Steady-state real-time optimization using transient measurements and approximated Hammerstein dynamic model: A proof of concept in an experimental rig

Abstract

An attractive method for steady-state real-time optimization (RTO) using transient measurements consists of a persistent parameter adaptation throughout a dynamic parameter estimation followed by a static economic optimization. The method is called RTO with persistent parameter adaptation (ROPA) or Hybrid RTO (HRTO). Although such a method avoids the long steady-state wait to trigger the optimizations, it adds extra complexity to the RTO project by including the requirement of developing a rigorous dynamic model, which can hinder the industrial-scale applicability of the methodology. To overcome this requirement, a Hammerstein dynamic approximation model has been proposed to replace the rigorous dynamic model in a Hammerstein ROPA (HROPA) framework. Previous work has demonstrated that the proposed model is adequate for use in RTO since it preserves the parameter observability property of the static model. In this paper, two HROPA approaches are implemented in an experimental rig and compared to a previously implemented ROPA approach. The experimental results confirmed that HROPA presents a similar economic performance to ROPA with much less effort required in the modeling stage. Finally, a comprehensive literature review of the topic is presented, and general guidelines for the Hammerstein dynamic matrix identification are provided.