Abstract: Geological information is difficult to acquire when using advance geological prediction in long, deep-buried tunnels with complex terrains. To address this issue, a 1-km horizontal directional drilling (HDD)-based advance geological prediction is designed and implemented in the Tianshan Shengli tunnel, forming an integrated workflow that combines HDD, downhole multi-source testing (coring, borehole TV, and comprehensive logging), and drilling parameter analysis. In the field, a C4S1000 rig and systems for while-drilling guidance, mud circulation, and downhole testing achieve trajectory control. During drilling, mud pressure, flow rate, weight-on-bit, penetration rate, bit rotation speed, and torque are recorded, and mechanical specific energy (M_SE) is computed to indicate rock strength and jointing. Based on logging parameters and laboratory uniaxial compressive strength, the basic quality (BQ) index of the rock mass is calculated, and rock mass classes are assigned. These yield a parameter dataset and classification results that support construction organizations and risk management. The results show the following: (1) The integrated workflow operates stably in the tunnel, successfully drilling for 1 km while acquiring data. (2) The drilling trajectory meets control requirements, creating a maximum spatial offset of 10.161 m. (3)M_SE trends are consistent with variations in rock strength, exhibiting a three-stage development: a low-value section from 0 to 200 m, an overall lifting section of >300 m, and a value of 9.68 MPa at 612 m. This development can identify strength variation and potential risks. (4) The BQ-based rock classes derived from logging agree well with the conditions revealed on the tunnel face. These results verify the engineering applicability and effectiveness of HDD for tunnel engineering and provide a reference for similar projects.

Key words: tunnel engineering, advance geological prediction, horizontal directional drilling, mechanical specific energy, comprehensive logging, rock-mass classification