Shield data batch processing and cleaning method based on physical mechanics constraints

Abstract

This invention discloses a batch processing and cleaning method for tunnel boring machine (TBM) data based on physical and mechanical constraints. The method includes: acquiring TBM data fragments; extracting the actual characteristic value range of parameters and matching it with a pre-constructed physical typical value range library; actively inferring and unifying the dimensions; adaptively calculating working condition thresholds based on the extreme value characteristics of the data; dividing physical tunneling working conditions and extracting pure steady-state data within the propulsion stage; performing multi-dimensional physical and mechanical constraint verification on the pure steady-state data, including anomaly detection in the spatial distribution area of ​​array sensors, multi-parameter physical coupling constraint verification, and long-time-series sensor gradual drift detection, generating system-level noise labels; calculating a data quality score including a physical reliability dimension based on the label; and persistently storing the cleaned data carrying the label and score. This invention can effectively identify and eliminate hidden system-level noise that violates engineering mechanical constraints.

Images