The course is aimed at providing the Students with practical concepts and instruments for modelling complex systems in their operative environment or asset, potentially simulating the entire life-cycle of an industrial system (including degradation phenomena) through its Digital-Twin.

The course will merge the multidisciplinary inputs from energy, mechanical, electrical, control and management engineering in a unique and coherent virtual framework that can be used for design, operation and monitoring optimizations.

The course will introduce students to the field of industrial system monitoring, specifically focusing on the role of models (and Digital-Twins) in the diagnosis and prognosis of industrial systems and components, exploiting statistical pattern recognition and machine learning to interpret changes in their measured features.

The course will provide additional insight to the modelling of energy plant systems, with laboratory experiences focused on realistic application scenarios, also involving national and/or international industrial representatives.

By the end of the course, students will learn contents and practices according to what defined in the leaning objectives.

In terms of acquired knowledge and understanding, students will be able to:

• Examine scenarios where Digital-Twin models can bring benefits
• Predict benefits and drawbacks of modelling methods in the framework of industrial system design, operation and monitoring optimization
• Interpret existing application studies in the literature and implement methodologies appropriate for solving complex problems, both systematically and creatively.

Concerning the ability to apply the acquired knowledge and understanding, students will be able to:

• Apply the methods covered during the course on realistic scenarios
• Demonstrate the pros and cons of the modelling strategy in relation to the application scenario
• Critically, independently and creatively solve problems with some originality in new or unfamiliar environments within the multidisciplinary context of Digital-Twins

Through a schedule of hands-on practices and laboratories, students will also acquire the skills to formulate a judgment, meaning to

• Analyse the modelling framework
• Formulate subjective expectations on the results
• Judge the results as a comparison with prior expectations
• Provide a sound judgments even on the basis of incomplete or restricted information

Furthermore, through the group laboratories, students will gain the ability to autonomously take initiative to identify and address learning needs for further knowledge, leveraging on the multidisciplinary background of the group. Additionally the students will improve their management and relational soft skills by working as a group of experts.

Finally, students will improve their communication skills, being able to communicate their conclusions and recommendations with the argumentation of the knowledge and rationale underpinning these, to both specialist and non-specialist audiences clearly and unambiguously.

The course contents are delivered through lectures and hands-on computer laboratory practices with professors and/or a tutor.

The course covers the following topics divided in modules as per the description below:

Module 1 - Introduction

Topic 1.1 - Introduction and motivation for the rapid growth of Digital-Twins for industrial applications. Real case examples will be provided as input to the students, thus setting the motivation and goal of the course. Specific attention will be devoted to the application of Digital-Twins for system health management and maintenance optimization.

Topic 1.2 - Design and maintenance criteria evolution (from safe life to damage tolerant design, from fault driven to predictive maintenance). Introduction of Health and Usage Monitoring Systems as a means toward process automatization.

Module 2 – Physics-based modelling

Topic 2.1 - Physics-based modelling: application of time-based simulation with MATLAB SIMULINK-SIMSCAPE. Hand-on practices on energy system simulation.

Topic 2.2 - Degradation and failure mechanisms: an overview of different degradation and failure mechanisms occurring on the most widely used components in the energy field of application is provided, including fatigue, creep, corrosion, pitting, lubricant degradation, battery State of Charge, including an overview of failure and degradation mechanisms. The students will particularly focus on the analytical and numerical modelling strategies for predicting damage progression.

Topic 2.3 - Time-based modelling of system degradation: simulation is used to predict in a virtual environment the effect of a potential damage over the system features observed by a sensor. 

Module 3 – Event-based modelling

Topic 3.1 - Event-based modelling: Discrete Event Simulation (DES) is introduced as a means to simulate systems depending on discrete time, with particular focus on event-based degradation modelling. Practical laboratories using MATLAB SimEvents (in combination to Simulink) will be provided, e.g. modelling scenarios typical of the operation research discipline.

Topic 3.2 - Decision logic modelling: concepts and instruments will be provided to simulate the decision logics in an operative framework. Practical laboratories using MATLAB State-Flow will be provided, simulating realistic scenarios of industrial systems’ management, e.g. the battery management system and the operation and monitoring of energy systems.

Module 4 – Surrogate modelling

Topic 4.1 – Surrogate modelling: for most application, the Digital-Twin should be fast enough to be run in real-time. This is not feasible with direct simulation, while surrogate models can be used for approximating input-output relations. Different methods based on Machine Learning (e.g. Artificial Neural Networks, Gaussian Processes, etc.) will be considered for the definition of surrogate models.

Module 5 – Model updating

Topic 5.1 - Basics of Monte-Carlo sampling: the concept of Monte-Carlo sampling will be recalled as a mean to implement repeated and efficient (forced sampling) simulation of the Digital-Twin.

Topic 5.2 – Digital-Twin updating: the Digital-Twin must be updated during service life based on observations by sensors. Various methods for parameter identification and tracking based on Bayesian inference and Monte-Carlo sampling will be implemented in the framework of damage identification and operation optimization, including the Metropolis-Hastings Monte-Carlo Markov Chain and the Particle Filter.

The theoretical aspects introduced in the course will be correlated with laboratories, for a direct implementation of the methods, e.g. including the implementation of a digital-twin for the Health and Usage Monitoring and Prognostic Health Management of energy systems and components, e.g. including pressure vessels, pumps, batteries, etc..

• The students will apply the concepts developed during the course to real-case studies.
• Students will analyze the industrial system, modelling its components in their operative scenario, including potential degradation and failure mechanisms and the logistic support to guarantee their operability.
• Students will perform trade-off analysis to support management decisions
• Finally, the students will provide a report in the form of a presentation of their results, that will be evaluated as part of their final score.

• SDG7 - AFFORDABLE AND CLEAN ENERGY
• SDG9 - INDUSTRY, INNOVATION AND INFRASTRUCTURE
• SDG12 - RESPONSIBLE CONSUMPTION AND PRODUCTION

The course describes methods for the operation optimization of energy systems, relying in the development of a digital-twin. To this aim, the course paves the way towards (i) a more reliable and modern energy produciton system (SDG7), (ii) a more resilient and sustainable energy production industry (SDG9-SDG12)

Students must have acquired the following skills:

• command of MATLAB working principles and basic scripting is appreciated, otherwise covered with some additional materials and instructions at the beginning of the course;
• basic knowledge of statistics for data analysis;
• fundamentals of theoretical and applied mechanics, solid mechanics, electrical systems and industrial plants, as provided in the B.Sc. in Energy Engineering

If necessary, students will be provided with references and material to fill the gap.

Students will pass the course after delivering a report of the practices and laboratories developed together with a tutor and an oral examination intended to assess the understanding of the concepts delivered through the reports.

[Report]

Via this exam, it will be possible to test students’ acquired knowledge and understanding, their capability to apply the acquired knowledge and understanding, and their capacity to synthetise them through short reports. A 3-page report is required for some specific practices and laboratories, including a short summary of the topic goal, the method used with boundary hypothesis for its application to the case study, the results and comments on the results. Groups of 2 students are allowed in making the reports.

[Oral exam]

Via an oral examination, it will be possible to test students’ contribute in the reporting, asking to provide comments on particular aspects, decisions, hypothesis, etc. made in the context of the practices. Through theory questions, it will also allow to test the student command on the main aspects treated during the course. Depending on the number of students enrolled in each exam session, theory questions might be posed in a written format.

Final Grading

To pass the course, students must successfully pass the oral test (grade higher or equal to 18/30). Reports will be assessed based on a A to E grade, being A Excellent, C sufficient and E definitely insuffient.