Abstract
In the project “Modular desktop machining centre with SMA actuation” within the SPP1476, the Chair of Production Systems has developed modular and standardised SMA actuators for use in a machine tool axis.
Keywords
- Machine Tool
- Shape Memory Alloy
- Pulse Width Modulation
- Shape Memory Alloy Wire
- Shape Memory Alloy Actuator
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.
1 Introduction
Within the scope of the priority program SPP1476 ‘‘Small machine tools for small work pieces’’, the Chair of Production Systems (LPS) at Ruhr-Universität Bochum is currently working on the subproject of “Modular desktop machining centres with shape memory alloy actuators”. The aim of this project is the development of a new drive for machine tools for micro manufacturing based on Shape Memory Alloys (SMA). The new drive should be smaller and cheaper to produce than conventional systems while having similar properties. The project started in September 2010. A roundup of the main research results are presented in this paper.
2 Shape Memory Technology
As part of a change in temperature (thermal shape memory effect) or under external mechanical tension (pseudo-elastic shape memory effect), some alloys show a reversible deformation. When heated, the material remembers the previously memorised form, even though it has been deformed significantly in the meantime. This shape memory characteristic, which has been known since the 1930s, is especially found in metallic alloys such as CuZnAl, CuAlNi and NiTi. Nickel-titanium alloys with a nickel proportion of approximately 50% are most frequently used.
The thermal shape memory effect is caused by the crystallographic transformation from the high temperature austenite phase to the easily deformable low temperature martensite phase. When the martensite is deformed (Fig. 1), the structure of the crystal lattice changes. The deformation remains when the external pressure is removed.
By heating the pre-stretched material, the conversion to high temperature austenite starts at the austenite start temperature (As) and is completed at the austenite finish temperature (Af). Thanks to its ability to rearrange the crystal lattice and to restore its original macroscopic shape, the material performs some valuable mechanical work which is made use of in numerous application fields.
Heating can be realised by external temperature fields or electricity (Joule heating). With a short and intense pulse of electric current, the transformation can be accomplished in a few milliseconds, depending on the volume of the shape memory alloy.
During the cooling phase, the material reverts back to the martensitic crystal lattice between the martensite start (Ms) and martensite finish (Mf) temperature. The depicted transformation process is hysteresis-afflicted, therefore, Ms and Af as well as As and Mf lie apart depending on the concrete alloy composition. For the Smartflex wires (from The SAES Getters Group) used in the following experiments, the hysteresis width is approx. 25 K.
The one-way effect, shown in Fig. 1, can be cyclically repeated with an external load application, for example, with a return spring, and is thus established for actuator purposes. This is called the extrinsic two-way effect. The phase transformation produces a usable contraction of up to 8% of the original length. Both the use of the maximum contraction and the increasing mechanical tension in the material lead to a decrease of possible work cycles. With the right design for the actuators, it is currently possible to achieve up to 106 work cycles. Due to the high energy density (3 MJ/m3), high loads can be realised even with geometrically small actuators. Besides the thermal effect, there is also the pseudo-elastic shape memory effect, which leads to super-elastic material behaviour and is therefore an appropriate resetting element for shape memory actuators, among other applications.
Shape memory alloys have been researched within the project SFB 459 at Ruhr-Universität Bochum since 2000. The Chair of Production Systems (LPS) supported this project by introducing the application-oriented topic of “computer-aided design of shape memory alloys” and the research field of “laser-based generative production of complex components made of NiTi(x)-powder”. The research results already published have confirmed the suitability of shape memory components in mechatronic systems [1,2,3]. Additional results, which have been published outside the project SFB 459, describe advanced applications such as adaptive wings in aeronautics [4], micro-valves with high energy density [5] or micro-grippers [6]. Studies of the Institute of Dynamics and Vibrations at the Leibniz Universität in Hannover also confirm the suitability of these actuators for control tasks [7].
Other current developments in SMA technology, especially for high precision tasks and machine tool applications are detailed in [8], as well as [9,10,11,12].
3 Overall Machine Concept
At the beginning of the project in 2010, existing machine tools for micro manufacturing were analysed and possible synergy potentials identified. In micro machining, ultra-precision machines are mainly used nowadays. These machines share many components and features with conventional machine tools for macro machining tasks. Due to the restricted kinematic capabilities of this traditional machine type, it is hardly possible to guarantee a high-precision processing of complex component geometries [13]. The drives in many commercially available processing centres are usually realised as powertrains, as shown in Fig. 2.
Every component and interface in this string may be responsible for inaccuracies. Therefore, every single component has to be made with high precision, which leads to high manufacturing costs. While developing a smaller and more cost-efficient machine, the number of components in the powertrain should be reduced.
In recent years, the use of direct drives in machine tools minimised some of these problems (see e.g. [14,15,16]) and increased the accuracy even for larger scale machines. Nevertheless, the potential for an inexpensive machine tool with a small footprint remains, especially for prototyping applications.
Since actuators, based on SMAs, most commonly produce a translational movement, every element in the conventional powertrain from the engine up to the spindle nut can be substituted. The advantages of such a direct drive, which does not require conversions and couplings between rotational and linear movements, are nowadays used by implementing piezo stacks for ranges of a few micrometres. For higher displacements, it is necessary to apply rotatory drives or expensive linear motors which require space.
Based on this analysis, the SMA actuators have been constructed in a modular way in order to realise both very fine and large travel motions by series connections of multiple actuators [17]. Through these actuator stacks, costs and installation space can be saved while, at the same time, the required precision is guaranteed. Figure 3 shows the first concept of this construction.
The individual actuators are identical except for the bolt, which defines the travel range of the individual actuator through a limit stop. If an identical housing is used, costs can be saved for a future series production. The combination of actuators with different travel ranges results in a good compromise between high precision and the overall travel range of the assembly [18]. After the clarification of the requirements within the SPP1476, the axis, which contains several actuators, should have a displacement of at least 10–15 mm with a maximum force of approximately 15 N and a minimum resolution of around 5 µm.
4 Design of the Modular Actuators
For the development of the individual actuators, the first step was to decide on the geometry of the used SMA material. Wires, ribbons, rods and springs were considered. Because of low prices, the good proportions between cross section and surface and the resulting cooling characteristics, SMA wires were selected for the actuators. A comprehensive foundation in design, control and regulation is available in technical literature for this type of SMA construction [1]. In addition, extensive preliminary studies were conducted at the Chair of Production Systems.
The next step was to test different principles for converting the contraction of an SMA wire into a translational movement. A V-shaped arrangement turned out to be particularly suitable. The wire is firmly clamped on both sides of the actuator housing. When activated, the wire contracts and pushes a bolt out of the housing. The activation takes place by heating the wire with an electrical current. This wire arrangement allows a relatively flat design of the actuators, whereby the height of the overall axis assembly can be kept low. The footprint of a single actuator is 90 × 30 × 10 mm. A first functional model of the actuator and axis concept is illustrated in Fig. 4.
In each housing, the SMA wire is screwed into a connector block made of aluminium and pushes a Teflon bolt out of the casing when activated. For isolation and future manufacturing purposes, the housing is made of plastic. First experimental results have shown that this design does not provide the desired maximum travel of 5 mm per actuator. This is partly due to the terminal blocks, which conduct heat out of the wire due to the choice of material and the large volume and therefore prevent a complete transformation. Investigations with a thermal imaging camera confirm this thesis [19, 17]. There was also evidence that the combination of the design and material of the housing is not ideal as it causes the actuator to expand slowly and the accuracy to decrease the longer the actuator is activated. The resetting of the wire is not implemented in the actuator itself, but rather as a combined counterforce for the entire axis. Thus space can be saved and the actuator design can be kept simple.
To meet the demanded requirements, further iteration steps in actuator design were performed. The design of the revised actuators includes various improvements which were derived from the first experiments.
The reworked housing (Fig. 5) is made of aluminium and has a significantly higher stiffness. The SMA wires are crimped at the ends and can therefore be easily inserted into the housing. Plastic inserts, which were made in a 3D fuse deposition modelling printer, provide thermal and electrical isolation between the housing and the wire. The bolt has several holes for the guidance of the wire in order to test different angular arrangements. In addition, the second prototype of the actuator is designed specifically for cooling experiments. The large housing volume allows to insert fittings for air guidance, while threads are provided for connecting a compressed air line at the sides. The right half of Fig. 5 shows one example of the tested installations.
The tests for an active cooling of the actuator by convection resulted in a significant increase in dynamics and the response behaviour. Under normal circumstances, the speed of the actuator during the deactivation of the SMA wire is only dependent on the difference between wire temperature and ambient temperature as well as the surface of the SMA. As a result of cooling, the speed of the actuator can now be controlled in both directions, during the activation by the induced electrical power (see Fig. 8) and in the deactivation phase by changing the air stream. In addition to the experiments for air cooling, the actuator was operated in distilled water in order to test the effect of a liquid cooling. The results of both experiments are shown in Fig. 6.
The diagram shows the activation and deactivation cycle of an actuator without cooling (red). After switching off the current (purple), a certain reaction time passes and the actuator does not reset immediately. This is due to the different transition temperatures during the phase transformation in the SMA material. Therefore, the wire within the actuator must first be cooled to a certain temperature before the transition starts. The conversion itself is slow and only approximately linear, as can be seen in the progression of the red curve. In the case of an active cooling, after switching off the current (green), the reaction time of the actuator is reduced to below 0.5 s, while the reconversion itself shows a nearly linear course with a significantly increased gradient.
The tests for water cooling (blue, dotted line) achieved even better results. When it is operated in the temperature controlled medium, the actuator can be activated differently as the SMA wire can be preheated or cooled by the medium. If the wire is preheated, a lower temperature difference for the conversion has to be bridged and the actuator can trigger faster. Thanks to the good thermal conductivity of the medium, the reaction time at its deactivation can be further shortened. Through the accelerated conversion, the actuator can reach speeds of about 6 mm/s for the return movement [17]. The only disadvantage of the operation in liquid is a slightly higher energy input. Due to the heat dissipation to the surrounding medium, the wire has to be heated with a higher current. In contrast, the shorter heating time leads to a shorter connection to the power supply (Fig. 6, blue).
From the experimental results of the first two functional actuator models, the final design has been derived. Among other things, it has been possible to optimise the manufacturability, costs, installation effort, cooling requirements, precision, durability, and installation space.
Figure 7 illustrates the CAD model and a functional model of the final actuator. The casing is made of fibre-reinforced plastic and is constructed in a way that maximises the stiffness in feed direction. The construction is mostly designed for machining, but in the case of future mass production, the housing may easily be manufactured by injection moulding. The internal dimensions of the actuator have been greatly reduced compared to the previous iteration stages in order to keep the cooling volume low around the wire. Cooling is achieved by nozzles below the wire. The compressed air is connected to the housing with one connection per side. The bolt is made of aluminium. Within the bolt, the wire is thermally and electrically insulated by an insert made of ceramic. By changing the stop on the bolt, the travel range of the actuator can be set in a range of up to 5 mm. To produce actuators with different travel ranges, only the bolts have to be replaced. The SMA wire is threaded through a ceramic washer at both ends of the actuator and held in place by crimping.
The outer dimensions of this actuator are 105 × 33 × 13 mm. One individual actuator is therefore larger than the first iteration. Because of the V-shape, the actuators can be packed tighter when used in an axis, which reduces the overall dimensions of the assembly.
5 Control Strategies
This final actuator type was used for the tests to check the displacement control. Special focus was placed on the setting of a defined speed during the activation of the actuator and the ability to reach and hold a defined position between the end stops. It is essential that no overshoot occurs when approaching a pre-set displacement and that the deviation, while holding the position, is as low as possible. The left diagram in Fig. 8 shows the experimental results.
A single actuator was activated in a closed loop control setup, with a laser triangulation sensor measuring the displacement. The SMA wire in the actuator had a diameter of 0.3 mm. The bolt in the actuator was designed for a maximum displacement of 2 mm, both approached positions are thus intermediate positions. The first point (0.5 mm) was approached at a rather high speed, without premature braking. The bolt overshot the desired position by 1.3 µm. At the second point (1 mm), the speed of the actuator was slowed down before reaching the desired value, whereby the overshoot was reduced to a value below the resolution of the sensor of 0.8 µm. The maximum deviation while holding a displacement was 1.6 µm for both values. Thus, the accuracy of the axis is not dependent anymore on the manufacturing accuracy of the actuators and limit stops as originally formulated in the project application, but can now be adjusted by a control technology. This opens up a variety of new applications for this type of actuator.
The control of the actuators is realised by changing the current or pulse width modulation. Both approaches will be provided via a specially designed electronics assembly and a self-programmed PC software and user interface. The control strategy is explained in further detail in [20]. In addition to holding positions, the control of the axis speed is crucial. The diagram in Fig. 8 on the right shows that the activation speed of the SMA actuators at constant ambient temperature directly depends on the value of the heating current. Through different current profiles, the actuation speed can be set. The present test situation shows that the achievable deviation is 1.6 µm.
Next, several actuators were combined to an axis and the collective control was implemented. The first decision in designing the axis assembly concerns the guidance of the individual actuators. Since the overall displacement is generated by placing the actuators in series connection, the components must be mounted independently and as smoothly as possible. For this purpose, sliding rods and matching precision linear ball bearings are used in the first prototype of the axis (Fig. 4, right).
Because of the two slide bars, the assembly either got jammed or a pronounced stick-slip effect occurred. Besides, it was also important to put the actuators’ force application point on the same level as of the mounting of the slide bars in order to avoid further negative effects. For the next experimental prototype of the axis (Fig. 9), high-precision linear guide rails and sleds were used, which, in the required size and configuration, produce a friction force of about 0.5 N, which the actuators can easily overcome with their maximum force of more than 10 N. The force application point of the axis is on top of the assembly, where the next axis, a clamping device or other peripherals can be connected. The connection plate moves on independent guide rails. This approach ensures that inaccuracies caused by the modular actuators can be minimised.
The new axis assembly (145 × 145 × 35 mm) was used to run further experiments that considered the simultaneous control of up to five standard actuators. Therefore, the different actuators were put in a series connection and thus treated as one big SMA actuator with just one controlled power supply and one overall displacement sensor. The power to activate the actuators was supplied and modulated via pulse width modulation. Figure 10 illustrates the experimental results.
A defined path has been chosen to make the results comparable to future experiments. It contains different gradients both for the activation and deactivation phase as well as holding positions. It also tests the whole range of motion of the actuators, as they are able to produce a maximum displacement of around 20 mm.
The plotted error in Fig. 10 shows that the biggest deviations appear when the actuators are activated from a holding position. At these points the error reaches its maximum of 1% of the displacement (at 6 mm displacement). Generally, it is not easy to create a path with sharp corners using SMA actuators because they do not have the best dynamic response. This can be avoided, when using SMAs in a machine tool axis, by choosing a path according to the actuators’ abilities. For the rest of the path, the maximum error is about ±20 µm (less than 0.2% of the corresponding displacement) and even lower for the holding positions.
In contrast to the positioning accuracy and the stroke, other evaluating parameters like stiffness are less promising. The axis’ stiffness is by design directly dependent on the elastic behaviour of the SMA wires which push the bolt out of each actuators’ housing (Fig. 7). This iteration of the axis should therefore only be considered for applications with low machining forces of up to 2 N.
Another restriction of SMA technology in an actuator setup of this size, is the dynamic response, during activation (Fig. 8, right) and during deactivation without active cooling (Fig. 6). As of now, the feed axis is not suitable for highly dynamic operations. Further experiments and subsequent design revisions are needed.
6 Development of the z-Axis
During the development of the x/y-axis and especially after discussions about the technical requirements with the project partners within the SPP1476, it was found that different specifications are needed for the z-axis compared to the horizontal axis. Indeed, virtually all feed axis projects within the priority program primarily focussed on the development of the x/y-axis, whereas there was a common need for a z-axis. Because of this, particular attention was paid to the development of the z-axis during the project. This vertical axis must provide greater forces to be able, depending on a top or bottom installation, to move a clamping device, the workpiece or the tool. In addition, the axis should provide a larger travel range to allow the user to open the working area while changing tools or work pieces.
The first concept of the z-axis had a similar setup to the x-axis, with a horizontal placement of the SMA actuators. The force of the actuators was extended via a toggle lever, while an additional SMA actuator, which was located below the standard actuators, provides larger displacements outside the working space.
The experimental results have proven that the lever mechanism is not sturdy enough to be used in a machine tool axis. In addition, the project group has agreed on installing the z-axis at the top of the assembly, which is thought to be more useful.
The final z-axis, shown in Fig. 11, consists only of two standard actuators, for a combined displacement of around 10 mm. Additionally, a manual adjustment for larger travel ranges has been incorporated to open the working space. The axis has adjustable counter weights to accommodate tools or other peripherals weighing 0.1–2 kg. The assembly measures 260 × 125 × 45 mm without the counter weight.
Figure 11 shows the axis, as exhibited at the Hanover fair 2016, installed in a cube with the GrindBall attached via the multi-functional interface. The control for the z-axis works in the same way as for the horizontal axes and it operates with the same high precision. With each tool change, the counter weights need to be adjusted to provide optimal working conditions for the SMA actuators. Therefore this concept is not suitable for an automated setup. This should be altered in future design updates.
7 Summary and Outlook
The project demonstrates that these SMA actuators are inexpensive to produce thanks to a simple construction and that they bear beneficial characteristics which can make them suitable for use in the feed axis of machine tools for micro manufacturing in future iterations. The target positioning accuracy of 5 µm has been achieved and surpassed. As for the path accuracy, it stays between ±20 µm for 5 actuators in series connection, which is accurate enough for the applications targeted in the SPP1476, but not yet suitable for the whole spectrum of micro machining.
Until the end of the project and beyond, the actuator control should be further investigated and optimised especially for series connections. In addition, the properties of the axis are to be characterised and improved in terms of its accuracy, repeatability, stiffness and behaviour in permanent operation. Results have already shown that the use of SMA actuators in feed axes of machine tools in micro-production bring about clear benefits in the fields costs, installation space, weight and modularity. During the previous cooperation in SPP1476, several demonstrators were designed together with other project partners. Here, the LPS has, in addition to the horizontal feed axes, introduced the z-axis and the SMA clamping device, and thus successfully contributed to the completeness of the common machine.
References
Chechowicz, A.: Adaptive und adaptronische Optimierung von Formgedächtnisaktorsystemen für Anwendungen im Automobil, Dissertation, Fakultät Maschinenbau. Ruhr-Universität Bochum (2012)
Czechowicz, A., Langbein, S.: Adaptive resetting of SMA-actuators. J. Intell. Mater. Syst. Struct. 23(2), 127–134 (2012). SAGE. doi:10.1177/1045389X114317
Oelschläger, L.: Numerische Modellierung des Aktivierungsverhaltens von Formgedächtnisaktoren am Beispiel eines Schrittantriebes. Dissertation, Ruhr-Universität Bochum (2004)
Mussolf, A.: Formgedächtnislegierungen—Experimentelle Untersuchungen und Auf-bau adaptiver Strukturen. Dissertation, Technische Universität Berlin (2005)
Yong, L., Kohl M.: Formgedächtnis-Mikroventile mit hoher Energiedichte. Dissertation, Forschungszentrum Karlsruhe (2003)
Just, E.: Entwicklung eines Formgedächtnis Mikrogreifers. Dissertation, Forschungs-zentrum Karlsruhe—Technik und Umwelt (2001)
Schiedeck, F.: Entwicklung eines Modells für Formgedächtnislegierungen im geregelten dynamischen Betrieb. Dissertation, Leibnitz Universität Hannover (2009)
Jani, J.M., Leary, M., Subic, A., Gibson, M.A.: A review of shape memory alloy research, applications and opportunities. Mater. Des. 56, 1078–1113 (2014). doi:10.1016/j.matdes.2013.11.084
Smith, S.T., Seugling, R.M.: Sensor and actuator considerations for precision, small machines. Precis. Eng. 30(3), 245–264 (2006). doi:10.1016/j.precisioneng.2005.10.003
Pagel, K., Drossel, W.-G., Zorn, W.: Multi-functional shape-memory-actuator with guidance function. Prod. Eng. Res. Dev. 7(5), 491–496 (2013). doi:10.1007/s11740-013-0463-5
Matsunaga, T., Totsu, K., Esashi, M., Haga, Y.: Tactile display using shape memory alloy micro-coil actuator and magnetic latch mechanism. Displays 34(2), 89–94 (2013). doi:10.1016/j.displa.2013.03.001
Follador, M., Cianchetti, M., Arienti, A., Laschi, C.: A general method for the design and fabrication of shape memory alloy active spring actuators. Smart Mater. Struct. 21(11), p. 115029 (2012). doi:10.1088/0964-1726/21/11/115029
Ehrfeld, W.: Handbuch Mikrotechnik, S. 83 ff. Hanser Verlag (2002)
Altintas, Y., Verl, A., Brecher, C., Uriarte, L., Pritschow, G.: Machine tool feed drives. CIRP Ann. Manuf. Technol. 60(2), 779–796 (2011). doi:10.1016/j.cirp.2011.05.010
Yang, X.J., Li, Y., Ma, C.F., Lv, D., Zhang, J., Zhao, W.H.: A novel evaluation method on the precision of linear motor feed system in high-speed machine tools. In: MSF 836–837, pp. 220–227 (2016). doi:10.4028/www.scientific.net/MSF.836-837.220
Itagaki, H., Tsutsumi, M.: Control system design of a linear motor feed drive system using virtual friction. Precis. Eng. 38(2), 237–248 (2014). doi:10.1016/j.precisioneng.2013.09.002
Meier, H., Pollmann, J., Czechowicz, A.: Cooling strategies for a SMA wire actuator in a feed axis. In: Proceedings of SMASIS 2012, Conference on Smart Materials, Adaptive Structures and Intelligent Systems, September 2012
Meier, H., Czechowicz, A., Langbein, S.: Concepts for standardized shape memory actuators for positioning applications. In: Proceedings of SMASIS 2010, Conference for Smart Materials, Adaptive Structures and Intelligent Systems, September 2010
Meier, H., Czechowicz, A.: Computer-aided development and simulation tools for shape-memory actuators. Metall. Mater. Trans. A 2011/Band 42. Springer. ISSN: 1073-5623
Meier, H., Pollmann, J., Czechowicz, A.: Design and control strategies for SMA actuators in a feed axis for precision machine tools. Prod. Eng. Res. Dev. 7(5), 547–555 (2013). Springer
Stöckel, D.: Legierungen mit Formgedächtnis. Kontakt and Studium, expert Verlag, Bd. 259. ISBN 3816903231 (1988)
Weck, M., Brecher, C.: Werkzeugmaschinen 3—Mechatronische Systeme, Vorschubantriebe, Prozessdiagnose, S. 210 ff. Springer (2006)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2017 Springer International Publishing AG
About this chapter
Cite this chapter
Pollmann, J., Meier, H., Kuhlenkötter, B. (2017). Modular Desktop Machining Centre with SMA Actuation. In: Wulfsberg, J., Sanders, A. (eds) Small Machine Tools for Small Workpieces. Lecture Notes in Production Engineering. Springer, Cham. https://doi.org/10.1007/978-3-319-49269-8_8
Download citation
DOI: https://doi.org/10.1007/978-3-319-49269-8_8
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-49267-4
Online ISBN: 978-3-319-49269-8
eBook Packages: EngineeringEngineering (R0)