Heat is transferred one of three ways by convection, conduction, and radiation. Convection is heat transfer through liquids or gases. Conduction is heat transfer by neighboring particles or particles in direct contact of each other. Radiation is unique in that it does require a medium to transfer its energy and can be done so through a vacuum via electromagnetic waves. The easiest analogy to describe each of the different types of heat transfer is to use a boiling pot of water as an example. The heat source of course is defined as the metal coils below the pot. Convection occurs when the water in the pot gains kinetic energy due to the water below it being heated by the pot. The water is not in physical contact with the heat source, but receives energy from neighboring water molecules. Conduction is the direct transfer of heat from the source itself. The pot directly touching the metal coils is an example of conduction. The heat will transfer from the coils to an unheated pot because energy flows from high potential to low potential or high temperature to low temperature. The metal coils are an example of radiation they typically emit light in the infrared range. This occurs when the metal coils are heated and the electrons of the metal gain energy which allows them to move from different energy levels in the atom. This causes light to emit or photons to occur which is electromagnetic radiation.

Fig 1. Temperature Distribution on Motherboard

So, to first access the heat sources of the fuzzy bear an infrared camera must be used. This will allow for precise determination of components generating the most heat. An example of this can be seen in the figure below in which temperature differences are seen by the color gradients from an example circuit board .

In regards to the fuzzy bear robot, I propose to use convection to cool, or transfer heat throughout the robot in order to potentially reuse the heat for another purpose. But this will rely heavily on the amount of heat the fuzzy bear’s electrical components generate in combination with the outer surface covering which will retain heat. Again, how much heat the total system will generate determines the course of action. As stated previously, the first step is to map out the bear using the infrared camera when the bear is in different scenarios such as moving, or while on for extended periods of time. So, essentially two scenarios will be possible depending on the total heat generation of the bear. The first scenario, is the electrical components including the enclosure do not generate enough heat to be reused by a proposed device such as the Seebeck or thermocouple. If this were the case, then the proposed goal would then become to efficiently cool the bear by means of convection. Essentially, circulating the air inside the bear via a micro-fan and duct system which will allow the hot air to be carried away from the hotter electrical components via convection. This hot air can then be used in one of two ways. First it can be redirected to a specific spot on the inside of the bear to create a warming effect, or it can be expelled outside through a discrete opening composed of a thinner outer material. The cooling system would then be controlled by an infrared sensor, or sensors such as the Melexis Contact-less Infrared Sensor – MLX90614 5V. An interrupt could be coded for the infrared sensor which could monitor the temperature inside the bear. If the temperature became to hot the fan would come on, or if the bear wanted to become warm for embrace then heat could be diverted via a separate duct. So possibly two ducts, one to expel the hot air and one to divert it. The cooling system could be used with deionized water, or mineral oil which is used for computer cooling. But for liquids there is always a risk of leakage, and mineral oil is expensive. Also, the components required for liquid cooling could potentially add significant weight which has already been an issue thus far in the project. The second scenario, would mean that the bear generates enough total heat to be reused by the thermocouple, so that the energy produced could perhaps recharge its batteries or power the micro-fan to cool the bear and incite the warming embrace effect of the bear. A Peltier block works by taking two good thermally conductive plates, and then sandwiching two semiconductors between them as seen in the figure below. Essentially, heat or energy travels from high potential to low potential. The plate labeled “Heat source” receives or has the high heat potential. The cool side plate has the low heat potential. The heat or electrons flow from high to low potential, but while doing so they create a voltage potential in the n and p type semiconductor materials. This occurs because the semiconductors plates are doped with different elements giving them different potentials when a current is ran through them. That potential can then be used to drive a load such as a fan or whatever is the case.

Figure 2. Heat Source and Cool Side

So, we can exploit this fact by strategically placing the TEG to absorb heat from the fuzzy bear electrical components, and then possibly using that energy for some other purpose. The TEGs are  are lightweight have a dimension of 40mm x 40mm x 3.4mm and are cheap. So utilizing them  for this project could be possible, but the temperature generation needs to be measured and taken into account.

In conclusion, I have proposed two methods of cooling either liquid or air convection. I prefer to use the air duct cooling system because of size and weight constraints, and no potential for fluid leakage.  Once adequate temperature generation is measured a TEG could then be used to power the cooling fan or recharge batteries which is dependent upon the amount of heat generation. The first step again would be to measure the temperature generation using the infrared camera to verify the heat generation. Overall, this aspect of the project has the potential to make the bear more energy efficient.