Improvements made since the break and what's to come
February 17, 2024 - By Hedda Grelz, Andres Calderon, Oselumenosen Ekhuemelo, and Adolfo Diaz
__________________________________________________________________________________
For the work period February 4th to February 17th, Team Navy has made significant progress in both machining the vent and conducting tests of PWM parameters that will be used for the active heating of the SMA wire. The team worked with Bryan Flores in the MECE printing lab to print out two parts of the vent so the testing of the vent could start earlier. The first 3D print of the vent case (with pulley system) and the vent lid are shown in Figure 1. The vent part (that will move up and down) has to be machined with low carbon steel as the weight of this vent is crucial for the SMA/spring analysis the team conducted and switching the material to ASA for this part would create issues in the testing. Note that the finalized vent will be 100% machined out of low-carbon steel, but the 3D-printed parts will be used to test the mechanism earlier.
Figure 1: First 3D print of the vent case, pulley system, and vent lid.
The initial print of the part encountered several issues, primarily related to thread tolerance, non-cylindrical extruded parts in the pulley system, and flaws in the guiding pins, complicating the assembly process. To resolve these challenges, the team has modified the thread tolerance and decided to print the pulley system separately for subsequent press-fitting into the vent case. Additionally, efforts are underway to sand down the guiding pins to ensure uniformity, facilitating the spring fittings. An adjusted CAD has been submitted to Bryan, and the team anticipates receiving a second iteration of the print by the middle of next week
The team has also worked on the active opening of the vent. After a discussion with Dr. Song and Dr. Agrawal, it has been determined that the team should not need a feedback control system for position stability since the vent will only have two states: open and closed. This simplifies the control system component of the project as the team will just use PWM open loop control. In this past work period, the team figured out how to perform PWM to heat the SMA wire using an Arduino, a variable power supply, and a DC/DC solid-state relay (SSR). The setup for wiring all components together is shown in Figure 2.
Figure 2: Experimental setup for duty cycle testing on the SMA wire.
The team was originally planning on using a Raspberry Pi, but with no feedback control needed, the team started testing with Arduino since it’s more user-friendly. If there is a need and time later in the semester to implement a more extensive control mechanism using the Raspberry Pi, the team will do so. The Arduino code written can vary the duty cycle and the frequency of the PWM which is then sent to the solid state relay (SSR), which acts as a decision point between the microcontroller and the SMA wire. If the input falls within an acceptable range for the SSR, then the current from the variable power supply heats the wire. The SSR is needed since the Arduino cannot send out the required voltage to heat the wire. The team has started initial experiments to determine ideal voltage values and PWM parameters. The results from the first testing session are displayed in Table 1. In this first experiment. The voltage was held constant at 5 V and the PWM frequency was set to 10 Hz from recommendation by Dr. Song.
Table 1: Results from the first duty cycle experiments with constant frequency (10 Hz) and constant voltage (5 V)
As seen above, when the voltage and frequency are held constant there is a clear inverse correlation between the duty cycle and the SMA response time; as the duty cycle increases, the response time reduces. In each iteration of the test, the SMA was able to lift the weights used in the experimental setup. However, as the duty cycle reached ~50%, there was a 1 [mm] increase in SMA-generated displacement. This shows that the lower duty cycles might be too small to contract the wire fully. For the higher duty cycle values, the actuation time has some jumps and does not increase with the same pattern as for the lower duty cycles. This discrepancy can be traced to human error, as the obtained values also depended on the testers’ reaction time and reflex capabilities. The team will continue testing next week to obtain more accurate results and test more voltages and duty cycle combinations to find the values needed to open the vent and then hold it. The most important thing with this first test was to see that all the electrical connections and the Arduino code worked to heat the wire.
Machining the parts has been the most tedious task, as it is hard to reach a common ground where the machinist can fabricate the parts while still keeping the integrity of the venting mechanism. The team has been in close contact with Jamar Murray (machinist) to ensure that the final product comes out as intended. The last blog talked about changes in thread sizes which would facilitate the fabrication of the assembly method. After having another meeting with the machinist more changes were made, these are displayed in Figure 3.
Figure 3: Updates to vent case and lid according to machinist meeting.
Since the machinist’s experience with machining threads in small components is limited along with the equipment at hand, he made some suggestions to facilitate the process. One of the modifications was getting rid of the threads completely. Instead, the new assembly method will be using set screws. This changes the design of the vent case and the vent lid. In Figure 3, the original parts are labeled as (a), and the modified parts are labeled as (b). The difference is clear as the modified parts are more robust than the original ones. This is because to be able to use set screws the walls needed to be thicker to have enough space for the set screws to be able to screw in and create a tight assembly. The set screws will be on the outer diameter of the vent lid reaching through to the inside diameter of the vent case. These holes will be machined by hand with a drill bit once all parts have been machined. The set screws will be flush with the outer diameter so they will not hinder the application onto a LIB battery pack. The only change this modification causes is an increase in the outer diameter of the mechanism from 5 cm to 6.1 cm which is still below the suggested standard of 7 cm diameter. Another subject discussed in the meeting was machining the vent part ahead of schedule or prioritizing it when working on our project. As said earlier on the blog, having the vent made out of low-carbon steel is essential for effective validation testing which is why this is important.
Over the next two weeks, the team will continue testing the control system and refining the PWM parameters such as duty cycle and applied voltage to ensure optimal response time, maximum opening, and positional stability when opening the vent. Additionally, we will use a displacement sensor (linear potentiometer) to investigate the integrity of the wire and stability of the vent, by collecting displacement readings from the linear potentiometer under different duty cycles to get more precise readings. The team is also going to implement a temperature sensor as the triggering mechanism for the control system. Team Navy is currently carrying out trials and researching to understand the temperature sensors (DS18B20) capabilities with Arduino, and the ensuing correct integration with the SSR. Since the DS18B20 is functionally and visually similar to a thermocouple, the team plans on leveraging their background in thermocouple use during a related University course “Experimental Methods”.
The team also hopes to have the battery case for demonstration purposes done within the next period of work. This should not present too much of a challenge as it is simply assembling a box using the UHMW polyethylene sheets ordered earlier in the year along with some aluminum angles and bolts. Inside the box, the team also plans on applying some high-temperature silicone in the edges to ensure a completely closed system so that no heat escapes through the crevices of the joints. Lastly, the team is communicating more with Bryan Flores to get the 3D-printed parts ready for testing of the device.
There are a couple of things that could create obstacles for the team in the next work period. Once the team has performed more testing on the PWM parameters and made a decision on the voltage, duty cycle, and frequency for both opening and holding the vent open, a temperature sensor will be integrated to trigger the active opening. The team needs to figure out how to measure the temperature rate of change using Arduino and only trigger the PWM when this rate of change value reaches 25 °C/min (or 0.42 °C/s). The code then needs to hold the set duty cycle while opening the vent, and then lower to another duty cycle to hold the vent open until the temperature sensor senses that the temperature has decreased to 50 °C or lower. Integrating the temperature sensor might cause an obstacle since it complicates the code and wiring needed for the active component. It could also pose a challenge if it does not work well with Arduino. If this is the case, the team will have to switch over to Raspberry Pi, which is not as user-friendly. The issues the team had with Rasberry Pi before was that it was difficult to connect to MATLAB and kept giving errors for unknown reasons. This could present a major roadblock if the team is not able to properly download the appropriate software to code the temperature sensor. However, from looking online the DS18B20 is known to be compatible with Arduino so it will most likely not be necessary to use Rasberry Pi for this reason. The team has already looked up many tutorials and read papers on how to use temperature sensors with a microcontroller and managed to connect a displacement sensor to Arduino in this past working period. As long as the team dedicates enough time to figuring out how to connect and code the sensor readings/triggering, it should not cause major issues. Dr. Song and the students in his lab have also offered to help if Team Navy runs into any issues.
Another issue that the team anticipates could come up in the next couple of weeks has to do with the spring resetting mechanism that will work to close the vent after activation when the battery has cooled down. The team did an analysis to determine the spring force needed to reset the SMA wire and four steel springs were selected accordingly. However, since the analysis had a small margin of error, the team is worried that the four springs might not provide enough force when testing the vent. To prevent having to re-machine the entire vent if this happens, the team is planning on integrating extra guiding pins such that extra springs can be easily implemented in the vent if the spring force is not strong enough to close the vent.
Lastly, the team is still brainstorming for an effective way to attach the SMA wire to the vent and through the pulley system. Some ideas are screwing it to the bottom side of the vent or feeding a wire through the vent to avoid having to use alligator clamps. The reason for this is that they would create hotspots that would fatigue the wire faster and would just not be as efficient as having a current flow through the wire seamlessly within the same pathline. Once the vent has been machined and the team has the physical part, it will be easier to work out a method of attachment since it will be easier to visualize how different integration methods would affect the part. Since the machinist will prioritize the vent part first, this trial-and-error process of attaching the SMA wire should have enough time to be completed before validation testing.
Comments