Esnek üretim sistemlerinin tasarımı ve çizelgelemesi

Abstract

Son yıllarda önemi gittikçe artan Esnek Üretim Sistemleri bu çalışmanın konusunu oluşturmaktadır. Çalışmada ilk önce, ürerim sistemleri genel olarak incelenmiş, daha sonra Esnek Üretim Sistemleri ile iligili tanımlar ve bu sistemlerde karşılaşılan çeşitli problemlerden bahsedilmiştir. EÜS'nin tasarımında çok önemli yer alan donanım seçimi ve çizelgeleme hakkında ayrı bölümlerde bahsedilmiş bu problemlerin çözümleriyle ilgili bilgiler verilmiştir. Son bölümde, bu sistemlerle ilgili bir simülasyon çalışması yapılarak, onüç parça tipinin işlendiği, beş CNC tezgah, bir parlatma tezgahı ve bir tesviye tezgahında oluşan bir sistem incelenmiştir. Sistem tasarımıyla ilgili olarak, iki alternatif incelenmiştir. Birinci alternatif, tek robotlu bir hücre, ikinci alternatif ise iki robotlu bir hücre olarak belirlenmiştir. Her iki alternatif için çeşitli robot çağırma öncelik kurallarrnm, sistem üzerinde etkileri incelenmiş ve her iki alternatif için en iyi sonuçlan veren durumlar kendi aralarında karşılaştınlmıştır. Yapılan çakşmanın sonunda, öncelik kurallarmdan, parça tipine göre öncelik sırası belirleyen kural her iki alternatif için de en iyi sonucu veren kural olarak tespit edilmiştir. Tek robotlu ve iki robotlu alternatiflerin sonuçlan karşılaştırıldığında, İM robotlu alternatifin daha iyi performans gösterdiği saptanmıştır.

Manufacturing has emerged in tins decade as one of the important keys to organizational success and, as a result of renewed emphasis on manufacturing methods, a number of comprehensive manufacturing strategies are receiving widespread attention. Computer integrated manufacturing (CIM), just-in-time manufacturing, factory automation, and flexible manufacturing systems (FMS) are some of the recurring themes. What the 'factory of the future' will ultimately become remains to be seen; however, it seems clear that these strategies will have a significant influence on manufacturing for a long time to come. One of these strategies, flexible manufacturing systems, is of particular interest since many manufacturing firms are currently making heavy investments in FMS technology. FMSs are automated manufacturing systems which are designed to produce different part types with the efficiently of mass production systems and the flexibility of job shops. FMSs are able to process a large volume of small to medium size batches of parts. A typical FMS consists of versatile numerically controlled machines, connected by an automated material handling system, all under a central computer control. Recent advances in microelectronics, robotics, and computer architecture make flexible manufacturing technology suitable for a large set of manufacturing tasks. FMS applications can be found in a diverse set of industries (aerospace, agriculture, defense, electronic, machine tool, etc.) throughout the United States, Japan, and Europe. Design problems have been strategically critical in any system operation. Optimal design of physical layout is one of the most important issues that must be resolved in the early stages of the system design. Cost consequences of the decisions related to layout of the machines can be observed not only during the implementation but also during the operation of the system. Good solutions to these problems provide a necessary foundation for effective utilization of the system. This particularly true for FMSs, in comparison with conventional manufacturing systems, for the following reasons: xl 1. Alternative (flexible) routing: This flexibility can be attributed to different characteristics of FMSs. The machines used are flexible enough to be able to perform different operations when properly tooled The material handling system (MHS) allows bypassing of manufacturing cells, and part movement between almost any pair of machines is possible. Finally, the set-up time required to change the tooling of the machine is small. These characteristics generate a large set of alternative manufacturing routes for each product, that adds to complexity of the manufacturing environment and leads to an overloading of the MHS. 2. Expensive hardware used for material processing and handling. FMS machining stations, which are able to perform tasks such as automated part and tool changing, tool and workpiece holding in magazines, entire tool magazines of tool rack swapping, as well as typical operations are acquired through huge capital investments. Any subutilization caused through inefficient layout design imposes significant cost penalties. On the other hand, The advanced MHSs used in FMS implementations are expensive not only in terms of acquisition but also operating cost. 3. FMS stations are tightly linked: The limited buffer size implemented between FMS workstations increases the station interdependency, which results in the need for better layout design and MHS utilization. The layout layout decisions deal with the arrangement of workstations and the way the workstations are connected through the transportation lines of the material handling system. For celluar manufacturing implementation of FMS, the so-called cell layout design is important. One of the most significant costs besides the obvious issue of capital investment arises in the design of FMS. By their very nature, FMSs are complex; the components selected for a particular FMS must be suitably matched and must operate well together over a wide range of production situation. In general, the FMS design problem has been divided into three sub- tasks; planning, design, and implementation. The planning sub-task is concerned with determining the products or components to be produced by the FMS, the processes to be utilized, and the quantities to be manufactured. The design sub-task is treated as two separate decisions. First, the equipment features of the FMS are specified by selected individual machines, robots, conveyors, and automated guided vehicles. Then appropriate machine tools, fixtures, computer control procedures are specified. The implementation sub- task is concerned with converting a proposed FMS design into an operational system xrt Simulation has become a widely used methodology for investigating the hardware specification decision in FMS design. This approach requires specification of an initial design and development of a simulation model of the initial design in an appropriate language. Then the model is executed and the designer analyses the output to determine whether design objectives are achieved. If the designer thinks that changes in the design would improve the performance of the system, the simulation model is altered to reflect the proposed changes and executed. This process ids repeated until satisfactory results are obtained. A 1980 survey by Shannon, indicated that among Operation Research (OR) practitioners in US industry, the second most widely used OR technique was simulation modeling. Of all OR techniques, simulation modeling was the one most respondents were most eager to learn more about. More recently, simulation modeling was identified by Society of Manufacturing Engineers (SME) as one of the important technologies for manufacturing engineers in the 21st century. In this study, the design and the scheduling of flexible manufacturing systems were considered In the first section, the conventional manufacturing systems and celluar manufacturing systems were determined and compared with each other. The conventional systems were divided into four groups, job shop, flow shop, project shop, and continues process manufacturing. The celluar systems were divided into two groups, manned celluar systems, and unmanned celluar systems. Then, the conventional and celluar manufacturing systems were compared in terms of machine utilization and time consumption. In the second section, the design of flexible manufacturing systems was discussed. First, the flexible manufacturing systems were determined. The system architecture of FMS, the vision of FMS process, the programmability of FMS were discussed subsequently. The flexibility of FMS is also determined in this section. Several problems which could be faced in the FMS design were considered. The FMS and conventional manufacturing systems were compared in terms of indirect costs, machine utilization, and lead times. The economical issues in the flexible manufacturing systems, the methods used for economical analyzing problems were also discussed in this section. In the third section, scheduling of FMS was discussed. First, the scheduling in FMS was determined. The scheduling methods in FMS and several appropriate solutions for FMS scheduling were determined. In the fourth section, the simulation in FMS was discussed. First, a definition of the simulation was made. Then, simulation techniques were Xlll considered Types of simulation, phases of a simulation study, simulation software, and physical simulator were considered Types of simulation tectoniques were compared in terms of visualty and summarizing level. The more visual are the tectoniques, the more real are they. The computer simulation integrated with animation is a very powerful tool to analyze the design decisions. It's in the middle of a range which starts with the mathematical programming and ends with the pilot plant. At the end of this section, simulation in FMS was discussed The most important advantage of the simulation is having chance to see problems like bottlenecks in the system before having the real system implemented The change to be able to educate people who are going to work with the system before having it implemented is another important advantage. Several kinds of what-if analyzes can be made about the system on the simulation model so that the system can be designed as efficiently as possible. In the sixth and final section, a simulation study about a FMS performed. The system on which the study was performed consists 5 CNC macMning stations and manufactures thirteen different parts. The machining stations are linked by robot(s) with each other. Two different alternatives were proposed for this system. One is a configuration with a robot linking stations with each other and the other is a two robot configuration. The cell with the former configuration has a U-shape and the other has a 2U-shape. A robot serves just to the machines which are located in the same U-shape. If we consider cell 2 as Ul and U2, robotl serves to the machines in Ul and robot2 serves to the machines in U2. In both models, some priority rules and their affects on the system were investigated The models were coded by SIMAN simulation language. It helped reducing model developing time. The results of the simulation runs were picked out by SIMAN in summary reports, then they were input to Quattro Pro spreadsheet program and the graphics were obtained by the help of Quattro Pro. The starting conditions of the simulation runs were determined by the method of Conway and the numbers of runs were determined by the method of Fishman. The starting conditions were reached after two runs. Therefore, the statistics started to be cellected after two runs, 2160 time units. According to Fishman's method, the number of runs was determined as three. It means, after three runs, the results can have a real meaning. The result obtained from the simulation are compared with each other. First, all the results for priority rules for each alternative were compared with each other and the best results for the both alternatives were matched afterwards in this comparison. The one-way ANOVA and hypothesis tests were used for comparing the results. At the end of the ANOVA and hypothesis test, it couldn't be seen any significant differences between the priority rules. However, it could be seen from the graphics that there was a difference between the priority rules. The results for xitf the alternative layouts were seen different according to ANOVA and hypothesis tests. As a result, the priority rule which assign the highest priority to the parts according to their types gave the best results for the both alternatives. As the one robot and two robot configurations were compared with each other, it was seen that the two robot configuration had a better performance. However, these kinds of analyzes are not enough to be able to decide which configuration would give the optimum solution. Therefore, some detailed economical analyses should also be performed. It can be proposed for the further researchs, that a detailed economical analyze might be done and the number of the priority rules which were investigated might be enlarged. In this study, the buffer sizes in the system couldn't be investigated because of model restrictions. It can also be proposed that the buffer sizes in the system might be investigated.