Earth pressure balance tunnel boring machine vibration-based classification charts for real-time ground type and composition detection

This paper presents a novel study on real-time vibration-based ground characterisation during earth pressure balance tunnel boring machine (EPBM) tunnelling. Vibration signals were collected while excavating through various rock types in Changsha City, China, namely, slate and granite with varying degrees of weathering, with the aid of a series of accelerometers deployed on the EPBM. Frequency domain analysis of these vibration signals revealed two key findings: the spectral amplitudes within 100–130 Hz strongly correlate with ground types; and the spectral profile tends to flatten when encountering mixed face ground conditions, due to the impulse signals generated by ground interfaces. These findings facilitate real-time detection of ground type and composition by way of fast Fourier transform analysis of vibration signals. Two classification charts were developed: one for full-face rock type detection using average spectral amplitude (within 100–130 Hz) and penetration rate, and another for ground composition detection in mixed face conditions using mean square frequency and spectrum flatness. The proposed charts yield error-free identifications for all the 1752 recorded points in the project.