Lung involvement is a frequent and clinically relevant comorbidity in a wide spectrum of diseases, including connective tissue diseases, Sjögren’s syndrome, rheumatoid arthritis, systemic vasculitides, and viral infections such as COVID-19. Despite their diverse etiologies, these conditions often converge toward interstitial lung disease (ILD), a major determinant of morbidity and mortality. Early identification of ILD remains a crucial yet challenging goal, as early clinical manifestations are often subtle and typically become evident only at advanced stages. Current diagnostic methods, particularly high resolution computed tomography (HRCT), are unsuitable for routine screening due to cost and radiation exposure. Physical examination, specifically lung auscultation, offers a simple, reproducible, and non invasive means for early ILD detection. “Velcro-like” crackles are characteristic of fibrotic involvement of the lung parenchyma and correlate with radiologic findings such as honeycombing and ground glass opacities. In recent years, the automated detection and classification of pathological lung sounds has become an active field of research, leveraging digital signal processing and machine learning. However, the lack of standardization in data acquisition and the limited sensitivity of current electronic stethoscopes restrict diagnostic performance. The scope of this thesis consists of designing, simulating, developing, and testing a novel vibrating membrane, also known as diaphragm, specifically optimized for pulmonary sound acquisition. The new diaphragm is intended to enhance the performance of electronic stethoscopes in terms of acoustic bandwidth and sensitivity. To this aim, an initial simulation model representing the current non-optimized membrane configuration was developed and validated against reference data, with a commercial diaphragm produced by Littmann used as a benchmark, designed as a compromise between cardiac and pulmonary auscultation. Once the model was verified, it was employed to determine the optimal design parameters maximizing the membrane’s dynamic performance. By improving the mechanical response, the proposed membrane enables the system to capture lung sounds with richer spectral content and higher signal quality, thus providing a more informative input for subsequent computational and diagnostic analysis. The research integrates analytical modeling, one-dimensional (1D) and two/three-dimensional (2D/3D) simulations, and experimental prototyping. The membrane geometry and material parameters were systematically optimized to extend frequency response and pressure gain while maintaining structural integrity. The developed prototypes have been tested in an experimental setup composed by a programmable shaker, an accelerometer and a laser vibrometer. The shaker provides the excitation to the vibrating membrane, whereas the sensors allow to acquire the force applied to the diaphragm and the related displacement. The tested prototypes demonstrated strong agreement with the simulated models, confirming the validity of the numerical approach and establishing it as a valuable design tool for future membrane development. In particular, the proposed design improves the quality of the acquired signal and lays the foundation for more reliable automated identification of pathological pulmonary sounds, ultimately contributing to earlier diagnosis of ILD and related disorders. This work thus bridges clinical needs and engineering innovation, offering a technological advancement that may support more effective, less expensive and non invasive screening strategies for interstitial lung disease.
Il coinvolgimento polmonare rappresenta una comorbidità frequente e clinicamente rilevante in un’ampia gamma di patologie, tra cui malattie del tessuto connettivo, sindrome di Sjögren, artrite reumatoide, vasculiti sistemiche e infezioni virali come il COVID-19. Pur avendo eziologie differenti, queste condizioni tendono spesso a evolvere verso la malattia polmonare interstiziale (ILD), uno dei principali determinanti di morbidità e mortalità. L’identificazione precoce dell’ILD è un obiettivo cruciale ma complesso, poiché i sintomi iniziali sono spesso lievi e diventano evidenti solo in stadi avanzati. Le attuali metodiche diagnostiche, in particolare la tomografia computerizzata ad alta risoluzione (HRCT), non risultano adatte allo screening di routine a causa dei costi e dell’esposizione a radiazioni. L’esame obiettivo, e in particolare l’auscultazione polmonare, offre un approccio semplice, riproducibile e non invasivo per la rilevazione precoce dell’ILD. I caratteristici “crepitii a Velcro” indicano il coinvolgimento fibrotico del parenchima polmonare e correlano con reperti radiologici come honeycombing e opacità a vetro smerigliato. Negli ultimi anni, il riconoscimento automatico dei suoni polmonari patologici è divenuto un campo di ricerca attivo grazie all’elaborazione digitale del segnale e al machine learning. Tuttavia, la mancanza di standardizzazione nella raccolta dei dati e la limitata sensibilità degli stetoscopi elettronici attuali ne riducono le prestazioni diagnostiche. Questa tesi ha come obiettivo la progettazione, simulazione, sviluppo e validazione di una nuova membrana vibrante (o diaframma) ottimizzata per acquisizione suoni polmonari, mirata a migliorarne le prestazioni in termini di larghezza di banda acustica e sensibilità. A tal fine, è stato sviluppato un modello di simulazione rappresentativo della configurazione attuale non ottimizzata, validato rispetto a dati di riferimento presi con la membrana Littmann progettata come compromesso tra suoni cardiaci e polmonari . Una volta verificato, il modello è stato impiegato per individuare i parametri di progetto ottimali, massimizzando la risposta dinamica della membrana. Migliorando il comportamento meccanico, la membrana proposta consente di acquisire suoni polmonari con contenuto spettrale più ricco e qualità del segnale superiore, fornendo così un input più informativo per le successive analisi computazionali e diagnostiche. La ricerca combina modellazione analitica, simulazioni monodimensionali (1D) e bi/tridimensionali (2D/3D) con attività sperimentale. La geometria e i parametri del materiale della membrana sono stati ottimizzati per massimizzare la risposta in frequenza e il guadagno in pressione, preservando l’integrità strutturale. I prototipi realizzati sono stati testati in un setup sperimentale composto da shaker programmabile, accelerometro e vibrometro laser: lo shaker fornisce l’eccitazione, mentre i sensori misurano la forza applicata e lo spostamento del diaframma. I risultati sperimentali hanno mostrato un’elevata concordanza con le simulazioni, confermando la validità dell’approccio numerico e il suo valore come strumento di progettazione per future evoluzioni del dispositivo. Il design proposto migliora la qualità del segnale acquisito e pone le basi per un’identificazione automatizzata più affidabile dei suoni polmonari patologici, favorendo una diagnosi più precoce dell’ILD e di patologie correlate. Questo lavoro rappresenta un ponte tra esigenza clinica e innovazione ingegneristica, offrendo un avanzamento tecnologico che potrà supportare strategie di screening per la malattia polmonare interstiziale più efficaci, economiche e non invasive.
Progettazione, simulazione, prototipazione e test di una nuova membrana vibrante adatta a migliorare l’acquisizione dei suoni polmonari negli stetoscopi elettronici / Marco Modena , 2026 May 22. 38. ciclo, Anno Accademico 2024/2025.
Progettazione, simulazione, prototipazione e test di una nuova membrana vibrante adatta a migliorare l’acquisizione dei suoni polmonari negli stetoscopi elettronici
MODENA, MARCO
2026
Abstract
Lung involvement is a frequent and clinically relevant comorbidity in a wide spectrum of diseases, including connective tissue diseases, Sjögren’s syndrome, rheumatoid arthritis, systemic vasculitides, and viral infections such as COVID-19. Despite their diverse etiologies, these conditions often converge toward interstitial lung disease (ILD), a major determinant of morbidity and mortality. Early identification of ILD remains a crucial yet challenging goal, as early clinical manifestations are often subtle and typically become evident only at advanced stages. Current diagnostic methods, particularly high resolution computed tomography (HRCT), are unsuitable for routine screening due to cost and radiation exposure. Physical examination, specifically lung auscultation, offers a simple, reproducible, and non invasive means for early ILD detection. “Velcro-like” crackles are characteristic of fibrotic involvement of the lung parenchyma and correlate with radiologic findings such as honeycombing and ground glass opacities. In recent years, the automated detection and classification of pathological lung sounds has become an active field of research, leveraging digital signal processing and machine learning. However, the lack of standardization in data acquisition and the limited sensitivity of current electronic stethoscopes restrict diagnostic performance. The scope of this thesis consists of designing, simulating, developing, and testing a novel vibrating membrane, also known as diaphragm, specifically optimized for pulmonary sound acquisition. The new diaphragm is intended to enhance the performance of electronic stethoscopes in terms of acoustic bandwidth and sensitivity. To this aim, an initial simulation model representing the current non-optimized membrane configuration was developed and validated against reference data, with a commercial diaphragm produced by Littmann used as a benchmark, designed as a compromise between cardiac and pulmonary auscultation. Once the model was verified, it was employed to determine the optimal design parameters maximizing the membrane’s dynamic performance. By improving the mechanical response, the proposed membrane enables the system to capture lung sounds with richer spectral content and higher signal quality, thus providing a more informative input for subsequent computational and diagnostic analysis. The research integrates analytical modeling, one-dimensional (1D) and two/three-dimensional (2D/3D) simulations, and experimental prototyping. The membrane geometry and material parameters were systematically optimized to extend frequency response and pressure gain while maintaining structural integrity. The developed prototypes have been tested in an experimental setup composed by a programmable shaker, an accelerometer and a laser vibrometer. The shaker provides the excitation to the vibrating membrane, whereas the sensors allow to acquire the force applied to the diaphragm and the related displacement. The tested prototypes demonstrated strong agreement with the simulated models, confirming the validity of the numerical approach and establishing it as a valuable design tool for future membrane development. In particular, the proposed design improves the quality of the acquired signal and lays the foundation for more reliable automated identification of pathological pulmonary sounds, ultimately contributing to earlier diagnosis of ILD and related disorders. This work thus bridges clinical needs and engineering innovation, offering a technological advancement that may support more effective, less expensive and non invasive screening strategies for interstitial lung disease.| File | Dimensione | Formato | |
|---|---|---|---|
|
Modena.pdf
Open access
Descrizione: Modena.Marco.pdf
Tipologia:
Tesi di dottorato
Dimensione
4.06 MB
Formato
Adobe PDF
|
4.06 MB | Adobe PDF | Visualizza/Apri |
Pubblicazioni consigliate

I metadati presenti in IRIS UNIMORE sono rilasciati con licenza Creative Commons CC0 1.0 Universal, mentre i file delle pubblicazioni sono rilasciati con licenza Attribuzione 4.0 Internazionale (CC BY 4.0), salvo diversa indicazione.
In caso di violazione di copyright, contattare Supporto Iris




