(744f) Statistical Analysis of Streamlined Process Simulation-Based Gate-to-Gate Energy Use for Chemical Reactions and Separations

Chemical process energy is a key component of chemical life cycle inventory (LCI) data. Energy used in manufacturing of chemicals account for 50%-80% of the environmental impacts associated with chemical production. Current chemical LCIs are typically either dependent on aggregated plant-based data that are not specific to the target chemical, or on estimations based on simplified design calculations and/or use of proxy data. Such estimation techniques miss the process-level details and advanced thermodynamic calculations of chemical process simulation tools commonly used in chemical engineering design. In this work, a streamlined process simulation-based methodology for process energy data generation is used to generate gate-to-gate energy use data for manufacturing of organic chemicals. Chemical process is divided into three major blocks, with the reaction at the core, energy for pretreatment before the reaction, followed by energy required for separation and purification of the product. Chemical process simulator Aspen Plus is used to model and simulate 150 chemical processes based on this methodology. Potential process heat integration is analyzed using Aspen Energy Analyzer, which uses Pinch Analysis to provide an estimate of target heating and cooling utilities. The extracted energy data for each chemical process are used to provide a time-efficient method for predicting process heating and cooling requirements through multi-variate statistical methods such as regression analysis. Process design-based predictor variable such as heat of reaction, type of reaction (addition, substitution, etc.), and unit process (hydrogenation, nitration, etc.), and phase were used in the analysis. Heat of reaction has a strong and positive correlation with reactor heat duty, with a R² of 0.8. Separation operations on average require 46% (SEM = 4%) of the heating energy and 57% (SEM = 4%) of the cooling energy. The models from this study can be used for predicting process energy requirements for chemical production in early design stages, and for chemical LCI data generation in the absence of detailed industry data.