Frekans cevabının iyileştirilmesi ve taşınabilecek yükün artırılması için şekil hafızalı alaşımların demet olarak kullanılması

Shape memory alloys (SMA), as the name implies, memorize their original shape, which is given at about 700°C. Severe deformations at low temperatures remain until alloy is heated up. In return of heating, alloy remembers and gets its original, pre-deformed shape. There are two main phases, called martensite and ostenite, causing shape changes of the alloy. Transition between these two phases happens based on temperature changes. Martensite phase occurs when material is cold and that is the phase where deformation takes place. Heating leads material to alter from martensite to ostenite. Then the alloy goes back to its pre-deformed shape by relocating its crystal structure with heat. So, heat is the way of managing of SMA phase transitions which is created either by an outside heat source or by current passing throughout the material. Thus, it is feasible to control position of SMA by controlling the current. In recent years, the need of lightweight and small actuation technology leads to shape memory alloys to put on the table that have quite promising properties such as high power to mass ratio, easy to establish and noise free working. They provide the opportunity to take conventional actuators' place in many applications; thus encouraging the development of advanced actuators with a noticeable decrease of cost and complexity. While targeting this, there are some drawbacks, such as poor efficiency, low weight lifting capabilities and poor frequency response. In addition, control of SMA becomes a challenge because of hysteresis behavior of the alloy. Handling couple of SMA drawbacks of it usage as an actuator or an artificial muscle, such as increasing weight lifting capability and improving frequency response, are the main aims of this research. There are two approaches for a heavier payload that SMA can lift: the easiest way is to use a wider wire. Since thick SMA wire requires more time for cooling after actuation, the wire will be ready seconds afterwards for the next actuation. Consequently, increasing the diameter of the SMA wire results in a good weight lifting capability, in contrast it makes frequency response even worse. Other approach of improving the payload capability, which is followed in this research, is to bundle an increased the number of SMA wire. The more SMA wire means the more weight lifting capability and the wider surface area. And the fruit of using wider surface area is a shorter cooling time, which results in a better frequency response. All in all, by increasing the number of wire, it is subjected in this research that to get a faster cooling time with the same payload capabilities with a wider SMA wire. There are three main parts of the experimental setup used in this work: electrical, mechanical and software parts. Electrical part consists of two SMA wire drivers which were already built for another research and a data acquisition card (DAQ). Wire driver is responsible from sending current to the wire based on the reference voltage received from DAQ card. In order the current to follow the reference value, a PI control was implemented. Controller gains are defined to provide 1 amperes output in return for a 10 volts input. The driver has a closed loop so the output, which is SMA wire actuation current, that follow the reference. However, the loop is closed within the driver and it is not sending back any current feedback to software. Instead, wire driver sends voltage drop value on the SMA wire to data acquisition card as a feedback. Since current value is already known, which is 1/10 of the reference value, it is easy to calculate wire resistance in software by dividing voltage drop value to current value. DAQ card is responsible from acquiring reference voltage data from software and send them to SMA wire drivers. In addition, it delivers voltage drop value from driver card to software. Mechanical part comprises SMA wires lifting a load (150 gr) and a LVDT (Linear Variable Differential Transformer) to measure displacement of load. Load is placed on a vertical linear guide and core of LVDT is mounted on load. When SMA wire is actuated, it causes the load to displace which in turn leads to core displacement so that LVDT can measure and sends the position data to software. Respectively, two SMA wires with 76 µm diameter, one wire with 100 µm diameter and one wire with 150 µm diameter is screwed carrying same load, 150 gr. Wire length is same for three wires, 30 cm. Software part built within LabVIEW, using graphical interface, and has two analog outputs: powering up LVDT and sending reference voltage value to the SMA wire driver via DAQ card and two analog inputs: reading position and voltage drop value of SMA. By utilizing these analog inputs and outputs, a sine wave signal at different frequencies (between 0.1 Hz and 7 Hz) is sent as reference signal to the driver. And Bode graph is plotted for three wires with different diameter to compare their frequency responses. Input signal magnitudes are chosen based on the approximate current value for 1 second given in the manufacturer website. Open loop and closed loop tests are carried out and open loop results show that, - Increasing the diameter of the wire results in a slow actuation, since the wire needs more time to cool down. - For heavier payloads, gain is increasing which means displacement of load becomes longer. However, as a drawback, the heavier the load leads to the larger delays. - An SMA wire with 76 µm diameter can carry 80 gr load. Doubling the number of the wire increases the payload lifting capability to 160 gr which is higher than 100 µm wire's weight lifting capability. On the other hand, 76 µm wire needs 0.8 second to cool down which is 72% of cooling time of 100 µm wire and 40% cooling time of 150 µm wire. Even 24 µm widening of the wire results in about 30% slower cooling time, which is more obvious at relatively high frequencies. - Phase delay is 180 degree at 5 Hz for 76 µm wire, 2 Hz for 100 µm wire under 150 gr loading. For the same load, 150 µm wire displaces the load 0.05 mm at 1.2 Hz. Since the wires are twofold, the variation of wire length is actually double of these displacements. Bearing mind that the narrow bandwidth of SMA wires, widening the diameter of SMA wire ends up with a drastically reduced bandwidth. - For a 0.2 Hz input signal, displacement of the load, which is 150 gr, is 4.3 mm for 76 µm wire, 5 mm for 100 µm wire and 0.8 mm for 150 µm wire. Using about 300 gr payload instead of 150 gr leads 150 µm SMA wire to displace 1.6 mm instead of 0.8 mm. - Two SMA wires with 76 µm diameters have the same payload lifting capability with one SMA wire with 100 µm diameter. In addition, as mentioned before, two 76 µm SMA wires are about 30% faster than 100 µm SMA wire. Additionally, in theory, it requires two more SMA wires with 76 µm to have the same weight lifting capability with one SMA wire with 150 µm diameter. On the other hand, outcomes of the closing the control loop are; - Controller makes the SMA wire contract faster. Since the controller first starts with maximum current to contract the wire rapidly, then decreases the actuation current while the error becomes smaller. Because of the fact that, no current flows during cooling cycle; closed loop control has no influence on cooling. - In open loop experiments, at 0.2 hertz, 76 µm and 100 µm wires respond to input signal with about 50 degrees of delay. 150 µm wire has a delay about 70 degrees. Closed loop control is eliminated these delays and the all three wires can follow the sine input without any delay at 0.2 hertz. - Phase delays occuring at 0.2 hertz in case of open loop experiments takes place at about 2 hertz for the closed loop controlled experiments. - Absolute gain is less than 0.5 for 76 µm and 100 µm wires and almost zero for 150 µm at 1 hertz. At the same frequency, it is about 1 for 76 µm and 100 µm wires and 0.3 for 150 µm at 1 hertz. It can be concluded that, closed loop control of the wires is quite beneficial for both pursuing input signal and displacement of the load. - Bundling the wires is very effective to accelerate the response and makes evident difference in open loop experiments. - In closed loop experiments, for small increase in diameter of the wire, bundling does not differ much. However, in case of 76 µm and 150 µm wires, instead of doubling diameter of the wire, it would be more useful, especially for longer displacement at high frequencies, to use 4x76 µm wires in parallel to carry the same load. Since 150 µm wire has a 320 gr payload capability and it is 80 gr for the 76 µm wire. - Payload has an effect on displacement in open loop experiments but closed loop controlled wire does almost not effected from the load. Usage of SMA wires in biomedical applications such as artificial muscle, instead of using a thicker wire, it is the most efficient way to bundle the increased the number of thinner wire in order to obtain a faster and powerful muscle. Mechanically parallel bundled SMA wires are more effective than one wire bundled using pulleys, that is, mechanically serial mounted. Since, in case of a damage of a wire does not make the whole actuator or muscle out of use.