Sample-efficient parameter exploration of the powder film drying process using experiment-based Bayesian optimization

Abstract

Parameter optimization is a long-standing challenge in various production processes. Particularly, powder film forming processes entail multiscale and multiphysical phenomena, each of which is usually controlled by a combination of several parameters. Therefore, it is difficult to optimize the parameters either by numerical-model-based analysis or by “brute force” experiment-based exploration. In this study, we focus on a Bayesian optimization method that has led to breakthroughs in materials informatics. Specifically, we apply this method to exploration of production-process-parameter for the powder film forming process. To this end, a slurry containing a powder, polymer, and solvent was dropped, the drying temperature and time were controlled as parameters to be explored, and the uniformity of the fabricated film was evaluated. Using this experiment-based Bayesian optimization system, we searched for the optimal parameters among 32,768 (85) parameter sets to minimize defects. This optimization converged at 40 experiments, which is a substantially smaller number than that observed in brute-force exploration and traditional design-of-experiments methods. Furthermore, we inferred the mechanism corresponding to the unknown drying conditions discovered in the parameter exploration that resulted in uniform film formation. This demonstrates that a data-driven approach leads to high-throughput exploration and the discovery of novel parameters, which inspire further research.

Automated odor-blending with one-pot Bayesian optimization

“…This would enable humans to understand these hazardous odors and dramatically reduce the number of accidents based on hazardous odors such as natural gas leaked from pipelines 5 and volcanic gases. 6 To establish an automated odor-blending system, we need to develop three elements: an olfactory sensor [7][8][9][10][11][12][13] to replace the human sense of odors, blackbox optimization [14][15][16] to determine the amount of blending, and a robotic system [17][18][19][20][21][22][23][24][25] to perform actual blending of samples.…”

Bayesian optimization of radical polymerization reactions in a flow synthesis system

“…License: CC BY-NC-ND 4.0 successfully applied BO to the identification of processing conditions that minimized the formation of defects in powder films while carrying out a limited number of experiments. 31 Given that BO is predicated on the acquisition of experimental data, numerous experiments are typically required for method validation and so it is advantageous to employ experimental techniques capable of generating abundant, highly reproducible data. In this context, high-throughput methods for materials exploration [32][33][34][35] and analysis [36][37][38] are increasingly being integrated with machine learning algorithms.…”

Exploring Metal Nanocluster Catalysts for Ammonia Synthesis Using Informatics Methods: A Concerted Effort of Bayesian Optimization, Swarm Intelligence, and First-Principles Computation

“…Nugraha et al 55 demonstrated that BO can efficiently aid in designing experiments to discover the optimal composition of metal precursors, yielding mesoporous ternary metal PtPdAu alloys with enhanced electrocatalytic activity in methanol oxidation. Nagai et al 56 applied BO to the drying process of catalyst inks for polymer electrolyte fuel cells to determine optimal drying conditions with a small number of trials. Okazawa et al 19 used a combination of BO and DFT calculations to find the optimal binary alloy catalyst for the nitrogen activation reaction.…”