Search

Water inrush has become one of the bottlenecks restricting tunnel construction. Among various advanced forecasting techniques, the direct current method is more cost-effective and sensitive to water-bearing structures. It has been widely used in exploring water inrush disasters in practical engineering. Although traditional resistivity linear inversion methods are reasonably practical, they usually suffer from volume effects and cannot accurately locate the location and morphology of water-bearing bodies. Therefore, nonlinear techniques such as deep learning have recently become popular to directly approximate the inversion function by learning the mapping of apparent resistivity data to the geoelectric model. This work presents a novel deep learning-based electrical approach that combines resistivity and polarizability to estimate water-bearing location and morphology. Specifically, we design an encoder-decoder network. A shared encoder extracts features from the input data, two encoders output resistivity, and polarizability models, respectively, and fine-tuned collinear regularization for both outputs reduces solutions’ multiplicity. Compared with traditional linear inversion methods and independent parameter inversion, our proposed joint inversion method shows superiority in locating and delineating anomalous bodies.

To better understand the shallow sea geological information and avoid the risk caused by potential geo-disasters, the efficient offshore geological exploration methods are required. Better detection resolution can be obtained by using a spark source. As the foundation, the observation system plays an important role for geological detection. The influence of the variation of the parameters of the observation system on the detection accuracy is analysed theoretically. Then, the numerical simulations based on the finite difference method was applied, and imaging characteristics of observation system with different parameter were studied. For submarine acoustic exploration, the spark source with frequency over 200 Hz can obtain the clear reflections for geological interpretation; besides, the receiver array with the interval of 5 m–10 m helps to obtain better wave signals when the buried depth of a geological body is less than 300 m, and the width is more than 10 m. Based on numerical simulations, the observation system was optimized and designed. The results of numerical examples show that the accurate position information of the structure can be obtained by using the observation system proposed in this paper. Different imaging performances are obtained by adjusting the parameters of the observation system. On this basis, combined with directivity analysis, the optimal observation system parameters are proposed. Finally, the proposed observation system is used to image the fault model. The research results of this paper can provide reference for the observation system design in similar projects.

In order to solve the problem that the traditional computer-aided teaching system is affected by communication technology, which leads to the inability to interact between teachers and students, the author proposes a research on a computer teaching system based on the Internet of Things and machine learning. The hardware structure is designed according to the functions of each module of the system, in which the student learning module is composed of a teaching coordination agent and a number of other agents, responsible for the presentation of specific teaching materials, problem solving, knowledge sharing through a collaborative mechanism, and providing personalized teaching basis for the system. The teacher’s teaching module mainly provides students with corresponding teaching strategies according to the learning requirements and uses its own reasoning mechanism to provide intelligent guidance to the problems encountered in the teaching process; the assessment module uses assessment rules to analyze student responses, comprehensive assessment of students’ learning behaviors, attitudes, effects, and abilities. The software function is designed with SQL Server 2000 as the database server, in the case of determining the data attributes, the data online evaluation is carried out, and the distance teaching is completed by combining the network technology. The experimental results show that when the time is 20s, the teaching efficiency of the traditional system is 61%, and the teaching efficiency of the artificial intelligence system is 91%. As a result, the teaching efficiency of the system based on the Internet of Things and machine learning is high, and it can provide equipment support for students’ learning.

This study investigates the permeability characteristics of unsaturated compacted loess by focusing on the anisotropy parallel and perpendicular to the compaction. Three tests are conducted on compacted loess: permeability test, soil-water characteristic curve test, and scanning electron microscope (SEM) test. Samples are maintained and tested at different dry densities under optimal moisture conditions. Test results showed that the horizontal saturated permeability coefficient of compacted loess was larger than in the vertical direction, showing obvious anisotropy. Based on the saturated permeability coefficient, the permeability coefficient of unsaturated compacted loess can be predicted according to the soil-water characteristic curve fitted using the van Genuchten model combined with the Childs and Collis-Geroge model. The relational formula was established between the unsaturated permeability anisotropy ratio and the matrix suction through introducing a pore connectivity-tortuosity parameter, which represents the relationship between unsaturated permeability anisotropy ratio and matrix suction. The difference of pore characteristics between horizontal and vertical directions of compacted loess is the main reason for the permeability anisotropy.