Running prosthetic feet are carbon fiber blades whose mechanical characterization, in terms of stiffness and energy dissipation, is essential for optimizing the configuration of the prosthesis and enhancing the performance of Paralympic athletes. To date, however, no standardized protocols exist for the objective dynamic characterization of their stiffness under real-use conditions: manufacturers assign stiffness categories based solely on the athlete's body weight, without providing measurable or reproducible mechanical values. This work aimed to assess the methodological feasibility of developing a lumped-parameter multibody model with three specific objectives: (i) to assess its capacity to estimate the experimental ground reaction force (GRF) curves and the characteristic distance , defined as the distance between the upper and lower extremities of the blade and used as a proxy for its compression, through quantitative comparison based on the normalized root-mean-square error (nRMSE) with respect to the experimental data; (ii) to identify, by means of Particle Swarm Optimization (PSO), the equivalent torsional stiffness K* [N·m/deg] and the equivalent damping coefficient D* [N·m·s/deg] for each prosthetic configuration and speed range; (iii) to assess the biomechanical plausibility of the identified parameters as a function of prosthetic category and running speed. The study was based on a set of experimental trials consisting of instrumented 60-metre sprints performed by two Paralympic athletes with transfemoral and transtibial amputation, using four prosthetic categories and two speed ranges. The model discretized the prosthetic foot into sixteen rigid segments interconnected by a lumped torsional spring-damper element placed between two blocks at the main curvature region. Kinematics was prescribed via 3D translation and pitch angle derived from experimental measurements. Results show satisfactory agreement between simulation and experimental data, with mean nRMSE of 6.56-14.08% for GRF and 7.46-10.06% for characteristic distance. Concerning K*, we found that it increases with prosthetic category, consistent with compression stiffness values comparable to bench-test literature data and decreases with running speed. D* is systematically higher in transtibial configurations; its presence is supported by the hysteresis loop observed between GRF and blade angular compression, indicating incomplete elastic energy return. Statistical analysis confirms highly significant differences between amputation types and prosthetic foot categories (p < 0.001). The proposed approach represents the first attempt at simultaneous identification of equivalent torsional stiffness and damping in running prosthetic feet under real dynamic conditions.
I piedi protesici da corsa sono lame in fibra di carbonio la cui caratterizzazione meccanica in termini di rigidezza e dissipazione energetica è determinante per ottimizzare la configurazione protesica e migliorare le prestazioni degli atleti paralimpici. Ad oggi, tuttavia, non esistono protocolli standardizzati per la caratterizzazione dinamica oggettiva della rigidezza in condizioni di utilizzo reale: i produttori assegnano categorie di stiffness basandosi esclusivamente sul peso corporeo dell'atleta, senza fornire valori meccanici misurabili e riproducibili. Il presente lavoro valuta la fattibilità metodologica di un modello multibody a parametri concentrati, sviluppato in ambiente MATLAB/ Simulink, con tre obiettivi specifici: ) ideare un modello a parametri concentrati e verificare la sua capacità di riprodurre le curve sperimentali di forza di reazione al suolo (GRF) e la distanza caratteristica, definita come la distanza tra le estremità superiore e inferiore della lama e utilizzata come indicatore della sua compressione, mediante confronto quantitativo tramite nRMSE; (ii) identificare, attraverso ottimizzazione Particle Swarm (PSO), la rigidezza torsionale equivalente K* [N·m/deg] e il coefficiente di smorzamento equivalente D* [N·m·s/deg] per ciascuna configurazione protesica e fascia di velocità; (iii) verificare la plausibilità biomeccanica dei parametri identificati in funzione della categoria protesica e della velocità di corsa. Le prove sperimentali consistono in sprint strumentati su 60 metri, eseguiti da due atleti paralimpici con amputazione transfemorale e transtibiale, con quattro categorie protesiche e due fasce di velocità. Il modello discretizza il piede protesico in sedici segmenti rigidi interconnessi, con una molla torsionale con smorzatore concentrati collocata tra due blocchi nella zona di curvatura. Il moto è imposto tramite traslazione 3D e angolo di pitch ricavati dalle misurazioni sperimentali. I risultati evidenziano un accordo soddisfacente tra simulazione e dati sperimentali, con nRMSE medio di 6.56-14.08% per la GRF e 7.4-10.06% per la distanza caratteristica. K* abbiamo riscontrato che aumenta con la categoria protesica, in modo coerente con valori di rigidezza a compressione confrontabili con dati di letteratura da test a banco, e diminuisce con la velocità di corsa. D* risulta sistematicamente più elevato nelle configurazioni transtibiali; la sua presenza è supportata dal ciclo di isteresi osservato tra la GRF e la compressione angolare della lama, indicativo di una restituzione incompleta dell'energia elastica. L'analisi statistica conferma differenze altamente significative tra tipologie di amputazione e categorie di piede protesico (p < 0.001). L'approccio proposto costituisce il primo tentativo di identificazione simultanea di rigidezza torsionale e smorzamento equivalente in piedi protesici da corsa in condizioni dinamiche reali.
Sviluppo di un modello multi-corpo per la stima dei parametri elastico-dissipativi in piedi protesici da corsa
USAI, CHIARA
2025/2026
Abstract
Running prosthetic feet are carbon fiber blades whose mechanical characterization, in terms of stiffness and energy dissipation, is essential for optimizing the configuration of the prosthesis and enhancing the performance of Paralympic athletes. To date, however, no standardized protocols exist for the objective dynamic characterization of their stiffness under real-use conditions: manufacturers assign stiffness categories based solely on the athlete's body weight, without providing measurable or reproducible mechanical values. This work aimed to assess the methodological feasibility of developing a lumped-parameter multibody model with three specific objectives: (i) to assess its capacity to estimate the experimental ground reaction force (GRF) curves and the characteristic distance , defined as the distance between the upper and lower extremities of the blade and used as a proxy for its compression, through quantitative comparison based on the normalized root-mean-square error (nRMSE) with respect to the experimental data; (ii) to identify, by means of Particle Swarm Optimization (PSO), the equivalent torsional stiffness K* [N·m/deg] and the equivalent damping coefficient D* [N·m·s/deg] for each prosthetic configuration and speed range; (iii) to assess the biomechanical plausibility of the identified parameters as a function of prosthetic category and running speed. The study was based on a set of experimental trials consisting of instrumented 60-metre sprints performed by two Paralympic athletes with transfemoral and transtibial amputation, using four prosthetic categories and two speed ranges. The model discretized the prosthetic foot into sixteen rigid segments interconnected by a lumped torsional spring-damper element placed between two blocks at the main curvature region. Kinematics was prescribed via 3D translation and pitch angle derived from experimental measurements. Results show satisfactory agreement between simulation and experimental data, with mean nRMSE of 6.56-14.08% for GRF and 7.46-10.06% for characteristic distance. Concerning K*, we found that it increases with prosthetic category, consistent with compression stiffness values comparable to bench-test literature data and decreases with running speed. D* is systematically higher in transtibial configurations; its presence is supported by the hysteresis loop observed between GRF and blade angular compression, indicating incomplete elastic energy return. Statistical analysis confirms highly significant differences between amputation types and prosthetic foot categories (p < 0.001). The proposed approach represents the first attempt at simultaneous identification of equivalent torsional stiffness and damping in running prosthetic feet under real dynamic conditions.| File | Dimensione | Formato | |
|---|---|---|---|
|
2026_Chiara_Usai_Executive_Summary.pdf
accessibile in internet per tutti
Descrizione: Executive Summary
Dimensione
654.84 kB
Formato
Adobe PDF
|
654.84 kB | Adobe PDF | Visualizza/Apri |
|
2026_Chiara_Usai_Tesi.pdf
accessibile in internet per tutti a partire dal 02/07/2027
Descrizione: Tesi
Dimensione
4.24 MB
Formato
Adobe PDF
|
4.24 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/261560