Logistics of Closed-Loop Textile Recycling

Post-consumer carpet accounts for more than a quarter of all textiles discarded at municipal solid waste centers. Therefore, recycling of post-consumer carpet may reduce pressure on landfills as well as direct secondary materials back into production. To make recycled materials competitive with virgin materials, the cost of recycled materials needs to be as low as possible. Due to the high bulkiness of carpet, the transportation cost of post-consumer carpet is high which makes carpet reverse logistics a significant portion of the total cost of recycled materials.;This research focuses on two aspects of the carpet reverse logistics problem in the US, the location of collection centers as well as the design of the recycling network. To be economically feasible, acquisition of old carpet has to rely on the willingness of consumers and flooring installers to bring old carpet to collection centers. A well-designed collection network is required, with these centers located in close proximity to highly populated areas. Such a network can provide sufficient volumes of carpet to take advantage of economies of scale at large recycling plants while requiring the minimum number of centers. The collection network problem is formulated as a set covering optimization model with partial coverage. In order to solve difficult instances of this NP-hard problem, a novel greedy randomized heuristic is created by combining and extending greedy approaches for similar problems available in the literature. The design of the heuristic allows it to work efficiently with large unicost covering problems that have sparse coverage matrices. Computational results show that the heuristic performs better than other greedy heuristics proposed in the literature for similar types of problems. By applying the heuristic, a set of nationwide collection networks utilizing different target collection rates has been designed. Two different cases are considered: one extends the current collection network and another builds a new collection network. The relationship between the target collection rate and the required number of collection centers is identified. Using the relationship, an appropriate target collection rate can be established by considering the effort and investment required to build the corresponding collection network.;In the second part of the research, the design of the recycling network for Nylon 6 carpet is determined. Location of reverse processing facilities in the recycling network, as well as identification of which reverse activities should be performed at each layer of the network can significantly reduce the logistics costs. Three alternative network designs for nationwide carpet recycling systems are developed and compared. In two scenarios, the networks include layers of local collection centers, recycling plants, and markets for recycled materials. In the third scenario, a layer of regional collection centers is inserted before the recycling plants to aggregate carpet for more efficient sorting and transportation. To find the optimal number and locations of the recycling plants (and regional collection centers) and the optimal flows among network facilities, a hierarchical facility location model is formulated. To solve large instances of the problem, a heuristic method based on the alternative location-allocation procedure is developed and a computational study is conducted to assess its performance. Three alternative configurations of a Nylon 6 carpet recycling network in the US are designed, and the scenario that includes the intermediate layer of regional collection centers reduces the total cost of the network significantly. In addition, the cost of recycled Nylon 6 is determined to be very sensitive to the utilization of the recycling plants, and in order to minimize cost, the recycling network should receive a sufficient volume of carpet to operate the recycling plants at full capacity.