Study On Scheduling Of The Slab Logistics And Reheating In Steel Industry

In the iron and steel production, the continuous logistics under high-temperature always requires large scaled equipments and energy consumption which cause the high operation cost. How to save the energy consumption while ensuring the high utilization factor and capacity of these equipments has become an urgent problem in the steel production. The problem is generally addressed in three levels:technologic process and equipment, automation and control system, and production and logistics management. Taking the reheat furnace, a typical process of high-energy, and its material supplier as research background, the dissertation studies the slab logistics scheduling in the slab yard, which is the material supplier of the reheat furnace, and the slab reheating scheduling in the furnace. The research can help to supply material for the furnace in time and to reduce the idle burning in the furnace and thus is significant to both production rate and energy saving.A series of appropriate slabs need to be selected from the slab yard when executing a rolling plan. Slab retrievals can cause slab stack shuffling (SSS) since the slabs are stored at different positions of different stacks. Modeling and optimization of two SSS problems considering slab delivery times and equilibrium of crane workloads, corresponding to the spatial and temporal balance requirements, respectively, are both researched in the dissertation. Due to the close coupling between the SSS and crane scheduling, they are integrated into one problem. The integrated problem is modeled and solved efficiently. For the slab scheduling problem in the reheat furnace, an Intelligent Optimization algorithm is proposed. The main contents are as follows.1) Modeling and optimization of the SSS problem considering delivery times is researched. Differing with the traditional SSS problem, the SSS problem requires that every selected slab shoule be taken out in time and the shuffled slab stay at the new position instead of returning the original stack. Based on the above characteristics, an integer programming model is formulated for the SSS problem with the objective of reducing the total crane workload. Treating each item in the SSS problem as a stage, a dynamic programming approach is proposed for the problem, which can only resolve the small scaled problem instances. For the practical scale, due to its intractability, a segmented dynamic programming (SDP) based heuristic is proposed, which utilizes the local similarity of the problem. Two interesting properties of the problem are also derived to speedup the SDP based heuristic approach. The experiment results show that the heuristic is very near to the optimum in average solution quality for the small-scale problem, obviously better than the ILOG CP Optimizer for the medium scale, and can reduce the crane workload by10.76%on average for the practical scale.2) Modeling and optimization of the slab stack shuffling problem considering equilibrium of crane workloads is researched. Differing from the existing researches, it considers the equilibrium of crane workloads, a new practical requirement. It is of more significance that the SSS problem is formulated into an integer linear programming (ILP) model by introducing the concept of stack-scheme which can transform the nonlinear feature of the SSS problem into linearity. The suitable condition of the proposed ILP model is also extended from the original problem, no common candidate slab between different rolling items, to the generalized problem, permitting common candidate slabs. Several properties of the problem are deduced to decrease the number of variables involved in the ILP model and thus, to a degree, reduce the computational complexity. The experiment results show that the optimization method, based on the proposed ILP model, can obtain the optimal solution for the practical scaled problems.3) Modeling and optimization of the integrated logistics scheduling of slab stack shuffling and crane routing is researched. The integrated problem needs to determine the spatial and temporal decisions on the slab selection for each rolling item, stack selection for the shuffled slab and crane scheduling simultaneously while ensuring the slab delivery times. For the integrated problem, an integer linear programming (ILP)-based heuristic approach is proposed, which decomposes the original problem into a slab selection master problem and a combined crane scheduling subproblem. The two problems are solved iteratively until a feasible crane schedule is obtained. Under certain conditions, a worst case analysis is provided for the proposed ILP-based heuristic. The performance of the heuristic is also verified by numerical experiments, in which the obtained objective value is compared with the obtained lowerbound and the average gap is3.16%. In addition, a variant of the problem is also studied and solved to the optimality.4) For a class of Job-Shop problems, an improved hybrid MILP/CP algorithm is propsoed. To overcome the low efficiency of the existing MILP/CP algorithm for a class of Job-Shop problems, two strategies, based on the introduced concepts of critical job assignment and critical cut, are proposed to generate more number of effective cuts for the purpose of speeding up the algorithm. The experiments show that the proposed strategies are efficient.5) Modelling and optimization of the reheat furnace scheduling problem is researched. The reheat furnace scheduling problem is to arrange a sequence of slabs into multiple furnaces and determine the respective feed-in times and residence times of the slabs such that their required drop-out times are met. The objective is to optimize the reheating quality and energy saving. Unlike the traditional scheduling problem with constant processing time, the residence time of each slab is a variable to decide and subject to its neighborhood slabs in the reheating sequence belonging to same furnace. The RFSP is formulated as a mixed integer programming model by considering the above features and requirements. A scatter search (SS) algorithm combined with constraint propagation (CP) is designed to solve the problem. The computational results show that the proposed SS with CP is relatively effective and efficient.