Guest lectures and seminars - Page 44

Maintenance plays a crucial role in ships and especially in the vital electric propulsion system. Intelligent predictive maintenance idealistically aims at preventing system failures and minimizing needless repairs, i.e., predicting failure likelihood and time to failure while providing the crew explainable predictions and recommending the best action for timely intervention. This presentation will cover a relevant work in collaboration with Sensor Systems in BigInsight, particularly a paper published under https://doi.org/10.1109/TII.2022.3144177. The failure prediction approach is driven by event logs, which include warnings, alarms, and operational information that describe all the happenings onboard the ship. The failure prediction objective is turned into classification and regression tasks; however, the training data pose three challenges. The events are irregular textual messages. The training data samples are not labelled. The datasets are extremely imbalanced, due to sparse failure events and multiple failure modes. The problem is casted into a weakly supervised machine learning framework. In a multiple instance learning process, the ungiven data labels are learned recursively while fitting the model parameters using deterministic annealing. The overall approach was tested on real ship data, and it successively forecasted few propulsion failures with explainable causes.

Abstract: We investigate the simulation of a rising bubble and a stationary droplet interaction to gain a better understanding of the rising dynamics and the morphology changing of bubble-droplet aggregate. A detailed study is conducted on the interaction process under different-size bubbles with various combinations of spreading factors. The current simulation framework consists of the conservative phase-field Lattice Boltzmann equation (LBE) for interface tracking and the velocity-pressure LBE for hydrodynamics. We simulate the contact line dynamics to confirm the method's accuracy. We further investigate the morphology changing of two contact droplets under different combinations of spreading factors and depict the final morphologies in a diagram. The separated, partially engulfed and complete engulfed morphologies can be replicated by systematically altering the sign of the spreading factors. The rising bubble and droplet interaction is simulated based on different final morphologies by adding a body force. The results show that the aggregate with double emulsion morphology can avoid distortion and maintain a greater terminal velocity than the aggregate with partially engulfed morphology.

Talk is online on Zoom. Please contact "timokoch at uio.no" for the Zoom link. This talk is part of the Mechanics Lunch Seminar series. Bring-your-own-lunch and lots of questions.

Deep learning (DL) has had unprecedented success and is now entering scientific computing with full force. However, current DL methods typically suffer from instability, even when universal approximation properties guarantee the existence of stable neural networks (NNs). In this talk we will show that there are basic well-conditioned problems in scientific computing where NNs with great approximation qualities are proven to exist, however, there does not exist any algorithm, even randomised, that can train (or compute) such a NN to even 1-digit of accuracy with a probability greater than 1/2. These results provide basic foundations for Smale’s 18th problem ("What are the limits of AI?") and imply a potentially vast classification theory describing conditions under which (stable) NNs with a given accuracy can be computed by an algorithm. We begin this theory by initiating a unified theory for compressed sensing and DL, leading to sufficient conditions for the existence of algorithms that compute stable NNs in inverse problems. We introduce Fast Iterative REstarted NETworks (FIRENETs), which we prove and numerically check (via suitable stability tests) are stable. The reference for this talk is: https://arxiv.org/abs/2101.08286 (to appear in Proc. Natl. Acad. Sci. USA).