Application of Surrogate Models as an Alternative to Process Simulation for Implementation of the Self-Optimizing Control Procedure on Large-Scale Process Plants—A Natural Gas-to-Liquids (GTL) Case Study

Abstract

High computational loads, time-consuming convergence, and simulation crashes are common when using process simulators for flowsheet optimization. In this paper, by replacing the large-scale physical process simulations by surrogate models, the optimization time and computational load are reduced significantly along with maintaining the accuracy and reliability. A gas-to-liquids (GTL) plant was used as a large-scale process plant case study. The multilayer perceptron neural network (MLP-ANN), radial basis function neural network, support vector machine, and adaptive neuro-fuzzy inference system models were selected as alternative surrogate models. These alternatives were investigated for implementation of the self-optimizing control procedure on the above case study to find the best individual and combined self-optimizing controlled variables. The MLP-ANN surrogate model showed the best performance in predicting the optimum points and for selecting the best self-optimizing CVs. In fact, it even performed better than using the full process simulator.

Echo state network-based metamodel for the problem of stagnant petroleum cooling in pipe-in-pipe flowlines

“…Concerning the computational costs, recently, surrogate models was used to optimize the design of aerodynamic shapes, significantly reducing the computational effort required for each design evaluation [16]. An application to the use of Gaussian process surrogate models for uncertainty analysis in environmental models and demonstration of the use of surrogate models for real-time control in robotics, enabling faster and more efficient control actions [17,18]. Qui et al (2022) [19] compared a surrogate model with a commercial simulator for pipeline gas transport, producing result with high accuracy, below 6%, and CPU cost 1250 times faster than LedaFlow software.…”

A Deep Learning Hybrid Framework Combining an Efficient Evolutionary Algorithm for Complex Many-Objective Optimization of Sustainable Triple CO2 Feed Methanol Production

“…These models possess the capability to approximate the mechanism model in a reduced dimension while ensuring accuracy, thereby mitigating the complexity of many-objective problems while exhibiting strong generalization properties to ensure the efficient convergence of the optimization process. Over the course of time, a multitude of surrogate models have been gradually proposed, including the support vector machine (SVM), , Kriging model, , and artificial neural network (ANN). − An ANN is a crucial component of contemporary computer technology that aims to simulate the structure and functionality of the neural network in the human brain. The simulation comprises a network of interconnected neurons that exhibit strong data processing capabilities and the ability to adapt and learn.…”