Reheat furnace production scheduling based on the improved differential evolution algorithm

Abstract

The reheat furnace, located between the continuous caster and the hot rolling mill, plays the role of buffer coordination zone, and is one of the most important production equipment in the hot rolling process. As reheat furnaces were the largest energy-consumer group in the hot rolling process, their schedule optimization was of great importance to achieve high production efficiency and reduce energy consumption. In this paper, a new reheat furnace production scheduling method with the target of minimum fuel consumption was proposed. First, the energy inputs and outputs from the reheat furnace were analyzed based on the first law of thermodynamics, then the equation for calculating of the fuel consumption was derived. Second, various production constraints were summarized to consider the actual characteristics of the dispatching plan in reheat furnaces, and the mathematical model of scheduling optimization was constructed with the minimum fuel consumption set as the optimization objective. The adaptive differential evolution algorithm and the tabu search algorithm were applied to obtain the optimal solution. The differential evolution algorithm could dynamically adjust the scaling factor and the crossover rate according to the change of the fitness function value of each generation of individuals, and this adaptive strategy could balance the ability of development and exploration of the algorithm. After the model was validated with actual production data, the feasibility and effectiveness of the algorithm were verified by nine groups of actual billet production cases. Furthermore, to explore the influencing factors of energy consumption of reheat furnace, two evaluation parameters, μ1 and μ2, were defined to quantify the matching degree of time series of the buffer waits and the heating processes to ideal production in reheat furnaces. According to the sensitivity analysis of the relationship between the fuel consumption and the two evaluation parameters, it was found that their sensitivity gradually decreased when the ratio of continuous casting billet arriving at the reheat furnace to hot rolling increased from 0.5 to 2.

Keywords

• reheat furnace
• scheduling optimization
• fuel consumption
• differential evolution algorithms
• sensitivity analysis