Skip to Main content Skip to Navigation

Hierarchical predictive control. Application to reheating furnaces in the steel industry.

Abstract : In steel industry, reheating furnaces are the biggest energy consumers after blast furnaces. As a result, reduction of energy consumption is the major concern of furnace control. In a walking-beam slab reheating furnace, steel slabs are heated by moving through successive zones from ambient temperature to a homogenous temperature profile of 1250°C at the furnace exit, to be rolled subsequently in the hot rolling mills. Temperature of slabs is controlled mainly by a two-level hierarchical structure, so called level 1 and level 2. The aim of this thesis is to improve the heating performance and consequently to reduce the energy consumption of the furnace by using hierarchical distributed model predictive control (MPC) strategy for both levels. In a first step, distributed model predictive controllers are developed for the level 1 in order to track zone temperature set-points. The distributed feature of the control law enables to consider coupling effects between zones while reducing the computation complexity compared to a complete centralized approach. The industrial results showed significant improvement on temperature tracking accuracy and an energy consumption reduction of 3%. In a second step, the hierarchical MPC is constructed for the level 2 in order to determine the optimal zones temperature setpoint from the slab temperature setpoint, based on a numerical heat transfer model of the furnace. The simulation results obtained with this strategy compared against industrial data show an energy consumption reduction of 5% and a better heating quality. The previous structure is finally extended to take into account and optimize the scheduling of the slabs within the MPC level 2 in order to increase productivity of the considered furnace. The simulation shows a potential increase of productivity of the furnace of 15 tons per hour while improving the slab heating quality.
Complete list of metadata

Cited literature [307 references]  Display  Hide  Download
Contributor : Didier Dumur Connect in order to contact the contributor
Submitted on : Monday, June 11, 2018 - 5:59:54 PM
Last modification on : Saturday, June 25, 2022 - 10:31:53 PM
Long-term archiving on: : Wednesday, September 12, 2018 - 9:29:25 PM


Files produced by the author(s)


  • HAL Id : tel-01812758, version 1


Xuan Manh Nguyen. Hierarchical predictive control. Application to reheating furnaces in the steel industry.. Engineering Sciences [physics]. CentraleSupélec, Université Paris-Saclay, 2015. English. ⟨tel-01812758⟩



Record views


Files downloads