The cardiovascular system is the system that operates for the nutrients and oxygen providing for the entire body tissues and cells which is obligatory in terms of homeostasis. It also affects maintaining the body temperature. Even though it consists of many active and passive elements, the heart is the fundamental element of this system as the actuator. Essentially, pumping of the blood to the whole body for gas and fluid exchange, and feeding the cells, and also it is the main element for conveying the oxygen-poor blood to the lungs to re-oxygenate it. In terms of the volumetric importance and operative load, the left ventricle can be named one of the most important chambers in the heart, and failures that are arising from them become the major cause of mortality nowadays. Therefore, diagnosing the possible behaviors and features of the diseases and pathologies which are sourced by heart failure, has become more important than in the past. Since in vivo experiments and applications are strictly hard to test in human body experiments, computational and mathematical models of the cardiovascular system also become one of the widest ranges of research fields. Current applications of the modeling might be based on the mammalian myocardium test parameters, so the capacitance and resistance variables of the physiological elements are mostly taken as constant parameters. System equations for the models are determined for the static and unchangeable test conditions. In this thesis, as a novel approach, the model is assessed in both hydraulic and electrical activities, by neural network application. Differently from the past studies which are assumed that all the energy given to the heart is actively converted into fluid pressure, within the derivation of the analogous mass – spring – damper model of the myocardium, which is based on the Hill muscle model, utilization of the black box feedforward neural network training model can be realized. To constitute the indicator of the dynamicity of the heart, it is coming up with an activation parameter α, and through training and testing the model on a customized mock loop system in the lab environment, a novel, accurately operated mathematical model for the usage of the different pathologic cases for the future research has been proposed
Il sistema cardiovascolare è il sistema che opera per i nutrienti e l'ossigeno fornendo i tessuti e le cellule dell'intero corpo, che è obbligatorio in termini di omeostasi. Influisce anche sul mantenimento della temperatura corporea. Nonostante sia composto da molti elementi attivi e passivi, il cuore è l'elemento fondamentale di questo sistema in quanto attuatore. In sostanza, pompa il sangue a tutto il corpo per lo scambio di gas e fluidi e alimenta le cellule, ed è anche l'elemento principale per convogliare il sangue povero di ossigeno ai polmoni per riossigenarlo. In termini di importanza volumetrica e carico operatorio, il ventricolo sinistro può essere definito una delle camere più importanti del cuore e gli insuccessi che ne derivano diventano oggi la principale causa di mortalità. Pertanto, è diventato più importante che in passato la diagnosi dei possibili comportamenti e delle caratteristiche delle malattie e delle patologie che sono all'origine dello scompenso cardiaco. Poiché gli esperimenti e le applicazioni in vivo sono rigorosamente difficili da testare negli esperimenti sul corpo umano, anche i modelli computazionali e matematici del sistema cardiovascolare diventano uno dei campi di ricerca più vasti. Le attuali applicazioni della modellazione potrebbero essere basate sui parametri del test del miocardio nei mammiferi, quindi le variabili di capacità e resistenza degli elementi fisiologici sono per lo più prese come parametri costanti. Le equazioni di sistema per i modelli sono determinate per le condizioni di prova statiche e immutabili. In questa tesi, come nuovo approccio, il modello viene valutato sia in attività idrauliche che elettriche, mediante l'applicazione di reti neurali. Diversamente dagli studi precedenti che presuppongono che tutta l'energia data al cuore sia attivamente convertita in pressione del fluido, all'interno della derivazione dell'analogo modello massa – molla – damper del miocardio, che si basa sul modello muscolare di Hill, utilizzo di il modello di addestramento della rete neurale feedforward della scatola nera può essere realizzato. Per costituire l'indicatore della dinamicità del cuore, sta escogitando un parametro di attivazione α e, attraverso l'addestramento e il test del modello su un sistema ad anello simulato personalizzato nell'ambiente di laboratorio, un nuovo modello matematico accuratamente gestito per l'utilizzo di sono stati proposti i diversi casi patologici per la ricerca futura
Identifying the heart dynamics parameters by a neural network assessing Hill model features of the myocardium : a study on mock-loop data
VURAL, BERKEM
2021/2022
Abstract
The cardiovascular system is the system that operates for the nutrients and oxygen providing for the entire body tissues and cells which is obligatory in terms of homeostasis. It also affects maintaining the body temperature. Even though it consists of many active and passive elements, the heart is the fundamental element of this system as the actuator. Essentially, pumping of the blood to the whole body for gas and fluid exchange, and feeding the cells, and also it is the main element for conveying the oxygen-poor blood to the lungs to re-oxygenate it. In terms of the volumetric importance and operative load, the left ventricle can be named one of the most important chambers in the heart, and failures that are arising from them become the major cause of mortality nowadays. Therefore, diagnosing the possible behaviors and features of the diseases and pathologies which are sourced by heart failure, has become more important than in the past. Since in vivo experiments and applications are strictly hard to test in human body experiments, computational and mathematical models of the cardiovascular system also become one of the widest ranges of research fields. Current applications of the modeling might be based on the mammalian myocardium test parameters, so the capacitance and resistance variables of the physiological elements are mostly taken as constant parameters. System equations for the models are determined for the static and unchangeable test conditions. In this thesis, as a novel approach, the model is assessed in both hydraulic and electrical activities, by neural network application. Differently from the past studies which are assumed that all the energy given to the heart is actively converted into fluid pressure, within the derivation of the analogous mass – spring – damper model of the myocardium, which is based on the Hill muscle model, utilization of the black box feedforward neural network training model can be realized. To constitute the indicator of the dynamicity of the heart, it is coming up with an activation parameter α, and through training and testing the model on a customized mock loop system in the lab environment, a novel, accurately operated mathematical model for the usage of the different pathologic cases for the future research has been proposedFile | Dimensione | Formato | |
---|---|---|---|
LM_BerkemVural_Thesis.pdf
accessibile in internet per tutti
Descrizione: thesis
Dimensione
2.11 MB
Formato
Adobe PDF
|
2.11 MB | Adobe PDF | Visualizza/Apri |
Executive_Summary.pdf
accessibile in internet per tutti
Descrizione: summary
Dimensione
927.89 kB
Formato
Adobe PDF
|
927.89 kB | 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/10589/189103