280 likes | 388 Views
Scheduling Problem in Steel Industry. SNU MAI LAB seminar 2000.10.20 Lee Hyoung Gon(hungry@ultra.snu.ac.kr). Synchronized Scheduling Method in Manufacturing Steel Sheets. RYOJI TAMURA, MEGUMI NAGAI Information Systems Department, Sumitomo Metal Industries, Fukuoka.
E N D
Scheduling Problem in Steel Industry SNU MAI LAB seminar 2000.10.20 Lee Hyoung Gon(hungry@ultra.snu.ac.kr)
Synchronized Scheduling Method in Manufacturing Steel Sheets RYOJI TAMURA, MEGUMI NAGAI Information Systems Department, Sumitomo Metal Industries, Fukuoka. IFORS, Vol. 5, No. 3, pp. 189~199, 1998
Introduction • New Scheduling System • Casting to a rolling process in steel sheets manufacturing • multiobjective assigning and sequencing problem with lots of constraints • Two-Stage algorithm composed of macro-scheduling and micro-scheduling
Steel Sheets Manufacturing Process(1/2) Cast-to-roll scheduling problem
Primary Production Scheduling at Steelmaking Industries H.S. Lee, S.S. Murthy IBM J. RES. DEVELOP. Vol 40 No. 2 March 1996
Overview of steelmaking and the primary steelmaking process(1/2) • Ironmaking • Combining the raw ingredients of steel into a generic intermediate product known as hot iron or pig iron. • Primary steelmaking • Accepts the supply of hot iron from the blast furnace and transforms it into semifinished products in a variety of grades(specific metallurgical compositions of steel), shapes, and dimensions
Overview of steelmaking and the primary steelmaking process(2/2) • Finishing • Consists of numerous operations that can be applied selectively to the semifinished products of the primary steelmaking process, to achieve customers’ specifications • Cold-rolling -> precise dimension, surface finish, mechanical property • Annealing : material’s grain size, ductility • Tempering : internal stresses, ductility • Pickling : clean the surface • Coatings : corrosion resistance
Scheduling issues in primary steel production • Utilization of manufacturing units • Allocation of production among parallel manufacturing • Specification of heats or heat groups • Specification of rolling groups • Sequencing of heats/rolling groups for manufacturing • Coordination of schedules between production stages • Rescheduling
Scheduling for continuous casters(1/3) • Simple mixed-integer linear programming model • “A model for sequencing a continuous casting operation to minimize costs”(1987) • Electric arc furnace, no change of width in continuous casters, single objective cost function
Scheduling for continuous casters(2/3) • Complex models using heuristics(LTV caster scheduling model) • Implemented in 1983, in support of the first continuous caster installed at LTV Cleveland Works. • Maximization of on-time delivery • Maximization of caster productivity • Satisfaction of quality requirements • Minimization of semifinished inventory “A Scheduling Model for LTV Steel’s Cleveland Work’s Twin-Strand Continuous Slab Caster”(1988) “Twin Strand Continuous Slab Caster Scheduling Model”(1990)
Scheduling for continuous casters(3/3) • Cooperative scheduling approach in which an expert system assists a scheduler(Scheiker Scheduling) • “Cooperative scheduling and its application to steelmaking processes”(1991) • Iteratively modifies an existing schedule through the user interface • Scheduler typically makes global changes that increase global efficiency • Scheduling engine uses a rule base to recognize violations of local constraints
Scheduling for hot strip mill • “Roll-A-Round program”(1991) • Uses heuristics based on the traveling-salesman problem and linear programming
BACK TO…Synchronized Scheduling Method in Manufacturing Steel Sheets RYOJI TAMURA, MEGUMI NAGAI Information Systems Department, Sumitomo Metal Industries, Fukuoka. IFORS, Vol. 5, No. 3, pp. 189~199, 1998
Steel Sheets Manufacturing Process(2/2) – some terms • “heat” : necessary amount of molten steel to be poured into the BOF once. • “try” : the molten steel is poured 5-7times consecutively, and this sequence of heat is called a “try”. • “chance” : rolls are exchanged after rolling one try amount of slabs, and this exchange interval is called a “chance”. • Thus one “try” in casting corresponds to one “chance” in rolling.
The Problem(1/3)- Three Main Objectives • To incorporate particular instructed orders/slabs into the schedule • To attain the target supply quantity for subsequent processes • To meet the rolling due-date
The Problem(2/3)- Constraints • Cast Constraints • The same ingredient orders must be gathered • The “first” heat requires about 270 tons • The last slab of a “try” requires low grade order • Roll Constraints • The strip width has to be changed from wider to narrower • The strip thickness should be transmitted smoothly • Particular orders requires positional limits in rolling sequence • Etc.
The Problem(3/3)- conventional scheduling • 24 hours, 5-7 roll “chances”, three shifts of workers(4) • Select and sequence 60-80 pieces among 5000 orders per roll chance • Consecutively, select 50-70 pieces among 3000 non-D/C slabs in the yard, and insert them between orders. • Some know-how is established to some extent, but for the most part it depends upon a trial-and-error method.
The reason of taking the two-stage approach • Main reason • Reduction of computing time by avoiding combinatorial explosion. • Following advantages • Creating a balanced schedule to avoid differences from the expected one. • Improving solution precision and easiness to check the schedule by user participability • Program flexibility for environmental changes.
Macro Scheduling • Generating a feasible route. -feasible block sequence. • Pseudo-assignment of slabs to the route. • Repetition for all feasible routes. • Determination of the optimum route.
Micro Scheduling • Selecting a block on the route. • Selecting a previous path sequence. • Generating a feasible path. • Pseudo-assignment of slabs to the path. • Linking a path to the previous path sequence and evaluating the new path sequence. • Repetition for all feasible paths. • Repetition for all previous path sequences. • Reduction of the path sequences. • Repetition of all blocks. • Determination of the optimum sequence and assignment of slabs.
User Interface • Preparation • Operators can refer to necessary information(order, slab spec, input parameter) • Modification • Refine the schedule(insert slabs, delete slabs or exchange the slab position)
Conclusion • Human-machine harmonization scheduling system • Now this system largely contributes to efficient scheduling, in Kashima Steel Works.