Unveiling the mechanical properties and damage mechanism of the complex composite structure, comprising backfill and surrounding rock, is crucial for ensuring the safe development of the downward-approach backfill mining method. This work conducts biaxial compression tests on backfill–rock under various loading conditions. The damage process is analyzed using DIC and acoustic emission (AE) techniques, while the distribution of AE events at different loading stages is explored. Additionally, the dominant failure forms of specimens are studied through multifractal analysis. The damage evolution law of backfill–rock combinations is elucidated. The results indicate that DIC and AE provide consistent descriptions of specimen damage, and the damage evolution of backfill–rock composite specimens varies notably under different loading conditions, offering valuable insights for engineering site safety protection.

Acoustic emission (AE) technology combined with acoustic tomography can realize the early detection for unknow velocity structure of rock mass. However, multi-factors limit the resolution of tomography in complex rock mass. To meet the high-efficiency and -resolution requirements of unknown structure detection in underground engineering, this paper develops a three-dimensional AE tomography method (AET) using velocity-free location (VFL) and time-difference adjoint tomography (TDAT). The proposed method adopts independent rather than alternate location and tomography system. The VFL is utilized to determine the position of AE sources without velocity measurement. The adjoint state system is introduced in TDAT to avoid ray tracing. Further, the quasi-Newton method is adopted to iteratively update the velocity model. The proposed AET reduces the strict demand on prior model and possesses outstanding inversion efficiency. Numerical and laboratory experiments were conducted to verify the effectiveness and feasibility. The results show that the proposed method can accurately estimate the position of AE sources, and exactly reconstruct the velocity anomaly and damage buried in complex rock mass. It not only realizes the non-destructive velocity structure reconstruction of complex rock mass using Ae events, but also provides new insight for other AE testing application fields.

Acoustic emission (AE) source localization provides important spatial information for damage characterization and fracture evolution analysis in construction materials, while its accuracy and applicability are strongly dependent on sensor network design. This study proposes an optimization framework for selecting an effective six-sensor network for velocity-free AE source localization in construction materials. The source coordinates are determined by solving a nonlinear inverse problem using the Levenberg–Marquardt algorithm, and candidate sensor subsets are evaluated by combining location error metrics with the number of effective localization results to quantify the effective monitoring range for damage characterization. The framework is investigated through numerical simulations and pencil-lead break tests on a 600 mm × 600 mm ceramic tile. Among different six-sensor configurations, the best-performing layouts place four sensors at the outer corners and two sensors at the horizontal or vertical inner corners. A benchmark comparison with the Fisher-information-based optimized sensor network and sensitivity analyses further show that the optimized sensor network maintains higher effective monitoring ranges under arrival-time noise, velocity uncertainty, and sensor coordinate perturbations. The proposed approach provides a useful reference for robust and cost-effective AE sensor network design in damage monitoring, fracture characterization, and nondestructive evaluation of construction materials.

Accurate acoustic emission (AE) source localization in complex structures remains challenging due to non-straight wave propagation paths and the difficulty of obtaining reliable wave velocity models. To address these issues, this study proposes a velocity-free AE source localization method based on an any-angle pathfinding algorithm. The method integrates the Anya algorithm to estimate geometrically optimal propagation paths and a velocity-free objective function formulated under a weak anisotropy assumption. By avoiding the directional limitations of conventional grid-based pathfinding, the proposed approach provides more accurate distance estimates between potential sources and sensors without requiring prior velocity information. The effectiveness of the method was validated through an AE pulse experiment conducted on a granite specimen containing multiple cylindrical holes. The results demonstrate that the average location error of the proposed method is 19.02 mm, which is less than 38.14 mm of the traditional method and 22.90 mm of the velocity-free method using a grid-based search algorithm. Therefore, the proposed framework is applicable to two-dimensional structures or extruded three-dimensional structures, offering a feasible approach for AE source localization in complex environments.