基于SVM 的装备费用估算

针对传统建模方法无法解决“小样本、高维度、非线性”类大型复杂装备系统装备的费用估算问题,提 出基于支持向量机(support vector machine,SVM)的方法。针对SVM 方法固有的缺陷进行优化改进。结果表明,该 方法为装备全寿命费用估算做了有益探索。     

激光处理对油松种子活力指数的影响

用激光分别处理油松干、湿种子,干种子发芽率与活力指数分别为对照组的 １３６％ 和 １６７％ ,浸水 ３６ｈ后的湿种子发芽率与活力指数分别为对照组的 １３７％ 和 １５９％ 。     

Mie散射理论测量粒子系复折射率的透射方法

粒子的光学常数并不等同于构成粒子材料的光学常数,而通过粒子光谱复折射率数据,研究粒子及聚集粒子系的辐射特性,不仅对辐射物性的研究具有较高的理论意义,而且具有广泛的实际应用价值。粒子的复折射率不能直接通过实验测量(没有直接测量的仪器),须由实验测定其他量,然后结合相应的理论模型反求,属反问题研究。用简化的Mie散射理论及Kramers-Kronig关系式,利用傅里叶红外光谱仪对Al2O3粒子及煤灰粒子进行透射率实验测量,结合相应的理论模型,反演Al2O3粒子及煤灰粒子的复折射率。并对透射率实验误差对反演结果的影响进行了分析。     

拓扑集成指数与TIBO类衍生物构效关系的研究

4H-甲基咪唑苯二氮(?)酮(TIBO)类衍生物是抗爱滋病的一种新药,分子连接性指数是经证明应用广泛、较为成功的一种指数,本文定义并计算了TIBO类衍生物原子的特征值δi,利用量子化学计算方法,建构新的拓扑集成指数G和分子连接性指数mX,基于多元回归技术建立的对TIBO类衍生物药物的油水分配系数,作出精确估算和预测的定量结构-活性相关关系,得到的多元回归方程为:logP=0.782 G-0.1430X 0.2312X-3.829,估算的平均相对误差为2.53%。为了检验模型的稳定性和预测能力,做了留一法交互校验,预测平均相对误差为3.40%。该模型相关系数高,稳定性好,预测能力强。     

程序升温条件下煤炭自燃特性

通过程序升温实验研究了煤炭的自燃特性,测定了不同粒度煤样在不同温度下产生CO,CO₂等气体的浓度.讨论了指标气体的选择,分析了CO,CO₂等气体浓度随温度的变化规律,得到了耗氧速度与煤温及煤体粒度的关系.结果表明,煤体粒度越小,温度越高,则煤氧复合的强度越大.本实验的研究结果对煤矿火灾的防治具有指导作用.     

Optimal robot positioning using task-dependent and direction-selective performance indexes: General definitions and application to a parallel robot

Performance evaluation is an important issue for optimal robot positioning within workcells. Performance indexes can provide essential contributions to such an evaluation, in particular if they can account for the specific task with respect to which the optimization is carried out.     

火力发电厂性能指标一致化问题的思考和建议

针对火力发电厂纯凝汽式机组主要性能指标计算的影响因素,结合国内外不同标准对于机组工况的定义,提出了性能指标计算取值过程当中存在的工况不一致的问题,分析了问题存在的原因,提出了解决问题的建议。     

Determination of sieve grading curves using an optical device

The grading curve of an aggregate is a fundamental characteristic for mix design that can easily be modified to adjust several mix properties. While sieve analysis remains the reference method to determine this curve, optical devices are developing, allowing easier and faster assessment of aggregate grading. Unfortunately, optical grading results significantly differ from sieve grading curves. As a consequence, getting full acceptance of these new methods requires building bridges between the various particle size measurement systems. In order to build the sieve grading curve of an aggregate sample from video image analysis, a new model using images captured by a single camera device has been developed. This model has been tested on different samples stemming from four quarries with various petrography and flakiness index. For the 10/12.5, 12.5/16, 10/16 and 10/25 size fractions, the difference between actual sieve grading curve and optical grading curve is smaller than 7%. For the 6.3/10 fraction, the maximum difference between the two grading curves is equal to 15%. This difference could mainly be explained by parasitical rotation of small particles before image capture.     

Consideration of the BDNF gene in relation to two phenotypes: Hoarding and obesity.

The gene coding for the brain derived neurotrophic factor (BDNF) has emerged as an interesting candidate for multiple brain and brain disorder-related phenomena. The primary aim of the present investigation was to consider the relationship between the BDNF Val66Met variant and two phenotypes: compulsive hoarding as a symptom dimension of obsessive–compulsive disorder (OCD), and body mass index (BMI). We examined the BDNF gene in a large (N = 301) clinical sample of probands with OCD. Participants were classified as hoarding or nonhoarding using a strict, multimeasure grouping approach. Results revealed that the Val/Val genotype was linked with hoarding classification and more severe hoarding behaviors, as well as greater BMI levels. Hoarding status was also associated with greater BMI scores, with individuals in the hoarding group being far more likely to be classified as obese compared with the nonhoarding group. Our findings may provide a distinct avenue through which hoarding and BMI could be linked. These findings are suggestive of a complex gene, body weight, and psychopathology relationship wherein a primitive, survival “thrifty gene” strategy may be conserved and represented in a subgroup of humans manifesting severe hoarding symptoms. (PsycINFO Database Record (c) 2011 APA, all rights reserved)     

Quantification of human error in maintenance using graph theory and matrix approach

Assessment of human error in maintenance requires identification of the contributing factors that lead to human error(s). These factors are called human error inducing factors (HEIFs), which take into consideration both the active and latent error contributing aspects related to man, machine and environment. A systems approach of the Graph Theory is applied in this paper for quantifying human error in maintenance activities that models the identified factors and their interactions/interrelationships in terms of human error digraph. The nodes in the digraph represent the HEIFs and the edges represent their interrelationships. The digraph is converted into an equivalent matrix and an expression based on this is developed, which is characteristic of the human error in maintenance. This expression is used to evaluate a human error index by substituting the numerical value of the factors and their interrelations. The index is a measure of the human error potential involved in the maintenance of systems. A higher value of index indicates that the error likelihood is more for the associated tasks, and more efforts are required to make the system less prone to human error. The proposed methodology is illustrated using a case study. The approach is anticipated to play a significant role in identifying sources of human errors and predicting their impact; and will help to integrate human factors during design stage with the objective of reducing human error in maintenance. Copyright © 2011 John Wiley & Sons, Ltd.