MultiphysicsAI, a novel technology from Ansys startup MaxWaves, is radically transforming the way piezoelectric micromachined ultrasonic transducers (PMUTs) are designed for biomedical applications. This novel and unique method uses cloud-based finite element method (FEM) simulations in combination with advanced neural surrogates. Consequently, it supersedes the fundamental trial-and-error approach and introduces systematic inverse optimization. That shift allows engineers to realize validated performance improvements in a tenth of the time it once took.
The groundbreaking application of MultiphysicsAI helps experts easily balance even the most challenging design trade-offs, such as sensitivity and bandwidth. With the aid of fast simulations and smart data analysis, designers can make informed decisions that maximize the functionality of innovative PMUTs.
Accelerated Performance Improvement
The productivity of MultiphysicsAI is best illustrated by its capacity to perform large-scale simulations in minutes instead of days. Using off-the-shelf cloud infrastructure, the technology delivers repeatable performance improvements with low engineering burden. A recent case study demonstrated this capacity by iteratively optimizing four geometric parametric across 10,000 coupled piezoelectric-structural-acoustic finite element simulations.
This technique eliminates the huge majority of time spent iterating in manual fashion. Rather, it opens the door to fast, easy, transparent, data-driven exploration in a matter of seconds. You trained on 10K random geometries. Effectively, the AI surrogates allow for a mean error of only 1%. This level of accuracy allows for rapid assessments of key performance indicators pertinent to PMUT functionality.
Key Performance Indicators and Optimization
The main KPIs analyzed using MultiphysicsAI are transmit sensitivity, center frequency, fractional bandwidth, and electrical impedance. Using Pareto front optimization, the technology is able to significantly improve these metrics. For example, it doubles fractional bandwidth from 65% to a stunning 100%, all while improving sensitivity by 2-3 dB.
In addition to minimizing passband ripple, the optimization process guarantees the center frequency to stay equal to 12 MHz, with a tolerance of ±0.2%. This degree of accuracy is critical for biomedical applications where small variations can affect functionality.
Future Implications for Biomedical Applications
MultiphysicsAI does more than improve performance. It represents a paradigm change in PMUT design for multimodal biomedical applications. By leveraging the power of cloud computing and AI, engineers can explore more intricate design possibilities without being hindered by previous limitations.
MultiphysicsAI’s systematic inverse optimization shortens the design process. It blazes a trail for exciting new applications of ultrasonic technologies. The medical industry is increasingly adopting more sophisticated ultrasonic applications. This technology fusion will greatly improve the efficiency of biomedical devices.