Abstract
With regard to the characteristics of long process, multi-units, cross-logistics and batch production mode, the order planning problem at the steelmaking shop is studied with consideration of the plant-wide process logistics balance. The problem is to allocate the orders to the continuous casters over plan periods such that the semi-finished materials produced by the steelmaking process can be delivered to the downstream hot-rolling and cold-rolling processes on time at required volume. Based on the detailed analysis of unit layout characteristics of the plant-wide process and the order composition information, a hierarchical planning method is proposed to decompose the large-scale planning problem into a set of sub-problems that are sequential solved. For the order batching subproblem, we develop an integer programming model to consolidate orders with small demand into cast-lots by considering the key batching rules. The model is efficiently solved by the optimization solver of CPLEX. For the cast-lot scheduling subproblem, a novel mixed integer programming model is developed by considering the technological constraints, the facility capacity constraints, the logistics balance constraints, the inventory constraints, and the due date constraints. An improved differential evolution algorithm is proposed to solve the cast-lot scheduling subproblem. The main improvement strategy is that the superior individuals within an external archive are further enhanced through a variable neighborhood search mechanism. Computational results on random and practical instances of different scale problems verify the effectiveness of the proposed method.
Highlights
A prominent feature of iron and steel production process is that physical and chemical reactions occur continuously in the process of material flow, and the state, nature and shape change constantly
Taking the actual production order data provided by a large steel company as an example, considering the planned period of 15 days, the total unfulfilled quantity of all orders in the steelmaking process is about 300,000 tons, and the number of cast-lots obtained by order splitting and order batching is about 300, the model includes at least 720,000 binary variables, 48,000 integer variables, 660 continuous variables, and 360,000 constraints
From the perspective of the plant-wide process, this paper studies the order planning problem which is to decide the allocation of orders on casters over plan periods with the objective of balancing materials flow from steelmaking shops to hot-rolling and cold-rolling shops
Summary
A prominent feature of iron and steel production process is that physical and chemical reactions occur continuously in the process of material flow, and the state, nature and shape change constantly. Tang and Jiang [1] study a batching problem of consolidating orders into heats each is a full furnace of molten steel They consider a set of process constraints such as the converter smelting capacity and the batching conditions in terms of size, steel grade, delivery date, and etc. The scheduling system Bellabdaoui and Teghem [12] study the steelmaking-continuous casting production scheduling problem in the context of the Belgian Arcelor Iron and Steel Group, and consider the characteristics of heats batching, continuous casting uninterruptible, dynamic processing time and others. By considering the batch production mode of large facilities such as converter and tundish, the problem decides how to allocate the orders to the casters over plan period to ensure that the technological constraints and management requirements of large-scale facilities are satisfied, the downstream logistics is balanced and the key orders are timely delivered. When making the order plan, the key orders should be specially considered to improve the on-time delivery capability
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
