This paper proposes an optimal control strategy for process parameters aimed at addressing layup defects, achieving defect recognition and localization, and maintaining temperature stability on the actual layup surface. Infrared thermography is employed to identify defects during the carbon fiber layup process, and by integrating path control, the identification and localization of defects during the layup are effectively achieved. The minimum edge detection rate on both sides of a single layer reaches 94.3%, and the interlayer gap measurement is below 10%. Considering that temperature instability on the layup surface can lead to defects, a dynamic temperature control model has been established, and the infrared lamp’s temperature rise coefficient is determined, reducing defects caused by temperature fluctuations during the layup process. The results indicate that infrared thermography technology is feasible for the detection and reduction of defects in CFRP composites.

Here, wire-arc additive manufacturing (WAAM) is employed to manufacture a super-invar alloy thin-wall rectangular component. The microstructure is characterized by cellular sub-grains with different morphologies inside the epitaxially grown columnar crystals. Based on the finite element simulation results, the value of the G (the temperature gradient)/R (the solidification rate) during the deposition process is calculated as 1.59 × 108 K·s·m−2, which is associated with the columnar cellular microstructure. The transfer mode of the droplet during the WAAM is liquid bridge transition. The mechanical properties of specimens are anisotropic, and the longitudinal samples are better than transverse samples; the UTS is 398.8 MPa, the YS is 291.4 MPa, and the elongation is 40.8%. The coefficient of thermal expansion (CTE) is measured to be 0.265 × 10−6 K−1 in the range of 20 °C to 100 °C. The findings provide a reference for the fast fabrication of super-invar alloy components through WAAM, which promotes the applications of super-invar alloy in aerospace.

In the context of rapid technological advancement, haptic human-machine interfaces (HMIs) enhance user experience by simulating touch. Electroactive polymers (EAPs) are smart materials with high responsiveness, flexibility, and tunability, making them suitable for haptic actuators and feedback applications. This review examines the role of EAPs in haptic interaction, analyzing driving mechanisms, structural design, functional materials, fabrication methods, and practical applications. We also address challenges like performance limitations and manufacturing complexities, while discussing future trends in material optimization, structural design, and innovative driving strategies. This review serves as a valuable reference for future research and technological advancements in EAPs.

Laser-based powder bed fusion (LB-PBF) as-built material properties; residual stresses and final component shape are dependent on the thermal history of the printed part. Design and optimization of the deposited material thus far depends on experimental trial and error. A systematic approach based on model informed optimization is missing; mainly due to the computational expense of resolving the scan path and performing the required transient simulations. In this work; the heat conduction equation is reformulated to enable accelerated simulations. The laser position and operating conditions are read from the build file. The laser trajectory throughout the component is resolved providing 3D temperature evolution. Simple demonstration parts exhibiting both hot and cold regions are used to induce different metallurgical responses of the deposited IN718. Thermocouples are used to validate calculated temperatures. CALPHAD based calculations utilize temperature predictions to obtain localized phase concentrations and predict the distribution of thermophysical properties throughout the build. Results are compared with hardness measurements confirming the accuracy of the overall modelling chain.

The integration of automation, connectivity, and advanced analytics in manufacturing enhances productivity but also increases vulnerability to cyber threats including sensor attacks. Sensors—critical to automation—are particularly susceptible to false data injection attacks that can disrupt operations and lead to system failures. Despite advancements in prevention and detection methods, effective post-attack recovery remains an underexplored area—critical to minimizing operational downtime in manufacturing. The research addresses this gap by introducing a novel agent-based recovery modeling approach tailored for manufacturing systems. A reinforcement learning-driven recovery strategy is developed to restore operations efficiently after sensor attacks. The approach is evaluated through two distinct sensor attack scenarios. The recovery agent's performance is benchmarked against a PID controller using key metrics: downtime, throughput, and efficiency. Results demonstrate significant improvements by enhancing the resilience and security of manufacturing systems against sensor attacks.

The profile of the melt pool is essential in selective laser melting (SLM) to control the process quality and avoid defects. Physics informed neural network (PINN) method is proposed to address challenges in various science and engineering problems when traditional numerical calculations are time-consuming, or deep learning (DL) methods have high demand for data. However, SLM process involves many complex physical phenomena. Low-fidelity data from low-fidelity models struggle to accurately reflect these phenomena, while high-fidelity data from high-fidelity models contains more physical equations, making it difficult for current PINN. This article proposed a transfer learning-enhanced PINN (TLE-PINN) method using high-fidelity data for precise and fast melt pool prediction. It contains the enhanced PINN (EPINN) and transfer learning framework. The EPINN model integrates the heat transfer law and boundary condition to loss function, imposing strong physical constraints on data. Then, the transfer learning framework, combining the concepts of PINN and DL, initially trains with PINN and then further fine-tunes it using DL method. Notably, it only uses a single model, which is more convenient to traditional methods that require two models. The developed solution demonstrates outstanding performance when compared with experiments and existing methods, showing significant potential for industrial applications.