(374e) Material Flow Optimization of Acetaminophen Production Network Via Hybrid Mechanistic Machine Learning Approach to Minimize Waste Generation

The growing demand for acetaminophen, a widely used analgesic and antipyretic drug, has led to a substantial increase in its production rate in the US. This growth poses significant environmental challenges, primarily concerning waste generation and resource utilization. This study presents a novel approach to mapping and optimizing the material flow of the acetaminophen production network, consisting of 6 industrial systems classified into 4 NAICS sector, using a hybrid mechanistic machine learning (ML) approach that effectively minimizes waste generation and enhances sustainability.¹

The research methodology focuses on the development of a comprehensive material flow map, identifying and quantifying the inputs, outputs, and interconnections of each industrial systems involved in the acetaminophen production network.² By utilizing Physical Supply & Use Tables (PSUTs), the material exchanges and interdependencies within the network are systematically quantified, enabling a thorough assessment of its performance and waste generation patterns. Furthermore, this material flow quantification offers valuable insights into potential areas for enhancement, such as process inefficiencies, bottlenecks, and waste reduction opportunities.³

The material flow optimization is proposed to be achieved through the application of a hybrid data-driven machine learning technique that employs the use of Sparse Identification of Nonlinear Dynamics (SINDy), a system identification and inverse problem-solving algorithm, on the mechanistic process flow model of the 6 industries.^4,5 This technique facilitates the identification of optimal operating conditions for each process stage, minimizing waste generation while maintaining desired production levels. The proposed optimization framework adopts a multi-objective approach, balancing economic, environmental, and social factors to ensure a sustainable acetaminophen production network.

The outcomes of this study will highlight the efficacy of the material flow mapping and optimization methodology in significantly reducing waste generation in the acetaminophen production network. By implementing the system-identified improvements, the acetaminophen industry can enhance its sustainability, mitigate its environmental impact, and promote responsible resource management. This research, therefore, contributes to the broader scientific discourse on sustainable industrial production and resource management and can be adapted for application in other chemical production networks as well.

References:

1. Tsolakis, N. & Srai, J. S. Mapping supply dynamics in renewable feedstock enabled industries: A systems theory perspective on ‘green’ pharmaceuticals. Oper. Manag. Res. 11, 83–104 (2018).
2. Vunnava, V. S. G. & Singh, S. Integrated mechanistic engineering models and macroeconomic input–output approach to model physical economy for evaluating the impact of transition to a circular economy. Energy Environ. Sci. 14, 5017–5034 (2021).
3. Fischer‐Kowalski, M. et al. Methodology and Indicators of Economy‐wide Material Flow Accounting: State of the Art and Reliability Across Sources. J. Ind. Ecol. 15, 855–876 (2011).
4. Brunton, S. L., Proctor, J. L. & Kutz, J. N. Discovering governing equations from data by sparse identification of nonlinear dynamical systems. Proc. Natl. Acad. Sci. 113, 3932–3937 (2016).
5. Kaptanoglu, A. et al. PySINDy: A comprehensive Python package for robust sparse system identification. J. Open Source Softw. 7, 3994 (2022).