Ψηφιακό δίδυμο κινητήρα οχήματος για την πρόβλεψη βλαβών με την χρήση τεχνητής νοημοσύνης
Digital twin of a vehicle engine for fault prediction using Artificial Intelligence
View/
Keywords
Ψηφιακό δίδυμο ; Μηχανική μάθηση ; Συσκευή OBD II ; Πρόβλεψη βλαβών ; Τεχνητή νοημοσύνη ; Digital twin ; Machine learning ; OBD II Device ; Fault prediction ; Artificial IntelligenceAbstract
This research work deals with the development of a digital twin thermal vehicle engine for fault
prediction using Artificial Intelligence. It includes the study, analysis and evaluation of the
measurements collected in files from the thermal engine, and used to create and train a
machine learning model, and the appropriate classification and prediction algorithms. In each
file, the measurements were recorded by a mobile phone application (Bimmerlink mobile app)
installed on a mobile phone interconnected via Bluetooth with an OBD II diagnostic device. The
OBD II diagnostic device was connected to the corresponding OBD port of a BMW passenger
vehicle. The data concern snapshots of measurements from 32 engine parameters and the
corresponding sensors that monitor the engine operation and the movement of the vehicle. The
files with the snapshots of the measurements were processed using the Miniconda Orange data
analysis application. In this application, a supervised machine learning model was created for
the purpose of predicting vehicle failures, since it provides a set of widgets that allow the
processing of data files, the creation of machine learning models, the training and evaluation of
classification algorithms, the creation and evaluation of predictions. The deliverables include the
analysis and processing of measurement data, in terms of parameter correlations, scatter plots,
and the statistical evaluation indicators for the classification algorithms used. From the results
obtained and in combination with studies from the international bibliography, conclusions have
been drawn for the dataset and an assessment of which algorithm is the most appropriate to be
used for data classification, failure prediction and preventive engine maintenance.
Postgraduate Studies Programme
Ψηφιακός Πολιτισμός, Έξυπνες Πόλεις, IoT και Προηγμένες Ψηφιακές ΤεχνολογίεςDepartment
Σχολή Τεχνολογιών Πληροφορικής και Επικοινωνιών. Τμήμα ΠληροφορικήςNumber of pages
82Language
GreekCollections
Except where otherwise noted, this item's license is described as
Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα
Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα