help the eparation rate. 4. Why can a eparation not be intentionally repeated on the ame car? Becaue, according to figure 5 & 7, the chance i 1-15% that a bad connection will reult for an inexperienced peron, i.e. 85-9% of the time, a bad connection will not be repeated. The deign of the double wide lip uccefully eliminated the chance of a mialignment. All the problem mentioned above, therefore, can and will be olved. The deign ha been reviewed by the AAR Brake Sytem Committee (BSC) and a tet allotment ha been aigned. There are currently a total of 1 rail car, a combination of bi-level and tri-level auto rack, on field tet. They originated in Portland, regon, where they were inpected and equipped with the double wide gaket and gladhand between the date of 4/3/9 and 5/15/9/. The car number have been recorded and the number of UDE they experience i being tracked. The propoed plan wa to keep a many vehicle running in group of 3 between Portland, regon (PNW) and Logitic Park in Chicago, IL. for one full year. Thi would allow the difference to be noted in variou weather condition. The car were only able to be kept together for a few cycle 5 Copyright 29 by ASME

6 between the two point a decribed in the plan. Since thi time, the vehicle have been interchanged off-line to other railroad and are now in ervice and mixed with other vehicle equipped with the tandard gaket. A of thi point, none of the vehicle equipped with the double wide gaket have been involved in a UDE. There have been three car that required air hoe upport adjutment. There have been no other finding at thi time. The vehicle will continue to be monitored over the next nine month and more data will be gathered. urther dicuion on thi topic will be taking place at the AAR Brake Sytem Committee meeting that i cheduled for the week of Augut 24 th. DISCUSSINS & CNCLUSINS After analyzing data upplied by the cla 1 railroad, it wa dicovered that the trolley arrangement wa reponible for the majority of the UDE. Extenive reearch wa performed to dicover the root caue of the UDE. The data gathered pointed to the fact that the force required to eparate the end hoe can vary greatly. It wa alway aumed that thi force wa around the 4-6 lb range. Teting dicovered that thi force wa eldom thi high. The real range i from 25 to 4 lb. And it can drop well below thi range. The root caue of thi variance wa found out to be mialigned gladhand gaket. Mialigned gaket can drop the force required for pull apart from the 25-4 lb range down to a low a 3lb. Thi dicovery led to the theory that mialigned gaket led to the UDE. The teting performed on the trolley movement force howed that the force to move a trolley can vary from - 63 lb. Thi teting did not even take into account trolley arrangement that were bent, had weld platter on the rod, or otherwie were not in good working condition [1]. The ummation of all the teting i that there i a range, from 3 lb. to 63 lb., that the force required to move the trolley arrangement could be greater than the force required to eparate the two end hoe. In the field, thi could lead to an UDE. The olution to thi problem wa to keep the force required to eparate the end hoe well above the force required to move the trolley arrangement. The double wide gaket remove the poibility of a mialigned gaket, and thu alway maintain the force required to eparate the end hoe to a level well above the trolley force. ield teting thu far ha hown that, when uing the double wide gaket, the number of UDE i greatly reduced. Thi validate the theory of the primary caue of UDE. By ignificantly lowering the occurrence of UDE the incidence of train toppage will be reduced, thu driving average velocity, and increaing the efficiency of the railroad. [3]AAR Manual of Standard and Recommended Practice, Section E, Standard S-427, Adopted 1969, Revied 25, pg. E- 243 [4]AAR Manual of Standard and Recommended Practice, Section E, Standard M-61, Adopted 193, Revied 26, p.p. E-22 E-24 [5]AAR Manual of Standard and Recommended Practice, Section E, Standard M-62, Adopted 1913, Revied 22, p.p. E-22 E-24 [6] Preon, B.N.J., 21, Theory of Rubber riction and Contact Mechanic, J. Chem. Phy. p.p REERENCES [1]Nilon, K. 29, Air Hoe Separation, ort Worth, TX, BNS Railway HQ [2]AAR Manual of Standard and Recommended Practice, Section E, Standard S-421, Adopted 1999, Revied 22, 25, 26, Copyright 29 by ASME