Concept and setup of a multi-component facility for force and torque in the range of 1 MN and 2 kN·m ACTA IMEKO June 2014, Volume 3, Number 2, 39 – 43 www.imeko.org Concept and setup of a multi-component facility for force and torque in the range of 1 MN and 2 kN·m Sebastian Baumgarten, Dirk Röske, Holger Kahmann, Dietmar Mauersberger, Rolf Kumme Physikalisch-Technische Bundesanstalt (PTB), Bundesallee 100, 38116 Braunschweig, Germany Section: RESEARCH PAPER Keywords: multi-component measurement, friction coefficient sensor, screws, clamp force, torque Citation: Sebastian Baumgarten, Dirk Röske, Holger Kahmann, Dietmar Mauersberger, Rolf Kumme, Concept and setup of a multi-component facility for force and torque in the range of 1 MN and 2 kN•m, Acta IMEKO, vol. 3, no. 2, article 10, June 2014, identifier: IMEKO-ACTA-03 (2014)-02-10 Editor: Paolo Carbone, University of Perugia Received February 7th, 2013; In final form June 28th, 2014; Published June 2014 Copyright: © 2014 IMEKO. This is an open-access article distributed under the terms of the Creative Commons Attribution 3.0 License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited Funding: (none reported) Corresponding author: Sebastian Baumgarten, e-mail: Sebastian.baumgarten@ptb.de 1. INTRODUCTION Screw joints are an essential constructional element in nearly all fields of economy and everyday life. The screw industry aims to improve screws and to optimize the interaction between the screw and the material to be fastened, in order to make this fastening element safer and better value. In this context, friction coefficient sensors have been used for a few years. These are multi-component transducers which can measure both an axial force and at least one torque. In the case of screws, the forces involved are: the prestressing force and the tightening torque or the friction torque between the screw head and the support as well as inside the thread. Industry entrusted the PTB with the qualification of the various commercially available friction coefficient sensors using a suitable calibration procedure to enable traceable measurements. In order to meet these needs and to fulfil its mandate (dissemination of the units), the PTB set up a novel measuring facility which enables the simultaneous generation of both a precise axial force and of a very accurately known torque into a friction coefficient measuring system. This paper presents the principle and the design of this multi-component measuring facility. 2. DESIGN AND CONCEPTION OF THE AUXILIARY DEVICE 2.1. Schematic setup To attain the forces required for the selected measuring range, it was decided to place a two-armed lever with a horizontal lever side plate into the force flow of PTB's 1-MN force standard machine (1-MN-K-NME). A force couple generating the required torque is to act at the ends of this lever. The force is transmitted to the lever arm by means of thin metallic bands which act tangentially and are placed parallel to each other. Two mass stacks with different load masses are located on either side of the device to generate force. The forces, which are realized as gravitational forces by means of load masses (principle of the direct deadweight effect), ensure that a sufficiently small measurement uncertainty can be achieved. The vertical force is transformed into a horizontal force by means of a converting device. To this end, an air bearing is used to minimize force losses due to friction. The limited space offers only little clearance for the positioning of the load masses. Lift tables located beneath the operation platform are the optimal solution. On each lift table a linear table and a rotation disc is mounted. The load masses rest on the rotations disc. The individual weights are applied ABSTRACT This article describes the design of a measuring facility which can be used to investigate and calibrate so-called "friction coefficient sensors". These measuring facilities are used to measure the prestressing force and the tightening torque, respectively the friction torque of screws. These measurements are aimed at optimizing screw joints. The measuring facility described here is part of a force standard machine (fsm). In addition to the force which this system can realize with a very small measurement uncertainty of 0.002 % (k = 2) (in the measuring range from 20 kN to 1 MN), it can also generate an extremely precise torque (objective: better than 0.005 % at k = 2) in the range from 20 N·m to 2 kN·m. ACTA IMEKO | www.imeko.org June 2014 | Volume 3 | Number 2 | 39 consecutively by lowering the lift table. Converting the force effect from a vertical gravitational force into tensile force acting horizontally at the end of the lever posed a problem. Figure 1 shows a simplified schematic drawing of the planned set-up. The drawing shows one mass stack only – a second one is located symmetrically on the opposite side. Figure 2 shows the planned set-up of the auxiliary device in the form of a CAD model. The device in question is a modular principle which can also be applied to other force standard machines. There are some similarities to an older torque calibration machine at the NPL [1]. With the additional facility that is being built, axial forces in the range from 20 kN to 1 MN and torques in the range from 20 N·m to 2 kN·m can be realized. 2.2. Torque Transmitting System To generate a pure torque Mz, one needs a force couple with the same force value but of opposite direction. Small deviations from the ideal orientation of the force vectors lead to additional bending moments and to a reduced value of the torque Mz. The force couple needed is realized by means of two identical load masses located at the end of levers facing each other. The exact coordination of the orientation of the mass stacks and of the force bands is particularly challenging. In order to ensure an optimal positioning of the bands during the loading phase, various sensors are integrated into the system, and the direction can, if necessary, be corrected by means of stepper motors. In the first orientation phase, the spatial position of the mass stacks and of the force bands is determined as a function of the lever position. The spatial coordinates are measured using the portable coordinate measuring machine in the form of an index arm, Figure 3. The index arm can be mounted at each pillar of the force standard machine frame. Single points are measured with a relative uncertainty (k = 2) of Δx = 10 µm. With a range of 1200 mm it is not possible to reach all characteristics of the entire system from one position so the measurement arms have to be mounted at every pillar. The software guarantees a coordinate transformation for the new point of origin. From the data of the different positions and orientation of the modules the necessary displacements and rotation angles can be calculated. First, one of the mass stacks is rotated until the band makes contact with the outer surface of the lever. This position is used as a reference; the coordinate measuring index arm is then mounted onto the opposite side. The coordinate measuring system is maintained, and the relation between the second mass stack and the lever is now determined. The force application point of both bands on the lever must be identical. The second mass stack is shifted on a linear table and rotated by means of a rotary table until this condition is fulfilled. Figure 4 is a schematic representation of the completed orientation shown from above. 2.3. Vectorial Representation The vectors and positions in space are illustrated in Figure 5. The point P0 is the centre of the lever. PB is the point in the middle of the rotor in which the band starts to roll off the rotor. The force application point is referred as PA. The points Pm1, Pm2, PB and P0 are measurement points which can be determined by means of the coordinate measuring arm. The points Pm1 and Pm2 span the vector 21 mm PP which describes the position in space of the axis of the rotor equipped with an air bearing. Point PB is – geometrically – equidistant from the points PM1 and Pm2. The point where the band runs off the air bearing shaft is located at the vertical distance rrotor (rrotor = radius of the air bearing shaft) from the axis. To Figure 1. Schematic representation of the set-up for a loading device, side view with mass stack, air bearing, metal band and lever arm in the set-up of the 1-MN-K-NME. Figure 2. CAD model of the set-up of the multi-component measuring facility and additional mass stacks. Figure 3. The figure shows the portable coordinate measuring arm, which is mounted at a pillar of the 1-MN-K-NME frame. ACTA IMEKO | www.imeko.org June 2014 | Volume 3 | Number 2 | 40 simplify the description, all points are considered as lying in the xy plane. In reality, this assumption can, however, be guaranteed to a certain extent only. When orientating the system, the objective is to come as close as possible to the ideal case ∆z = 0 between all measurement points. In order to determine the rotation angle of a mass stack in such a way that the band can be applied tangentially to the lever front surface, the following steps are necessary: determining the vector 21 mm PP and the vector BPP0 . From this, it is possible to calculate the angle γ1 under which the two vectors cross each other. The triangle PAPBP0 must be right-angled. The length of the vector APP0 is identical with the half of the lever length llever. With the lever radius r and the value of the vector BPP0 being known, the angle γ2 can be calculated. The rotation angle by which the mass stack must be rotated around PB is yielded from the difference between the two angles. The point PA is the only point that cannot be measured. It represents the force application point and must be determined from the other geometrical quantities. 2.4. Control Sensors During the measurement, displacements of the band, of the air bearing shaft or of the load stack may occur. The individual sensors are there to detect distance changes. With the signals they provide, various stepped motors are controlled which can then compensate for these displacements. The sensors A-D are installed in such a way that it is possible to check the orientation of the metallic bands also during a loading operation of up to 1 MN-K-NME. In more detail, sensor A is an optical distance sensor with a wavelength of 670 nm. The height change between the end of the lever and a fixed reference point at the frame of the force machine is determined with a laser. The resolution is in the range of 7 µm with a linearity error lesser then 1%. During the loading and under its own weight, the lever end may underlie a deformation, the force vector thus changing by an additional z component (see Figure 6). Sensor A controls a stepped motor which can vary the height of the air bearing head by the same difference in a range of 16 mm with a relative uncertainty of 6.25 · 10-4 mm. The operating principle and wavelength of Sensor B are identical to that of Sensor A. The resolution is in the range of 20 µm with a linearity error less than 1%. It is used to determine the distance between the topmost weight and a fixed reference point at the frame of the load stack. Due to a change in the length of the band caused by thermal or mechanical influences, a change in the length of the band could cause individual weights not to be applied correctly or even not to be applied at all. The data transmitted by sensor B are used to control the lowering effected by the motor. If during the loading operation the force vector BA PP is no longer perpendicular to the air bearing axis 21 mm PP , this leads to an additional force effecting on the air bearing shaft. The shaft will then slowly drift laterally inside the air bearing stator. To detect a displacement of the air bearing shaft, sensor C is used to measure the distance capacitively and sensor D to determine the inclination (see Figure 5). The inclination sensor is custom-made. It has a measuring range of ± 0.2° and a resolution of ~ 1 · 10-6 °. The measuring range of the capacitive sensor is 2 mm with a relative uncertainty (k = 2) of 4.5 · 10-2 mm. The two sensors control stepper motors which adjust the position of the roll by varying the inclination and by displacing the complete air bearing head. The optimal position is attained when no more drift displacements can be observed. Figure 4. Schematic top view of the lever system and bands once the orientation of the mass stacks has been completed. Figure 5. Representation of the measurement points in the xy plane and of the vectors and angles that can be calculated on this basis and that are necessary to orientate the system. Figure 6. Sensors A, C and D, used to determine the position and the inclination of the air bearing roll. Displacements are measured by optical or capacitive means. ACTA IMEKO | www.imeko.org June 2014 | Volume 3 | Number 2 | 41 3. MODULES OF THE MULTI-COMPONENT MEASURING FACILITY 3.1. Axial Force of the 1-MN Force standard Machine The axial force is realized by the 1-MN force standard machine – a dead-weight machine. The measurement uncertainty of the 1-MN force standard machine is known sufficiently well and has been investigated on different occasions [2,3,4]. The relative measurement uncertainty (k = 2) of the 1-MN force standard machine is 2 · 10−5. The deviation from the ideal force vector development was investigated in [5]. This investigation showed that, at an axial force Fz = 80 kN, shear forces Fx = Fy ~ 10 N can be expected. Furthermore, a 6- component measuring platform was implemented in the upper part of the force standard machine in order to, in future, be able to measure also parasitical influences. 3.2. Lever arm The lever used to generate the torque has a length of 2 m and an uncertainty ∆llever = 1 µm. Figure 7 shows a version made of aluminium. The shape has been chosen in such a way that a torque can be generated in both directions (left and right), hereby exploiting the space of installation optimally. In order to reduce changes in torque due to the dilatation of the lever arm, INVAR steel was chosen for the material, since it has a very low linear thermal expansion coefficient of < 2 · 10−6 K−1 at room temperature. If the temperature changes by 1 K, the length of the lever changes by ∆llever = 3.7 · 10-3 mm. The torque Mz will be increased by a value of 3.7 · 10-6 N·m. 3.3. Masses The additional mass stacks located on either side are composed of 10 individual weights. One stack has load masses for the following weight forces: 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 approx. 100 kg. Each mass stack thus generates a couple of 1000 N. The material chosen for the weights is X2CrNiMoN 18-14. This is a corrosion-resistant steel which is not easily magnetisable and is mainly used in maritime constructions. For the masses, a relative uncertainty of 5 · 10-6 can be stated. For a realistic estimation of the resulting force, also ambient influences, such as temperature, pressure, and humidity, have to be taken into account. The air buoyancy contributes to reducing the effective mass. The value of the force can be calculated by means of the following formula:       −⋅⋅= m air1 ρ ρ gmF (1) T ehp T air + ⋅⋅−⋅ = ⋅ 15,273 009024,034848,0 0612,0rρ (2) where m is the mass, g is the acceleration due to gravity (9.81253 m · s-2), p the air pressure, hr the air humidity, T the temperature and ρm the density of the weights. In order to keep ambient influences, such as changes in temperature, pressure and humidity, as low as possible, the complete measurement set-up is located in a fully air- conditioned hall. The temperature varies over a range of ∆T = 0.2 K, and ∆hr = 5 % for the humidity. The value of the gravitational acceleration inside the measurement hall was determined with g = 9.81253 m·s-2 with a relative measurement uncertainty of 0.0001 %. 3.4. Force-Torque-Sensor For the calibration of the two-component measuring equipment, a transfer standard with a sufficiently small relative measurement uncertainty of < 1 · 10-4 is required. The transfer standard is a two-component transducer covering a measuring range with axial forces Fz ~ 500 kN and torques Mz ~ 500 N·m. The selected force/torque transducer is realized in the form of a build-up system (Figure 8). The sensor is calibrated separately with regard to the axial force and with regard to torque. The adaption parts allow it to be mounted in the 1-MN force standard machine for an axial load with a relative measurement uncertainty of 2 · 10−5 as well as in the 1- kN·m torque standard machine (1-kN·m-DmNME) [6] with a relative measurement uncertainty (k = 2) of 2 · 10−5. Afterwards, the cross-talk of the transducer when loaded simultaneously with Fz and Mz can be investigated in the two- component measuring equipment. 4. VECTORIAL ERROR ANALYSIS Generating a torque by means of a couple requires a vectorial approach (see Figure 5). The direction of both the force vector and the distance vector must be determined. Deviations from the orthogonal orientation of the two vectors in the xy plane make it necessary to consider vectorial error Figure 7. The figure shows the first version of the aluminium lever mounted on the 1-MN force standard machine. The shown shape has been retained for the INVAR lever. One can also see the measurement points used to determine the position towards which the coordinate measuring machine later moves. Figure 8. Force-Torque-Transducer as a Build-Up-System of a 500 N·m Torque-Transducer on the top and a 500 kN Force-Transducer between these and the pressure plate under it. ACTA IMEKO | www.imeko.org June 2014 | Volume 3 | Number 2 | 42 analysis. If only one side of the system is taken into account, the torque obtained then results from the vector product of the distance and the force vectors: F PP PP PPPPPPFrM BA BA ABAA ⋅×=×=×= 00  . (3) In order to obtain a statement with regard to the individual moments, one needs the complete data concerning the position of the vectors APP0 and BA PP . The vector product is solved in Equation (4):           =           − − − = z y x BAABAA BAABAA BAABAA BA M M M PPPPPPPP PPPPPPPP PPPPPPPP PP F M xyyx zxxz yzzy 00 00 00 ·  . (4) The problem is that the exact position of the force application point PA is not known. It is, therefore, not possible to measure directly with the index arm. The point PA can, however, be calculated from points P0 and PB, from the knowledge of the value of the vector APP0 , but also from the fact that the triangle PAPBP0 must be right-angled. The values for xA PP0 , yAPP0 and zAPP0 can be determined with the aid of the software of the coordinate measuring arm – an uncertainty statement is, however, not given. To obtain an uncertainty for the point PA, the internal calculation must be reconstructed. A detailed calculation of APP0 , of the individual components Mx, My and Mz, as well as of their uncertainty will be the subject of another publication. For the torque, a relative measurement uncertainty in the range of 5 · 10-5 is aimed at. An accurate analysis of the measurement uncertainty budget from sources of errors such as, e.g., the mass, lever length and bending, orientation and length of the metallic band, and ambient influences will also be the subject of a subsequent, detailed analysis. 5. OUTLOOK The relative measurement uncertainty (k = 2) of the 1-MN force standard machine is 2 · 10−5. According to first estimations, we are expecting a relative measurement uncertainty in a range smaller than 5 · 10−5 for torque generation. The measurement uncertainty budget will be considered to a greater extent after the facility has been commissioned. The aim is to reduce the measurement uncertainty by optimizing the measurement cycle with regard to the acquisition of the spatial coordinates. It is expected that this novel measurement facility will lay the basis for the calibration of friction coefficient sensors. A first step towards this is calibration by means of a transfer standard. In addition, a comparison of diverse commercial friction coefficient sensors is aimed at. Furthermore, the development of the facility is to be pursued with regard to a possible introduction of bending moments into multi-component transducers. REFERENCES [1] A. Robinson, “The Commissioning of the First UK National Standard Static Torque Calibration Machine“, 19th Conference on Force, Mass and Torque Measurement, Cairo, Egypt, 2005, pp.1-6. [2] W. Weiler, M. Peters, H. Gassmann, H. Fricke, W.Ackerschott, “Die 1-MN-Normalmeßeinrichtung der PTB Braunschweig.”, VDI-Z 120, 1978, pp.1-6. [3] A. Sawla, W. Weiler, M. Peters, A. Bray, R. Levi, M. Vattasso “A comparsion of force standards between the IMGC and the PTB”, 6th Conf IMEKO TC3, Odessa, pp.1-13. [4] W. Weiler, A. Sawla “Force Standard machines of the National Institutes for metrology”, PTB Bericht Me-22, 1978. [5] C. Ferrero, “The measurement of parasitic components in national force standard machines”, Measurement, Vol. 8, No. 2, Apr-Jun 1990, pp.66-76. [6] K. Adolf, D. Mauersberger, D. Peschel “Specifications and Uncertainty of Measurement of the PTB’s 1kNm Torque Standard Machine”, Proc. of the 14th IMEKO TC3 Conference, September 5-8, 1995, Warsaw, Poland, pp.178-182. ACTA IMEKO | www.imeko.org June 2014 | Volume 3 | Number 2 | 43 Concept and setup of a multi-component facility for force and torque in the range of 1 MN and 2 kN m