Laboratory for Water and Turbine Machines (LVTS)
Are you a motivated researcher ready to work at the intersection of acoustics, cavitation, high-power lasers, and Artificial Intelligence? We are seeking a PhD candidate to join the Laboratory for Water and Turbine Machines (LVTS) at the University of Ljubljana to develop "smart" sonochemical reactors. This project will be conducted in close collaboration with the Budapest University of Technology and Economics (BME). The Vision: A New Paradigm of Sonochemical Reactors The extreme conditions within collapsing bubbles—temperatures up to 12,000 K and pressures of 10,000 atm—enable chemical reactions impossible by conventional means. However, traditional reactors are inefficient and suffer from the erosive effect of bubbles. This project aims to move from fixed-frequency operation to a dynamic system where Reinforcement Learning (RL) agents manipulate complex acoustic pressure fields in real-time to organize bubbles into structured arrays or homogeneous clouds in order to maximize chemical yields.
Your Profile We are looking for a candidate who meets the following criteria:
- Education: A Master’s degree, preferably in Physics, Electronic Engineering, or Mechanical Engineering.
- Aptitude: A strong passion for experimental work and hands-on laboratory implementation.
- Skills: Experience in programming and electronics is preferred but not mandatory.
- Fluent English language required. What You Will Learn During this doctoral program, you will gain expertise in diverse advanced technologies and scientific fields:
- Acoustics & Bubble Dynamics: Non-linear radial oscillations and the forces trapping the bubbles as acoustic “clamps”.
- Hardware & Electronics: Designing signal generation systems using programmable boards (e.g. Teensy/FPGA).
- Signal processing: Use mathematical methods to analyze complex acoustic signals (e.g. FFT, Phase-Amplitude Coupling) *Optical Systems: Manipulation of high-power Nd:YAG lasers for bubble seeding and ultra-high-speed imaging up to several millions frames per second.
- Artificial Intelligence: Implementation of real-time control loops using machine learning, supported by experts from Budapest.
- Chemistry: Assessing chemical reactions and hydroxyl radical yields in sonicated liquids to develop real life industrial applications.
Experimental Work:
Your research will focus mostly on the experimental implementation of the project:
1.Reactor Design and Assembly: You will design and build a customized acoustic resonant chamber and implement a 3D bubble tracking system using an orthogonal video camera array.
2.Control of a Few Bubbles: You will use the RL agent to displace a few stable bubbles along arbitrary 3D paths while monitoring its radial dynamics and shape stability with high-speed video.
3.Cluster Manipulation: You will manage groups of up a few hundred bubbles, using laser-induced seeding to form structured arrays. 4.Yield Optimization: You will conduct experiments to maximize targeted chemical reactions, e.g. hydroxyl radical production, and use chemical analysis to provide feedback for the control algorithm and further train the reinforcement learning agent.
You will have access to state-of-the-art equipment at the LVTS laboratory and Budapest BME:
- Ultra-High-Speed Cameras: The Kirana 7M, capable taking 180 frames at an astonishing speed of 7 million frames per second. Other Photron cameras will cover ranges between 50 fps to 1 million fps to record an extended number of frames.
- Lasers: Pulsed Nd:YAG lasers for precise bubble generation.
- Instrumentation: Needle hydrophones for acoustic mapping and high-frequency amplifiers driven by a tailored programmable multi-channel signal generator.
- Computational Power: Local Nvidia A5000 GPUs for real-time AI inference and control of the bubble clouds.
Are you ready to pioneer the use of Reinforced Learning in Cavitation and shape the future of "smart" bubble dynamics? For more information on the application process and deadlines, please contact Dr. Juan Manuel Rosselló at the University of Ljubljana JuanManuel.Rossello@fs.uni-lj.si
Za vsebino oglasa, nadaljnji proces izbire kandidata in obliko medsebojnega sodelovanja med kandidatom in delodajalcem odgovarja delodajalec. Več informacij je na voljo tu
