(392bo) Optimal Detergent Design and Wash Cycle Configuration for Domestic Automatic Dishwashers

The consumer products market has become increasingly competitive, driven by rising consumer expectations and the effects of globalization ¹. Today’s consumers not only seek products that align with their preferences and needs ², but they also prioritize shorter delivery times, higher quality standards, and sustainable, health-oriented options. Formulated products, such as detergents, involve mixing various ingredients to achieve specific physicochemical properties ³. The functionality of the different ingredients that make up the formulation of a detergent product promotes the interaction between ingredients, which may compromise the effectiveness of the product. An example of this is found in the enzymes. This biological catalyst can be sensitive to interactions with bleaching agents, chelants, and surfactants, leading to potential decomposition or inactivation during the washing process ^4,5. Additionally, wash cycle conditions such as temperature, or duration of the wash cycle can influence the efficiency of the detergent ^6,7. This reason highlights the need to develop methodologies for detergent design where the interaction mechanisms between ingredients and washing conditions are considered simultaneously with the cleaning mechanisms involved in washing cycles.

In this work, an optimization framework it is presented based on machine learning to facilitate the integrated design of detergents together with the operating condition of dishwasher cycles. Initially, the critical stages of the washing process were identified for modelling, focusing on the primary physico-chemical principles relevant to each stage. Additionally, the drying stage was incorporated considering surface tension and contact angle models derived from experimental data. Following the characterization of these key stages, an algorithm was proposed to unify them to build the entire washing cycle. This algorithm effectively manages the events that define each stage, simultaneously predicting the physical and chemical mechanisms involved. The result is a comprehensive cleaning model that forecasts essential metrics for users, including stain removal efficiency, cycle duration, and operational costs. These cleaning metrics are optimized incorporating the integrated cleaning model into an optimization tool based on genetic algorithms in order to search better alternatives for formula composition and wash cycle configurations which are able to increase the customer satisfaction.

The effectiveness of this tool was assessed through a theoretical enhancement of a reference formula and washing cycle. The results indicate that the solution suggests a product formulation and cycle configuration that can reduce washing and drying durations and operating costs by 28%, 7%, and 58.86%, respectively, while improving cleaning performance by 4.65%. In this way, the optimization framework enhances detergent efficiency by identifying the optimal combination of ingredients and wash conditions, contributing to reduce energy consumption and supporting the Sustainable Development Agenda.

Acknowledgments

This work was supported by funding to José Enrique Roldán San Antonio under the call for predoctoral contracts USAL 2021, co-funded by Banco Santander.

We would like to thank the Procter & Gamble Newcastle Innovation Centre (UK) for providing the experimental data as well as software licenses required in the research.

References

(1) Litster, J.; Bogle, I. D. L. Smart Process Manufacturing for Formulated Products. Engineering 2019, 5 (6), 1003–1009. https://doi.org/10.1016/j.eng.2019.02.014.

(2) Tijskens, P.; Schouten, R. Modeling Quality Attributes and Quality-Related Product Properties. In Postharvest handling; Elsevier, 2022; pp 99–133. https://doi.org/10.1016/B978-0-12-822845-6.00004-X.

(3) Mladenovic, N. Pooling Problem: Alternate Formulations and Solution Methods. Manage. Sci. 2004, 50 (6). https://doi.org/10.1287/mnsc.1030.0207.

(4) Broze, G. Handbook of Detergents, Part A: Properties; CRC Press, 1999.

(5) Schick, M. J.; Hubbard, A. T. Liquids Detergent; CRC Press, Ed.; 2005.

(6) Johansson, I.; Somasundaran, P. Handbook for Cleaning/Decontamination of Surfaces; Elsevier, 2007.

(7) Santos, A. M. P.; Oliveira, M. G.; Maugeri, F. Modelling Thermal Stability and Activity of Free and Immobilized Enzymes as a Novel Tool for Enzyme Reactor Design. Bioresour. Technol. 2007, 98 (16), 3142–3148. https://doi.org/10.1016/j.biortech.2006.10.035.