Super-sampling-frequency structural response prediction based on compressive sensing and frequency-domain evaluation index

Structural response prediction methods have been proven effective in forecasting uncollected responses, which offers a strategy to mitigate the burden on data transmission and storage. However, the existing methods may need excessive data volume as they can only predict responses from the same sampling frequency responses, limiting the application scope of these methods. To overcome this limitation, this paper introduces a new prediction scheme, namely super-sampling-frequency structural response prediction, for forecasting high-sampling-frequency responses from low-sampling-frequency collected responses. To realize this prediction scheme, a method based on Compressive Sensing (CS) and a novel Frequency-domain Evaluation Index (FEI) is proposed, named Compressive Sensing with Frequency-domain Evaluation Index (CS-FEI). This FEI is introduced to address the lack of indices for determining the frequency atom number in CS. Accelerations collected from a numerical model and a real-world 600-meter-tall skyscraper are used for verification and comparison. The results show that CS-FEI can realize super-sampling-frequency response prediction with high performance. It offers practical potential as an effective response prediction method, alleviating the substantial data demands inherent in current methods. © 2025 Elsevier Ltd.