Based on the shield tunnel engineering in weathered granite stratum in Xiamen, Stochastic calculations, by combining the random field theory and the finite difference analysis together with Monte Carlo simulation, are used to carry out the change law of the characteristics of surface deformation curve and surface deformation model. Results show that with the increase of the vertical scales of fluctuation, the decrease of the transverse scales of fluctuation or the increase of the coefficient of variation, the low peak distribution characteristics of the location of the maximum surface settlement induced by shield tunneling become more obvious, and the randomness and chaos of the shape of surface deformation curve gradually increase. The diversity of surface deformation model is affected by parameter correlation and randomness. Under the condition of small transverse scales of fluctuation and large vertical scales of fluctuation, the sensitivity of coefficient of variation to surface deformation mode is limited.

Aiming at the problem of deformation of adjacent masonry structures caused by the surrounding soil disturbance in the construction of twin shield tunnels, taking the spatial variability of soil parameters as the breakthrough point and the surface deformation mode as the basis, a program is compiled to search the position of characteristic points of the settlement curve, and a safety assessment method of masonry buildings above twin tunnels based on the random field theory is obtained. According to the reliability theory, the concept of the potential damage zone is defined. The probability that the maximum tensile strain of the building exceeds the allowable ultimate tensile strain is not less than 5% in the potential damage zone. The assessment method mentioned above, together with the Monte Carlo simulation, is used to evaluate the safety of buildings and the range of potential damage zones induced by twin tunneling with different relative locations and different construction sequences, taking into account configurations and construction sequences of twin tunnels, the scales of fluctuation of elastic modulus of the soil, and the size of the building. The research results show that, in the parallel and overlapping twin tunnels, the randomness of the surface deformation mode caused by construction increases with the decrease of the transverse scales of fluctuation or the increase of the vertical scales of fluctuation. In overlapping and shoulder-mounted twin tunnels, the effect of the construction sequence on the safety of buildings is the opposite. When the overlying tunnel is excavated first in shoulder-mounted condition, the safety of buildings is more excellent. Considering the effect of the spatial variability of parameters, the probability of higher damage levels in the buildings is significantly increased. The greater the scales of fluctuation, the greater the damage probability. The potential damage zones of the buildings will be expanded, and the safety distance of the buildings will be wider than the deterministic analysis result by 1.5∼2.5 m.

Cardiovascular disease (CVDs) is one of the universal deadly diseases, and the detection of it in the early stage is a challenging task to tackle. Recently, deep learning and convolutional neural networks have been employed widely for the classification of objects. Moreover, it is promising that lots of networks can be deployed on wearable devices. An increasing number of methods can be used to realize ECG signal classification for the sake of arrhythmia detection. However, the existing neural networks proposed for arrhythmia detection are not hardware-friendly enough due to a remarkable quantity of parameters resulting in memory and power consumption. In this paper, we present a 1-D adaptive loss-aware quantization, achieving a high compression rate that reduces memory consumption by 23.36 times. In order to adapt to our compression method, we need a smaller and simpler network. We propose a 17 layer endto-end neural network classifier to classify 17 different rhythm classes trained on the MIT-BIH dataset, realizing a classification accuracy of 93.5%, which is higher than most existing methods. Due to the adaptive bitwidth method making important layers get more attention and offered a chance to prune useless parameters, the proposed quantization method avoids accuracy degradation. It even improves the accuracy rate, which is 95.84%, 2.34% higher than before. Our study achieves a 1-D convolutional neural network with high performance and low resources consumption, which is hardware-friendly and illustrates the possibility of deployment on wearable devices to realize a real-time arrhythmia diagnosis.

Aiming at the special geological conditions in Karst areas, the pile-soil stress ratio of composite foundation is analyzed. According to carry out the field test of composite foundation in Karst area, soil pressure boxes are arranged to measure the stress of piles and soil at the same time. As the load increased, the variation trend of the pile-soil stress ratio of composite foundation is analyzed. On the basic, the pile-soil stress ratio is compared with non-karst area. The test result showed that both the pile-soil stress ratio of PHC piles and high pressure jet grouting piles increase. The increase of the PHC piles is more obvious. The pile-soil stress ratio of PHC piles is larger than high pressure jet grouting piles. The main reason is that the strength of cement-soil mixed piles is lower. The high pressure jet grouting piles plays an important role in improving soil bearing capacity between piles and supporting the side abutment pressure of PHC piles in composite foundation. It is shown that the stress concentration is more obvious in karst area.

Reversible data hiding in encrypted images is an effective technique for data hiding and preserving image privacy. In this paper, we propose a novel schema based on polynomial arithmetic, which achieves a high embedding capacity with the perfect recovery of the original image. An efficient two-layer symmetric encryption method is applied to protect the privacy of the original image. One polynomial is generated by the encryption key and a group of the encrypted pixel, and the secret data is mapped into another polynomial. Through the arithmetic of these two polynomials, the purpose of this work is achieved. Furthermore, pixel value mapping is designed to reduce the size of auxiliary data, which can further improve embedding capacity. Experimental results demonstrate that our solution has a stable and good performance on various images.Compared with some state-of-the-art methods, the proposed method can get better decrypted image quality with a large embedding capacity.