Intelligent robotic disassembly optimisation for sustainability using the bees algorithm

Hartono, Natalia ORCID: 0000-0003-2314-1394 (2023). Intelligent robotic disassembly optimisation for sustainability using the bees algorithm. University of Birmingham. Ph.D.

Preview

Hartono2023PhD.pdf
Text - Accepted Version
Available under License All rights reserved.
Download (17MB) | Preview

Abstract

Robotic disassembly plays a pivotal role in achieving efficient and sustainable product lifecycle management, with a focus on resource conservation and waste reduction. This thesis discusses robotic disassembly sequence planning (RDSP) and robotic disassembly line balancing (RDLB), with a specific emphasis on optimising sustainability models. The overarching goal was to enhance the efficiency and effectiveness of disassembly processes through intelligent robotic disassembly optimisation techniques. At the heart of this research lies the application of the Bees Algorithm (BA), a metaheuristic optimisation algorithm inspired by the foraging behaviour of honeybees. By harnessing the power of the BA, this research aims to address the challenges associated with RDSP and RDLB, ultimately facilitating sustainable disassembly practices. The thesis gives an extensive literature review of RDSP and RDLB to gain deeper insight into the current research landscape. The challenges of the RDSP problem were addressed in this work by introducing a sustainability model and various scenarios to enhance disassembly processes. The sustainability model considers three objectives: profit, energy savings, and environmental impact reduction. The four explored scenarios were recovery (REC), remanufacture (REM), reuse (REU), and an automatic recovery scenario (ARS). Two novel tools were developed for assessing algorithm performance: the statistical performance metric (SPM) and the performance evaluation index (PEI). To validate the proposed approach, a case study involving the disassembly of gear pumps was used. To optimise the RDSP, single-objective (SO), multiobjective (MO) aggregate, and multiobjective nondominated (MO-ND) approaches were adopted. Three optimisation algorithms were employed — Multiobjective Nondominated Bees Algorithm (MOBA), Nondominated Sorting Genetic Algorithm - II (NSGA-II), and Pareto Envelope-based Selection Algorithm - II (PESA-II), and their results were compared using SPM and PEI. The findings indicate that MO-ND is more suitable for this problem, highlighting the importance of considering conflicting objectives in RDSP. It was shown that recycling should be considered the last-resort recovery option, advocating for the exploration of alternative recovery strategies prior to recycling. Moreover, MOBA outperformed other algorithms, demonstrating its effectiveness in achieving a more efficient and sustainable RDSP. The problem of sequence-dependent robotic disassembly line balancing (RDLBSD) was next investigated by considering the interconnection between disassembly sequence planning and line balancing. Both aspects were optimised simultaneously, leading to a balanced and optimal disassembly process considering profitability, energy savings, environmental impact, and line balance using the MO-ND approach. The findings further support the notion that recycling should be considered the last option for recovery. Again, MOBA outperformed other algorithms, showcasing its capability to handle more complex problems. The final part of the thesis explains the mechanism of a new enhanced BA, named the Fibonacci Bees Algorithm (BAF). BAF draws inspiration from the Fibonacci sequence observed in the drone ancestry. This adoption of the Fibonacci-sequence-based pattern reduces the number of algorithm parameters to four, streamlining parameter setting and simplifying the algorithm’s steps. The study conducted on the RDSP problem demonstrates BAF’s performance over the basic BA, particularly in handling more complex problems. The thesis concludes by summarising the key contributions of the work, including the enhancements made to the BA and the introduction of novel evaluation tools, and the implications of the research, especially the importance of exploring alternative recovery strategies for end-of-life (EoL) products to align with Circular Economy principles.

Type of Work:

Thesis (Doctorates > Ph.D.)

Award Type:

Doctorates > Ph.D.

Supervisor(s):