Computational dynamics in industry and healthcare : from heavy machinery to human locomotion analysis
Jiang, Dezhi (2026-06-12)
Väitöskirja
12.06.2026
Lappeenranta-Lahti University of Technology LUT
Acta Universitatis Lappeenrantaensis
School of Energy Systems
School of Energy Systems, Konetekniikka
Kaikki oikeudet pidätetään.
Julkaisun pysyvä osoite on
https://urn.fi/URN:ISBN:978-952-412-470-6
https://urn.fi/URN:ISBN:978-952-412-470-6
Kuvaus
ei tietoa saavutettavuudesta
Tiivistelmä
Computational dynamics provides a mathematical approach to simulating complex physical systems, valuable in both engineering and healthcare. This dissertation applies this discipline to two areas: improving heavy machinery simulation using game-based environments and analyzing human gait for physical rehabilitation monitoring. The subject work addresses overly simplified hydraulic component models and the need for practical tools to remotely track patient walking progress.
In the first application, a real-time multibody excavator model is integrated with gamification to enable user-driven parameterization of hydraulic valve flow characteristics. Sixteen participants performed tasks across three difficulty levels, with quantitative and qualitative data collected. Results showed 35% average improvement in points and productivity from level one to two, which demonstrating effective skill acquisition. The platform enabled exploration of quick-opening, linear, and equal-percentage valve characteristics, revealing distinct effects on machine behavior.
The second application develops a simple mechanical model that estimates the mechanical cost of transport per step using only center of mass kinematic data. Analysis of synthetic trajectories revealed that pendular motion achieves the lowest cost (0.11), level walking achieved the highest (0.22), with sinusoidal motion as an intermediate (0.19). Experimental data yielded dimensionless mechanical cost values of 0.14 to 0.17, aligning with reported dimensionless metabolic cost of transport. A narrative review confirmed parallel trends between mechanical and metabolic costs in rehabilitation contexts, particularly when interventions provide net positive work.
Unifying these applications is the principle that appropriate model simplification, guided by physical understanding, yields practical tools. The frameworks developed offer path-ways toward continuous, data-driven assessment in operator training and remote physical rehabilitation monitoring.
In the first application, a real-time multibody excavator model is integrated with gamification to enable user-driven parameterization of hydraulic valve flow characteristics. Sixteen participants performed tasks across three difficulty levels, with quantitative and qualitative data collected. Results showed 35% average improvement in points and productivity from level one to two, which demonstrating effective skill acquisition. The platform enabled exploration of quick-opening, linear, and equal-percentage valve characteristics, revealing distinct effects on machine behavior.
The second application develops a simple mechanical model that estimates the mechanical cost of transport per step using only center of mass kinematic data. Analysis of synthetic trajectories revealed that pendular motion achieves the lowest cost (0.11), level walking achieved the highest (0.22), with sinusoidal motion as an intermediate (0.19). Experimental data yielded dimensionless mechanical cost values of 0.14 to 0.17, aligning with reported dimensionless metabolic cost of transport. A narrative review confirmed parallel trends between mechanical and metabolic costs in rehabilitation contexts, particularly when interventions provide net positive work.
Unifying these applications is the principle that appropriate model simplification, guided by physical understanding, yields practical tools. The frameworks developed offer path-ways toward continuous, data-driven assessment in operator training and remote physical rehabilitation monitoring.
Kokoelmat
- Väitöskirjat [1213]