Orhan Torkul

An industrial visual inspection system that uses inductive learning

This paper presents an industrial visual inspection system that uses inductive learning. The system employs RULES-3 inductive learning algorithm to extract the necessary set of rules and template matching technique to process an image. Twenty 3×3 masks are used to represent an image. Each example consists of 20 frequencies of each mask. The system was tested on five different types of tea or water cups in order to classify the good and bad items. The system was trained using five good cups and then tested for 113 unseen examples. The results obtained showed the high performance of the system: the efficiency of the system for correctly classifying unseen examples was 100%. The system can also decide what type of the cup is being processed.

An integrated real time MRP and group technology system

The approach taken in this paper is twofold. First manufacturing environment is simplified for the purposes of planning and control without losing any of the essential characteristics. Second, a simple GT model is applied to the shop floor area and real time MRP is applied to the assembly area. The aim of this study is to develop and compare with a simulation of similar proposal except that jobshop is used in the shop floor area instead. The variable factors in both models were the set up time to operation time ratio and the intensity of the loading on the machines. In the highly loaded situations, the GT model faired better than the job shop model. However, for low loaded situations the performances of the two models were similar.

A real-time inventory model to manage variance of demand for decreasing inventory holding cost

Re-order point transformed into, a time driven model, real-time inventory model.Developed real-time inventory model eliminates the safety stock.Real-time model is proven to be better by some equations.Real-time model is compared with re-order point model by a simulation experiment. Variance of demand is one of the inevitable problems in the manufacturing environment. Market conditions and competition force companies reducing costs. Different approaches and methods have been developed to remedy these problems. Common points of these approaches are utilizing resources and responding in the shortest time possible. To achieve better inventory management, inventory-holding cost is tried to be decreased under the same service level. Re-order point is arranged dynamically regarding couple of factors in some studies. We try to arrange the re-order point in a time based manner by transforming the re-order point into re-order time, which eliminates the safety stock; and it makes the inventory model real-time.Recent production planning studies focus on real-time planning and dynamic scheduling to increase utilization and robustness. In these studies, real-time data is used for planning, but in most of them manufacturing systems and planning methodology are not transformed into a real-time system approach. The novel aspect of this study is the presentation of a model in which the manufacturing system has been designed as a real-time system that consists of real-time planning activities working with real-time data, and eliminating the safety stock.

Automatic generation of variants depending on changes of product properties in a flexible manufacturing environment

Automatic BOM formulation procedure is defined for variant generation.How CAD/CAM data can be used to construct dynamic BOMs is handled.Automatic variant generation steps and procedure are defined.Genetic algorithm is used for the system test.Data of 6228 BOMs is generated with definition of 665 BOMs, and 89% improvement is achieved by the number of BOM. Competition forces manufacturing systems to be flexible and to increase product variety and process complexity. These tasks depend on the flexible design of a bill of materials (BOM), one of the most important inputs in manufacturing planning and control systems. Product variety forces systems to generate BOMs with regard to product properties through a BOM pattern. A variant bill of materials provides a structure to manage product variability. In this study, an algorithm is designed to build a BOM pattern using computer-aided design and computer-aided manufacturing (CAD/CAM) data, and another algorithm is designed to generate variants with regard to product specifications. Genetic algorithm is used to generate new products to provide high product variability for testing algorithms. After the test, both algorithms are applied to a real industry problem. The BOM pattern is built automatically using CAD/CAM data, and variants are generated with regard to the pattern, and the results are discussed.

Applications of Inductive Learning to Automated Visual Inspection

In recent years, there has been a growing amount of research on inductive learning and its applications to different domains. Out of this research a number of promising algorithms have surfaced. Inductive learning algorithms are domain independent. In principle, they can be used in any task involving classification or pattern recognition. In this paper a number of applications of RULES family of induction algorithms to visual inspection are presented. The main advantages of using induction for visual inspection are: (a) The systems does not suffer from orientation problem which is very important for digital image processing. (b) The pattern does not have to be stored in the memory in graphics form because they are represented by rules. This saves memory space. (c) The decision can be reached in short time because the number of conditions in each rule and the total number of rules are not big. (d) It is easy to develop a software and design a hardware for these systems as they are not complicated.

An Improved Fuzzy C-Means Algorithm for Cell Formation Problems with Alternative Routes

Cell formation is one of the most important problems faced in designing cellular manufacturing systems. Fuzzy c-means FCM has been successfully used to solve a variety of the cell formation problems because it allows the representation of uncertain information. Many products or parts to be manufactured in the real world have alternative routes. To ignore the routes is not a realistic approach. However, most FCM approaches used to form a cellular system in the literature have ignored or avoided using the alternative routes because of its complexity. In this paper, an improved FCM algorithm has been proposed to overcome the computational complexity of the alternative routes. The improved algorithm presents an easy and practical way to solve the cell formation problems with alternative routes. An experiment was designed to test and compare the performance of the improved algorithm. The results of the experiment have shown that most of the obtained results are close to the test problems and better than the conventional crisp methods in the literature.

Gelecegin Fabrikasına Endüstriyel Bir Yaklaşım

Bilgisayar Butunlesik Imalatta (CIM) otomasyon, farkli imalat fonksiyonlari icin bir cok bilgisayar yazilim paketleriyle desteklenir.