A Digital Twin for Deep Vibro Ground Improvement-Video

Video Transcription

Video Summary

Negin presents a data-centric digital twin for detecting anomalies in deep ground improvement, developed with Keller. It analyzes multivariate sensor data from vibro-replacement rigs alongside geospatial information, since the subsurface work cannot be inspected directly. The system calculates more than 70 expert-defined KPIs in real time and offers maps, smoothed models, heatmaps, and point-level reports to help engineers identify unusual patterns and investigate results. Combining expert knowledge with supervised and unsupervised machine learning helps catch anomalies more reliably. The approach aims to improve quality control, reduce costs and risks, and support digitalization across construction processes.

Keywords

digital twin

anomaly detection

deep ground improvement

multivariate sensor data

machine learning