机械镀锌-铝复合镀层的结构分析

采用金相显微镜、扫描电镜、能谱仪分析了混粉法机械镀锌-铝复合镀层的结构、形貌和化学成分分布。结果表明，机械镀锌-铝镀层主要是由尺寸不等的锌粉颗粒相互参杂、填充，形成致密结构体；锌粉颗粒之间的间隙填充着更小尺寸的锌粉和夹杂物，镀层中部分锌粉颗粒发生变形；片状铝粉在镀层中不均匀分布，铝粉和锌粉没有明显的包裹、缠绕；镀层中Fe，Sn，Zn的含量分布正常，且Al的含量比施镀前混粉中Al的含量低得多。     

Betweenness estimation in OLSR-based multi-hop networks for distributed filtering

In traditional networks special efforts are put to secure the perimeter with firewalls: particular routers that analyze and filter the traffic to separate zones with different levels of trust. In wireless multi-hop networks the perimeter is a concept extremely hard to identify, thus, it is much more effective to enforce control on the nodes that will route more traffic. But traffic filtering and traffic analysis are costly activities for the limited resources of mesh nodes, so a trade-off must be reached limiting the number of nodes that enforce them. This work shows how, using the OLSR protocol, the centrality of groups of nodes with reference to traffic can be estimated with high accuracy independently of the network topology or size. We also show how this approach greatly limits the impact of an attack to the network using a number of firewalls that is only a fraction of the available nodes.     

Dynamics stability in wireless sensor networks active defense model

Wireless sensor networks, which are widely used in military, industrial and transportation fields, are vulnerable to various kinds of attacks, since they are mostly deployed in a relatively open environment. Based on the evolutionary game theory, this paper proposes a proactive defense model for wireless sensor networks, in which we emphasize that the node has a limited ability to learn the evolution of rationality from different attack strategies of the attacker, and can dynamically adjust their strategies to achieve the most effective defense. Following this approach, the cost (e.g., energy consumption and wastage of machinery) has been greatly saved and the life cycle of the nodes has been extended as well. By employing the proposed model, the whole wireless sensor network can be implemented in an effective way.     

Anticipatory network models of multicriteria decision-making processes

In this article, we will investigate the properties of a compromise solution selection method based on modelling the consequences of a decision as factors influencing the decision making in subsequent problems. Specifically, we assume that the constraints and preference structures in the (k?+?1)st multicriteria optimisation problem depend on the values of criteria in the k-th problem. To make a decision in the initial problem, the decision maker should take into account the anticipated outcomes of each linked future decision problem. This model can be extended to a network of linked decision problems, such that causal relations are defined between the time-ordered nodes. Multiple edges starting from a decision node correspond to different future scenarios of consequences at this node. In addition, we will define the relation of anticipatory feedback, assuming that some decision makers take into account the anticipated future consequences of their decisions described by a network of optimisers ? a class of information processing units introduced in this article. Both relations (causal and anticipatory) form a feedback information model, which makes possible a selection of compromise solutions taking into account the anticipated consequences. We provide constructive algorithms to solve discrete multicriteria decision problems that admit the above preference information structure. An illustrative example is presented in Section 4. Various applications of the above model, including the construction of technology foresight scenarios, are discussed in the final section of this article.     

An order-picking operations system for managing the batching activities in a warehouse

Nowadays, customer orders with high product variety in small quantities are often received and requested for timely delivery. However, the order-picking process is a labour-intensive and costly activity to handle those small orders separately. In such cases, small orders are often grouped into batches so that two or more orders can be served at once to increase the picking efficiency and thus reduce the travel distance. In this paper, an order-picking operations system (OPOS) is proposed to assist the formulation of an order-picking plan and batch-handling sequence. The study integrates a mathematical model and fuzzy logic technique to divide the receiving orders into batches and prioritise the batch-handling sequence for picking, respectively. Through the proposed system, the order-picking process can be managed as batches with common picking locations to minimise the travel distance, and the batch-picking sequence can be determined as well. To demonstrate the use of the system, a case study in a third-party logistics warehouse is presented, and the result shows that both the order-picking activity and labour utilisation can be better organised.     

Towards adaptive learning with improved convergence of deep belief networks on graphics processing units

In this paper we focus on two complementary approaches to significantly decrease pre-training time of a deep belief network (DBN). First, we propose an adaptive step size technique to enhance the convergence of the contrastive divergence (CD) algorithm, thereby reducing the number of epochs to train the restricted Boltzmann machine (RBM) that supports the DBN infrastructure. Second, we present a highly scalable graphics processing unit (GPU) parallel implementation of the CD-k algorithm, which boosts notably the training speed. Additionally, extensive experiments are conducted on the MNIST and the HHreco databases. The results suggest that the maximum useful depth of a DBN is related to the number and quality of the training samples. Moreover, it was found that the lower-level layer plays a fundamental role for building successful DBN models. Furthermore, the results contradict the pre-conceived idea that all the layers should be pre-trained. Finally, it is shown that by incorporating multiple back-propagation (MBP) layers, the DBNs generalization capability is remarkably improved.     

悬浮法合成固体粒状丙烯酸树脂

研究了悬浮法合成固体丙烯酸树脂颗粒。讨论了不同软硬单体配比对丙烯酸树脂颗粒的玻璃化转变温度、相对分子质量及其分布的影响以及不同的引发剂及悬浮剂对其聚合过程、最终产物形态及上述性能的影响。制得的丙烯酸树脂的数均相对分子质量用凝胶渗透色谱(GPC)测定为3.9×104～6.2×104,相对分子质量分布为2.04～2.32;用红外光谱图表征了制得的丙烯酸树脂结构。差示扫描量热法(DSC)测定了聚合物的玻璃化转变温度(Tg)。     

Measurement of longitudinal distribution of solids holdup in a three-phase fluidized bed by ultrasonic technique

A method of measurement of solids holdup in a three-phase reactor by analyzing the shape and the phase lag or lead of an ultrasonic wave has been proposed. The solids holdup in a solids-dilute three-phase fluidized bed could be measured without being affected by the presence of gas bubbles. The experimental values of solids axial dispersivity measured by the ultrasonic technique are in reasonable agreement with an empirical correlation of previous data obtained by other methods. The possibility of the application of the ultrasonic technique to the measurement of solids holdup in a three-phase fluidized bed has been suggested.     

逆凝聚包覆法制备硫黄调节型粉末氯丁橡胶 Ⅰ.粒径分布

采用逆凝聚包覆法制备硫黄调节型粉末氯丁橡胶(PCR)并考察加料方式、搅拌转速和胶乳加人流速对产物粒径分布的影响。建立氯丁胶乳的逆凝聚粉末化模型并采用均匀试验设计的方法对该模型进行了验证。验证结果表明，该模型可以有效地模拟PCR的逆凝聚包覆氯丁胶乳的粉末化过程。PCR产物中粒径1～3mm的中等粒子的质量分数仅与搅拌转速有关，表明湍流爆裂是产生中等粒子的主要原因，而带状拉伸则是产生较粗颗粒和细小粒子的主要原因。扫描电子显微镜观察发现，采用逆凝聚包覆法制备的PCR粒子为单颗连续粒子，即产物粒子由单颗胶乳液滴凝聚而成。     

Synthesis and characterization of epoxidized styrene–butadiene rubber/silicon dioxide hybrid materials

Hybrid materials were synthesized from epoxidized (68, 43, or 14%) styrene–butadiene rubber (SBR) and the hydrolysis product of tetraethoxysilane (TEOS) in situ under ultrasonic irradiation. The products were characterized with thermal analysis (differential scanning calorimetry and thermogravimetric analysis), stress–strain tests, scanning electron microscopy (including energy‐dispersive spectrometry), and swelling in tetrahydrofuran and water. The most transparent were those prepared from SBR with the highest degree of epoxidation, whereas those obtained from less epoxidized SBR and with larger amounts of TEOS showed distinct phases that could be considered two hybrid phases (one rich in TEOS and another rich in SBR). © 2004 Wiley Periodicals, Inc. J Appl Polym Sci 92: 798–803, 2004