Zhenhua Nie

A critical challenge in structural health monitoring (SHM) is the extreme scarcity of labeled damage data from real-world bridges. Consequently, domain adaptation approaches that require target domain data with damage situation become nearly infeasible for real-world engineering applications. To address this limitation, this paper proposes a structural damage detection (SDD) method based on domain generalization, which constructs a highly generalizable damage recognition model by integrating style transfer and adversarial training. A latent domain with diverse styles is generated by mixing style features, such as mean and variance from different samples, followed by label generation through clustering. Adversarial training is then introduced to encourage the feature extractor to learn domain-invariant damage-sensitive features, thereby enabling the model to focus on essential features that are indicative of damage states. To validate the effectiveness of the proposed method, a simply supported beam model was established as the benchmark structure. Simulation results demonstrate that, compared to conventional damage identification algorithms, the proposed method exhibits superior damage localization performance on unseen target domain data. Furthermore, laboratory experiments confirm that the approach achieves excellent results in practical scenarios.

This study introduces a new bridge damage detection method using adversarial discriminative domain adaptation transfer learning. The method aims to transfer damage information between numerical models and real-world bridges across domains. The method uses an adversarial approach to teach the target domain’s feature extractor to mimic the feature mapping of a pretrained source domain extractor, aligning target domain features with the source’s feature space. The innovation is asymmetric mapping for adversarial learning, which represents real-world bridge data features within the finite element model data space and promotes cross-domain damage information transfer. The method’s efficacy was tested through numerical simulations and steel beam-cutting experiments. Moreover, the simulation tested accuracy and stability as domain shifts increased. The lab experiment cut steel beams to induce damage, treating the damaged structure as the target model. Both confirmed the method’s strong performance amidst errors and uncertainties.

This study proposes a baseline-free damage detection method for small-medium span bridges using dual-sensor transfer entropy under moving loads. In this method, acceleration responses are processed using frequency-adaptive moving windows, with the window length determined by the bridge's fundamental frequency and the sampling rate. The windowed signals undergo L1-norm compression, generating damage factors through transfer entropy quantification. Damage localization is achieved by identifying peaks in the damage factor curve when the moving load traverses damage locations. Validation via numerical simulations and experimental testing confirms robustness and sensitivity for single and distributed damages across varying speeds, masses, damage quantities, and sensor layouts. The paradigm establishes a novel dynamics-informed monitoring framework achieving three key advances: (i) Baseline-free diagnosis eliminating historical data dependencies, (ii) Full-capacity damage localization with only two sensors, (iii) Enhanced sensitivity to minute damage through vehicle-entropy synergy, enabling precise bridge assessment and overcoming key limitations in traditional vibration-based SHM.

Existing vibration‐based damage identification methods face significant challenges in practical engineering applications due to the reliance on a large number of sensors. To overcome the limitation, this paper proposes a novel structural damage identification method based on moving phase space reconstruction, which can use only a single sensor for effective damage identification. The proposed method introduces the concept of a subphase space and adopts an improved damage indicator to analyze the degree of topological change of phase trajectories at different time points. The parameters required for this method are completely determined by the vibration signal, without the need for detailed geometric or material data of the structure. To verify the effectiveness and reliability of the proposed method, the numerical simulation of a simply supported beam is carried out, and then the experimental verification is conducted on a hollow box girder. The results show that the proposed method can successfully identify and locate structural damage using only the vibration data of a single sensor and can effectively evaluate the degree of topological change in phase trajectories over time.

Traditional bridge detection methods require a large number of sensors, which restricts the application in practical engineering. In order to reduce the cost of the health monitoring system and improve the accuracy of bridge damage detection, a bridge damage detection method based on principal component analysis and discrete wavelet transform is proposed. This method performs principal component analysis on the displacement under moving load and obtains the first principal component composed of dynamic component and modal shape. The first principal component is separated by the low-pass filter to receive the dynamic component of the first mode. Then the dynamic component is analyzed by the discrete wavelet transform to obtain the wavelet coefficient. Then the wavelet coefficient is used to reconstruct the signal. Finally, damage index is defined by using the reconstructed signal to detect the damage. When the structure is damaged, the damage index peaks at the damage location. In order to verify the effectiveness of the method, a beam bridge is simulated and a steel beam bridge is tested in the laboratory. Both simulation and experimental results show the proposed method can accurately identify the damage location in various scenarios with limited sensors.

The construction of new engineering disciplines and the “Double Tops” initiative have put forward new requirements for the training of engineering professionals. Taking the course of C language programming as an example, this paper analyzes the pain points in current course teaching, combines the characteristics of students’ learning situation, takes students as the center, and based on the knowledge graph, an effective tool for organizing resources, carries out a modular teaching practice of C language programming design under the constructivist learning paradigm. Integrating the OBE (Outcomes-Based Education) concept into teaching, it enhances students’ practical ability, cultivates their computational thinking, and highlights the training of students’ ability to analyze and solve complex engineering problems and creative thinking. The exploration and practice of this course teaching can provide reference and reference for the teaching design of other programming courses.

This paper investigates the thermal battery as a research topic. We conducted an in-depth analysis of various thermal battery aspects, such as the cathode material CoS₂ and electrolyte material morphology, crystal type, and interface state changes before and after service. The aim was to explore the core reaction and main failure mechanisms of the thermal battery. Prior to the reaction, the thermal battery cathode and electrolyte material consisted of pure-phase CoS₂ and a composition of MgO-LiF/LiBr/LiCl. After service, the cathode and electrolyte of the single thermal battery exhibited significant morphological alterations caused by the presence of a molten state. The cathode transformed from CoS₂ to Co₃S₄ and Co₉S₈ together with the presence of a marginal quantity of Co monomers visible throughout the discharge process, which was confirmed by means of XRD and XPS analyses. After the reaction, the electrolyte material was primarily made up of LiF, LiBr, and LiCl while the crystal components remained largely unaltered, albeit with apparent morphological variations. As was deduced from the thermodynamic analysis, the cathode material’s decomposition temperature stood at 655 °C, exceeding the working temperature of the thermal battery (500 °C) by a considerable margin, which is indicative of outstanding thermal durability within the thermal battery’s operational temperature range. Furthermore, the discharge reaction of the positive electrode was incomplete, resulting in reduced CoS₂ residue in the thermal battery monomer after service. The reaction yielded a combination of Co₃S₄, Co₉S₈, and small amounts of Co monomers, indicating possible inconsistencies in the phase composition of the pole piece during the reaction process. In this study, we examine the distribution of residual stress in the thermal battery under various operating conditions. The simulation results indicate that exposure to a 70 °C environment for 2 h causes the maximum residual stress of the battery, which had an initial temperature of 25 °C, to reach 0.26 GPa. The thermal battery subjected to an initial temperature of 25 °C exhibited a maximum residual stress of 0.42 GPa subsequent to a 2-hour exposure to a temperature of −50 °C.