Based on the similarity of the community detection methods, the Givern-Newman (GN) algorithm is fast and accurate but has a higher running time. In order to improve the efficiency of GN Algorithm, this study presents a semi-supervised GN algorithm based on node similarity. By making full use of the constraint set of the prior knowledge must-link and cannot-link, the prior information is extended by the derived rules, and the extended information is verified by the method of distance measurement. Using new annealing maximisation algorithm to calculate node similarity iteratively, and validated using artificial and real networks. It proves that the proposed algorithm reduces the GN algorithm's running time and improves efficiency.

This study presents the implementation of a new unsupervised machine learning based system called a buck converter controller using unsupervised machine learning (ABCML) to control the operation of a type of switching voltage regulators, commonly called buck converters. The system uses the voltage and current state variables of buck converters to evaluate four switching-based clustering algorithms, namely a Gaussian mixture model, a self-organising mapping (SOM), a k-means clustering and a hierarchical clustering algorithm. Step response results of the controller implementation with the buck converter show that the SOM controller improves the performance of the voltage regulator system by 99.77% in terms of voltage overshoot and by 63.24% in terms of fall time, whereas the hierarchical clustering algorithm improves the settling time of the output voltage by 3.91%. This finding of optimal machine learning implementation approaches and their comparison could be used to improve the stability and the performance of switching voltage regulation systems which are widely used in electronic systems of today.

The circular micro-texture is constructed on the inner surface of the gap seal hydraulic cylinders. The coupling relationship of the surface roughness and micro-texture is studied and solved by the equivalent flow method. At the same time, the tribological experiment is carried out with 45 steel materials commonly used as hydraulic cylinders, and the testing results are compared. The results show that the roughness has a very significant effect on the frictional properties of the annular textured surface. The coupling effect is generated and makes the friction coefficient first decrease and then increase. There are optimal roughness and optimal gap which make the further improvement of the frictional properties on the cylinder surface. This helps to improve the response frequency of the hydraulic cylinder, thereby improving the hydraulic system efficiency. Also, it has positive effects on the improvement of the working efficiency in the hydraulic system.

This research studied the wringing effect present on a piston with two contact surfaces in a downhole tool and investigated piston design guidance to prevent this occurrence. A piston on the downhole drilling tool is designed to be moved by hydraulic forces on-demand in downhole operation. The piston failed to move during an experiment at a depth of 8772 m in an oil drilling well located in the Gulf of Mexico. The wringing effect was suspected as the root cause of this failure, and subsequent laboratory experiments confirmed that wringing presence caused non-functional piston. Then, potential influencing factors causing wringing effects were analysed and investigated in a series of laboratory experiments aiming to eliminate the wringing adhesive force between the contact surface of the piston and its mandrel. The investigated factors included the roughness and area of the contact surfaces and pre-compression force. Based on the findings, the piston design was optimised and re-tested, and it was found that the optimised piston eliminated wringing adherence. With experimental data, semi-empirical design guidance regarding the influencing factors was summarised for critical piston design to eliminate possible wringing.

The correct topological relationship is crucial in the low-voltage distribution network, as the actual topological structure of the low-voltage distribution network changes frequently and tremendously due to the need of operation and maintenance, and it cannot be correctly reflected upon failure in timely updating of data, low circulation, poor quality etc. therefore, it is necessary to identify the topology. The knowledge graph technology can clearly reflect the existing relationship between data, deducing and mining hidden knowledge, suitable for topology identification of the low-voltage distribution network. In the study, the knowledge graph technology was employed for topology identification: firstly, analyse the construction method of the knowledge graph, integrate data in multiple low-voltage distribution network information systems based on the knowledge graph technology, deduce missing data, find out the relationship between data, and then build the knowledge graph of low-voltage distribution network topological structure, and finally, based on ‘Typical Design Specification of Low-Voltage Distribution Network Infrastructure Project’ and semantic segmentation, identify the user–transformer relationship in low-voltage distribution network information system. The test results of the examples were very satisfactory, showing the theoretical values and practical application values of the identification method proposed in this study.

To realise the distributed control of the hybrid energy storage system (HESS) in an islanded AC microgrid, a dynamic HESS power allocation strategy based on the virtual impedance (VI) for supercapacitor (SC) and the battery is proposed. Dynamic power-sharing of two kinds of energy storage devices can be achieved without real-time measuring of load power. The state of charge (SOC) recovery of SC is achieved with a SOC loop integrated into the VI loop. The principle of dynamic power-sharing of HESS is derived based on the establishment of HESS equivalent circuit model. By analysing the influence of the voltage and current double loop controller parameters on the output impedance and the VI of the DC/AC converter, the method of setting the controller parameters for power distribution is presented. A simulation model is established with the designed equivalent power fluctuation conditions. The simulation results show that under various working conditions, the SC loop can compensate for the high-frequency part of the equivalent power fluctuation and the battery loop absorbs the low-frequency part of the power fluctuation. The dynamic power-sharing of HESS with the SC SOC recovery is realised effectively.

Adding polypropene (PP) fibre can enhance the mechanical properties of coal gangue ceramsite concrete (CGCC), which is a kind of green concrete for solid waste utilisation. However, there is a lack of data concerning the effects of PP fibre content on the micro-pore structure characteristics and macro-mechanical properties of CGCC. This investigation aims at exploring the optimum PP fibre content and establishing the relationship between micro-pore structure characteristics and macro-mechanical properties, which provides a reference for optimising the mixture ratio. Based on low-field nuclear magnetic resonance technology and mechanical strength test, micro-pore structure characteristics and macro-mechanical properties of CGCC with different PP fibre content of 0, 0.6, 0.9, and 1.2 kg m⁻³, respectively, were studied. The results showed that when the fibre content was 0.6kg m⁻³, the mechanical strength was the highest and the increase of splitting tensile strength was the largest. Moreover, PP fibre can effectively reduce the most probable pore diameter and can transform harmful pores to harmless pores with the optimum content of 0.6 kg m⁻³. In conclusion, appropriate PP fibre addition has a significantly improving effect on the micro-pore structure characteristics and macro-mechanical properties of CGCC.

Low-dose computed tomography (LDCT) image often contains mottle noise and streak artefacts, which seriously interfere with clinical diagnosis. In this study, the separation-based (SEPB) method is proposed for mottle noise and streak artefacts suppression and structure preservation. In it, the LDCT image is decomposed into the structural image with residual mottle noise and the streak artefacts image with residual structural details by the image decomposition structural-preserving image smoothing method. The structural image is filtered by the K-singular value decomposition algorithm to remove the residual mottle noise, and the structural details in the streak artefacts image are extracted by the morphological component analysis theory. The extracted structural details are added to the filtered structural image to get the LDCT result image. Meanwhile, in the process of extracting the structural details, the streak artefacts dictionary learned from the streak artefacts image is corrected by the local intuitional fuzzy entropy to remove its structural atoms. The experiments are conducted on the modified Shepp–Logan phantom, the pelvis phantom and the clinical abdominal data to evaluate the proposed SEPB method. Compared to several comparative denoising methods, the experimental results show that the SEPB method has better performance in subjective visual effect and objective indicators.

This study deals with the error probability optimisation problem in three-phase bidirectional amplify and forward relaying systems. Different from the common two-phase strategy, in the three-phase counterpart, the relay power is split into two different terminals in terms of data error probability. With such a three-phase division, the authors present a unified optimisation algorithm based on the exact symbol error probability (SEP) and its upper bound expression. In addition, by virtue of the dual decomposition algorithm, the authors derive the optimal power allocation and demonstrate its uniqueness at both terminals and the relay node. Moreover, the unified optimisation algorithm can be reduced to some special cases, such as the maximum minimisation problem and unbalanced link problem. The proposed results reveal that the proposed algorithm has a significant impact on SEP reduction.

Aiming at solving problems such as non-linear, time-varying parameters, and weak digital computing capability in the underground drainage system, the monitoring system is built based on a digital signal processor (DSP). Adopting DSP28335 as its control core, the hardware circuit and software program are designed in this monitoring system. The single-neuron fuzzy proportional, integration, and differential (PID) control algorithm with feedforward proportional and differential compensation is proposed in this study. The simulation results are compared and analysed with a traditional PID algorithm, fuzzy PID algorithm, and single-neuron fuzzy PID algorithm with feedforward in Matlab/Simulink, and the experimental platform is built to verify the application effect of three control algorithms. The simulation and experiment results show that the single-neuron fuzzy PID algorithm with feedforward has significant advantages such as shorter adjustment time, good adaptivity, and strong anti-interference ability, which could effectively improve the work efficiency of an underground mine drainage monitoring system.

Recently, the deep convolutional neural networks (CNNs) have shown great success and for facial expression recognition (FER). These CNN-based approaches have made breakthroughs in the accuracy by using deeper networks. As the amount of train data is very less, these models easily fall into the overfitting in the training. In this study, a novel multi-scale deep residual attention network (Ms-RAN) is proposed for FER. The proposed Ms-RAN is mainly based on the multi-scale residual attention unit, which consists of two different scale sub-units. Each sub-unit is composed of the convolutional layers, parametric rectified linear units (PReLUs), and the residual attention connection. By focusing on the relationship between channels and automatically learning the importance of different channel features, the proposed Ms-RAN can make the proposed model pay more attention to the most informative channel features, while suppressing those unimportant channel features. Owing to the unique design of Ms-RAN, the combination of various levels features can be enhanced in the proposed method, and valuable and different ranges of expressive information can also be provided for recognition. The experimental results demonstrate that the proposed method achieves superior performance than other state-of-the-art approaches on five databases, CK+, Oulu-CASIA, BU-3DFE, BP4D+, and MMI.