Idea: Ice on Tracks
Hypothesis:
- The idea was inspired by sports on ice (ice skating, luge, etc.), and used the same mechanism (very low friction sliding between ice and metal).
- The very low friction greatly reduces energy loss, conserving most of the energy for the movement of the vehicle.
After testing:
- Rapid conduction of heat from the metal track to the ice block caused rapid melting at the points of contact. This resulted in major grooving of the ice block during travel.
- Lack of viable refreezing mechanism
- Peltier module: to produce sufficient temperature difference, the module itself heats up too much, which means another cooling unit is required to cool down the peltier module; the size and mass of the total set-up are too large for our small-scale setups.
- Dry ice: see below
- Liquid nitrogen: see below
Insights:
- Without refreezing mechanism, this idea is not particularly sustainable.
- Not feasible with constant contact at same area and hence we were led to our next idea of segmented tracks.
Idea: Segmented Tracks
Hypothesis:
- Melting and grooving of ice occurs only when the ice block is in contact with the metal tracks.
- Segmenting tracks allows the points of contact to change rapidly and therefore reduce the melting at any part of the ice, reducing the formation of grooves.
After testing:
- 10cm segments were still too long and caused grooving significant enough to hinder transition of ice block from one segment to the next. This hindrance greatly reduced the speed of the ice block.
Insights:
- The transition between segments were too significant even for a remarkably short distance and hence this idea is not feasible unless the tracks are cooled which would take a significantly more energy which we hope to be able to conserve.
Idea: Flat Tracks
Hypothesis:
- Even pressure and contact on all parts would eliminate grooving.
- Grooving introduces more planes of contact; eliminating grooving might reduce friction.
After testing:
- Lack of groove causes the ice to travel in a non-linear fashion, thus longer distance and longer travel time.
- When testing with a groove in the center for stability and to direct the ice, melted water on the tracks still created an adhesive effect on the ice block, slowing it down significantly.
Insights:
- This method requires constant drying of the tracks to work well and to do this, we need to heat the tracks. However, heating will cause the ice to melt even faster and this was not ideal. Hence, we realised that this was not a viable idea.
Idea: Starting Mechanism (elastics)
Hypothesis:
- When deformed, elastic materials converts most of the energy used to deform them as elastic potential energy. When properly designed, this potential energy can be reconverted into kinetic energy of the system.
- The consistency of elastic materials also means that by standardising the amount of deformation, we can standardise the amount of energy used to push off the ice block in each experiment, allowing us to evaluate the effects of the other variables on the movement of the ice block.
Reason we abandon idea:
- We are no longer using ice blocks, and our transport system is also no longer powered only by an initial start-off mechanism. Thus this idea is not relevant anymore.
Insights:
- To make this system effective, we need a long distance of spring with low spring constant to conserve as much energy as possible and to start and stop the ice smoothly. Furthermore, we need stiff backing that can provide the force to hold the spring in place without collapsing or moving backwards as that will further contribute to a loss in energy.
Idea: Hybrid Ice (Boiled and Pykrete)
Hypothesis:
- As the ice blocks we had been using often break upon impact, we decided to attempt making ice using boiled water and using pykrete for our ice blocks.
- Boiled water: boiling would purge tap water of air and reduce the formation of cavities in the ice block.
- Pykrete: the numerous layers of tissue paper would act as barriers of cracks, preventing any large cracks that would break the ice block.
After testing:
- Boiled water: cavities due to air was insignificant and ice blocks were still just as brittle.
- Pykrete: while the ice became nearly impossible to break upon impact, the ice was no longer smooth and could not slide along the tracks.
Insights:
- We realise that air bubbles in freezing normal water was less significant than we expected and that there is a trade-off between strength of ice and low friction for pykrete.
Idea: Hybrid Ice (Liquid Nitrogen and Dry Ice)
Hypothesis:
- To reduce the melting of ice, we can use cold materials to absorb the heat from the ice and the environment. This can also help refreeze the ice.
- Dry ice and liquid nitrogen may be able to imitate the Leidenfrost effect to allow the vehicle to “float” and reduce the friction significantly.
After testing:
- Dry ice
- Poor heat conduction through ice hence using dry ice cannot maintain a significantly thick layer of ice.
- Ice frozen with dry ice weakens the structure even more with air bubbles.
- Liquid Nitrogen
- Flows too any cracks too quickly to be trapped and to imitate the Leidenfrost effect properly because each droplet experiencing its own Leidenfrost effect.
- To allow significant amount of liquid nitrogen through the ice, the slits in the ice further weakens the ice structure, causing it to break into two (or more) pieces as soon as the mold is removed.
- Mold was difficult to remove even with methods attempting to rectify this problem. Of the three methods we used, the taped mold was the easiest to remove, followed by aluminium wrapped and lastly cling wrapped.
Insights:
- Liquid nitrogen is unable to provide any sort of lift to the ice at all and also makes the ice more brittle when it is extremely cold. Dry ice, however, is able to provide some form of lift and we decided that dry ice would be more feasible as a base as using dry ice, we would remove many complications that came with ice such as the longevity of the ice and the grooving due to heat conduction of metal.
Idea: Hardware for Dry Ice Hovercraft
Experience:
- Importance of checking for all the parts we need in advance
- If we use remote controller directly, ESC needs to be able to power receiver and will need a BEC
- Can look up or consider alternatives such as 4-in-1 or 2-in-1 ESCs
- Connectors need to match or need to purchase connectors necessary before assembly
- If using external controller on top of remote controller such as flight controller or even Arduino, need suitable remote controller (either with PPM or need a PPM encoder)
Idea: 3D Printing Vehicle Design
Hypothesis:
- Need to be able to secure the dry ice firmly with rods and with specific dimensions.
- Need to insert the motors in orientations different from the usual drone.
- 3D printing allows us to design and have items specific to our requirements.
After testing:
- Able to put together 2 vehicles – one for only one dry ice and one that can house two.
- Although we incorporated offsets in our designs to allow fitting, each part fitted a little differently, some more tightly than others, etc.
Insights:
- Each printer has its own error margin and as such, we should try to print fitting parts in one sitting or at least with the same printer as much as possible.
Idea: Using Flight Controller to Control Motors
Hypothesis:
- Unable to control the four motors with one channel each as it was not intuitive hence the need for a controller to control the ESC instead of the remote control directly.
- Flight controller (such as Pixhawk) can control more than one motor with one channel depending on the program uploaded to its software and might be applicable to our use.
After testing:
- Most of the programmes in the software we used (Mission Planner) had settings that were not applicable to us:
- Drone – we were able to make the motors turn but because of the settings, even a slightest tilt of angle in the ice causes the motors to try to “fix” it by having one spin at higher rate than the other which instead caused our vehicle to turn despite commands to move straight.
- Rover – the motors we were using were drone motors and this mode was unable to detect the correct motors and could not work.
- Submarine – the remote controller we used were not suitable for submarine and the programme could not detect our remote controller.
Insights:
- We needed to be able to programme the Pixhawk directly to control our motors accordingly but with limited guidance and under tight time constrain, we decided to try simpler methods.
Idea: Using Arduino UNO to Control Motors
Hypothesis:
- Arduino can be programmed directly by us to read data from the remote controller and send instructions to the ESC accordingly.
After testing:
- Arduino is able to move the motors according to what we require.
- However, the ground that is smooth enough for vehicle to start is often tiled or has rough patches, resulting in the vehicle turning when it shouldn’t as the motors cannot adjust accordingly.
Insights:
- Because Arduino alone cannot sense the direction the vehicle is moving and counter it accordingly such that it moves straight when we want it to regardless of external stimulus, in future work, we could either install a sensor into the Arduino or attempt to programme a Pixhawk programme from scratch.