Proposal
Plastic pollution, amounting to 12 million tons annually, necessitates sustainable alternatives to single-use plastics. Compostable thermoplastic starch (TPS) films show promise but lack strength and durability compared to traditional plastics. This study employs an adaptive design of experiments (DoE) approach to enhance TPS films by optimizing testing points. The research focuses on varying concentrations of plasticizers (acetic acid and glycerol) in a water and potato starch mixture, aiming to identify the optimal ratio maximizing tensile strength and % elongation at break. Gaussian process regression (GPR) with uncertainty estimation and Bayesian optimization (BO) utilizing an acquisition function (AF) are employed. The AFs are compared to determine the best-suited one, iteratively updating the model until minimal improvement is predicted. Results indicate TPS films with 1.06% plasticizer weight exhibit the highest tensile strength (2.33 MPa) and % elongation at break (93.45%). Further testing is recommended to validate GPR and BO models and explore additional factors like adjusting the potato starch to plasticizer ratio for enhanced TPS film properties.
Recommended Citation
Marks, Theresa; White, Gracie; Lohman, Scott; and Malhotra, Mayank
(2024)
"Optimization of TPS Films Using An Adaptive Design of Experiments Approach in a Bayesian Optimization Framework,"
The Journal of Purdue Undergraduate Research:
Vol. 14, Article 7.
DOI: https://doi.org/10.7771/2158-4052.1663
