|
Title
|
|
|
|
Generic industrially applicable algorithms to design mechanisms with minimal energy usage
| |
|
Author
|
|
|
|
| |
|
Abstract
|
|
|
|
| Energy consumption is a critical issue, particularly in industrial sectors where electric motors account for nearly 45% of global electricity use. This substantial share underscores the need for innovative energy efficiency solutions, driven by environmental concerns and the aim to reduce operational costs. While hardware upgrades, such as efficient motors or energy-recovery devices, are commonly pursued, they are often costly and limited by design constraints. A promising alternative is design optimization, which focuses on refining mechanical geometries to achieve energy savings. For instance, modifying mechanism link lengths in a case study reduced RMS motor torque by 71%. Despite its potential, existing design optimization methods are often complex, relying on analytical models that are not scalable or practical for industrial use. These methods are constrained by intricate dynamics and tend to prioritize precise motion over energy efficiency. Additionally, many optimization algorithms fail to guarantee global optimality for the proposed designs. This thesis addresses these limitations through four key objectives. First, it demonstrates the effectiveness of CAD-based simulations for optimizing system geometry, offering a scalable alternative to analytical models. A novel approach integrates CAD simulations with optimization algorithms to explore design configurations and reduce energy consumption. Second, the study develops efficient methods for guiding optimization within feasible design spaces, using Gaussian process models to ensure feasible simulations and probabilistic insights. Third, to address the computational demands of CAD-based optimization, Sparse Interpolation (SI) and Bayesian Optimization (BO) are employed, improving efficiency without compromising accuracy. Finally, global optimization search strategies are adopted to ensure comprehensive insights into solution optimality. Together, these contributions culminate in a generic, computationally efficient global design optimization methodology applicable to various industrial mechanisms. Future work explores co-optimization of geometry and motion profiles, revealing further energy savings through appropriate optimization architectures. The methodology demonstrates broad industrial applicability, particularly in manufacturing, robotics, and automation. Case studies—including an emergency ventilator mechanism, a weaving loom component, and a toggle press—showcase the robustness and versatility of the approach, achieving RMS motor torque reductions of 71%, 49.2%, and 49.2%, respectively. These findings highlight the method's value for industries seeking to reduce energy consumption while maintaining system precision and reliability. |
| |
|
Language
|
|
|
|
English
| |
|
Publication
|
|
|
|
Antwerpen
:
Universiteit Antwerpen
,
2024
| |
|
DOI
|
|
|
|
10.63028/10067/2109730151162165141
| |
|
Volume/pages
|
|
|
|
xvi, 162 p.
| |
|
Note
|
|
|
|
:
Derammelaere, S. [Supervisor]
:
Cuyt, A. [Supervisor]
| |
|
Full text (open access)
|
|
|
|
|
The publisher created published version Available from 19.12.2026
|
| |
|