Tonga volcano explosion generated Tsunamis in the Pacific ocean

The Hunga Tonga – Hunga Ha’apai submarine volcano exploded at 4:10AM UTC on 15^th January 2022 on the Tongan archipelago. The eruption was so violent that it created a plume of approximately 500 km in diameter and propelled ash 55 km up into the mesosphere and produced a pressure shock wave, inducing atmospheric pressure disturbances that have been captured for several days by weather stations around the Globe. The front of this atmospheric pressure wave travelled at approximately 0.9 Mach (1,127 km/h) and circled the Earth several times, decaying progressively. Shortly after the explosion, tsunami waves were detected in and along the coasts of the Pacific Ocean. The initial arrival time was earlier than expected based on the propagation speed of long gravity waves and coincides with the arrival time of the atmospheric pressure wave. Furthermore, tsunami waves were also detected in water bodies, too far away from the source for the tsunami waves to reach directly. Similar fluctuations in the form of tsunami or long waves captured by tidal gauges in an extremely far field had already been reported in previous volcanic explosion events, as for example those recorded in France and UK for the Krakatoa eruption in 1883. In this talk we will first illustrate how the moving air pressure front generates tsunamis and demonstrate the tsunami characteristics at far field, using several idealized problems. We will then use the existing data from a global numerical simulation of the atmospheric pressure wave to force the shallow water wave model and study the sea level responses in the Pacific Ocean and beyond. Numerical results are compared with the DART buoy data, which require corrections because of the presence of atmospheric pressure. The numerical model predicts well the arrival time and the magnitudes of the leading forced waves. Moreover, the numerical can demonstrate the appearance of trailing free wave trains behind the leading forced waves.

AI-Assisted Design-on-Simulation Technology for Advanced Packaging

Several design parameters affect the reliability life of area array type advanced packaging, such as upper and lower pad sizes, solder volume, buffer layer thickness, chip thickness, etc. An essential factor in the design of advanced packaging is how to fast and accurately predict the reliability life of packaging structure. Traditionally, the accelerated thermal cycling test (ATCT) is used to evaluate the reliability life of electronic packaging. However, optimizing the design parameters through ATCT is time-consuming and expensive. If the simulation approach is not adopted, it usually takes several ATCTs to design an electronic packaging structure and make it can pass the required mean-time-to-failure cycles. Each ATCT will take several months to get the reliability life results, which increases development time considerably. Hence, reducing the number of ATCT experiments becomes necessary.

Finite element simulation has proven to be the most effective method to reduce design cycles. However, the simulation results may vary according to the designer's domain knowledge, capabilities, and experience. This shortcoming can be overcome with artificial intelligence (AI). In this new approach, finite element analysis (FEA) is combined with machine learning algorithms, e.g., artificial neural network (ANN), support vector regression (SVR), kernel ridge regression (KRR), K-nearest neighbor (KNN), and random forest (RF), to create an AI model for predicting the reliability life of electronic packaging. AI-assisted design-on-simulation technology provides an efficient way for new electronic packaging design.