湍流燃烧二阶矩模型的大涡模拟验证

对美国Sandia国家实验室测量的甲烷-空气湍流射流火焰进行了三维大涡模拟(LES),其统计的时均量以及脉动量符合实验测量值,且和RANS代数二阶矩湍流燃烧模型模拟结果大致符合。LES统计得到的化学反应率系数脉动和甲烷浓度脉动的互关联与对应的时均量梯度的乘积分布趋势相同,大小幅值的位置也相对应,与二阶矩湍流燃烧模型的封闭假设相符,验证了二阶矩代数燃烧模型的合理性。     

Development of models for prediction of the antioxidant activity of derivatives of natural compounds

Antioxidants are important for maintaining the appropriate balance between oxidizing and reducing species in the body and thus preventing oxidative stress. Many natural compounds are being screened for their possible antioxidant activity. It was found that a mushroom pigment Norbadione A, which is a pulvinic acid derivative, shows an antioxidant activity; the same was found for other pulvinic acid derivatives and structurally related coumarines. Based on the results of in vitro studies performed on these compounds as a part of this study quantitative structure–activity relationship (QSAR) predictive models were constructed using multiple linear regression, counter-propagation artificial neural networks and support vector regression (SVR). The models have been developed in accordance with current QSAR guidelines, including the assessment of the models applicability domains. A new approach for the graphical evaluation of the applicability domain for SVR models is suggested. The developed models show sufficient predictive abilities for the screening of virtual libraries for new potential antioxidants.     

混合优化RBF-BP网络的板形缺陷识别

针对传统板形模式识别方法存在精度低、鲁棒性弱的问题,提出了一种混合优化RBF-BP组合神经网络板形模式识别方法。首先利用自组织映射网络(SOM)对样本聚类,利用聚类后的网络拓扑结构确定RBF的中心,并计算RBF的宽度,克服了传统聚类算法随机选取中心导致聚类结果不稳定的问题。然后利用遗传算法(GA)良好的全局搜索能力优化整个网络的权值。RBF-BP组合神经网络是由一个RBF子网和一BP子网串联构成的,该网络同时具备BP神经网络能较好地预测未知样本的能力以及RBF神经网络的逼近速度快的优点。并以某900HC可逆冷轧机板形识别为应用背景,在MATLAB2010a环境下进行仿真实验,结果表明混合优化RBF-BP组合神经网络的板形模式识别方法能够识别出常见的板形缺陷,提高了板形缺陷识别精度并具有较好的鲁棒性,可以满足板带轧机高精度的板形控制要求。     

基于多维状态参量和时间序列对电机的异常检测

针对传统对电机的异常检测常常出现误判或滞后的情况,采用基于时间序列对电机单一状态参量用AR拟合,同时利用SOM神经网络无监督的方式量化电机数据。然后,利用得到的量化序列结合AR曲线得到序列的转移概率,及早发现某种状态参量的异常变化;之后,DBSCAN算法挖掘多维参数之间特征关系来确定电机是否出现异常。最后结合实例说明该方法的检测过程,并对比验证了该方法的优越性。     

一种基于遗传优化的BP神经网络的测光红移估计算法

除了星系的光谱红移之外,星系测光红移的估计也对研究宇宙大尺度结构及演变有着重要的研究意义。利用斯隆巡天项目最新发布的SDSS DR13的150 000个星系的测光及光谱数据,在红移值Z<0.8范围内,先使用SOM自组织神经网络对星系样本进行早型星系和晚型星系的聚类,然后用遗传算法优化后的BP神经网络对星系的测光红移进行估算。估算结果与作为标准的已知星系光谱红移进行比对,早型星系的红移估计最小均方误差约为0.001 3,晚型星系最小均方误差约为0.001 7。实验结果表明,遗传优化的BP算法在精度上优于BP神经网络算法,且效率上优于K近邻、核回归等传统测光红移估计算法。     

基于SOM&SVM组合分类器的客户细分方法实证研究

为解决传统的RFM客户细分方法还不能很好地刻画客户行为,同时也没有就RFM指标权重进行分析这一问题,在RFM指标的基础上扩充了客户细分的指标体系,并提出了基于AHP的RFM指标权重确定策略.鉴于传统的单一分类器存在的很多缺陷,提出基于SOM&SVM的组合分类器模型,充分利用SOM和SVM单一分类器各自的优点,综合两种分类器的分类信息,避免单一分类器可能存在的片面性,从而提高分类的准确性.最后通过实例对上述模型的有效性进行验证.     

基于粗糙集和SOM网络的客户市场细分方法

市场细分对提高企业核心竞争力具有重要作用。针对目前一些市场细分方法的不足,提出了基于粗糙集和SOM网络的市场细分方法。该方法首先对客户消费数据库进行离散化处理和属性约简,然后采用SOM网络进行自动聚类,形成稳定的客户群,客户群的数量可通过设定SOM网络竞争层神经元个数     

水稻土可见-近红外-中红外光谱特性与有机质预测研究

土壤有机质是农田肥力评估的重要指标,要实现快速获取大面积土壤有机质的含量需要建立高效、稳健的预测模型。光谱技术能够快速诊断土壤有机质,以水稻土为例,从校正样本选择方法的对比,研究了可见-近红外、中红外和可见-近红外-中红外三种不同波段光谱对土壤有机质的预测能力。可见-近红外和中红外区域的光谱反射率转换成吸收率后通过Savitzky-Golay平滑法去噪,通过三种校正样本选择方法建立相应的偏最小二乘回归预测模型。通过Rank-KS法建立的三种波段的有机质预测模型均优于Rank法和KS法,中红外波段光谱的模型预测能力强于可见-近红外和可见-近红外-中红外波段的预测模型,基于Rank-KS法建立的中红外波段有机质预测模型取得了最好的预测效果,RMSE_P仅为3.25 g·kg-1,RPD达到4.24,依据VIP得分筛选出可见-近红外和中红外波段的水稻土有机质重要建模波段。因此,中红外光谱建模技术能够对水稻土有机质进行快速定量分析,Rank-KS法可提高模型的预测能力,为今后农田肥力评价和科学施肥提供技术支持。     

Target training with soft computing tools

The personal capabilities and intentions of employees indicate their performance within their organization. It is important for the organization to capture this kind of tacit knowledge since the workforce are the true experts in perceiving the organization's current reality and evaluating which assets require development – including themselves as knowledge assets. The collective inner voice of the workforce helps the organization's management to steer the company and its assets in a sustainable direction.This article presents how the collective inner voice of the workforce can be captured and how it can be used for the benefit of the organization and its employees. The objective is to support individuals’ personal aspirations, as well as to save the money, time and resources that an organization spends on personnel training.The focus of this article is on demonstrating a possible soft-computing method used for competency simulation. The process starts with a linguistic self-evaluation conducted by employees, where individuals’ own perception of current and target competence levels is captured. The self-evaluation is conducted with the help of fuzzy logic. Clusters are formed from the result dataset using an unsupervised neural network clustering method: self-organizing maps. A demonstrator tool is then used to perform a “what-if” type of analysis/simulation on the clusters in the results. With the demonstrator tool, employees can roughly test the impact of alternative training scenarios for themselves. For individuals this may open up new directions for self-development, and for organizations this may allow the efficient use of training resources. We tested the approach with a dataset from a real human resource development project among nuclear power plant operators.The case study reveals the potential of soft-computing based collective competency simulation as one part of personnel development projects in the future. Yet the techniques and the demonstrator tool used in this experiment are far from being products that employees could easily use as part of their training project. Possible benefits of the proposed approach are demonstrated in this article.