Researchers and engineers alike in biomedical applications are witnessing a revolution. Our innovative MultiphysicsAI platform is changing the way piezoelectric micromachined ultrasonic transducers (PMUTs) are designed. This cloud-based solution integrates finite element method (FEM) simulation with advanced neural surrogates, allowing specialists to efficiently explore intricate design trade-offs between sensitivity and bandwidth. This novel technology embraces an overall efficient PMUT optimization process. It provides guaranteed performance gains within minutes, as opposed to the days required for each manual iteration.
MultiphysicsAI greatly accelerates the PMUT design process, using a training dataset of 10,000 randomized geometries. This, paired with an unprecedented training dataset of 2 million survey responses, enabled the AI surrogates to reach an awe-inspiring mean error rate of only 1%. The surrogates provide sub-millisecond inference for critical performance indicators (KPIs). These indicators are transmit sensitivity, center frequency, fractional bandwidth, and electrical impedance. Today, MultiphysicsAI accelerates and automates the design process. This frees up engineers to focus on innovation rather than spending hours on tedious calculations.
Enhanced Performance Through Optimization
Another standout feature of MultiphysicsAI is its capacity for Pareto front optimization. This optimization technique increases the fractional bandwidth of PMUTs from 65% to 100%. Simultaneously, it increases sensitivity by 2-3 dB. Remarkably, this method achieves a center frequency of 12 MHz within ±0.2% tolerance. Such capabilities are game changing for engineers focused on developing PMUTs of emerging performance levels needed in high-performance, high-demand biomedical applications.
Beyond performance, the optimization process serves to decrease engineering overhead. With MultiphysicsAI, weeks of deterministic, manual iteration become seconds of streamlined, data-driven exploration. This smart engineering tool arms engineers with quality data and science to make informed decisions early in the design process. This kind of efficiency is absolutely essential in such an industry where time-to-market can make or break the company’s competitive edge.
Case Study Validates Effectiveness
A newly released use case from Lam Research provides some additional context about the power of MultiphysicsAI. This unfortunately required optimizing four geometric parameters over 10,000 coupled piezoelectric-structural-acoustic simulations. The outcomes demonstrated that the platform rapidly produced validated performance improvement. It demonstrated its capacity to cut cumbersome, complicated, manual, bureaucratic processes into a simple, efficient digital process.
MultiphysicsAI delivers tremendous power and flexibility through the use of industry standard cloud infrastructure. This democratizes its use among a myriad of professionals and removes the barrier of needing specialized resources. The need for next-generation PMUTs is climbing at an equally fast pace. Tools such as MultiphysicsAI that dramatically improve design efficiency and performance are now available.