马钢设计院综合管理信息系统建设实践与思考

从管理的角度，介绍马钢设计院以项目管理为核心的管理信息系统建设的早期实践，并分析存在问题。     

420t/h无烟煤锅炉燃烧稳定性的探讨

云浮发电厂2台420t／h燃用无烟煤的锅炉在投产初期，锅炉存在燃烧稳定性差、结焦严重、易灭火等问题，经过不同阶段的改造，2台锅炉的燃烧稳定性得到了改善。文章结合2台锅炉的不同改造情况及实际运行状况进行比较，探讨如何增强锅炉燃烧稳定性，并提出相应措施，为此类型锅炉的改造及稳定运行提供一定的指导。     

基于方向梯度局部方差的分层遥感图像融合

为了充分利用各种遥感图像的信息．本文对以往图像融合技术进行了简要的概述，分析了各种方法的优缺点．给出了一种新的基于方向梯度局部方差的金字塔图像融合方法。将各源图像进行金字塔分解，得到多分辨率描述．建立方向梯度金字塔．采用局部方差准则融合源图像的各分解层．再进行金字塔分解的逆变换，得到融合图像。此外．利用信息熵等准则和其他融合方法进行比较．实验结果表明用此方法能得到更好的融合效果。     

基于多智能体系统的面向对象本体研究

智能体间的信息交互和行为协调是共同完成被委托任务的必要条件,论文提出了在多智能体系统中智能体本身必须建立领域模型的技术要求,即用本体支持运行时的语义交互。为此,文中用面向对象的知识表示方法描述并建立本体,并以此为基础形成领域操作代数系统和智能体服务描述语言。结合开放购买的仿真案例,表明在一个完整的情景语义交互中,服务提供方需要以智能体服务描述语言表述自己提供服务的方法和过程,而接受服务方必须在理解智能体服务描述语言的基础上,获取某一具体服务。     

Wear characteristics of commercial gas turbine components

Wear is one of the common degradations in a commercial gas turbine. A stainless steel grade 304 (SS 304) fuel nozzle and its collar made of nickel-based superalloy (Hastelloy X) are two vital components that normally suffer from rapid wear. During an installation, the fuel nozzle is slotted into the collar of the combustion liner. In operation, these two surfaces are subjected to high pressure from the fuel combustion, hence continuously rubbing against each other causes vibrations. During the start–stop operation, these surfaces had undergone a large relative motion. The vibration is the main cause of the wear occurrence on the surfaces. Physical properties of the worn surfaces were obtained through visual observations: wear measurement, hardness and microstructure examination. Through visual observations, fretting wear was mainly suspected as the dominant wear mode, particularly after 8,000 of running hours at high temperature and vibration. In short, this paper discusses the preliminary findings of wear on the fuel nozzle and its collar. It also discusses the changes in the mechanical properties before and after the operation. Solutions for mitigating the problem were discussed.     

Hierarchical steganography using novel optimum quantization technique

Security and imperceptibility are fundamental issues for steganography. This paper presents a multi-level steganography system based on concealing and deception. Basically, the secret image is embedded in a cover image, which in turn is embedded in another higher-level cover image. Each steganography level provides additional security, with lower levels implemented using stronger steganography methods. Furthermore, in the event that the higher level steganography is compromised, it would include misleading information to deceive the attacker that he/she is successfully eavesdropping. One challenge of hierarchical system is retaining hidden data in lower level in a higher level stenographic technique. To mitigate this issue, we designed a novel technique which combines discrete cosine transform and least significant bit (LSB). The technique objective is to maximize the capacity and invisibility of the secret image with minimal modification to the cover image (at most k-bits per block). The performance (capacity and peak signal-to-noise ratio) of the proposed method has superior performance compared with LSB since it combines the benefits of 1-bit and 3-bit LSBs. Furthermore, the paper demonstrates that the secret message is successfully embedded and extracted in two-level steganography system with peak signal-to-nose ratio greater than 20 dB.     

Static and dynamic compressive properties of polypropylene/zinc oxide nanocomposites

The effect of strain rate is widely recognized as an essential factor that influences the mechanical properties of polymer matrix composites. Despite its importance, no previous work has been reported on the high‐strain rate behavior of polypropylene/zinc oxide nanocomposites. Based on this, static and dynamic compression properties of polypropylene/zinc oxide nanocomposites, with particle contents of 1%, 3%, and 5% by weight, were successfully studied at different strain rates (i.e., 0.01 s^?1, 0.1 s^?1, 650 s^?1, 900 s^?1, and 1100 s^?1) using a universal testing machine and a split Hopkinson pressure bar apparatus. For standardization, approximately 24 nm of zinc oxide nanoparticles were embedded into polypropylene matrix for each of the tested polypropylene/zinc oxide nanocomposites. Results show that the yield strength, the ultimate strength, and the stiffness properties, of polypropylene/zinc oxide nanocomposites, were greatly affected by both particle loading and applied strain rate. Furthermore, the rate sensitivity and the absorbed energy of all tested specimens showed a positive increment with increasing strain rate, whereas the thermal activation volume showed a contrary trend. In addition, the fractographic analysis and particle dispersion of all composite specimens were successfully obtained using a field emisission scanning electron microscopy. POLYM. ENG. SCI., 54:949–960, 2014. © 2013 Society of Plastics Engineers     

Optimum mechanical behavior of calcium phosphate cement/hydroxyl group functionalized multi-walled carbon nanotubes/bovine serum albumin composite using metaheuristic algorithms

Injectable calcium phosphate cements have been introduced as adjuncts to internal fixation for treating selected fractures. These cements harden without producing much heat, develop compressive strength, and are remodeled slowly in vivo. The main purpose of the cement is to fill voids in metaphyseal bone, thereby reducing the need for bone graft. However, such cements may also improve the holding strength around metal devices in osteoporotic bone. This paper presents the optimum mechanical behavior of calcium phosphate cement/hydroxyl group functionalized multi-walled carbon nanotubes/bovine serum albumin (CPC/MWCNT-OH/BSA) composites in terms of compressive strength using well-known metaheuristic optimizers. The process parameters studied were wt% of MWCNT-OH (0.2–0.5 wt%) and wt% of BSA (5–15 wt%). The obtained results from metaheuristic algorithms were compared with the results from the response surface methodology (RSM) in the literature. The results obtained from metaheuristic algorithms outperformed the results given by the RSM in terms of less error percentage and high compressive strength.