Search for: All records

Abstract. The Vacuum-Assisted Resin Infusion Molding (VARIM) process is widely used in wind turbine blade manufacturing due to its cost-effectiveness and reliability. However, challenges such as prolonged curing cycles and defects caused by non-uniform cure remain persistent. To address these issues, multizone heating systems have been developed to enable independent temperature control across blade sections. Yet, optimizing the temperature profile in each zone is computationally intensive, requiring detailed modelling of curing kinetics and heat transfer mechanisms. To overcome these challenges, in this work, a machine learning (ML) based digital twin of the VARIM process was developed using a time-distributed long short-term memory (LSTM) network trained on data generated by a high-fidelity multiphysics solver. The model achieved a predictive accuracy of 96.7 % in replicating the resin curing behavior. Its time-distributed architecture effectively captures the spatial – temporal dependencies across multiple zones, allowing precise prediction of the degree-of-cure evolution. Paired with a gradient-free optimization algorithm, the digital twin reduced curing time by 12.5 % while improving cure uniformity. This AI-driven framework eliminates costly trial-and-error experimentation, and provides a scalable, adaptive solution for improving both quality and productivity in wind turbine blade manufacturing, with strong potential for extension to other composite manufacturing processes. 

Abstract The vacuum-assisted resin infusion mold (VARIM) process is widely used in wind blade manufacturing for its cost-effectiveness and reliability. However, the current method faces challenges such as long curing times and defects due to nonuniform heating across the blade structure. To address this, a multi-zone heated bed setup tailored to blade thickness has been considered. However, determining an optimal temperature for each zone poses a computational challenge, which can be tackled with a novel machine-learning approach. Using a digital twin based on a high-fidelity multiphysics solver, a time-distributed LSTM model was trained to understand complex resin curing dynamics. This eliminates the need for costly lab experiments, as the model learns heating patterns and curing behavior efficiently. Once trained, the ML model acts as a digital twin by predicting the degree of cure for a given temperature setpoint with 96.73% accuracy. This model, when used as a surrogate for a Nelder-mead optimization workflow, improves the curing time by roughly 12.5% and presents a more uniform curing rate throughout the part. 

In this paper, a novel hybrid wire arc additive manufacturing (WAAM) and ultrasonic nanocrystal surface modification (UNSM) on porosity manipulation and surface properties of aluminum 5356 alloys was studied. The goal is to improve the quality of the WAAM-built part by eliminating bigger pores and reducing its size, reducing surface roughness, and increasing surface hardness. The as-built WAAM and WAAM-UNSM-treated samples were quantitatively studied for porosity using an X-ray micro-computed tomography (μ-CT). The surface roughness was measured on the surface profile of the same samples before and after UNSM treatment. Followed by the Vickers micro-hardness tests to evaluate the hardness modified by the influence of the UNSM treatment. It was found that the bigger pores in the as-built WAAM samples were eliminated and the medium-sized pores were shrunk to almost half the size after the UNSM treatment. Further, the UNSM treatment showed a significant improvement in both surface roughness and hardness on the WAAM Al5356 samples. This experimental work demonstrates the critical advantages of hybrid WAAM-UNSM in improving the qualities of the WAAM processed parts. 

Natural muscles show tensile actuation and realize torsional rotation by combining with the skeleton, which integrate with sensing and signaling function in a single element to form a feedback loop. The currently developed artificial muscle and sensing devices always work upon external stimuli, and a separate controlling and signal transmission system is needed, increasing the complexity of muscle design. Therefore it is highly desired to develop flexible and compact fiber artificial muscles with large strain for advanced soft robotic systems. In this paper, twisted elastomer fiber artificial muscles with tensile and torsional actuations and sensing function by a single electric signal are developed, by using twisted natural rubber fiber coated with a buckled carbon nanotube sheet. The twisted natural rubber fiber can be electrothermally actuated to show contraction and rotation by entropic elasticity. The buckled carbon nanotube sheet can transmit electric current, and the contact area between the buckled carbon nanotube sheets increased during actuation, resulting in resistance decrease by thermo-piezoresistive effect. A feedback circuit was designed to connect or disconnect the electric current by measuring the resistance change to form a feedback loop to control on/off of the muscle. The current study provides a new muscle design for soft robotics, controllers, and human-machine integration. 

Abstract The brick-and-mortar structure inspired by nature, such as in nacre, is considered one of the most optimal designs for structural composites. Given the large number of design possibilities, extensive computational work is required to guide their manufacturing. Here, we propose a computational framework that combines statistical analysis and machine learning with finite element analysis to establish structure–property design strategies for brick-and-mortar composites. Approximately 20,000 models with different geometrical designs were categorized into good and bad based on their failure modes, with statistical analysis of the results used to find the importance of each feature. Aspect ratio of the bricks and horizontal mortar thickness were identified as the main influencing features. A decision tree machine learning model was then established to draw the boundaries of good design space. This approach might be used for the design of brick-and-mortar composites with improved mechanical properties. 

We report a mechanical metamaterial-like behavior as a function of the micro/nanostructure of otherwise chemically identical aliphatic polyurea aerogels. Transmissibility varies dramatically with frequency in these aerogels. Broadband vibration mitigation is provided at low frequencies (500–1000 Hz) through self-assembly of locally resonant metastructures wherein polyurea microspheres are embedded in a polyurea web-like network. A micromechanical constitutive model based on a discrete element method is established to explain the vibration mitigation mechanism. Simulations confirm the metamaterial-like behavior with a negative dynamic material stiffness for the micro-metastructured aerogels in a much wider frequency range than the majority of previously reported locally resonant metamaterials.