We had a virtual progress meeting with Dr Ho over Zoom, where we presented our first prototype. We discussed the various reasons why the temperature of the cold air exiting the set-up was insufficiently cold. This was largely due to the fact that our initial set-up made use of heat sinks, which required the air to come into direct contact with the heat sink. We thus began looking into other ways of cooling the air more efficiently. Below are a list of reasons and potential solutions that we discussed.

1. There was insufficient time for the air to come into contact with the cold surface, leading to the air not being cooled to a sufficient extent before exiting the set-up
   • We considered passing the air through a copper pipe, with the copper pipe being in direct contact with the Peltier module. This increases the distance that the air has to travel, allowing it to be in contact with the cold surface longer and thus be cooled to a larger extent before being blown out of the set-up. However, this was ultimately rejected because the pressure needed to keep the air inside the pipe flowing at an appreciable speed is quite high, and cannot be achieved simply by attaching fans to the ends of the pipe to direct and control the air flow.
   • We contemplated using evaporative cooling, where the air is passed through a wet sponge. Energy from the air would be used to evaporate water in the sponge, thus decreasing the temperature of the moving air. This idea seemed promising as water has a high specific heat capacity, making it an efficient method to decrease air temperature. Unfortunately, as we researched on the types of sponges that were commonly used for evaporative cooling, we found that we needed porous sponges to allow for the air to flow through. These sponges are not readily available and cannot be adequately substituted with kitchen sponges commonly found in our local supermarkets.
2. Our dissipation of rejected heat was inefficient, resulting in the temperature of both the cold and hot side of the Peltier being high
   • We explored the idea of water-cooling for the hot side of the Peltier, where the heat sink is submerged in a reservoir of water and the water is circulated around the container using a pump. This solution posed a few concerns for us as this required waterproofing the set-up so as to separate the circuitry from the water.
3. There was insufficient cooling capacity to match the volumetric airflow of the fans
   • We discussed using more Peltier modules to allow for greater cooling capacity. This idea meant that we would require a much larger power consumption, making our product less energy consumption friendly than we initially hoped for. From our initial set-up, each heat sink could only be attached to one Peltier due to size constraints of the area for thermal contact. Increasing the number of Peltier modules used meant that we might need to increase the number of heat sinks used in our product, making the product not only larger but also very much heavier.
   • We also looked into other Peltier modules, such as the TEC1-12715 module which allows for higher current rating. This meant that the amount of power passing through the Peltier modules would be larger, increasing the cooling capacity and therefore allowing for lower temperatures to be reached.

 

After much discussion and research, we decided to use radiative cooling to cool the air.

We headed down to Sim Lim Square on Thursday to buy various secondhand fans and CPU heatsinks of different sizes and shapes so that we could test the different types and see if they corroborated with our theoretical simulations. From our set-up, we expected a decrease in temperature of ambient air to cold air by about 6°C.

The next day, we did the set-up according to the preliminary design from our COMSOL simulations and tested the setup both inside the MnT lab and outside.

 Fig.1: Prototypical set-up

 

On that day, the MnT lab had an ambient temperature of 24°C while the outdoor area was at 29°C. We measured the temperature of the air exiting our model on both the hot side and the cool side, and compared the temperature change between the exiting air and ambient temperatures. We found that temperatures decreased by 1.5°C when we took the set-up outside, indicating that using one Peltier plate to cool the air was insufficient. The draft from the fans was also inadequate. Based on our calculations, an estimated 137W cooling capacity is required to cool 40cfm of air by 6°C. This requires more than 4 TEC1-12706 modules and a much smaller volumetric capacity than our fans were already providing.

After this, we carried out some minor experiments where we varied the voltage and current of the Peltier plate to see the effect on the temperature of the hot and cold side of the plate. From this, we were able to optimise the voltage and current at 12V and 4A.

During this week, our group decided on the project that we wanted to focus on and conducted market research to better understand the products available. We read up on cooling methods and in particular, the Peltier effect used in thermoelectric coolers.

Bing Hong and Deborah ran some simulations using COMSOL to try various set-ups involving one Peltier plate, two heat sinks and two fans, one set for each side of the Peltier plate. From the results, we understood that we needed pin heat sinks to maximise surface area of the heat sink in contact with the air and the amount of volumetric flow through the fan that would be optimum for heat dissipation. We also optimised the set-up of items to allow for maximum cooling of air flowing through the model.

Justin and Zhen Xuan headed down to the MnT lab to check the items that were already present in the MnT lab and discuss with Tony the feasibility of our idea.