DESIGN OF A FACILITY FOR THE PRECISE SIMULTANEOUS GENERATION AND MEASUREMENT OF FORCE AND TORQUE

Transcription

1 XX IMEKO Wold Congess Metology fo Geen Gowth Septembe 914, 2012, Busan, Republic of Koea DESIGN OF A FACILITY FOR THE PRECISE SIMULTANEOUS GENERATION AND MEASUREMENT OF FORCE AND TORQUE S. Baumgaten/Pesente, D. Röske, H. Kahmann, D. Mauesbege and R. Kumme Physikalisch-Technische Bundesanstalt, Dept. 1.2, Gemany, Bundesallee 100, D Baunschweig, Gemany Abstact: This aticle descibes the design of a measuing facility which can be used to investigate and calibate so-called "fiction coefficient sensos". These measuing facilities ae used to measue the pestessing foce and the tightening toque, esp. the fiction toque of scews. These measuements ae aimed at optimizing scew joints. The measuing facility descibed hee is pat of a foce standad machine (fsm). In addition to the foce which this system can ealize with a vey small measuement uncetainty of % (k = 2) (in the measuing ange fom 20 kn to 1 MN), it can also geneate an extemely pecise toque (objective: bette than % at k = 2) in the ange fom 20 N m to 2 kn m. Keywods: multi-component measuement, fiction coefficient senso, scews, clamp foce, toque 1. INTRODUCTION Scew joints ae an essential constuctional element in nealy all fields of economy and eveyday life. The scew industy aims to impove scews and to optimize the inteaction between the scew and the mateial to be fastened, in ode to make this fastening element safe and bette value. In this context, fiction coefficient sensos have been used fo a few yeas. These ae multi-component tansduces which can measue both an axial foce and at least one toque. In the case of scews, the foces involved ae: the pestessing foce and the tightening toque o the fiction toque between the scew head and the suppot as well as inside the thead. Industy entusted the Physikalisch-Technische Bundesanstalt with the qualification of the vaious commecially available fiction coefficient sensos using a suitable calibation pocedue to enable taceable measuements. In ode to meet these needs and to fulfil its mandate (dissemination of the units), the Physikalisch-Technische Bundesanstalt set up a novel measuing facility which enables the simultaneous geneation of both a pecise axial foce and of a vey accuately known toque into a fiction coefficient measuing system. This pape pesents the pinciple and the design of this multi-component measuing facility. 2. DESIGN AND CONCEPTION OF THE AUXILIARY DEVICE 2.1 SCHEMATIC SET-UP To attain the foces equied fo the selected measuing ange, it was decided to place a two-amed leve with a hoizontal leve side plate into the foce flow of PTB's 1- MN foce standad machine (1-MN-K-NME, [1]). A foce couple geneating the equied toque is to act at the ends of this leve. The foce is tansmitted to the leve am by means of thin metallic bands which act tangentially and ae placed paallel to each othe. Two mass stacks with diffeent load masses ae located on eithe side of the device to geneate foce. The foces, which ae ealized as gavitational foces by means of load masses (pinciple of the diect deadweight effect), ensue that a sufficiently small measuement uncetainty can be achieved. The vetical foce is tansfomed into a hoizontal foce by means of a conveting device. To this end, an ai beaing is used to minimize foce losses due to fiction. Fig. 1: Schematic epesentation of the set-up fo a loading device, side view with mass stack, ai beaing, metal band and leve am in the set-up of the 1-MN-K-NME

2 coodinate measuing machine Hexagon Infinite 2.0 in the fom of an index am. Fist, one of the mass stacks is otated until the band makes contact with the oute suface of the leve. This position is used as a efeence; the coodinate measuing index am is then mounted onto the opposite side. The coodinate measuing system is maintained, and the elation between the second mass stack and the leve is now detemined. The foce application point of both bands on the leve must be identical. The second mass stack is shifted on a linea table and otated by means of a otay table until this condition is fulfilled. Fig. 2: CAD model of the set-up of the multi-component measuing facility and additional mass stacks The limited space offes only little cleaance fo the positioning of the load masses. Lift tables located beneath the cosshead ae the optimal solution. The load masses ae ested on the lift tables. Two floo panels of the cosshead wee emoved, in ode to allow the loading mechanism to pass though it. The individual weights ae applied consecutively by means of a steppe moto. Conveting the foce effect fom a vetical gavitational foce into tensile foce acting hoizontally at the end of the leve posed a poblem. Fig. 1 shows a simplified schematic dawing of the planned set-up. The dawing shows one mass stack only a second one is located on the opposite side. Fig. 2 shows the planned set-up of the auxiliay device in the fom of a CAD model. The device in question is a modula pinciple which can also be applied to othe foce standad machines. With the device that is being built, axial foces in the ange fom 20 kn to 1 MN and toques in the ange fom 20 N m to 2 kn m can be ealized. 2.2 TORQUE TRANSMITTING SYSTEM To geneate a pue toque M z, one needs a foce couple with the same foce value but of opposite diection. Small deviations fom the ideal oientation of the foce vectos lead to additional bending moments and to a educed value of the toque M z. The foce couple needed is ealized by means of two identical load masses located at the end of leves facing each othe. The exact coodination of the oientation of the mass stacks and of the foce bands is paticulaly challenging. In ode to ensue an optimal positioning of the bands duing the loading phase, vaious sensos ae integated into the system, and the diection can, if necessay, be coected by means of steppe motos. In the fist oientation phase, the spatial position of the mass stacks and of the foce bands is detemined as a function of the position of the leve. The spatial coodinates ae measued using the potable Fig. 3: Schematic top view of the leve system and bands once the oientation of the mass stacks has been completed. Fig. 3 is a schematic epesentation of the completed oientation shown fom above. The leve plane epesents the x-y plane. The point of oigin is defined as the cente of the leve am. 2.3 VECTORIAL REPRESENTATION The vectos and positions in space ae illustated in Fig. 4. The points P M1, P M2, P B and P 0 ae measuement points which can be detemined by means of the coodinate measuing am. The points P M1 and P M2 span the vecto b which descibes the position in space of the axis of the oto equipped with an ai beaing. P 0 lies at the cente of the leve. Point P B is geometically equidistant fom the points P M1 and P M2. The point whee the band uns off the ai beaing shaft is located at the vetical distance ( = adius of the ai beaing shaft) fom the axis. To simplify the desciption, all points ae consideed as lying in the x-y plane. In eality, this assumption can, howeve, be guaanteed to a cetain extent only. When oientating the system, the objective is to come as close as possible to the ideal case z = 0 between all measuement points.

3 the height of the ai beaing head by the same diffeence. Senso B (see Fig. 1) of type FT-50 RLA-40_S_L4S (manufactue: Sensopat) has an extended measuing ange; its opeating pinciple is identical to that of Senso A. It is used to detemine the distance between the topmost weight and a fixed efeence point. Due to a change in the length of the band caused by themal o mechanical influences, the sequence of the application of the mass has to be adjusted. The weights ae loweed and applied onto a cental suspension ba. If the moto always dove down exactly the same distance, a change in the length of the band could cause individual weights not to be applied coectly o even not to be applied at all. The data tansmitted by senso B ae used to contol the loweing effected by the moto. Fig. 4: Repesentation of the measuement points in the x-y plane and of the vectos and angles that can be calculated on this basis and that ae necessay to oientate the system. In ode to detemine the otation angle of a mass stack in such a way that the band can be applied tangentially to the leve font suface, the following steps ae necessay: detemining the vecto b fom P M1 and P M1, and the vecto a on P 0 and P B. Fom this, it is possible to calculate the angle unde which the two vectos coss each othe. The tiangle 0 must be ight-angled. With the leve adius R OA and the value of the vecto a being known, the angle can be calculated. The otation angle by which the mass stack must be otated aound P B is yielded fom the diffeence between the two angles. The point P A is the only point that cannot be measued. It epesents the foce application point and must be detemined fom the othe geometical quantities. 2.4 SENSORS A-D Duing the measuement, displacements of the band, of the ai beaing shaft o of the load stack may occu. The individual sensos ae thee to detect distance changes. With the signals they povide, vaious stepped motos ae contolled which can then compensate fo these displacements. The sensos A-D ae installed in such a way that it is possible to check the oientation of the metallic bands also duing a loading opeation of up to 1 MN. In moe detail, senso A is an optical distance senso (manufactue: Sensopat) with a wavelength of 670 nm, type FT-50 RLA-20_S_L4S. The height change between the end of the leve and a fixed efeence point is detemined with a lase. Duing the loading, the leve end may dop, the foce vecto thus changing by an additional z component (see Fig. 5). Senso A contols a stepped moto which vaies Fig. 5: Sensos A, C and D, used to detemine the position and the inclination of the ai beaing oll. Displacements ae measued optically o capacitatively. If duing the loading opeation the foce vecto F ab is no longe pependicula to the ai beaing axis b, this leads to an additional foce effecting on the ai beaing shaft. The shaft will then slowly dift lateally inside the ai beaing stato. To detect a displacement of the ai beaing shaft, senso C, model BAW M12MG2-IAC208 BP03 (manufactue: Balluff), is used to measue the distance capacitatively, senso D to detemine the inclination (see Fig. 5). The inclination senso is custom-made (manufactue: Seika); it has a measuing ange of ± 0.2 and a esolution of ~ The two sensos contol stepped motos which adjust the position of the oll by vaying the inclination and by displacing it along the ai beaing axis. The optimal position is attained when no moe dift displacements ae obseved. 3. DETAILS 3.1 AXIAL FORCE OF THE 1-MN FORCE STANDARD MACHINE The axial foce is ealized by the 1-MN foce standad machine a dead-weight machine. The measuement uncetainty of the 1-MN foce standad machine is known sufficiently well and has been investigated on diffeent occasions [2, 3, 4]. The elative measuement uncetainty (k = 2) of the 1-MN foce standad machine is The deviation fom the ideal foce vecto development was investigated in [6]. This investigation showed that, at an

4 axial foce F z = 80 kn, sheaing foces F x = F y ~ 10 N can be expected. Futhemoe, a 6-component measuing platfom was implemented in the uppe pat of the foce standad machine in ode to, in futue, be able to measue also paasitical influences. 3.4 RESULTING FORCE Fo a ealistic estimation of the esulting foce, also ambient influences, such as tempeatue, pessue, and humidity, have to be taken into account. The ai buoyancy contibutes to educing the effective mass. The value of the foce can be calculated by means of the following fomula: ai F m g 1 ai m 0,34848 p 0, h e 273,15 T 0,0612 T (1) (2) m - mass g m s -1 p - pessue h - humidity T - tempeatue m - density of the weights Fig. 6: The figue shows the fist vesion of the aluminium leve mounted on the 1-MN foce standad machine. The shown shape has been etained fo the INVAR leve. One can also see the measuement points used to detemined the position towads which the coodinate measuing machine late moves. 3.2 LEVER ARM The leve used to geneate the toque has a length of 2 m and an uncetainty l = 1 µm. Fig. 6 shows a vesion made of aluminium. The shape has been chosen in such a way that a toque can be geneated in both diections (left and ight), heeby exploiting the space of installation optimally. In ode to educe changes in toque due to the dilatation of the leve am, INVAR steel was chosen as a mateial, since it has a vey low linea themal expansion coefficient of < K 1 at oom tempeatue. If the tempeatue changes by 1 K, the length of the leve changes by l = mm. The toque would then change by z = N m. 3.3 MASSES The additional mass stacks located on eithe side ae composed of 10 individual weights. One stack has load masses fo the following weight foces: 1 x 10 N, 2 x 20 N, 1 x 50 N, 3 x 100 N and 3 x 200 N. The maximum load mass is appox. 100 kg. Each mass stack thus geneates a couple of 1000 N. The mateial chosen fo the weights is X2CNiMoN This is a coosion-esistant steel which is not easily magnetizable and is mainly used in maitime constuction. Fo the masses, an uncetainty of m = kg can be stated. In ode to keep ambient influences, such as changes in tempeatue, pessue and humidity, as low as possible, the complete measuement set-up is located in a fully aiconditioned hall. The tempeatue vaies in the ange of T = 0.2 K, and h = 5 % fo the humidity. The gavitational acceleation inside the measuement hall was detemined as g = m s TWO-COMPONENT FORCE-TORQUE SENSOR Fo the calibation of the two-component measuing equipment, a tansfe standad with a sufficiently small elative measuement uncetainty of < is equied. The tansfe standad must be a two-component tansduce coveing a measuing ange with axial foces F z ~ 500 kn and toques M z ~ 500 N m. The selected foce/toque tansduce is ealized in the fom of a build-up system fom the company GTM, Gassmann Testing and Metology GmbH. The two tansduces ae a pecision foce tansduce KA-K-250kN-F-1mV/V [6], and a custom-made toque tansduce with the designation Dm-M-500_N m-2mv/v. The senso is calibated sepaately with egad to the axial foce and with egad to toque. The adaption pats allow it to be mounted in the 1-MN foce standad machine fo an axial load with a elative measuement uncetainty of as well as in the 1-kN m toque standad machine (1-kN m-dmnme) [7, 8] with a elative measuement uncetainty (k = 2) of Aftewads, the coss-talk of the tansduce when loaded simultaneously with F z and M z can be investigated in the two-component measuing equipment. 4. VECTORIAL ERROR ANALYSIS Geneating a toque by means of a couple equies a vectoial appoach (see Fig. 4). The diection of both the foce vecto and the distance vecto must be detemined. Deviations fom the othogonal oientation of the two vectos in the x-y plane make it necessay to conside vectoial eo analysis. If only one side of the system is

5 taken into account, the toque obtained then esults fom the vecto poduct of the distance and the foce vectos. M 0A F 0A F In ode to obtain a statement with egad to the individual moments, one needs the complete data concening the position of the vectos R 0A and R. The vecto poduct is solved in Equation (3). M F 0 Ay 0 Az 0 Ax z x y 0 Az 0 Ax 0 Ay y z x M M M x y z The poblem is that the exact position of the foce application point P A is not known. It is theefoe not possible to measue diectly with the index am. The point P A can, howeve, be calculated fom points P 0 and P B, fom the knowledge of the value of the vecto R 0A, but also fom the fact that the tiangle 0 must be ight-angled. The values fo R 0Ax, R 0Ay and R 0Az can be detemined with the aid of the softwae of the coodinate measuing am an uncetainty statement is, howeve, not given. To obtain an uncetainty fo the point P A, the intenal calculation must be econstucted. A detailed calculation of P 0A, of the individual components M x, M y and M z, as well as of thei uncetainty will be the subject of anothe publication. Fo the toque, a elative measuement uncetainty in the ange of is aimed at. An accuate analysis of the measuement uncetainty budget fom souces of eos such as, e.g., the mass, leve length and bending, oientation and length of the metallic band, and ambient influences will also be the subject of a subsequent, detailed analysis. (3) (4) REFERENCES [1] PTB s 1 MN Foce Standad Machine on the web: 121/measuing-devices.html [2] W. Weile, M. Petes, H. Gassmann, H. Ficke, W.Ackeschott, Die 1-MN-Nomalmeßeinichtung de PTB Baunschweig., VDI-Z 120, 1978, pp. 1-6 [3] A. Sawla, W. Weile, M. Petes, A. Bay, R. Levi, M. Vattasso A compasion of foce standads between the IMGC and the PTB., 6th Conf IMEKO TC3, Odessa, pp [4] W. Weile, A. Sawla Foce Standad machines of the National Institutes fo metology, PTB Beicht Me-22, 1978 [5] C. Feeo, The measuement of paasitic components in national foce standad machines, Measuement Vol. 8 No. 2, Ap-Jun 1990, pp [6] GTM Gassmann Testing and Metology GmbH, podukte/datenblaette/de/datenblatt_seie_k.pdf [7] PTB s 1 kn m Toque Standad Machine, 121/measuing-devices.html [8] K. Adolf, D. Mauesbege, D. Peschel Specifications and Uncetainty of Measuement of the PTB s 1kNm Toque Standad Machine, Poceedings of the 14 th IMEKO TC3 Confeence, Septembe 5-8, 1995, Wasaw, Poland, pp OUTLOOK The elative measuement uncetainty (k = 2) of the 1- MN foce standad machine is Accoding to fist estimations, we ae expecting a elative measuement uncetainty in a ange smalle than fo toque geneation. The measuement uncetainty budget will be consideed to a geate extent afte the facility has been commissioned. The aim is to educe the measuement uncetainty by optimizing the measuement cycle with egad to the acquisition of the spatial coodinates. It is expected that this novel measuement facility will lay the basis fo the calibation of fiction coefficient sensos. A fist step towads this is calibation by means of a tansfe standad. In addition, a compaison of divese commecial fiction coefficient sensos is aimed at. Futhemoe, the development of the facility is to be pusued with egad to a possible intoduction of bending moments into multi-component tansduces.