Artificial Intelligence and Computational Intelligence for Engineering

Research Group Leadership

Group Leader / Principal Investigator: Prof. Mauro Tucci, Full Professor, University of Pisa.

Prof. Mauro Tucci leads a research group at DESTEC, University of Pisa, working on Artificial Intelligence and Computational Intelligence for engineering applications. He supervises PhD students and postdoctoral researchers and acts as scientific coordinator and PI-equivalent research leader for research projects and collaborations.

Research Group

I am a Full Professor of Electrical Engineering at the Department of Energy, Systems, Territory and Constructions Engineering (DESTEC), University of Pisa.

I lead a research group working on Artificial Intelligence, Machine Learning and Computational Intelligence for engineering applications. I coordinate the research activities of PhD students, postdoctoral researchers and young researchers, and serve as scientific coordinator and principal investigator, or PI-equivalent research leader, in research projects and collaborations with academic and industrial partners.

My research activity lies at the intersection of Artificial Intelligence, electrical engineering, energy systems and computational electromagnetics.

Main research areas

The main research activities of the group include:

Artificial Intelligence and Machine Learning for engineering applications
Energy forecasting and renewable energy systems
Condition monitoring, diagnostics and predictive maintenance
Deep Learning and surrogate modelling
Physics-Informed Machine Learning
Computational electromagnetics and inverse problems
Optimization of electrical and electromagnetic systems

Main research areas – detail

Artificial Intelligence for energy forecasting

A major research activity of the group concerns Machine Learning and Deep Learning methods for forecasting problems in electrical and energy systems.

Current and recent research includes:

photovoltaic power generation forecasting;
wind power forecasting at plant and regional level;
modelling and forecasting of hydroelectric systems;
electrical load forecasting;
electricity price forecasting;
exploitation and selection of meteorological and environmental information for renewable-energy forecasting;
analysis of the effect of data availability and forecast horizon on Machine Learning models.

The research addresses both methodological aspects and applications to real energy systems.

Condition monitoring, diagnostics and predictive maintenance

A second major activity concerns data-driven methods for condition monitoring, fault detection and predictive maintenance.

Machine Learning, Deep Learning and statistical techniques are investigated for detecting anomalous behaviour and predicting incipient faults from monitoring and sensor data.

Applications have included:

wind and photovoltaic power plants;
hydroelectric plants;
electrical devices and transformers;
railway systems and pantograph–catenary monitoring;
industrial monitoring systems.

Part of this research has been carried out in collaboration with industrial partners, including major Italian companies in the energy and railway sectors.

Artificial Intelligence for computational electromagnetics

The group develops Machine Learning and Deep Learning techniques for the analysis, modelling and optimization of electromagnetic problems and devices.

Recent research includes:

Deep Learning surrogate models for electromagnetic simulations;
Physics-Informed Neural Networks;
hybrid Boundary Element Method / Physics-Informed Neural Network approaches;
inverse electromagnetic problems and source identification;
synthesis of boundary conditions using neural networks;
topology optimization of electromagnetic devices;
model-order reduction and fast surrogate modelling;
AI-assisted modelling of biomedical electromagnetic problems.

These approaches aim to reduce the computational cost of conventional numerical simulations while maintaining high accuracy and exploiting available physical knowledge.

Recent applications also include efficient modelling and optimization techniques for transcranial magnetic stimulation (TMS).

Computational Intelligence and optimization

My research activity in Computational Intelligence started with Machine Learning and neural approaches and also includes evolutionary computation and optimization.

I have developed and investigated evolutionary optimization algorithms for complex engineering problems and applied them to the design and optimization of electromagnetic devices, including special electromagnetic systems and wireless power-transfer applications.

More recently, this line of research has evolved towards the integration of optimization, Deep Learning and surrogate modelling.

AI for signal-based diagnostics and inverse problems

Another research direction concerns extracting information about complex physical systems from indirect measurements.

Applications include:

signal-based diagnostics;
source localization and inverse problems;
lightning localization and estimation from induced electrical quantities;
identification of physical system parameters from measurements.

The common objective is to combine engineering models with data-driven methods to infer quantities that are difficult or expensive to measure directly.

Research approach

A common theme of our research is the integration of engineering knowledge, physical models and data-driven Artificial Intelligence.

Rather than treating Machine Learning as a purely black-box tool, we are particularly interested in approaches in which physical knowledge, numerical models, measured data and Artificial Intelligence complement each other.

The research group works on the complete methodological pipeline, including data analysis, model development, numerical implementation, validation on real or simulated systems, and application to engineering problems.

Research group and supervision

I lead and coordinate a group of young researchers, PhD students, postdoctoral researchers and research collaborators working on these topics at the University of Pisa.

I supervise research activities in Artificial Intelligence applied to engineering and coordinate research collaborations with academic institutions and industrial partners.

Students and researchers interested in the following topics are welcome to contact me regarding PhD research, Master’s theses and research collaborations:

Artificial Intelligence and Machine Learning for engineering;
Deep Learning and generative models;
energy forecasting;
predictive maintenance and condition monitoring;
Physics-Informed Machine Learning;
surrogate modelling;
computational electromagnetics;
engineering optimization.

Group Leader

Prof. Mauro Tucci
Full Professor of Electrical Engineering
Department of Energy, Systems, Territory and Constructions Engineering – DESTEC
University of Pisa