PhD in Physics-Informed and Explainable AI for Wind Turbine Diagnosis

• An indefinite contract from Monday to Friday, along with a flexible schedule and a top-tier calendar.
• Possibility to telework up to two days per week and even a third day from one of our locations closest to your home.
• Develop a doctoral thesis applied to a real industrial challenge relevant to the energy transition.
• Train in advanced artificial intelligence technologies, physics-informed models, explainability, and industrial monitoring.
• Work jointly with specialists from IKERLAN and researchers from the UPC.
• Access to expert knowledge, simulation models, and experimental and computing infrastructures.
• Participate in international conferences and publish your research results in high-impact scientific journals.

To thrive in this team, you will need:

• Completed Master's degree, or nearing completion, in Industrial Engineering, Mechanical Engineering, Mechatronics, Automation, Data Science, Artificial Intelligence, Mathematics, Physics, or related degrees.
• Programming knowledge, preferably in Python.
• Knowledge of machine learning, artificial intelligence, or data analysis.
• Interest in the physical modeling of systems, machinery monitoring, and predictive maintenance.
• Ability to understand and work with mathematical models and sensor data.
• Sufficient level of English to read scientific documentation, write publications, and participate in international conferences.
• Interpersonal skills, work organization, initiative, autonomy, and teamwork.
• Oral and written communication skills.
• Motivation to develop a research career applied to industrial problems.
• And, above all, a passion for technology.

Would you like to develop an artificial intelligence that not only detects anomalies, but is also capable of explaining what is physically happening in a machine?

You will join the Data Analytics Engineering (DAE) specialization team at IKERLAN, where we develop and productize solutions based on data, models, and digital twins for the diagnosis, prognosis, and optimization of industrial systems.

Your thesis will be developed within the framework of the PhIE-WinTurCoM project, in collaboration with the Universitat Politècnica de Catalunya (UPC). You will research new physics-informed and explainable artificial intelligence techniques to improve the diagnosis of wind turbine drivetrains. The objective will be to overcome the scarcity of failure data, estimate physical parameters associated with degradation, quantify the severity of anomalies, and provide useful engineering explanations regarding their root cause.

You will combine physical models, SCADA data, and vibration signals from CMS systems to turn AI into a reliable, interpretable, and applicable diagnostic tool in real industrial conditions.

• Analyze the state of the art in physics-informed artificial intelligence, explainable diagnosis, and rotating machinery monitoring.
• Research hybrid models that integrate operational data, vibration signals, and physical knowledge of the system.
• Develop inverse models based on physics-informed neural networks to estimate unobservable parameters related to degradation, such as changes in stiffness, damping, or heat generation due to friction.
• Research techniques to quantify the uncertainty of estimates and provide confidence intervals associated with the diagnosis.
• Develop explainable AI methods that identify which signals and variables have caused an anomaly and translate the results into concepts understandable for engineering.
• Contribute to the development of the diagnostic software framework, integrating physical models, AI algorithms, and analysis tools.
• Validate methodologies using real data, synthetic data, and test benches.
• Participate in scientific dissemination through publications and international conferences.
• Collaborate in transfer and outreach activities with companies.
• Generate results transferable to other rotating systems and industrial sectors.
• And, of course, write and defend your doctoral thesis.