RELIABILITY MODELING IN DESIGN FOR REMANUFACTURE

Remanufacture involves the production-batch disassembly, cleaning, replacement and refurbishment of worn parts in defective or obsolete products. For appropriate products, remanufacture offers significant economic and ecological advantages over other end-of-life options. Since the essential goal of remanufacture is part reuse, the reliability of components is important. The goal of this work is to consider reliability effects on life-cycle costs to enable design for reuse. A reliability model is developed to describe systems that undergo repairs performed during remanufacture or maintenance. First, the behavior of the model and preliminary experimental verification of the model are described. The model allows replacement of failed parts with both the same and different types of parts. An example simulation applied the reliability model to compare the effects on life-cycle cost of various combinations of mechanical components in a series system. MOTIVATION In addition to resource conservation, design for product endof-life is compelled by existing and impending product takeback laws that place product end-of-life responsibility on the manufacturer. Given this responsibility, the manufacturer may choose to pay increasingly higher fees for landfill or incineration, or have the product reused, repaired, remanufactured, or recycled for scrap material. While many design-for-end-of-life guidelines emphasize facilitating scrapmaterial recycling, significant resources are consumed during the recycling process. Furthermore, material degradation often results due to molecular breakdown and contamination, both of which are frequently characteristic of current recycling technologies. Individual product repair and maintenance are limited by the high labor costs that tend to cause discarding of repairable products. Lund (1983) observed that by recycling at the component instead of the material level, remanufacturing avoids the possibly unnecessary resource consumption of scrapmaterial recycling while preserving the value added to the component during manufacture. Also, the production-batch and off-site nature of remanufacturing results in a labor cost significantly lower than that required for individual repair. While remanufacture is not suitable for all products, it is particularly appropriate for technologically stable items, where a large fraction of components can be reused after refurbishment. Product design that facilitates any of the steps of remanufacture, namely disassembly, sorting, cleaning, refurbishment, reassembly and testing, will facilitate remanufacture. However, the essential goal in remanufacture is part reuse. If a part cannot be reused as is or after refurbishment, the ease of disassembly, cleaning or reassembly will not matter. Refurbishment activities aim to return a part to a like-new or better condition, and include, for example, reboring out-of-round cylinders or fitting cylinders with sleeves. When parts are to be reused, in either remanufacture or maintenance, the reliability of the part is very important. Collaboration was initiated with three companies that remanufacture a variety of products to learn about the remanufacture process and how products can be designed to facilitate remanufacture. These companies are Eastman Kodak, a manufacturer and remanufacturer of photocopiers, single-use cameras, and medical analysis equipment; Nashua Cartridge Products, a remanufacturer of toner cartridges; and Arrow Automotive Industries, a remanufacturer of automotive aftermarket parts. This collaboration offered insights on reliability issues across the companies and needs for reliability modeling. Existing reliability models are unsuitable for describing systems that undergo repairs performed during remanufacture. The goal of this research is to develop and verify reliability models to be used in life-cycle cost estimations of systems where reuse of working components is possible. These calculations help explore initial part design and remanufacture process plan alternatives in the context of other life-cycle concerns. Currently, this model is used in a genetic-algorithm Proceedings of The 1996 ASME Design Engineering Technical Conferences and Computers in Engineering Conference August 18-22, 1996, Irvine, California