Mathematical Problems in Engineering

Most selective maintenance research strategies ignore the comprehensive evaluation of numerous following missions in favor of focusing just on the reliability of the next mission. In the same circumstance, however, assessing simply impending missions differs from considering the overall system’s maintenance planning outcomes. In this study, a decision-making model for selective maintenance is developed by evaluating the total system reliability over a finite planning horizon. The purpose is to calculate the maintenance activities for each system component during each break and the best number of maintenance interruptions for the planning horizon to achieve maximum system reliability. Consequently, the selective maintenance problem is formulated as a max–min optimization model. Also, a hybrid imperfect maintenance model is used to formulate the component improvement after maintenance. Finally, simple case illustrations of maintaining the production system in coal transportation are given based on the assumed data.

Recently, the same authors provided a switching algorithm for the preferred clock skew estimator appropriate for the precision time protocol (PTP) scenario. The algorithm chooses the one-way delay (OWD) clock skew estimator for the Forward path or the Reverse path or the two-way delay (TWD) clock skew estimator that has the best performance in the mean square error (MSE) perspective. However, the switching algorithm applies only to the Gaussian scenario. In a real system, the packet delay variation (PDV) can be characterized as an fractional Gaussian noise (fGn) process where the Hurst exponent parameter can also have values higher than 0.5. Thus, the Gaussian case-switching algorithm may not apply in a real system where after a small set of PTP measurements, the switching algorithm should be able to switch effectively to the preferred clock skew estimator with the best performance in the MSE perspective. In this paper, the PDV is characterized as an fGn process where the Hurst exponent (H) is in the range of 0.5≤H<1. We estimate the unknown PDV variances and the Hurst exponent parameters of the Forward and Reverse paths. These estimated parameters are used for switching to the preferred clock skew estimator from the MSE perspective, even in the presence of asymmetry in the PDV or in the Hurst exponent parameters. We also detect and alarm for the unexpected load, cyber-attack, or clock skew deviation that can occur in a real system.

In this paper the role of solid nonlinearity on the fluid–structure interaction (FSI) is studied. First a simple mass-spring-damper system with a third-degree spring is coupled with a linear sloshing system to study the effect of solid nonlinearity. It shows that even both systems show similar behavior in free vibration tests, but nonlinearity caused 50% less on the amplitude of vibration. Then a higher order solid system coupled with a sloshing tuned liquid damper which is modeled by boundary element method (BEM) is examined. Finally, the coupling of a nonlinear solid with the sloshing system calculated at high-resolution method of smoothed particle is analyzed. The results reveal that considering surface nonlinearity in BEM has not improved the prediction significantly and the full NavierStokes calculations is in another phase and amplitude to a step response of the system even after two periods of the system in comparison with linear estimation. As shown the nonlinearity in solid part can cause a big difference in response of coupled system that pointed the necessity of the use of accurate fluid models such as BEM or smoothed particle hydrodynamics method as well as solid system identification before the coupling.

For person reidentification (ReID), most slicing methods (such as part-based convolutional baseline (PCB) and AlignedReID) introduce a lot of background devoid of pedestrian parts, resulting in the cross-aliasing of features in the deep network. Besides, the resulting component features are not perfectly aligned with each other, thus affecting model performance. We propose a convolutional neural network (CNN) with embedding transformers (CET) person ReID network architecture based on the respective advantages of CNN and transformer. In CET, first, the residual transformer (RT) structure is first embedded in the backbone network of CNN to obtain a feature extractor, named transformers in CNN. The feature aliasing phenomenon is improved by utilizing transformer’s advantage in grasping the relevance of global information. Second, a feature fuse with learnable vector structure for fusing the output vector is added to the output of the transformer at the end of the network. A two branches loss structure is designed to balance the two different fusion strategies. Finally, the self-attention mechanism in transformer is used for automatic part alignment of human body parts to solve the part alignment problem caused by inaccurate detection frames. The experimental results show that CET network architecture achieves better performance than PCB and some other block-slicing methods.