Date on Master's Thesis/Doctoral Dissertation


Document Type

Master's Thesis

Degree Name



Mechanical Engineering

Degree Program

Mechanical Engineering, MS

Committee Chair

Brehob, Ellen

Committee Co-Chair (if applicable)

Kelecy, Andrea

Committee Member

Kelecy, Andrea

Committee Member

Segura, Luis

Author's Keywords

heat transfer; baking; thermal energy; linear regression; browning


Residential ovens are complicated thermal environments capable of delivering convection, conduction, and radiation heat transfer to food. The amount and mode of heat transfer can change based on the design of the oven cavity, cooling systems, and oven cycle algorithms. Studies have shown that the changes in one heat transfer mode can have an impact on the quality of baked goods. The many variables involved make designing residential ovens a time consuming and costly process. The goal of this study is to adapt thermal energy sensing technology to collect energy data from a residential oven and develop correlations to quality characteristics of baked cakes. The quality characteristics measured in this study are browning, porosity, rise height, and mass loss. Testing was limited to white cakes baked in a traditional bake mode with no convection fan. The correlation models developed should reliably predict the final quality measurements of baked cakes using thermal energy data collected in an oven cycle of matching parameters. The sensors used in this study can detect incoming thermal energy and break it up into component heat transfer modes of conduction, convection, and radiation. The sensors were placed in small cake tins to match the conditions of baking cakes. A test plan was performed to collect thermal energy data at a variety of oven temperature setpoints, rack positions, and bake times. Each test run was repeated, replacing the sensors with cakes, and the quality characteristics of the baked cakes were recorded. The thermal energy sensors performed well, splitting energy absorbed into conduction, convection, and radiation components. The thermal data was repeatable, showing less than two percent variation after ten minutes of testing on the middle rack at any given temperature setpoint. It was confirmed from previous studies that conduction dominates a traditional bake mode, contributing about 59% of the total energy absorbed. Second is radiation and finally convection, due to the lack of an active convection fan during a traditional bake mode. It was found that top and bottom browning L* color values correlated linearly with radiation and conduction energy absorbed, respectively. The coefficient of determination (R2) for these models was 0.97 for bottom browning and 0.95 for top browning. Validation data was limited, so further testing would improve the confidence of these models. The porosity and rise height metrics were not able to be correlated with the data collected. It is possible that the measurement methods of these cake quality metrics were not robust enough for reliable data, or the thermal energy data collected was not relevant to changes in porosity or rise. Future testing could investigate alternate thermal data to determine what correlative factors can be used to predict porosity and rise. Finally, mass loss was well correlated to a polynomial regression of total energy with a 0.98 R2 value.