Selecting informative data for defect segmentation from imbalanced datasets via active learning

The storage and labeling of industrial data incur significant costs during the development of defect detection algorithms. Active learning solves these problems by selecting the most informative data among the given unlabeled data. The existing active learning methods for image segmentation focus on studying natural images and medical images, with less attention given to industrial images, and little research has been performed on imbalanced data. To solve these problems, we propose an active learning framework to selecting informative data for defect segmentation under imbalanced data. In the initialization stage, the framework uses self-supervised learning to initialize the data so that the initialization data contain more defect data, thereby solving the cold-start problem. During the iterative stage, we design the main body of the active learning framework, which is composed of a segmentation learner and a reconstruction learner. These learners use supervised learning to further improve the framework’s ability to select informative data. The experimental results obtained on public and self-owned datasets show that the framework can save 70% of the required storage space and greatly reduce the cost of labeling. The intersection over union value proves that the designed framework can achieve the equivalent effect of labeling the whole dataset by labeling partial data.     

A new supervised multi-head self-attention autoencoder for health indicator construction and similarity-based machinery RUL prediction

Remaining useful life (RUL) prediction plays a significant role in the prognostic and health management (PHM) of rotating machineries. A good health indicator (HI) can ensure the accuracy and reliability of RUL prediction. However, numerous existing deep learning-based HI construction approaches rely heavily on the prior knowledge, and they are difficult to capture the key information in the process of machinery degradation from raw signals, thereby affecting the performance of RUL prediction. To tackle the aforementioned problem, a new supervised multi-head self-attention autoencoder (SMSAE) is proposed for extracting the HI that effectively reflects the degraded state of rotating machinery. By embedding the multi-head self-attention (MS) module into autoencoder and imposing the constraint of power function-type labels on the hidden variable, SMSAE can directly extract the HIs from raw vibration signals. As the current HI evaluation indexes don’t consider the global monotonicity and variation law of HI, two improved monotonicity and robustness indexes are designed for the better evaluation of HI. With the proposed HI, a two-stage residual life prediction framework based on similarity is developed. Extensive experiments have been performed on an actual wind turbine gearbox bearing dataset and a well-known open commercial modular aero-propulsion system simulation (C-MAPSS) dataset. The comparative results verify that the constructed SMSAE HI has better comprehensive performance than the typical HIs, and the proposed prediction method is competitive with the state-of-the-art methods.     

不同钙基吸附剂捕集CO2后的硫酸化反应特性研究

钙基吸附剂进行多次CO₂捕集后,碳酸化效率会大幅衰减,此时的吸附剂能否高效脱硫利用是值得重点关注的问题。鉴于此,筛选了高性能合成钙基吸附剂和天然石灰石吸附剂,通过热重分析仪分析对比其在多循环CO₂捕集后的碳酸化和硫酸化反应性能,采用微粒模型研究其硫酸化反应动力学特征。结果发现,高性能合成钙基吸附剂的碳酸化反应速率和CO₂吸附能力明显高于石灰石吸附剂。在长达500循环的CO₂捕集试验后,高性能合成钙基吸附剂的CO₂吸附能力比石灰石高10倍以上,其SO₂吸附能力相较于石灰石提升约40%。经历多次CO₂捕集反应循环后,2种吸附剂的硫酸化能力均有提升:其中,石灰石吸附剂的提升幅度更大,硫酸化转化率从26%提升到35%,而高性能合成钙基吸附剂的硫酸化转化率则从38%提升到43%。通过微粒模型计算发现,2种吸附剂的硫酸化反应均是与SO₂浓度相关的一级反应,多循环捕集CO₂反应后,石灰石吸附剂的硫酸化反应活化能下降接近30%,而高性能合成钙基吸附剂的硫酸化反应活化能只下降了5%。研究结果说明2种不同钙基吸附剂在进行循环CO₂捕集后,脱硫能力得到了不同程度的提高,且均可以较好地应用于SO₂的脱除。     

船舶凝水调节管道新型节流孔板设计研究

船用凝水管道上节流孔板设计不合理,易导致孔板下游流体发生汽蚀,并促使流体高速冲击管道,进而诱发管道剧烈振动和冲刷腐蚀失效,极大影响了船用凝水管道的安全性和稳定性。基于多孔消声器减振降噪机理,提出了具有"防止孔板汽蚀,减缓管路冲刷"功能的新型斜孔对冲式节流孔板,引入伯努利方程和阻塞压差计算方法,进行节流孔板孔径、级数和厚度的校核计算,得到适用于凝水调节管道的两级斜孔式节流孔板,并利用数值模拟手段对理论设计的单级和二级斜孔式孔板节流特性进行分析,计算结果表明单级节流孔板下游局部流域诱发大尺度涡流和汽蚀现象,易导致凝水管路振动;二级节流孔板不仅可以规避孔板诱发汽蚀,还能实现流体对冲射流效应,促使冲击能量相互抵消,减缓流体对管路的冲刷,数值模拟结果与理论设计数据基本吻合。因此该计算方法可为船舶凝水系统管道节流孔板结构设计提供一定理论支撑。     

Numerical and experimental fatigue crack growth analysis in mode-I for repaired aluminum panels using composite material

Crack-front shape is an important parameter influencing the stress intensity factor and crack propagation rate in asymmetric repaired panels. In this study, the numerical and experimental fatigue crack growth behaviour of centrally cracked aluminum panels in mode-I condition repaired with single-side composite patches are investigated. It is shown that the crack growths non-uniformly from its initial location through the thickness of a single-side repaired panel. There is a good agreement between the propagated crack-front shapes obtained from finite element analysis with those obtained from the experiments for various repaired panels with different patch thicknesses. Furthermore, effects of plate and patch thickness on the crack growth life of the repaired panels are investigated. The experimental results show that the crack growth life of thin panels may increase up to 236% using a 16 layers patch. However, for thick panels, the life may extend about 21–35% using a 4 layers patch. Implementing of 8 and 16 layers patches has not a significant effect on the life extension of thick panels with respect to the 4 layers patch life.     

Systematic analysis on combined heat and water transfer through porous materials based on thermodynamic energy

A new thermodynamic energy ‘water potential’ based on the principles of chemical potential of an element of mixed gas is defined as the driving force of gaseous phase water flux. Adhesive power or ‘capillary action’ and a portion of the water potential, is confirmed as the driving force of liquid phase water flux. A numerical model of combined heat and water transfer using the water potential is introduced and influences of forces, such as gravity and pressure on water flux are incorporated from the viewpoint of thermodynamics. A way to estimate diffusivities of gaseous and liquid phase water through porous materials is also shown. Accuracy of the numerical model is demonstrated through a comparison between calculation and experiment for different temperature gradients and water content in a porous material.     

低相对分子质量窄分布的聚丙烯酸RAFT聚合

采用可逆加成–断裂链转移自由基聚合(RAFT)选择水溶性的SS′二(αα二甲基–α′′乙酸)三硫代碳酸酯(TRIT为RAFT链转移剂,以水为溶剂,合成了相对分子质量为5 000~70 000的聚丙烯酸。GPC凝胶渗透色谱)测试结果表明合成的聚丙烯酸具有很窄的相对分子质量分布,多分散系数控制在1.0~1.8。     

Action recognition by joint learning

Due to the promising applications including video surveillance, video annotation, and interaction gaming, human action recognition from videos has attracted much research interest. Although various works have been proposed for human action recognition, there still exist many challenges such as illumination condition, viewpoint, camera motion and cluttered background. Extracting discriminative representation is one of the main ways to handle these challenges. In this paper, we propose a novel action recognition method that simultaneously learns middle-level representation and classifier by jointly training a multinomial logistic regression (MLR) model and a discriminative dictionary. In the proposed method, sparse code of low-level representation, conducting as latent variables of MLR, can capture the structure of low-level feature and thus is more discriminate. Meanwhile, the training of dictionary and MLR model are integrated into one objective function for considering the information of categories. By optimizing this objective function, we can learn a discriminative dictionary modulated by MLR and a MLR model driven by sparse coding. The proposed method is evaluated on YouTube action dataset and HMDB51 dataset. Experimental results demonstrate that our method is comparable with mainstream methods.     

Composite cure simulation scheme fully integrating internal strain measurement

This study proposes a new approach to determine key material parameters for stress/strain calculation of curing composite laminates and validate the simulation. Specifically, fiber Bragg grating (FBG) strain sensors are embedded in a composite laminate and the two key parameters for simulation, composite shrinkage strain and stiffness change during curing, are simultaneously determined from in-situ measurements by the embedded sensors. Furthermore, the simulation is validated using internal strain change during curing. This paper begins by presenting an overview of the proposed simulation scheme and by comparing it with previous approaches to highlight its advantages. Material parameter determination using a shear-lag effect at the edge of the embedded sensors is then described and the practical procedure to obtain the key parameters is demonstrated using a carbon/epoxy laminate. Finally, cure simulation is conducted for validation. Further extension to more general cases including thermoplastic composites is also discussed.     

基于板块元的水面舰船远场声目标强度预报方法

提出了一种基于板块元的水面舰船水下远场声目标强度数值预报方法,首先根据水面舰船水下船体及附体的型线进行等效建模,采用板块元方法进行网格划分,建立了水面舰船特有的复杂附体二次反射、海面散射效应的数值模型,将各板块元的声散射视为一种"滤波",各板块元的散射声传输函数具有不同幅度、相位、时延,将其叠加之和进行反傅氏变换后,获得水面舰船波形级的远场声目标强度及方位分布特征。通过分析比较典型声呐探测信号下水面舰船的声目标强度,验证了该方法的有效性,为优化与控制水面舰船目标强度,分析水面舰艇水声对抗的作战效能提供了物理依据。