Search

Thermoplastic vulcanizates (TPVs) based on ethylene-acrylic acid copolymer (EAA)/nitrile-butadiene rubber (NBR) and zinc dimethacrylate (ZDMA) reinforced EAA/NBR blends were prepared by dynamical v...

Mullins effect during uniaxial cyclic compression tests, together with its reversibility, of ethylene–acrylic acid copolymer (EAA)/nitrile–butadiene rubber (NBR) thermoplastic vulcanizates (TPVs) were investigated systematically. The results showed that EAA/NBR TPVs had excellent mechanical properties when the weight ratio was 40/60. Morphology studies showed that sphere-like NBR particles were dispersed evenly in the etched TPVs surface with diameters of 5–8 μm. The experimental results of Mullins effect indicated that a stress softening phenomenon in the stress–strain curves of EAA/NBR TPVs during the uniaxial loading–unloading cycles could be observed obviously; moreover, the reversibility of Mullins effect could be significantly enhanced by increasing heat treatment temperature.

With the swift development of Internet and artificial intelligence, recommender systems have become more and more important as useful information has a risk of submerging in huge amounts of data or smart services always provide a behavior predictor to meet diversified user’s needs. User-session based recommendations are commonly applied in many modern online platforms. Graph Neural Networks (GNNs) have been shown to have a strong ability to address the problem of session-based recommendation with accurate item embedding. However, there are a lot of application scenarios that have already provided user profiles. We propose a model based on Hierarchical GNNs for personalized recommendations, which evolves both item information in sessions and user profiles. The experiments on two industry datasets show the superiority of our model over the state-of-the-art methods.

To further improve the operational stability of the substation DC system and realize the lean operation and maintenance, by analyzing the common problems existing in and new requirements of the DC power management, this paper proposes a design scheme for DC power management system. This design scheme is expected to realize the remote intelligent support, avail future extension of the system, realize integrated processing of the equipment management business information of the DC system, and allow mutual communication of the on-site status monitoring information and history record information of the equipment (including defects, inspection and detection test information) so as to jointly provide complete information for daily management and assistant decision-making of the DC system.