Abstract: The worsening of hidden diseases of old tunnel, which are of variety and complexity, would cause harm to tunnel structure and traffic safety. Although ground penetrating radar (GPR) technology has been widely applied to the investigation of hidden diseases, no uniform understanding or systematic conclusion of anomaly features in GPR images has been reached. In this article we sum up the types and causes of old tunnel diseases, and set up corresponding earth-electricity models to identify anomaly features through forward modeling and comparing the GPR images with field measurements. Moreover, we establish an evaluation system for old tunnel disease by introducing the concept of membership degree in fuzzy mathematics. Results show that the dielectric constants of lining crack and cavity are significantly different due to different filling mediums, reflected by event dislocation and local improvement of diffraction patterns. The non-compacted and impervious layer gradually develops into a mixed structure involving water, concrete, and residues with multiple reflection interfaces, and the corresponding GPR image is cluttered with local strong reflection. Application practice prove that the presented model could well assess the safety levels of old tunnels based on GPR detection of potential disease areas.

Key words: old tunnel, hidden diseases, ground penetrating radar (GPR), anomaly features, disease evaluation system