|Tipo di tesi||Tesi di laurea magistrale|
|Autore||VITTORI ANTISARI, IRENE|
|Titolo||Implementation and validation of fault-detection and fault-diagnosis indicators in an internal combustion engine exploiting residual generation techniques|
|Titolo in inglese||Implementazione e validazione di indicatori di rilevamento e diagnosi anomali in un motore a combustione interna sfruttando tecniche di generazione di residui|
|Struttura||Dipartimento di Ingegneria "Enzo Ferrari"|
|Corso di studi||Advanced Automotive Engineering (D.M.270/04)|
|Data inizio appello||2021-10-21|
|Disponibilità||Accessibile via web (tutti i file della tesi sono accessibili)|
L’obiettivo della tesi, sviluppata presso il dipartimento di testing dell’azienda HPE COXA, è quello di creare uno strumento che possa identificare ed isolare nella maniera più precisa possibile un’anomalia del sistema di aspirazione di un motore a combustione interna durante le prove di test al banco. La necessità nasce dal momento che i sistemi di allarme già implementati sono volti alla sola salvaguardia del motore; una volta segnalata l’anomalia, all’operatore è lasciato capire quale componente stia provocando l’allarme, attraverso analisi dati, test visivi etc.
The objective of the thesis, developed at the testing department of the HPE COXA company, is to create a tool that can identify and isolate as precisely as possible a fault in the intake system of an internal combustion engine during the tests at the bench. The need arises because the alarm system already implemented is aimed only at safeguarding the engine; once the fault is reported, the operator needs to understand which component is causing the alarm, through data analysis, visual tests etc. The first part of the thesis describes the different types of faults that can be encountered and the various detection and diagnosis methods. The work environment in which the job was developed and the main characteristics of the engine analyzed are then described. In particular, it is a high-performance spark ignition engine, turbocharged, with variable valve timing (VVT) and direct injection (GDI). The actual development of the thesis is reported on the following chapters. The idea developed is as follows: a model that describes the operation of the engine intake system has been built, by creating maps that represent the behavior of each component when there are no anomalies. The maps have been created starting from the data acquired during a bench test which took place after the conclusion of the calibration work. The created maps link together some data acquired on the bench, following, where possible, the physical laws that govern the functioning of the analyzed air intake system; the objective of the model is to exploit these physical laws to predict the value of some quantities starting from the measured data. These predicted values are then compared with those just acquired to understand if the system is behaving in the same way or differently from what is expected by the model. Comparison often consists in subtracting the modeled value from the measured value, thus generating a residual. The first comparison was made between the data that were used to build the model and the model itself. Statistical methods were used, in particular the probability density function (PDF) and the cumulative distribution function (CDF) to analyze the results and, together with the definition of the diagnostic method, define the limits within which the residuals are considered. not abnormal. Subsequently, the measured values were artificially altered, to record how the model reacts to a possible malfunction of a sensor. Finally, the data recorded at the time when an anomaly was actually present were introduced into the model. The maps have been generated through a MATLAB code while the model was implemented in Simulink.