Due to sensor failure, noise interference and other factors, the data collected in the structural health monitoring (SHM) system will show a variety of abnormal patterns, which will bring great uncertainty to the structural safety assessment. This paper proposes an automatic data anomaly diagnosis method for SHM based on a multimodal deep neural network. In order to improve the detection accuracy, both two-dimensional and one-dimensional features of the sensor data are fused in the multimodal deep neural network. The network consists of two convolutional neural network (CNN) channels, one a 2D-CNN channel for extracting time–frequency features of sensor data and the other a 1D-CNN channel for extracting raw one-dimensional features of sensor data. After convolution and pooling operations for the sensor data by the 2D channel and 1D channel separately, the two types of extracted features are flattened into one-dimensional vectors and concatenated at the concatenation layer. The concatenated vector is then fed into fully connected layers for final SHM data anomaly classification. In order to evaluate the reliability of the proposed method, the monitored data lasting for one month of a long-span cable-stayed bridge were used for training, validation, and testing. Six types of training conditions (missing, minor, outlier, over-range oscillation, trend, and drift) are studied and analyzed to address the issue of imbalanced training data. With an accuracy rate of 95.10%, the optimal model demonstrates the effectiveness and capability of the proposed method. The proposed method shows a promising future as a reliable AI-assisted digital tool for safety assessment in structural health monitoring systems.

To solve the problem of waste slag treatment of slurry shields, a reuse scheme in which waste shield mud is used to replace traditional karst cave grouting filling material which is proposed; thus, a new type of underwater cementitious filling material (NWC‐FM) is developed. NWC‐FM is convenient, inexpensive, and environmentally friendly. Its mix proportion is designed, and its mechanical performance is tested. According to the characteristics of karst caves, a semiclosed and pressurized karst cave simulation box is designed and manufactured to simulate grouting filling processes at construction sites, and an NWC‐FM grouting simulation test is carried out. The results show that the fluidity of the NWC‐FM slurry is good, and the strengths of the samples in the three groups of designed mix proportions meet the requirements of field construction. The underwater poured NWC‐FM shows good fluidity, cohesion, nondispersibility, self‐leveling, and self‐compacting. After each pouring of NWC‐FM material, a 2∼3 cm isolation layer eventually forms on the surface due to the action of the additives, which can effectively block the contact between water and NWC‐FM and ensure the flatness of the underwater poured surface of the NWC‐FM material. The average compressive strength of the NWC‐FM consolidated core samples at 7 d, 14 d, and 28 d are 0.56 MPa, 0.72 MPa, and 0.79 MPa, respectively, meeting the technical strength requirements of construction sites. NWC‐FM has strong underwater nondispersibility and moderate strength, which can well meet the requirements of karst cave filling treatment during shield construction of urban subways. Additionally, as a low‐cost and environmentally friendly material, NWC‐FM will greatly reduce the project cost and minimize environmental pollution.