|Institution:||University of Michigan|
|Department:||Industrial and Operations Engineering|
|Keywords:||Modeling and optimization of disassembly systems; Industrial and Operations Engineering; Engineering|
|Full text PDF:||http://hdl.handle.net/2027.42/111570|
Remanufacturing is a promising product recovery method that brings new life to cores that otherwise would be discarded thus losing all value. Disassembly is a sub-process of remanufacturing where components and modules are removed from the core, sorted and graded, and directly reused, refurbished, recycled, or disposed of. Disassembly is the backbone of the remanufacturing process because this is where the reuse value of components and modules is realized. Disassembly is a process that is also very difficult in most instances because it is a mostly manual process creating stochastic removal times of components. There is a high variety of EOL states a core can be in when disassembled and an economic downside due to not all components having reuse potential. This thesis focuses on addressing these difficulties of disassembly in the areas of sequence generation, line balancing, and throughput modeling. In Chapter 2, we develop a series of sequence generation models that considers the material properties, partial disassembly, and sequence dependent task times to determine the optimal order of disassembly in the presence of a high variety of EOL states. In Chapter 3, we develop a joint precedence graph method for disassembly that models all possible EOL states a core can be in that can be used with a wide variety of line balancing algorithms. We also develop a stochastic joint precedence graph method in the situation where some removal times of components are normal random variables. In Chapter 4, we further advance the analytical modeling framework to analyze transfer lines that perform routing logics that result from a high variety of EOL states, such as a restrictive split routing logic and the possibility that disassembly and split operations can be performed at the same workstation.