(42b) Unleashing the Power of Hybrid Modeling for Process Design and Optimization.

Process optimization via mathematical modeling is a powerful tool in the chemical process industry for developing and implementing process improvements. For example, mathematical models can be used to simulate different process scenarios and identify the optimum operating conditions that will maximize productivity while minimizing energy usage. This can lead to significant energy and cost savings, increased capacity, and profitability for the industry. However, developing accurate first principles (physics-based) models can be complex, especially when modeling large and complex processes with multiple physical and chemical phenomena. Hybrid models, which combine physics-based models with machine-learning models or data-driven models, can be used to address some of these limitations. By combining different modeling techniques, hybrid models can provide a more accurate and comprehensive representation of the process while reducing the computational cost of the model. This work presents the use of Aspen Hybrid Models^TM at Dow in a highly integrated process flowsheet. First-principles-driven hybrid models were developed in Aspen Plus V12.1 for reactors with complex reaction networks coupled with distillation columns and multiple recycle streams. Methodology for development of the hybrid model will be shared along with high level results of model validation versus plant performance and model optimization opportunities.