My Fridge Experiment
The experiment I conducted was designed to analyse the change in temperature for different parts of a fridge, in order to determine the optimal location for items that I usually store in the fridge. The key findings were that there are:
Approximately 1.5 degrees celsius lower temperature between the upper and lower parts of the fridge
A noticeable difference between the front and back of the fridge
These insights will allow me to place items in the correct part of the fridge in future.
Research Question
“Is there a measurable difference in average temperature in different parts of the refrigerator over an hour?”
My motivation to explore this question is an interest in storing my items inside the fridge in the right location to ensure the optimal conditions for the particular item. A secondary interest was to investigate the ‘common sense’ notion that there is a measurable difference in temperature in certain sections of the fridge. For example, I am constantly being reminded by my partner to store meat and fish on the bottom shelf at the back of the fridge to preserve the life of those items. While this is known to be the coldest location as cold air sinks and it is the closest to the cold air source, the fridge manual makes no reference to a difference in temperature based upon the shelf of the fridge. The manufacturer claims that the fridge has a multi-flow capability which maintains a stable temperature throughout the fridge, regardless of where items are located. The time period for this experiment is approximately an hour for each section of the fridge, which produces enough readings to determine the average temperature of the fridge. The fridge stayed closed during this period and contained almost the same amount of food and drink during the experiment.
The system map below describes the behaviour of the fridge. The desired temperature is set by the fridge’s computer based upon user input. The controller maintains this temperature by increasing or decreasing the air flow within the fridge. Increasing the flow of cold air will decrease the temperature of the fridge, while reducing the flow will increase it. The temperature sensor monitors the internal temperature and communicates this temperature back to the controller in a closed loop.
In determining the research question I modelled the expected behaviour of the fridge as the sensor was moved to the various locations. The purpose of this model was to capture the hypotheses I had about the behaviour of the fridge in addition to being an easily readable representation of the mental model I had about the fridge and its behaviour.
Methods
The experiment was conducted using the Adafruit Circuit Playground Express (CPE), which has a temperature sensor as a way of recording the required data. The temperature was collected via Circuit Python code in the temp_log.py file supplied as part of the homework materials and renamed to run as Code.py . This device was selected as while the fridge has a temperature sensor, it does not measure the various sections of the fridge, nor does it record data about the change in temperature over time.
The CPE was initially placed on the top shelf at the back of the fridge for an hour. The CPE was then moved to the front of the fridge on the same shelf for a similar duration. The CPE was then moved to the middle shelf at the back, then the front of the fridge. The process was repeated with the bottom shelf. The temperature was recorded at 1-minute intervals. This approach was chosen because it covered the main ‘common sense’ judgements regarding fridges, that the coldest part of the fridge is the lower shelf as well as at the back away from the door. By this logic, the bottom shelf at the back of the fridge is the coolest part of the fridge, and the top self at the front of the fridge will be the warmest. While the experiment could have been extended to the vegetable drawer, or the compartments inside the fridge door, any data here, while interesting, would not have helped in determining where I place my items. These locations already have prescribed functions, namely for dairy, drinks and storing vegetables.
The system map below describes the expected behaviour of the fridge during the experiment. Decreasing the shelf level will decrease the temperature. Moving the CPE towards the front of the fridge will increase the temperature.
While conducting this experiment I made the following assumptions:
The fridge temperature would be unaffected by the time of day running the experiment
The CPE could be in a section longer than an hour, the important factor was in recording at least one hour of data.
The temperature of the fridge is stable across the fridge, i.e. the location left or right on a shelf is immaterial.
The CPE would record the temperature of the new section almost immediately upon being moved there, and therefore the previous temperature would not affect the quality of the data being recorded.
These assumptions were in part verified with various trail runs of placing the CPE across the fridge; recording data at different times of day and taking the CPE inside and outside the fridge. No noticeable differences were recorded in each of these scenarios.
Analysis
The results were as per the hypotheses and are articulated below:
The blue line on the y Axis represents the average temperature in degrees celsius recorded over the hour. The x Axis describes the section the CPE was placed during the recording. As the chart above shows the temperature in the fridge decreased as the CPE was placed upon lower shelves. Each time the CPE was moved from the back to the front of the fridge the temperature increased. The total difference in temperature from top to bottom was 1.55 degrees celsius. The maximum observable increase from back to front was 1.26 degrees.
While I did not encounter any significant measurement problem there were problems in getting the CPE to function as intended upon occasion. On a number of occasions I placed the CPE within the fridge, returned an hour later only to find that it had not been recording. I was unable during the experiment time to find the root cause of this issue. I re-wrote the code, decreased and increased the time intervals, but this did not impact the problem described above, which continued to occur intermittently. Therefore, if I was running this as a ‘proper analysis’ I would choose a different temperature sensor. I would also have multiple sensors located across the fridge so that variable could be removed as a factor, as well as accounting for any malfunction/deviation from the device.
Results / Discussion
The data was clear enough in my view to confirm the ‘common sense’ view that a fridge is colder at the bottom than at the top, and that you should store items towards the back if you wish them to be colder. What I did find surprising is the extent of the difference in temperature between the levels of the fridge as well as front to back. I was expecting a difference but not by 1.55 degrees. I was also not expecting that this difference between the levels would be greater than by front to back.
My system map has changed from the one presented in the methods section as the evidence gathered has led me to conclude that there may also be a material difference between the centre of the fridge and left and right. While early data testing did not indicate this and partially validated my assumption, the recorded data suggests further analyses of this should be included in any fridge sectional analysis
References
Katrina Proust & Barry Newell, CONSTRUCTING INFLUENCE DIAGRAMS & CAUSAL LOOP DIAGRAMS