1、05FTM20Dual Drive Conveyor Speed ReducerFailure Analysisby: M. Konruff, Rexnord Geared ProductsTECHNICAL PAPERAmerican Gear Manufacturers AssociationDual Drive Conveyor Speed Reducer Failure AnalysisMike Konruff, Rexnord Geared ProductsThe statements and opinions contained herein are those of the au
2、thor and should not be construed as anofficial action or opinion of the American Gear Manufacturers Association.AbstractWith increasing requirements, many conveyor systems utilize dual drive arrangements to increase output.Dual drives can provide an economical solution by utilizing smaller, more eff
3、icient, system designs. However,multiple drive conveyors must proportion the load between drives and load sharing without some type ofcontrol is difficult to achieve. This paper presents a case study on a failure analysis of a coal mine dual driveconveyor system that experienced gear reducer failure
4、s between 2 to 18 months. Physical and metallurgicalinspection of failed gearing did not indicate material or workmanship defects, but indicated overload. In orderto determine the cause of the failures, strain gage load testing was performed. The testing of the conveyordrives revealed load sharing p
5、roblems which that will be reviewed.Copyright 2005American Gear Manufacturers Association500 Montgomery Street, Suite 350Alexandria, Virginia, 22314October, 2005ISBN: 1-55589-868-81Dual Drive Conveyor Speed Reducer Failure AnalysisMike Konruff, P.E., Rexnord Geared ProductsIntroductionWith todays ev
6、er increasing requirements for big-ger and faster equipment, many conveyor systemsutilize dual drive arrangements to increase mineoutput. Dual drives can provide an economical solu-tion by utilizing smaller components and more effi-cient system designs.A downside to multiple drive conveyors is that
7、theconveyor load must be proportioned betweendrives. Load sharing without some type of control isdifficult to achieve. This paper presents a casestudy on a failure analysis of a dual drive conveyorsystem.A coal mine was experiencing speed reducer fail-ures on the dual drive main line conveyors comin
8、gout of the mine. The reducers operated from 2 to 18months before mechanical failure. All failures wereon primary pulley drives. Three of the four failuresoccurred on the same primary pulley. Physical andmetallurgical inspection of failed speed reducersshowed no signs of material or workmanship de-f
9、ects, but indicated that loading was higher than thereducer rating. In order to aid in determining the rootcause of the failures, strain gage load testing wasperformed. Load testing on the conveyor drives re-vealed a load sharing problem which contributed tothe failures.The Conveyor SystemThe subjec
10、t mine had recently replaced its main lineconveyors. Components of the new conveyor sys-tem consisted of both used and new equipment. Allmain line conveyors were dual drive arrangements.Figures 1 and 2 show the conveyor layout of themain line dual drives before the slope belt. Thespeed reducers were
11、 coupling connected to thepulley shaft with grid type couplings. Prime moverswere connected to the reducers with elastomericcouplings. The drive base was fabricated from steelplate and structural steel members. The base waswelded to the pulley support structure. Informationon the speed reducers and
12、prime movers is pro-vided below.Speed ReducersTriple reduction bevel helical models22.84 RatioRPM 1750Service Power 250 HPService Factor 1.36Prime Movers449T Frame TEFCCrusher Duty Design B250 Hp1790 rpm575 Volt, 3 ph / 60 HzNameplate Amps 225Locked Rotor Torque 210%Breakdown Torque 320%Class F insu
13、lation3900 ftBelt DirectionTakeupSecondary PulleyHead PulleyTail PulleyPrimary PulleyFigure 1. Conveyor Layout2Secondary DrivePrimary DriveGridCouplingsMotorMotorPulleysStrain GageTorquemeterStrain GageTorquemeterSpeed sensorSpeedsensorElastomericCouplingElastomericCouplingFigure 2. Drive Arrangemen
14、tTo start the conveyor, both motors are energized atthe same time with a single solid state starter. Theconveyor is brought up to speed, which takes 4 sec-onds with an unloaded belt. A vacuum bypass con-tactor engages after the drives are at full speedwhich puts the motors across the same line andsh
15、uts off the solid state circuit.Mine personnel stated peak mine design capacityon the main line was 1000 ton/hr. Coal travels fromthe mine sections to the slope belt via the main linebelts. All coal exiting the mine is transported on theconveyor system under test. During the measure-ment program, ob
16、served mine output levels (asmeasured by the slope scale) were between 200 to500 tons/hr with occasional instances of 800 tons/hr. Therefore average observed output was lessthan half of the system capacity. Mine output for daytwo was 8800 tons/21 hours of operation, which isfar less than the design
17、capacity. Typical duty cyclefor the conveyor system is about 18 hours. At thetime of testing, one miner was being utilized to minethe section. It was anticipated that after the super-section was started, the mine would be operatingcloser to design capacity.Failure AnalysisThe failure investigator mu
18、st be like a detective andgather as much evidence as possible to determinethe root cause of the failures. Failed reducers werereturned to the manufacturer for analysis. The re-ducers were photographed and bearing settingmeasurements recorded before disassembly. Eachcomponent of the reducer was caref
19、ully inspected,removed and their appearance documented. Hous-ings dimensions were checked on a CMM machineand verified against drawing specifications. Dam-age to the rotating elements was in varied degreesof distress. A common thread was the appearanceof the low speed pinions. The pinions were sever
20、elyspalled which can be attributed to either high loadsor inadequate lubrication. Some of the low speedpinions also had broken teeth. Component ratingspredict the low speed pinion as the limiting compo-nent in the reducer. Also, material removed fromthe pinions during operation caused debris damaget
21、o other components.The OEM claimed that current readings were belownameplate full load rating and that the drives werelightly loaded. The OEM also claimed that the cur-rent between primary and secondary drives wasbalanced by normal standards. Since the informa-tion supplied by the OEM did not agree
22、with theanalysis of the failed drives, further investigationwas required to help with determining the rootcause. A site visit was scheduled and an indepen-dent testing service was hired to measure loads.While the cost of load testing may appear to be high,it will often result in a lower total cost b
23、y enabling theinvestigator to determine the failures root causefaster than other analysis methods. The point atwhen to perform load testing is based on the experi-ence of the investigator.3Before going to the site, it is best to get as manyquestions answered as possible. This will help theinvestigat
24、or prepare for the trip and may provide anarea of focus. Having questions answered beforehand will also save valuable time at the site. Typicalquestions include:What is the duty cycle?Is the conveyor started loaded or unloaded?How many times has it started loaded?Are there operations records?How man
25、y starts per day?What is the starting procedure?What type of controls are used?What are the typical operating conditions (fullyloaded, shock, etc)?Has there been a report of any abnormal oper-ating conditions?What is the maintenance schedule and arethere records?Is it a new application or a retrofit
26、 into usedequipment?Along with other questions dependant on the situa-tion.It is always beneficial to go and see the conveyordrives and observe them under operation. Cluesthat were not evident during the physical analysis onthe failed reducers shed some light on potential con-tributors to the failur
27、es. The reducer attached to thesecondary pulley was not securely fastened to themounting plate. The reducer was acting like a shaftmounted drive. It was orbiting about the pulley shaftwith approximately .06” of movement. The primarydrive also showed signs that it had not been proper-ly mounted to it
28、s base. Shims from under one of thereducer mounting feet were missing. This can leadto a soft foot condition and incorrect gear tooth con-tact patterns due to induced housing deflection. Theprimary pulley had an excessive runout, which ap-peared to be about 1 inch on diameter at the lag-ging. Pulley
29、 runout tolerances for common convey-or applications should be less than .188 for this sizepulley per CEMA. Engineered pulleys with high mo-dulus belts should have less than .03” runout.The scope of the load test was to measure torquelevel and shaft speed on the primary and secondarygear drives on t
30、he #1 conveyor. Torque was mea-sured with strain gages attached to the shaft sur-face. Under torsion, shafts begin to twist. Theamount of twist is dependent on the amount oftorque. Strain gages are thin pieces of metal config-ured so that when the shaft is subject to torsion andits surface deforms,
31、the electrical resistance of thestrain gage will change. The electrical signal pass-ing through the gage is directly correlated to theforce applied.Since the failures were occurring at the low speedpinion, the preferred location for the strain gageswould have been on the low speed shaft. Becausethe
32、low speed shaft did not provide enough room forthe gages, the reducer input shaft was strain gaged.Two gages were place 180 degrees apart. A telem-etry system was used to send the torque signal to asignal conditioner. Shafts were wired for a positiveoutput with clockwise rotation. The torque mea-sur
33、ed on the primary drives speed reducer inputshaft (primary drive torque) is negative on thegraphs as it is spinning counter clockwise. Thetorque measured on the secondary drives speed re-ducer input shaft (secondary drive torque) is posi-tive as it rotates clockwise. To measure speed, fourmagnets we
34、re glued to the motor couplings. By re-cording torque values of both drives at the sametime, load sharing and interactions between thedrives can be observed. Hall effect sensors wereused to couple the speed signals to a frequency toanalog converter.Conveyors experience many different load condi-tion
35、s throughout their daily operation. Unloadedstartups and shutdowns will be significantly differentthan loaded startups and shutdowns. It is beneficialto record as many of these load conditions as pos-sible. To collect data for varied loading, the test pro-gram consisted of recording torque data duri
36、ngstartup, steady operation and shutdowns withloaded and unloaded belts.Analysis of Test DataAnalysis of the test data showed that during startup,the primary drive torque level was substantiallyhigher than the secondary drive with peaks exceed-ing speed reducer catalog rating (See chart 1 and2). Als
37、o, the peak of the primary torque increases asproduct on the belt increases. Chart 2 shows thatthe load on the primary reducer in the time interval of10125 to 10130 has a peak input speed shaft torqueof 1100 lb-ft. Speed reducer rating is 1021 lb-ft.4Chart 1. Reducer torque for unloaded start and st
38、opThe measured speed of the primary motor showsan unusual event on Chart 2 at 10128 sec. Thetorque readings significantly drop during this time.This event is the belt slipping on the primary pulley.The secondary drive also experienced severaltorque reversals, which is evidence that the secon-dary re
39、ducer was being pulled by the primary pulleyand not contributing to the required system torque.Back flank contact on secondary reducer gears fur-ther substantiates this phenomenon.At steady state operating conditions, the primarydrive takes significant more loading than the secon-dary reducer (Refer
40、 to Charts 3 and 4). Under mostloading conditions, the secondary drive was nearzero torque. Recorded current readings on the pri-mary and secondary drive motors show a 3:1 differ-ence and substantiates the load sharing imbalanceindicated by the test results.A good conveyor design will provide accept
41、able per-formance under all operating conditions at a sensi-ble cost. Multiple drives are common for high powerconveyor systems. There are many benefits of us-ing multiple drives over a single drive system. A ma-jor consideration when designing conveyors is belttension. Belt tension is dependent on
42、angle of wraparound the drive pulley and the coefficient of friction.Use of dual drive systems provides more wrap thana single drive and therefore can lead to less maxi-mum tension in the belt. Also, a single high powermotor would require a high in-rush current at start-up. By dividing the power req
43、uirement into two ormore smaller drives, motor starts can be staggeredwhich will reduce the starting current. Smaller com-ponents can also lead a less costly and more effi-cient design.5Chart 2. Torque and speed for partially loaded startA downside to multiple drive conveyors is that theconveyor loa
44、d must be proportioned betweendrives. A common method of load sharing is to haveone drive act as a master and one as a slave. Thetorque output of the slave is controlled by the mas-ter. Load sharing without some type of control is dif-ficult to achieve. Under steady state operating con-ditions, a sm
45、all difference in motor speed creates alarge difference in motor torque. Electric motors arerated to the nearest 5 rpm. Unless motors arematched by the manufacturer, there are no guaran-tees that two “identical” motors will carry the sameload. For direct-coupled motors rated at 1790 rpm,there could
46、be a load mismatch of up to 33%.Another potential load sharing problem is caused byone of the drive pulleys operating on the clean sideof the belt while the other runs on the dirty side. Aspreviously discussed, tension is dependent on thecoefficient of friction which will vary if one of the pul-leys
47、 operates on the dirty side. To get both drives tooperate on the clean side, a snub pulley is used.6Chart 3. 12 hour torque spectrumAlso, during some operating conditions, primary re-ducer loading was at or near reducer rating with re-petitive peaks that caused the reducer to operatewith a service f
48、actor of near unity. Based on data inChart 4, load conditions during time interval of22415 to 24450 has an approximate average inputshaft torque of 940 lb-ft. Speed reducer rating is1021 lb-ft, therefore SF=1.09 during this time peri-od. The load is oscillating at the pulley rotational fre-quency wi
49、th a magnitude of 55 lb-ft. The repeti-tive peaks cause the reducer to operate with aneffective service factor of 1.02.During the test program, the mine was operating atabout 50% of the stated peak mine design capacity.Based on the load data, primary reducer torque in-creased with belt loading. Therefore, it is expectedthat when the mine is operating at capacity, the pri-mary speed reducers will be operating with a ser-vice factor less than unity. AGMA recommendsthat speed reducer selections for heavy duty uni-formly loaded belt conveyors be selected with aminimum service f
