Esnek üretim sistemlerinde çizelgeleme

Abstract

Bir Esnek Üretim Sistemi bir anabilgisayar altında organize edilmiş ve merkezi bir taşıma sistemiyle bağlanmış üretim araçları topluluğudur. Esneklik ise; ürün, proses, yükleme, işleme başarısızlıkları gibi faktörlerde meydana gelen herhangi bir değişikliği düzeltmek amacıyla geliştirilmiş sistem yeteneğidir. EÜS de kullanılan esneklikler makina, malzeme taşıma, operasyon, proses, ürün, rota, hacim, kapasite arttırın program, üretim ve pazar esne kliğidir. EÜS'lerin donanımı NC'Iü tezgahlar, Malzeme Taşıma ve Depolama Siste mi ve Bilgisayarlı Kontrol Sistemi olmak üzere üç bileşenden oluşmuştur. Esnek Üretim Sistemleri çizelgeleme probleminde gerçek zamanda EÜS sistemlerinin çalıştırılması söz konusudur. Çizelgeleme probleminin konusu, belli bir zamanda hangi parçanın sisteme girişi yapılıp işleneceğine karar verilmesi, hangi sıranın optimum sırayı vereceğinin saptanmasıdır. Çizelgeleme problemleri sistemin genel performansında önemli bir rol oynar. Esnek Üretim Sistemlerinin Çizelgeleme Sınıflandırılması iki ayrı şekilde yapılabilir. Bunlardan birincisi çizelgeleme faktörlerine göre sınıflandırma diğeri ise öncelik verilen kaynaklar açısından çizelgelemedir. Yapılan çalışmalar incelendiğinde periyod dışı planların çizelgelenecek parça sayısının çok olması durumunda en uygun çizelgeleme yöntemi olduğu söylenebilir. Bu çalışmada ayrıca Esnek üretim Sistemlerinin Çizelgelenmesi ile ili bir uygulama yapılmış, bu uygulamada montaj ve işleme alt sistemlerini içeren bir EÜS'nin iki aşamalı çizelgelenmesi gerçekleştirilmeye çalışılmıştır, ilk aşama iki makinalı akış tipi atölye problemine benzetilebilir. (Toplu Çizelgeleme) İkinci aşamada ise işler parça ve ürün önceliklerine göre çizelgelenmişlerdir. (Detaylı Çizelgeleme) Yapılan uygulamada işleme ve montaj sistemleri OKA sistemiyle birbirlerine bağlanmıştır. Önce parçalar işlenmekte ve daha sonra montaja gön derilmektedir. Sunulan yaklaşım optimallik konusunda kesin garanti vermesede oldukça pratik olması nedeniyle gelecekteki Esnek Üretim Sistemlerinde kullanılması önerilebilir.

Flexible Manufacturing Systems (FMS) are regarded as one of the most efficient methods to employ in reducing or eliminating problems in manufactu ring industnes. FMS is the latest level of automation a long an evolutionary road to achieve ever more productivity and flexibility from manufactunng equipment. A FMS consists of a group of machining centers, interconnected by means of an automated material handling and storage system and controlled by an integrated computer system. So a FMS system can be defined as a computer controlled production system capable of processing a vanety of part tipes. Two factors are very important to any flexible manufacturing system. The first one is computer integration. Computers perform several functions in these systems; Scheduling and monitoring operations, handling material control and taking appropriate actions in case of sudden changes in the system. The second important factor is flexibility of the system. Flexibility is the ability of the system to quickly adjust to any changes in relevant factors like product, process and machine failures. With computer integration, it may be possible to equip the system with a certain degree of self-diagnostic and adaptive control abilities. Browne attempt to classify flexibility in to categones belonging to machine, process, product, routeing, volume, expansion, operation and production flexibi lities. Most of these categories are interdependent. The major components of a flexible manufacturing systems are: - CNC machines - Material Handling System - Computer Control System CNC provided the technological foundations on which FMS could be built It is being centred around a small minicomputer and it is applied to a wide varie ty of manufacturing processes. FMS is to ensure that the transformation of raw material to finished parts is as rapid, efficient and controlled. Transport of the parts between the FMS and its Automated Storage and Petneval System (ASRS) or between different FMS might be carried out using an Automated Guided Vehicle System (AGVS). The AGVS perform a vanety of functions in flexible cells and systems consists of: - Transporting parts, tools and fixtures to and from processing, queuing and build stations - Delivering raw meterial to the cell or system - Transporting finished parts from the system to assembly areas - Delivering parts, tools, and fixtures to and from an ASRS - Automatically raising and lowering pallets to registration positions on processing and queuing station shuttle mechanisms for loading and unloading. The robots are reprogrammable multifuctional manipulators that move matenal, parts and tools. The pnmary function of robots in flexible cells and sy stems is to load and unload parts. Automated stroge and retrieval systems (ASRSs) contain tall vertical sto rage racks, narrow aisles, and stacker cranes and are coupled with computers for automatically storing tracking and retrieving material implementing an ASRS can successfully reduce operating costs and gain control over the material sto rage and retrieval process. The main purpose of computer control systems is to achieve the required timing, quantity and quality of the production programme. The major compo nents of control are: a) Controlling of CNC machines b) Controlling of Matenal handling systems c) Controlling of activities of workparts in the system d) Controlling of in formation which is used to measure the performance of the system. The scheduling problem is complicated by subsystem constraints, routing flexibility and the need for system wide coordination. Scheduling in FMS can be described by a hierarchical structure ranging from top level decision making to detailed level scheduling decision. Top level scheduling emphasizes planning for production and plant operations over extended periods of time which may in clude, for example, resource planning and generation of sequences of operations. The objective at this level is the coordination of activities for multiple function areas. At the detailed level, scheduling controls demand over the course of each day and provides a means to achieve the production targets. It attempts to find the optimal routing of jobs combined with making efficient use of expensive resources which are subject to environmental and procedural con straints. These objectives are accomplished by allocating resources such as machine tools, fixtures and raw metarials according to time (due dates) and rou ting constraints. Jobs are assigned to specific work stations on a weekly, daily or hourly basis taking in to acccount the type, quantity and placement of resour ces as well as any associated time values and processing priorities. FMS pro duction scheduling, at the detailed level involves the considerd tion of two sepa rate but related activities loading and dispatching. Loading problem is to allocate the operations and associated cutting tools of the selected set of part types among the machine groups subject to the tech- xi nological and capacity constraints of the manufacturing systems. The loading problems are also sometimes called a priori scheduling problems. After loading, the next step is running the FMS'in real time and sche duling operations one at a time as reach machine becomes avaible, according to the appropriate dispatching rules. It is a challenging promlem in several ways. For example: - After a part is loaded, which machines should it visit and at what time should it be processed on specific machines? - If a machine fails, how should the schedule be adjusted for each part in the system - If a high prionty is loaded, then how should the schedule be adjusted for each part so as to accommodate the high priority job? FMS scheduling problem is very difficult and complex. The complexity aries from the added flexibility of various machine types in the production system. The complexity of the decision making process is further increased by the fact that a multitude of objectives. There are six general objectives of FMS scheduling problems: (i) Balancing the assigned machine processing times (ii) Minimizing the number of moverments from machine to machine (iii) Balancing the workload per machine for a siystem of groups of pooled machines of equal sizes. (iv) Unbalancing the workload per machine for a system of groups of pooled machines of unequal sizes. (v) Filling the tool magazines as densely as possible (vi) Maximizing the sum of operation priorities. A classification framework strengthens a review of FMS scheduling proce dures. There are two categories used the framework; the first involves three fac tors that define the type of system, and the second includes three factors that define the scheduling scheme. In the first category the three factors that define the types of FMS are the number of different part types the system processes, predominant flow pattern of these parts and demand pattern the system encounters. FMS that process fif teen or fewer parts are classified as dedicated systems that process more than one hundred parts are classified as random. Those systems that processes bet ween fifteen and one hundred parts are classified as intermediate. Another im portant factor that defines system types is the predominant flow patlem of parts through the system. FMS is classified as either job shops or flow shops. Job shop type FMS typically process many different part types, each with small annual demands. Flow shop type FMS typically process a small number of diffe- xii rent part types, each with large annual demands. The demand pattern a system encounters is the third factor in this category. By definition periodic demand pat tern indicates that a number of jobs arrive periodically. A continuous demand pattern indicates that orders for individual jobs arrive continually over time. In the second category, the next theree classifying factors define the types of scheduling schemes. The first factor in this category is the scheduling prom- lems the scheduling scheme addresses; input sequencing and detailed sche duling are the levels of this factor. The input sequencing decision determines what sequence a set of parts uses to enter the system in the upcoming period. Deciding on oppropnate detail schedule involves determining which machine a part should visit and when the visit should take place. The number of detailed scheduling decisions made at a given point in time is the second factor in this category. This factor involves making one decision at a time or many. Deciding upon which method is appropriate affects the choice of a scheduling scheme. One scheduling approach is to make all decisions for a group of jobs prior to the entry in to the system. This approach is classified as the off-line scheme. Another approach is to delay individual scheduling decisions until the last moment, which requires making one decision at a time frequently throughout the period. This method is the real time scheme. Some researchers suggest that using an off-line scheme for making detailed scheduling decisions improves system performance because it considers the condition of the entire shop. They arque that real-time schemes are rather myopic; they only consider current in formation about one small area of the system. Other researchers arque that off line schemes are impractical; that subsystem capacity constraints make such an approach difficult; that machine breakdowns rquire the frequent creation of new schedules; and that system delays (such as material handling delays) would in validate a fixed schedule. The overall effect of the last two arguments is that a spstem that uses an off-line scheme would have to reschedule frequently. Some researchers conclude that the heavy computational requirement of developing a single schedule makes the off-line scheme impractical. The last factor in this category concerns characteristics of the system. The number of characteristics considered affects the complexity of the scheduling scheme. The six levels of this factor are: machine breakdowns, material hand ling capacity, tool magazine capacity, pallet and fixture capacity, in-system sto rage capacity and routing flexibility. The trend toward more intermediate and random job shop systems with periodic demand will affect the chocice of a scheduling scheme for the average system in the future. Both the input sequencing and detailed scheduling prob lems are relevant in this environment. Either real time or off line schemes can be used; however, the following discussion suggests that of line schemes are more appropriate. Off-line schemes can take greater advantages of a situation in which a number of parts are to be scheduled for the upcoming period. Systems with per- xiii iodic demand provide such an opportunity. In addition, some researchers found that the performance difference of off-line scheduling schemes versus real-time schemes for random job shop systems with periodic demand incerased as rou ting flexibility increased. These two arguments support the appropriateness of off line schemes for the average system in the future. However, arguments have been made against the use of off-line schemes in existing FMS. The primary, argument is that the computational requirements of optimal off-line schemes are too large. This is partly due to complexities caused by subsystem capacity constraints and partly due to the complexities caused by routing flexibility. However, heuristic off-line procedures in systems with large amounts of routing flexibility have been found to provide good, fast performance. The second argument is that machine breakdowns will require frequent rescheduling, causing the performance of off-line schemes to diminish. Some researchers demonstrated that the performance of off-line schemes does not diminish even when the frequency of break downs is severe. In addition, more reliability is predicted for future systems. Therefore, offline schemes seem most appropriate for the average system of the future. In this study; a scheduling aigoritm for a flexible machining and assembl- my systems was presented. The machining and assembly systems are linked by the Automated Guided Vehisle System. Parts are machined first and then sent for assembly. The FMS scheduling problem is decomposed in to two subpro- blems: 1 ) Aggregate scheduling problem 2) Detailed scheduling problem. The aggregate problem generates a schedule for products, and it is equivalent to the two machine flow shop scheduling problem. Machining system corresponds to machine 1 and the assembly system corresponds to machine 2 'in the two machine flow shop problem. Johnson's algorithm can be applied to the aggregate scheduling problem. Before the detailed scheduling problem, we must know product priority, job priority and job schedulability status. There is number of constraints of FMS related to availability of machines, tools, fixtures and pallets. Each Job Jj is ready for scheduling when its schedulability status X-, = 1. Based on this consi derations a scheduling algorithm for the detailed level is presented.