Failure load prediction and optimisation for adhesively bonded joints enabled by deep learning and fruit fly optimisation

Adhesively bonded joints have been extensively employed in the aeronautical and automotive industries to join thin-layer materials for developing lightweight components. To strengthen the structural integrity of joints, it is critical to estimate and improve joint failure loads effectually. To accomplish the aforementioned purpose, this paper presents a novel deep neural network (DNN) model-enabled approach, and a single lap joint (SLJ) design is used to support research development and validation. The approach is innovative in the following aspects: (i) the DNN model is reinforced with a transfer learning (TL) mechanism to realise an adaptive prediction on a new SLJ design, and the requirement to re-create new training samples and re-train the DNN model from scratch for the design can be alleviated; (ii) a fruit fly optimisation (FFO) algorithm featured with the parallel computing capability is incorporated into the approach to efficiently optimise joint parameters based on joint failure load predictions. Case studies were developed to validate the effectiveness of the approach. Experimental results demonstrate that, with this approach, the number of datasets and the computational time required to re-train the DNN model for a new SLJ design were significantly reduced by 92.00% and 99.57% respectively, and the joint failure load was substantially increased by 9.96%.     

Influences of chitosan on freeze–thaw stability of Arenga pinnata starch

This study aims to investigate chitosan (CS) with five different molecular weight (Mw) on freeze–thaw stability of Arenga pinnata starch (APS) gel subjected to five freeze–thaw cycles (FTC). The syneresis of APS gels was reduced by adding CS and the APS gel with high Mw CS had lower syneresis duo to a higher water holding capacity (P < 0.05). The addition of CS significantly decreased the hardness and molecular ordered structure of APS gel. In addition, CS could improve the microstructural stability. The results suggested that CS could effectively improve the freeze–thaw stability of APS gel, and CS with higher Mw might have more practical utility to improve stability of APS gel.     

Comparison of cellulolytic enzyme complexes of different fungal origin

The cellulolytic culture filtrates of Trichoderma reesei QM 9414, Aspergillus terreus OKI 16/5 and Penicillium verruculosum WA 30 were purified and separated into components by one-step preparative isoelectric focusing (IF). The culture filtrates, the mycelia and the separated components were investigated for cellulolytic (filter-paper degrading (FPA), carboxymethylcellulose degrading (C_x), cotton hydrolyzing (C₁), cellobiase) and proteinase B activities. The molecular weights of the cellulolytic fractions were determined by SDS-electrophoresis. The culture filtrates differed in the proportions of the various cellulolytic activities. The analytical IF separation of the enzyme complexes resulted in a total of 28–31 protein fractions, mostly glycoproteins. The complexes of the different culture filtrates were separated, by preparative IF, into 28 fractions each. The greatest part of the activities could be recovered in fractions 8–18. The activities in the fractions were recovered to different extents. The main protein components in the enzyme complexes were found to be endoglucanases (C_x) with pI values between 3 and 6. Most of the endoglucanases proved to be glycoproteins. FPA and cellobiase activities were recovered to various extents in the fractions of the three culture filtrates. Only small parts of the C₁ activities were recovered in the Trichoderma and Penicillium fractions. The originally low C₁ activity of the Aspergillus culture filtrate was recovered to 76% in the separated fractions. Most of the C_x activities were found to be lost. However, the separation losses were found to be reversible: on combining the fractions, the activities originally present in the culture filtrate were nearly entirely restored. The molecular weights of the fractions of the different culture filtrates covered the range 10 000 to 70 000 D. The IF fractions of similar activity patterns also showed similar molecular weight distributions. The relatively large number of isoenzymes is assumed to be the result of the endocellular proteinase activity present, which splits off the cellulase enzyme components from proenzymes of high molecular weight during the fermentation process.     

拓扑特征映射神经网络的学习收敛性分析

本文利用神经交互作用函数描述拓扑特征映射神经网络，探讨了这种网络的学习收敛性。本文首先给出一个网络收敛的一般性结论，并利用该结论证明网络输入满足平均分布时的收敛性。     

三次均匀有理B样条曲线的权因子优化光顺算法

给出了一种使三次均匀有理Ｂ样条曲线光顺的权因子优化算法，通过优化计算，得到了光顺曲线的权因子。本文采用了非线性优化技术光顺曲线的权因子。     

An Experimental Comparison of the Nearest-Neighbor and Nearest-Hyperrectangle Algorithms

Algorithms based on Nested Generalized Exemplar (NGE) theory (Salzberg, 1991) classify new data points by computing their distance to the nearest generalized exemplar (i.e., either a point or an axis-parallel rectangle). They combine the distance-based character of nearest neighbor (NN) classifiers with the axis-parallel rectangle representation employed in many rule-learning systems. An implementation of NGE was compared to thek-nearest neighbor (kNN) algorithm in 11 domains and found to be significantly inferior to kNN in 9 of them. Several modifications of NGE were studied to understand the cause of its poor performance. These show that its performance can be substantially improved by preventing NGE from creating overlapping rectangles, while still allowing complete nesting of rectangles. Performance can be further improved by modifying the distance metric to allow weights on each of the features (Salzberg, 1991). Best results were obtained in this study when the weights were computed using mutual information between the features and the output class. The best version of NGE developed is a batch algorithm (BNGE FW_MI) that has no user-tunable parameters. BNGE FW_MI's performance is comparable to the first-nearest neighbor algorithm (also incorporating feature weights). However, thek-nearest neighbor algorithm is still significantly superior to BNGE FW_MI in 7 of the 11 domains, and inferior to it in only 2. We conclude that, even with our improvements, the NGE approach is very sensitive to the shape of the decision boundaries in classification problems. In domains where the decision boundaries are axis-parallel, the NGE approach can produce excellent generalization with interpretable hypotheses. In all domains tested, NGE algorithms require much less memory to store generalized exemplars than is required by NN algorithms.     

基于直觉模糊解析面积的聚类算法

针对现有的直觉模糊集聚类算法对权重的忽视或误用,提出一种基于直觉模糊解析面积的聚类算法.同时给出了直觉模糊集的解析面积和属性权重的计算方法,然后构造了聚类算法的目标函数,并给出聚类算法的详细步骤.算例验证了所提出的算法的合理性与可行性.     

多粒度犹豫模糊语言环境下未知权重的多属性群推荐方法

针对群推荐中存在的多粒度、犹豫性、模糊性语言信息问题, 提出多粒度犹豫模糊语言环境下未知权重的多属性群推荐方法. 首先, 提出多粒度犹豫模糊语言术语集的概念, 定义其距离公式; 然后, 在多粒度犹豫模糊语言环境下, 针对属性权重完全未知的情况, 建立目标规划模型, 利用拉格朗日方程求解, 针对属性权重不完全未知的情况, 建立线性规划模型求解; 最后, 通过算例计算和分析表明了上述模型求解权重问题是有效的.

     

Nonlinear sequential fusion estimation for clustered sensor networks

This paper is concerned with distributed fusion estimation problem for discrete‐time nonlinear systems with asynchronous sampling data in the clustered sensor networks. A nonlinear sequential fusion method consisting of a sequential measurement fusion (SMF) method and a sequential state fusion (SSF) method is proposed. It is shown that the SMF estimator using the Unscented Kalman Filtering (UKF) method can handle the asynchronous measurement data sequentially and the SSF estimator which uses sequential matrix weighting method can get a close performance as the centralized batch matrix weighting method but has a lower computational complexity. The proposed measurement fusion method is able to deal with asynchronous measurements and has lower computational complexity as compared with the augmentation method. A simulation example shows the effectiveness of the proposed nonlinear sequential fusion method.     

多源数据融合的中国人口数据空间化研究

人口数据通常是以行政区为单元进行统计,在与其他空间数据进行综合分析时往往存在尺度不一等诸多问题。人口数据空间化研究是探究人口空间分布规律的过程,其最终成果可为人口分布与其他数据的空间关联研究奠定基础,同时也可直接应用于灾害影响评估等领域,具有重要的学术意义和应用价值。基于人口分布和土地覆盖、交通路网数据的空间关系构建人口空间化模型,将人口划分为城镇人口和农村人口,通过相关权重法（Relative Weight）对土地覆盖类型赋权,利用交通路网数据修正,对农村人口进行空间展布;分析不同等级交通路网数据影响方式的不同,对城镇人口进行展布,最后通过规划求解等方法将不同类型数据叠加融合。人口数据空间化结果表明：我国人口分布对海拔地形十分敏感,低海拔平原和台地面积仅占我国国土面积的16.49%,而生活在其上的人口却占总人口的54.88%,相反,我国山地丘陵面积占比为55.78%,但是仅分布22.11%的人口;东部人口集中于黄淮海地区、四川盆地、长江中小游以及东北平原等地;中国城镇人口呈现典型的“点轴”分布特点,沿主要交通线的集中分布比较明显。