Manufacturing is currently experiencing a renaissance due to automation, robots, and AI, which will produce higher, and lower cost and higher productivity products. In this chapter a hybrid deep learning system is created with combinations of Convolutional Neural Networks with Long Short-Term Memory Neural Networks using the Whale Optimization Algorithm. The CNN monitors sensors and uses robotic vision to inspect products for defects while LSTM uses time series data to identify anomalies and predictive analytics. WOA defines the parameters for improved flexibility and performance. The findings show a 30% reduction in downtime, a 25% improvement in diagnosis of problems, and a 20% decrease in costs over 1 year. This model can help develop environmentally conscious, flexible, intelligent production systems.
Zero-carbon islanded microgrids (ZC-IMGs) operating without fossil-fuel generators, face major challenges in maintaining frequency and voltage stability due to the lack of synchronous inertia and reliable voltage reference. The integration of inverter-based grid-forming energy storage systems (GFM-ESSs) provides a viable solution; however, their coordination requires communication-efficient and scalable control strategies. To address these issues, this paper proposes a novel event-triggered distributed control framework for ZC-IMGs. First, a detailed nonlinear state-space model is developed to accurately characterize the dynamic behavior of the microgrid and provide a solid foundation for control design. Then, adaptive event-triggering and distributed parameter estimation mechanisms are introduced, which significantly reduce communication requirements, ensure minimum inter-event intervals, and eliminate Zeno behavior without centralized coordination. Subsequently, a distributed priority-based charging protocol is designed to guarantee fair and stable state-of-charge management across multiple GFM-ESS units. Simulation results show that the proposed method achieves excellent frequency and voltage regulation, fast recovery from disturbances, and effective state-of-charge management. At the same time, it reduces communication events by 92% compared with traditional continuous distributed control. The proposed framework therefore provides a scalable solution with significantly reduced communication requirements.
Background and purpose Current trends show an increase in the use of mobile application-assisted teaching (MA-AT) in physical education, but there are still few reports from previous studies on the effects of MA-AT on physical activity and functional fitness among university students. This study aims to evaluate the effects of MA-AT on physical activity and functional fitness among university students. Material and methods We adopted an experimental method with a pretest-posttest design. Eighty participants were involved and divided into MA-AT (n = 40) and CG (n = 40). Results Repeated measures ANOVA revealed significant effects of the condition factor (p < 0.001), the time factor (p < 0.001), and the condition × time interaction factor (p = 0.003) for physical activity. We found similar results for strength components, with significant effects of condition (p < 0.001), time (p < 0.001), and condition × time interaction (p = 0.005). There were significant effects of condition (p < 0.001), time (p < 0.001), and condition × time interaction (p = 0.007) on the power component. There were significant effects of condition (p < 0.001), time (p < 0.001), and condition × time interaction (p = 0.004) in the balance component, and a significant effect of the condition factor (p < 0.001), time factor (p < 0.001), and condition × time interaction factor (p = 0.005) was found in the agility component. Conclusions Thus, our study highlights that the application of MA-AT is proven to be effective in increasing physical activity and functional fitness among students.
Ushbu maqolada kuch avtotransformatorlarining texnik holatini baholash uchun transformator moyining fizik-kimyoviy xususiyatlariga asoslangan keng qamrovli simulyatsion model ishlab chiqilgan va amaliyotga tatbiq etilgan. Model izolyatsiya eskirishi va moyning degradatsiya jarayonini aks ettiruvchi asosiy diagnostik ko‘rsatkichlar – kislotalilik soni, namlik miqdori, dielektrik yo‘qotish tangensi hamda fazalararo taranglik kabi parametrlarni o‘z ichiga oladi. MATLAB dasturining Fuzzy Logic Toolbox muhiti asosida yaratilgan model “AGAR–UNDA” (IF–THEN) mantiqiy qoidalar tizimiga tayangan holda ushbu parametrlar bilan transformatorning texnik holat indeksi (HI) o‘rtasidagi noxatolik va o‘zaro bog‘liqliklarni tavsiflaydi. Ushbu intellektual tizim noaniq va to‘liq bo‘lmagan ma’lumotlarni aniq diagnostik xulosalarga aylantirish imkonini berib, transformator holatini monitoring qilishda qaror qabul qilish sifatini oshiradi. Ishlab chiqilgan yondashuv izolyatsiyaning eskirishi, issiqlik yuklanishi va ifloslanish jarayonlarini uzluksiz baholash, erta nosozliklarni aniqlash hamda prognozli texnik xizmatni ta’minlash imkonini beradi. Umuman olganda, model zamonaviy elektr energetika tizimlarida avtotransformatorlarning ishonchliligi, xavfsizligi va samaradorligini oshirishga xizmat qiladi.
Knowledge distillation is an highly effective machine learning technique for transferring knowledge from a large model, i.e., teacher, to a smaller model, i.e., student, improving model efficiency without sacrificing performance. This paper proposes a novel resource-aware federated learning framework Rafale that incorporates an adaptive client clustering mechanism based on computational resource availability and model complexity. High-resourceb clients train complex teacher models, and low-resource clients leverage distilled knowledge from teachers to train lightweight student models. Further, to optimize the aggregation process by introducing a clustering-based weightsharing mechanism that minimizes communication overhead while maintaining model accuracy. In this paper, a revolutionary Rafale framework that greatly reduces communication cost and separates model training from architectural constraints is proposed. To increase scalability and robustness, the suggested method makes use of resource-aware client selection, adaptive information sharing, and effective distillation techniques. In comparison to state-of-the-art FL approaches, experimental data show that RA-FKD delivers greater accuracy, decreases communication costs by up to 76%, and improves system efficiency. A cosine similarity loss KDC is introduced in knowledge distillation to improve the alignment between teacher and student models, enhancing knowledge transfer. Experimental evaluations on CIFAR-10 and CIFAR-100 datasets demonstrate that proposed framework achieving student test accuracy of 91.15% with reduction of 76.08% communication cost.
