(382g) Advanced Biomimetic Control Approach Integrated with Multi-Agent Optimization for Nonlinear Chemical Processes

A Biologically-Inspired Optimal Control Strategy, denoted as BIO-CS, has been recently developed. This strategy has been implemented on various advanced energy systems applications such as the Integrated Gasification Combined Cycle (IGCC), the Hybrid Performance (HYPER) system and a fermentation process (Mirlekar et al., 2017; 2017a; Lima et al., 2016; Li et al., 2016). In such applications, BIO-CS has shown to have features for handling process model nonlinearities as well as flexibility for employing different optimal control solvers and termination criteria. Also, in the past, multi-agent-based optimization techniques that mimic ant-colony-based behavior with improved efficiency have been studied under the name of Efficient Ant Colony Optimization (EACO) for molecular design and solvent selection case studies (Gebreslassie & Diwekar, 2015). The abilities of heuristic-based methods such as EACO, Efficient Genetic Algorithm (EGA) and Efficient Simulated Annealing (ESA) were also explored to develop homogenous Multi-agent Optimization (MAO) techniques by establishing communication protocols between the algorithm procedures and the global information sharing environment (Gebreslassie & Diwekar, 2017). However, the combination of biomimetic control strategies and agent-based optimization methods for nonlinear systems has not yet been addressed in an integrated fashion. To bridge this gap, in this work, BIO-CS is integrated with MAO to design a novel framework that leads to optimal process operations. The developed framework yields optimal setpoints or a trajectory of setpoints for a nonlinear, multivariable system considering an overall process objective by employing MAO. This system is then optimally controlled by BIO-CS to achieve such desired output setpoints.

The applicability of the proposed method is demonstrated using a nonlinear, multivariable fermentation process model (Li et al., 2016) for bioethanol production. In this multivariable system, finding the optimal production rate (or profitability) as well as performing the simultaneous control of product concentration and temperature of the fermentor at their setpoints are critical for optimal performance. The proposed framework is implemented on the fermentation process to address all these challenges. Specifically, implementation scenarios of setpoint tracking and plant-model mismatch are considered. The results of the developed framework are compared to a gradient-based Sequential Quadratic Programming (SQP) technique and a classical proportional-integral (PI) controller in terms of optimization and control studies, respectively. As an additional control development, BIO-CS is cast as a Model Predictive Controller (MPC) and the resulting formulation (BIO-CS as MPC) is compared to the agent-based BIO-CS approach in terms of computational time, time to reach steady state and tracking performance. The implementation results of the proposed framework will be discussed to show its potential for simultaneous optimization and control of advanced energy systems.

References:

1. G. Mirlekar, S. Li, and F. V. Lima. Design and implementation of a Biologically-Inspired Optimal Control Strategy (BIO-CS) for chemical process control. Ind. Eng. Chem. Res., 56: 6468–6479, 2017.
2. G. Mirlekar, P. Pezzini, M. Bryden, D. Tucker, and F. V. Lima. A Biologically-Inspired Optimal Control Strategy (BIO-CS) for hybrid energy systems. In Proceedings of the American Control Conference, 4821-4826, 2017a.
3. F. V. Lima, S. Li, G. Mirlekar, L. N Sridhar, and G. J. Ruiz-Mercado. Modeling and advanced control for sustainable process systems. In Sustainability in the design, synthesis and analysis of chemical engineering processes, G. Ruiz-Mercado and H. Cabezas (eds.), Elsevier, 2016.
4. S. Li, G. Mirlekar, and F. V. Lima. Development of chemical process design and control for sustainability. Processes, 4(3):23, 2016.
5. B. H. Gebreslassie and U. M. Diwekar. Efficient ant colony optimization for computer aided molecular design: Case study solvent selection problem. Computers and Chemical Engineering, 78:1-9, 2015.
6. B. H. Gebreslassie and U. M. Diwekar. Homogenous multiagent optimization for process systems engineering problems with a case study of computer aided molecular design. Chemical Engineering Science, 159:194-206, 2017.