PORTALE DELLA DIDATTICA

Ricerca CERCA
  KEYWORD

Process control data analysis for anomalies characterization and predictive maintenance for smart manufacturing

keywords ARTIFICIAL NEURAL NETWORKS, INDUSTRIAL PROCESSES, INDUSTRY 4.0, SMART MANUFACTURING

Reference persons ANDREA ACQUAVIVA, ENRICO MACII

Research Groups ELECTRONIC DESIGN AUTOMATION - EDA

Thesis type EXPEIRMENTAL, IN COMPANY

Description Plants of the future are expected to be able to quickly detect and characterize anomalies that can prevent product quality. Recently developed learning algorithms and complex systems analysis techniques can be successfully applied to this purpose specially in quality-driven manufacturing processes that are consistently monitored and for which a large amount of process control data are available.

The main objective of this thesis is to design a framework for anomalies detection in industrial plants. More specifically, exploiting data coming from process control technologies, the target is to fully characterize anomalies by correlating with operating conditions and with external factors or environmental conditions. The final purpose is to build an anomaly characterization framework, that will provide information about the possible causes of the anomaly as well as the dependencies between the controlled process and external factors (i.e. different from controlled parameters). The outcome of the framework can be used also for predictive maintenance and process control optimization.

Required skills programming, artificial neural networks or machine learning tools (main elements)

Notes The thesis will be in collaboration with FCA and includes a period in the company plant to learn and characterize the industrial processes to be analyzed and optimized


Deadline 31/12/2017      PROPONI LA TUA CANDIDATURA




© Politecnico di Torino
Corso Duca degli Abruzzi, 24 - 10129 Torino, ITALY
m@il