|CIMINO CHIARA||Cycle: XXXIII |
Section: Systems and Control
Tutor: BASCETTA LUCA
Advisor: LEVA ALBERTO Major Research topic
:Dynamic modelling, simulation and control for optimised production management in the Industry 4.0 contextAbstract:
The Industry 4.0 (I4.0 for short) revolution, sometimes also called "interconnected manufacturing", brings about improvements in the quality and efficiency of manufacturing processes and assets by exploiting the so called Internet of Things (IoT), i.e., a network where all the "things" (machines, processes, and so forth) can talk to each other and share information.
The mentioned improvements can be specialized in many ways, taking e.g. the form of energy consumption limitation, fault reduction and
mitigation, fault prediction and predictive maintenance, or any suitably weighed combination thereof. Also, I4.0 entails improvements in process monitoring, including real-time analysis, and as a results eases and enhances all the types of analysis that can help to improve production.
There is a vast literture supporting the ideas just sketched
A less frequently addressed -- but by no means less important -- aspect, is however that I4.0 also inherently fosters an integration of design,
operation and management throughout the life cycle of a manufacturing asset, and more specifically, that the key to success from this
standpoint is a pervasive use of dynamic models with boundaries and complexity tailored to answer any specific question about the said
asset, but at the same time structurally made consistent to one another thanks to the above mentioned shared basis of information.
The problem to address is that the involved dynamic models are heavily heterogeneous, ranging for example from first-principle physical ones in
the continuous time, to identified discrete-time ones for prediction, to discrete-event ones for both control design and assessment, up to data-
and learning-based ones for high-complexity management tasks; also, throughout the necessarily over-simplified taxonomy above, both
deterministic and stochastic paradigms come into play.
Besides establishing communications among the involved objects, a mathematical problem of paramount importance for I4.0 is therefore to
establish some holistic point of view and framework for the coordinated creation, use and maintenance of models. In this respect, the
dissertation aims at providing a contribution by relating different types of dynamic simulation models. The focus shall be set primarily on
detailed discrete-event models versus synthetic probabilistic ones, in a view to isolate uncertainty and allow for scalable-detail studies on
e.g. queue-like systems. For completeness, however, case studies shall be also considered in which the addressed manufacturing system comprises
nontrivial continuous-time phenomena.