Integration Of Billet Inventory Operation And Rolling Planning Under The Cold Charge Technology

In the steel mills, especially under the cold charge process, the billet inventory plays an important role in receiving the incoming billets from continuous casting line and delivering raw material (billets) to the hot rolling process. Logistic operations in billet inventory are characterized by large scale, high-dimensional, thus the operation for inventory operations has important impact on the production of hot rolling process. This dissertation studies the integrated management problem between billet inventory operation and rolling planning besed on the Malian stack way used in special steel mills for billet inventory and considering the real situation on planning and performing hot rolling. Forthermore by analyzing the joint policy of billet inventory operation and rolling planning, the models and the associated algorithms for the following four key problems are developed:entering and out of inventory operation optimization for steel billets under considering the production demand of hot rolling process; hot rolling compilation and adjustment based on billet inventory information. The integrated management of inventory operations and rolling planning can ensure the continuity and stability of hot rolling production and reduce the inventory cost.The main contributions and innovations are given as follows:(1)The billet stacking problem with delivery date is introduced to ensure the continuity and stability of hot rolling production. A mathematical model for this problem is established to minimize the billet shuffles, and then a heuristic algorithm based on dynamic adjustment strategy is designed. First, the billets staged on the roller are formed to a billet stacking batch; and then the billet batches are assigned to the stack positions. Based on the initial assigned planning, a dynamic adjustment approach is built to adjust each top batches in order to reduce the total shuffles of stacking plan. The results in numerical examples show that the designed algorithm can not only reduce the stack shuffles, but also improve the space utilization.(2)In order to decrease the time and cost of inventory during the rolling plan process, a rolling planning problem with billet shuffling optimization is addressed. It corresponds to match the best inventory billet for each rolling position on the basis of rough rolling plan and then sort the rolling units to form an executable rolling plan. A multi-objective integer programming model is designed to minimize the shuffles of rolling billets and rack switching numbers. Then a Partheno-genetic algorithm based on billets matching is constructed to solve the proposed model. The simulation results demonstrate the feasibility and effectiveness of the proposed model and algorithm.(3)Considering the real-time dynamism of billet inventory, a rolling planning and adjusting problem is further studied to ensure the continuous implementation of rolling plan. The effect factors of billet inventory on the rolling planning are discussed and three dynamic strategies are presented:namely, the rolled-billet replacement, the rolling sequence adjustment of rolling units and the billets cancel rolling adjustment. Based on the above three dynamic adjustment strategies, a weighted average mathematical model by minimizing the adjust penalties and shuffles is established. Then a three-stage heuristic algorithm is designed to provide a practical solution for implementing the rolling plan.(4)To improve the billet shuffling effectiveness of implementing rolling plan, a billet shuffling and falling problem is proposed based on rolling billets shuffling optimization and reassigning the best locations for the obstacle billets, which is different from the previous models where obstacle billets must go back. Then a two-stage mathematical model including billet shuffling optimization model and falling optimization model is established. A two-stage algorithm is designed to solve the proposed problem, where an improved multi-loop heuristic algorithm is used to match the best billets for the rolling plan in the first stage; in the second stage, a search algorithm based on radiation neighborhood is designed to assign the best falling locations for the obstacle billets during implementing rolling plan. Finally, numerical examples are given to illustrate the applicability of the proposed models and algorithms.