Search

This paper examines the effects of thermo-solid coupling and the influence of braking parameter changes on the vibration characteristics of high-speed train disc brakes. A multi-flexible body dynamics model of high-speed train disc brakes considering thermo-solid coupling was established to study the vibration characteristics of high-speed train disc brakes during service. The results show that the uneven distribution of temperature and stress produced during the brake disc’s service was the primary cause of the warping deformation of the brake disc, which prevented the brake disc and the brake pads from making sufficient contact and caused vibration while braking. By comparing the analytical findings of whether the model was subject to the coupling effect or not, the influence of thermo-solid coupling on the braking procedure was demonstrated from the standpoint of energy distribution. The severity of the high-speed train brake disc vibration gradually increased along with the braking pressure and initial speed. In addition, vibration aggravated the instability of the braking process, which could lead to thermoelastic instability and is harmful to the braking performance of the brake. These findings provide theoretical support for designing and manufacturing disc brakes for high-speed trains.

Hydrogen permeation reduces a material’s properties and increases the risk of brittle fracture, which causes a potential safety hazard. A workpiece’s hydrogen permeation resistance could be improved by improving its surface integrity through surface processing. This paper studies low-alloy steel’s surface integrity and its hydrogen permeation resistance in a hydrogen production reactor, using the electrochemical cathodic hydrogen-charging method to carry out electrochemical hydrogen-charging experiments. After the specimens were pretreated using different surface-grinding methods and shot peening pressure strengthening, they were hydrogen-charged on a self-designed and built electrolytic hydrogen charging platform. Before and after hydrogen charging, the specimens’ section hardness and tensile strength were tested, and the fracture morphology of the specimens was observed. The influence laws of surface roughness and surface residual compressive stress on the distribution of material hardness along the depth, the variation in material hardness, the fracture morphology, and the decline in the tensile properties of the low-alloy steel specimens after 5 h of hydrogen charging were analyzed. The reasons for the influence of surface integrity indexes on the hydrogen permeation resistance of the specimens were also analyzed. Based on the experimental results, a series of mechanical processing parameters were proposed to improve the material’s permeation resistance, which provides a theoretical and practical basis for the processing of materials with high surface integrity and hydrogen permeation resistance. Through the experiments, it was found that the hydrogen permeation resistance of the Ra 0.17 μm surface roughness specimen was the best of all specimens with different surface roughness values, and its hydrogen embrittlement sensitivity index was 20.96%. The specimen had the best hydrogen permeation resistance under 336 MPa surface residual compress stress, and its hydrogen embrittlement sensitivity index was 16.45%.

Corrosion can hurt the quality and service life of a workpiece, which could create potential safety hazards. However, improving the workpiece’s surface integrity through surface treatment could improve the effect of corrosion resistance. To study the effect of surface integrity on the corrosion resistance of low-alloy steel for hydrogen reactors, electrolytic corrosion experiments were carried out on specimens that were processed using different grit sandpapers or using different shot peening pressures using a self-built electrolytic platform. The influence regulation of initial surface roughness and surface residual stress on the corrosion rate of the low-alloy steel for a hydrogen reactor under different lengths of corrosion time, surface roughness after corrosion, and the tensile property degradation of the specimens after 4 h of corrosion were analyzed, respectively. In this paper, based on experimental research, we obtain the influence regulation of the processing parameters on the corrosion resistance of low-alloy steel for hydrogen reactors and provide processing parameters that could improve the corrosion resistance of low-alloy steel, which guides the corrosion resistance processing of hydrogen production reactors. It was found through experiments that with an increase in the initial surface roughness of the specimens, the corrosion rate of the specimen tends to decrease with the increase in corrosion duration; the surface roughness of the specimen after corrosion first increases and then decreases with the increase in corrosion time; and the tensile strength of specimen with the initial surface roughness of Ra 0.168 μm is relative good after 4 h of corrosion. With the increase in residual compressive stress on the surface of the specimen, the corrosion rate of the specimen decreases with the increase in corrosion time; the surface roughness of the specimen after corrosion first increases and then decreases with the increase in corrosion time; and the tensile strength of specimen with the surface residual stress of −335.64 MPa is relative good after 4 h of corrosion.

Based on the associated network of passenger and freight transport in the New Western Land-Sea Corridor, the structure and characteristics of passenger and freight transport network in different years are analyzed using passenger volume and freight volume, and the resilience of the network is analyzed using the disturbance model. Finally, the coordinated development of passenger transport and logistics is discussed by applying a coupling coordination model. The results show that, from 2000 to 2018, the New Western Land-Sea Corridor has realized the evolution of the spatial pattern of passenger transport network from “two center and numerous branches” to “three centers and three auxiliaries”. Moreover, the evolution of the spatial pattern of the freight transport network changed from “two centers and one auxiliary” to “one center and multiple auxiliaries”. In terms of network tenacity, the disturbance strategy based on node degree value has the strongest damage to the performance of passenger and freight transport networks. The coupling coordination value is gradually rising from a low level and finally develops to an ordered trend. Therefore, it is proposed that passenger transport development could be supported from the “growth pole” and “zone” development modes, while the logistics development would adopt the “axis” development pattern and focus on the protection of the vital city nodes in the corresponding network. In this way, it can realize the effective connection of multiple paths, improve the tenacity of network and boost hub economy, consummate the regional coordination mechanism, and finally promote the advancement of the New Western Land-Sea Corridor.

For the unique structural characteristics of ventilated brake discs and the complex problem of energy conversion during braking, a calculation method for energy conversion of the ventilated brake disc based on simultaneous heat generation and heat dissipation is proposed. The transient heat transfer model of the ventilated brake disc for high-speed trains is established. Considering the control equations of heat generation–heat dissipation and plate–cylinder convection heat transfer, the virtual simulation of the energy change of the ventilated brake disc during the braking process is carried out. The temperature and stress distribution of contact friction surface and clearance structure of the ventilated brake disc are analysed from the perspective of function conversion. The results show that the heat generated by the ventilated brake disc increases nonlinearly, and the heat dissipated increases linearly. The heat of ventilated brake disc increases with the increase of braking time, but its growth rate decreases continuously. The maximum temperature of the ventilated brake disc is 268 °C, which appears on the friction surface. After braking, its heat is 6.636 × 106 J. The analysis results and methods provide a basis for optimizing the structure of ventilated brake discs.

In recent years, the proposed Deep Interest Network (DIN), Deep Interest Evolution Network (DIEN) and Deep Session Interest Network (DSIN) have further developed click-through rate prediction models. The above three models mainly focus on the evolution and development of the user’s historical behavior sequence. To a certain extent, the influence of environmental vectors on the user’s choice of the advertisement for the item to be recommended is ignored. As a result, click-through rates cannot be predicted more accurately when items have strong environmental attributes. To solve this problem, we propose a new model based on DIN, called Deep Interest Context Network (DICN). DICN combines two local activation units. It adaptively learns the user’s interest representation from the user’s historical behavior data concerning an advertisement and the context in which the advertisement is located (i.e., environmental factors). The experimental results show that DICN significantly improves the performance and model expression ability of advertisements with strong environmental attributes.

Depending on the type of information network, information network embedding is classified into homogeneous information network embedding and heterogeneous information network (HIN) embedding. Compared with the homogeneous network, HIN composition is more complex and contains richer semantics. At present, the research on homogeneous information network embedding is relatively mature. However, if the homogeneous information network model is directly applied to HIN, it will cause incomplete information extraction. It is necessary to build a specialized embedding model for HIN. Learning information network embedding based on the meta-path is an effective approach to extracting semantic information. Nevertheless, extracting HIN embedding only from a single view will cause information loss. To solve these problems, we propose a multi-view fusion-based HIN embedding model, called MFHE. MFHE includes four parts: node feature space transformation, subview information extraction, multi-view information fusion, and training. MFHE divides HIN into different subviews based on meta-paths, models the local information accurately in the subviews based on the multi-head attention mechanism, and then fuses subview information through a spatial matrix. In this paper, we consider the relationship between subviews; thus, the MFHE is applicable to complex HIN embedding. Experiments are conducted on ACM and DBLP datasets. Compared with baselines, the experimental results demonstrate that the effectiveness of MFHE and HIN embedding has been improved.

In order to study the influence of workpiece surface topography on deformation and residual stress during shot peening, the paper built a rough surface single-shot peening simulation model and designed a rough surface workpieces shot peening experiment to verify the correctness of the built simulation model, which differed from the smooth surface simulation model and was closer to the actual processing situation. Based on the W-M fractal dimension theory, the shot peening model of the target with a rough surface is established, and the simulation analysis is carried out by finite element software. The effect of different impact positions of the pellet on surface residual stress and surface displacement were also simulated. The simulation found that with the increase of surface complexity, the surface deformation and residual stress would both increase after shot peening. Through shot peening tests on workpieces with different surface roughness, it is found that the error between simulation and experiment results on surface deformation and residual stress are both less than 20%, which verifies the validity of the rough surface single-shot peening model and that the established model could provide a new method on shot peening simulation analysis.

In recent years, the recommendation system and robot learning are undoubtedly the two most popular application fields, and the core algorithms supporting these two fields are deep learning based on perception and reinforcement learning based on exploration learning, respectively. How to combine these two fields to better improve the development of the whole machine learning field is the dream of numerous researchers. The Deep Reinforcement Network (DRN) model successfully embedded reinforcement learning into the recommendation system, which provided a good idea for subsequent researchers. However, the disadvantage is also obvious, that is, the DRN model is built for news recommendations, meaning that the DRN model is not transferable, which is also the defect of many current recommendation system models. Meanwhile, the agent learning method adopted by the DRN model is primitive and inefficient. Among many models and algorithms that have emerged in recent years, we use the newly proposed deployment efficiency to measure their comprehensive quality and found that few models focus on both efficiency and performance improvement. To fill the gap of model deployment efficiency neglected by many researchers and to create a model of reinforcement learning agents with stronger performance, we have been exploring and trying to complete research on the Gate Attentional Factorization Machines (GAFM) model. Finally, we successfully integrated the GAFM model and reinforcement learning. The Deep Reinforcement Factorization Machines (DRFM) model proposed in this paper is based on the combination of deep learning with high perception ability and reinforcement learning with high exploration ability, centered on improving the deployment efficiency and learning performance of the model. The GAFM model is modified and upgraded using multidisciplinary techniques, and a new model-based random exploration strategy is proposed to update and optimize the recommendation list efficiently. Through parallel contrast experiments on various datasets, it is proved that the DRFM model surpasses the traditional recommendation system model in all aspects. The DRFM model is far superior to other models in terms of performance and robustness, and also significantly improved in terms of deployment efficiency. At the same time, we conduct a comparative analysis with the latest deep reinforcement learning algorithm and prove the unique advantages of the DRFM model.

With the rapid development of recommendation systems, models and algorithms supporting the core of recommendation systems have emerged one after another, and researchers have attempted to optimize them. However, the structure of these models is complex. Popular deep neural networks often achieve the highest utilization of data by increasing the number of hidden layers, ignoring the problems of exploding and vanishing gradients and even the entire degradation of the networks. However, researchers pay too much attention to algorithms and models and do not consider the dataset itself. Methods for processing data and finding possible connections between the data and models have become new explorable points. Cold start is also a problem that researchers have been trying to solve and optimize since the birth of the recommendation system. Recent studies also provide good ideas for solving cold start, but the problem is that researchers still do not focus on datasets. In order to fill the gap in the exploration and research of datasets, this paper takes the long tail distribution and cold start problems that are common in recommendation systems such as the starting point, combines the residual network in computer vision with deep learning, and proposes the aging mechanism of datasets. In this paper, a multi-layer residual network based on aging mechanisms called Aging Residual Factorization Machines (ARFM) is proposed. Parallel experiments with other model algorithms are carried out on three datasets of different sizes and categories. Experimental results show that ARFM achieve performance advantages under the premise of different recommendation tasks.

Nowadays, to deal with the increasing data of users and items and better mine the potential relationship between the data, the model used by the recommendation system has become more and more complex. In this case, how to ensure the prediction accuracy and operation speed of the recommendation system has become an urgent problem. Deep neural network is a good solution to the problem of accuracy, we can use more network layers, more advanced feature cross way to improve the utilization of data. However, when the accuracy is guaranteed, little attention is paid to the speed problem. We can only pursue better machine efficiency, and we do not pay enough attention to the speed efficiency of the model itself. Some models with advantages in speed, such as PNN, are slightly inferior in accuracy. In this paper, the Gate Attention Factorization Machine (GAFM) model based on the double factors of accuracy and speed is proposed, and the structure of gate is used to control the speed and accuracy. Extensive experiments have been conducted on data sets in various application scenarios, and the results show that the GAFM model is better than the existing factorization machines in both speed and accuracy.

Nowadays, as the number of items is increasing and the number of items that users have access to is limited, user-item preference matrices in recommendation systems are always sparse. This leads to a data sparsity problem. The latent factor analysis (LFA) model has been proposed as the solution to the data sparsity problem. As the basis of the LFA model, the singular value decomposition (SVD) model, especially the biased SVD model, has great recommendation effects in high-dimensional sparse (HiDs) matrices. However, it has the disadvantage of requiring several iterations before convergence. Besides, the model PID-incorporated SGD-based LFA (PSL) introduces the principle of discrete PID controller into the stochastic gradient descent (SGD), the learning algorithm of the SVD model. It could solve the problem of slow convergence speed, but its accuracy of recommendation needs to be improved. In order to make better solution, this paper fuses the PSL model with the biased SVD model, hoping to obtain better recommendation result by combining their advantages and reconciling their disadvantages. The experiments show that this biased PSL model performs better than the traditional matrix factorization algorithms on different sizes of datasets.

Eco-friendly pervious concretes containing fly ash (FA) and blast furnace slag (BFS) were prepared in this study. The compressive strength and hydraulic behaviors were investigated to explore the effect of replacement content of FA and BFS. Rheological tests of cementitious pastes were first conducted and the results showed that FA could increase the apparent viscosity and BFS did not change the rheology performance. Compared to traditional concretes, FA and BFS both decreased the compressive strength of pervious concrete at 28 d, while pervious concrete incorporated with FA and/or BFS presented comparable strength at 60 d. Compared to the control concrete mix, at the same replacement rate, FA changed the compressive strength more obviously than BFS. FA and BFS both decreased the effective porosity and permeability coefficient of pervious concrete. However, when the replacement rate (30%) was the same, concretes with ternary blends presented obviously larger porosity than binary blends. The relationships between porosity and permeability, and strength were also established.