基于人工神经网络的混合气体定量检测方法

将气体传感器阵列与人工神经网络相结合，设计了用于混合气体定量检测的人工嗅觉系统；构造了局部连接的前馈神经网络的结构和基于递推预报误差的网络训练算法；通过实验对CO，H2，CH4三种气体的混合物进行了分析，实验结果证明气体传感器阵列和人工神经网络用于混合气体定量检测是可行的。     

曙光油田超稠油井套管损坏的机理和防治

超稠油特殊的油藏特点决定了其生产周期短、注汽频繁、出砂、汽窜、超覆现象严重，以及套管质量等因素的影响，套管损坏较为严重。对曙光油田超稠油井套管损坏的规律和机理进行了研究，并从钻井和完井工艺的完善、套管钢级和射孔工艺的选择、采油生产参数的确定、注汽管柱隔热效果的提高、防治砂工艺等方面，对套管损坏提出了预防措施和对策。并对套管损坏井制定了修复再利用方案。     

The Locus of Redundant-Targets and Nontargets Effects: Evidence From the Psychological Refractory Period Paradigm.

In target detection tasks, responses are faster when displays have 2 targets (redundant-targets effect; RTE) and slower when they have no targets (nontargets effect; NTE) relative to displays with a single target. The psychological refractory period paradigm was used to localize these effects. In Experiment 1, participants classified tones as high or low and then classified letters as targets or nontargets after a short or long stimulus onset asynchrony (SOA). The magnitudes of the RTE and NTE did not depend on SOA. In Experiment 2, the order of the tasks was reversed, and at short SOAs the RTE and NTE had similar magnitudes for both tone discrimination and target detection responses. These findings suggest that the RTE and NTE arise during response selection. Interactive effects of tone pitch with the number and type of target features were also observed, and these were tentatively interpreted as synesthetic effects. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

压力计探测液面法探讨

针对油田地质特征复杂，易结蜡，测试中易出现假液面等情况，为准确快速测得液面资料，提出压力计测试液面法。该方法可作为回声仪测液面方法的有效补充。     

伪彩色数字高程模型在工业CT断层图像检查中的应用

在工业CT断层图像缺陷检测应用中，提出了将数字高程模型DEM与伪彩色技术一起应用于二维灰度图像缺陷检测领域的算法流程。应用该算法流程，能够充分利用CT断层灰度图像信息，获得清晰的视觉检查效果，从而大大增强了系统检测缺陷的能力。在大型工业CT图像重建与检查分系统中采用该算法流程已经取得了较为理想的效果。     

Modelling of fatigue damage progression and life of CFRP laminates

A progressive fatigue damage model has been developed for predicting damage accumulation and life of carbon fibre‐reinforced plastics (CFRP) laminates with arbitrary geometry and stacking sequence subjected to constant amplitude cyclic loading. The model comprises the components of stress analysis, fatigue failure analysis and fatigue material property degradation. Stress analysis of the composite laminate was performed by creating a three‐dimensional finite element model in the ANSYS FE code. Fatigue failure analysis was performed by using a set of Hashin‐type failure criteria and the Ye‐delamination criterion. Two types of material property degradations on the basis of element stiffness and strength were applied: a sudden degradation because of sudden failure detected by the fatigue failure criteria and a gradual degradation because of the nature of cyclic loading, which is driven by the increased number of cycles. The gradual degradation of the composite material was modelled by using functions relating the residual stiffness and residual strength of the laminate to the number of cycles. All model components have been programmed in the ANSYS FE code in order to create a user‐friendly macro‐routine. The model has been applied in two different quasi‐isotropic CFRP laminates subjected to tension–compression (T–C) fatigue and the predictions of fatigue life and damage accumulation as a function of the number of cycles were compared with experimental data available in the literature. A very good agreement was obtained.     

Two-stage fatigue loading of woven carbon fibre reinforced laminates

A brief review of the models used to predict the cumulative fatigue damage in FRP composites is presented. Two‐stage fatigue loading of a [0/90,± 45₂,0/90]_s quasi‐ isotropic woven carbon fibre/epoxy resin laminate was evaluated at stress ratio R = 0.05 and the failure mechanisms investigated using x‐radiography after each loading stage. The results are presented in terms of fatigue strength and damage growth and are compared with those in the literature. A low‐to‐high loading sequence is more damaging than a high‐to‐low one and the Palmgren‐Miner linear damage rule may no longer be valid for this kind of material, as previously reported.     

利用非线性自适应技术检测微弱的瞬时信号

本文根据局部流形理论,提出了一类非线性自适应预测滤波模型。仿真结果显示,本文提出的自适应预测滤波模型具有很强的非线性近似能力,可以检测淹没在超混沌或杂波环境中的瞬时信号。     

疲劳/蠕变复合作用下聚苯乙烯的交互损伤研究

探讨了在疲劳／蠕变复合作用下聚苯乙烯的损伤交互作用，结果表明，在疲劳／蠕变复合作用下聚苯乙烯存在疲劳和蠕变的交互损伤，其断裂寿命比纯疲劳或纯蠕变的断裂寿命低；断裂机制是疲劳循环载荷松动和活化了分子链或链段，从而促进蠕变运动和断裂，并且，疲劳／蠕变的交互损伤程度与温度密切相关。     

双谱估计在导引头信号检测与处理中的应用

双谱估计是一种应用广泛的高阶谱估计。将双谱估计应用于超宽带导引头信号检测与处理系统中,为导引头信号处理中的平滑和滤波增添了一种新的方法。仿真结果表明该方法是有效的。