BP神经网络在毛织物风格评价中的应用研究

评述了人工神经网络和织物风格评价的特点，提出人工神经网络在毛织物风格评价中的工作原理和网络构建方法。基于反向传播神经网络模型(BP)建立了3层网络系统用于处理织物风格信号。经过学习训练，系统具有较强的自适应预测能力。     

低度浓香型酒机械冷冻除浊工艺的探讨

低度酒在生产销售中,遇低温易出现失光、混浊等现象,利用低度酒香味成分的低温析出特性,对半成品低度酒进行机械式冷冻处理。根据处理后半成品口感及理化指标变化对其进行调味。     

涤棉竹混纺仿毛织物风格评价

采用KES-F织物风格仪,测试了5种涤棉竹混纺仿毛织物的拉伸和剪切性能、弯曲性能、压缩性能、表面性能等16个力学指标,并以日本川端法计算其基本和综合手感值,考察其风格优劣程度。结果表明:实验用织物基本符合秋冬男士西服面料的要求,其中5#面料具有最佳风格且具有较高的仿毛感。     

棉织物涂料印花

涂料印花一直存在牢度、手感和鲜艳度等问题。通过试验，筛选了部分涂料印花色浆及配套助剂产品，并研究其在棉织物上的应用性能。生产结果表明，采用这些涂料色浆及配套助剂印制的产品手感好、牢度佳，游离甲醛含量极低，达到涂料印花产品在手感和牢度等方面的均衡性。     

新型简易单相五级逆变器

探讨了将中点箝位三级逆变器的一桥臂用于单相逆变桥,来作简易单相五级逆变器用,改善输出电压波形。其特点是元器件用量少、成本低、可靠性高。经Matlab/simulink仿真实验证明,设计方案是可行的。     

汽车JK把手注射模设计制造

分析了汽车JK把手塑件的结构及成型工艺,详细阐述了注射模结构及其主要成型零件的设计制造方法,介绍了注射成型工艺要点。实践证明,模具结构合理,运行平稳,塑件质量和生产效率达到了预期要求。     

一种改进的三电平空间电压矢量调制算法

提出了一种新的三电平空间电压矢量调制算法,该算法通过直接对三相参考电压进行比较从而判断出合成参考矢量所在区域,并基于伏秒平衡原理计算得到各个标准矢量的作用时间.通过与传统算法进行比较可以看出,提出的算法与传统算法控制效果相同,但其不需要进行正余弦计算,能够大量节省计算时间.最后通过实验验证了该算法的有效性.     

声学分析测定织物手感

与织物的视觉外观相同,织物的手感也是消费者购买纺织品的评价指标。影响织物手感的因素很多：纤维、纱线及其表面结构和后处理。测定纺织品特定手感性能的目的是用客观数值的方式表征其手感特点。介绍了一个可以定量描述织物手感特点和柔软性的测试系统,该测试系统是在声学分析的基础上进行测试的。     

Prediction of fabric handle value using ordinal regression model

Bed sheet fabric as a kind of home textile has been used since many years ago. Bed sheet is very significant because of being in direct contact with body consecutively for a long period of time. Bed sheet surplus qualitative parameters such as fiber substance, method of printing, finishing, etc., have a significant parameter called handle. In this paper, we proceeded to consider the relationship between fabric handle as a qualitative parameter and physical parameters which influenced the fabric handle using statistical modeling. The statistical model used was ordinal regression model. The modeling was done by SPSS V.19 software. We used 15 bed sheet fabrics. For subjective evaluation of 15 bed sheet fabrics, we selected 55 persons randomly as sample members according to Cochran’s formula. Population was selected from senior BS students and MS students at Isfahan University of Technology (IUT). We asked persons to classify bed sheet fabrics based on their preference of fabric handle from 1 (lowest) to 5 (highest). Physical parameters values were obtained through standard experiments. Finally, we analyzed obtained data through SPSS V.19 using ordinal regression model. Results showed a satisfying match between extracted data from the software and the real data from person’s evaluation.     

Design and analysis of an LC trap/LCR output filter for a single-phase NPC three-level inverter

This article presents a filter design approach and the efficiency analysis of a single-phase Neutral Point Clamped (NPC) three-level inverter with a combined LC trap and LCR filter system. The multilevel inverter system using NPC type inverter is suitable for high-power application. The single-phase NPC three-level inverter operates normally at low switching frequency. The proposed LC trap filter is comprised of a conventional LCR output filter. By adding an LC trap filter, the need for a low-value damping resistor and low LC cut-off frequency is eliminated. Use of the high-value damping resistor increases the efficiency of the output filter system. The controller used is DSP (TMS320C31). The effectiveness of the proposed system is verified by SPICE simulation and experimental results.