Abstract: The current dilemma in evaluating sizing quality lies in the different definitions of sizing yarn quality indicators, the lack of correlation research among indicators, and the limitations of existing evaluation methods. This paper tried to use equal sizing rate technology to detect and control the sizing rate of sizing yarn online. The goal is to achieve a scientific evaluation method that comprehensively reflects the quality of sizing yarns, and to conduct horizontal comparisons between different types of sizing agents.
To achieve this goal, a modified Maxwell Garnett (MG) equivalent dielectric constant model was established for the yarn structure, and a capacitive detection system was used to detect the sizing rate of the sizing yarn online. Three types of sizing agents including starch, poly vinyl alcohol (PVA), and modified sophora bean powder were selected to size cotton yarns on a sizing machine. A capacitive sizing rate detection system was set up to ensure every sizing sample have equal sizing rates, by adjusting the sizing process. The entropy weight Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) model was introduced into the comprehensive performance evaluation of sizing yarns by sorting the degree of closeness. By normalizing and standardizing the initial matrix X, a standard matrix of yarn performance was obtained. The initial matrix consists of six sets of samples, with each set containing eight performance indicators.
Results of the experiments shows that the relative error of sizing rates between the capacitive detection system method and the desizing method is less than 10%, indicating that the capacitive sizing rate detection system can meet the needs of practical applications. Secondly, under the same conditions of solid content slurry and sizing process, the sizing rate of fine yarns is higher than that of a roving yarns. The information entropy for the hairiness reduction rate is minimum, indicating that this performance indicator is the most important in the overall quality evaluation of sizing yarns. The information entropy for the enhancement rate is -0.86, indicating that it is relatively less important. In the experiment, when the same sizing agent is used, the comprehensive quality of the sizing yarn improves with the increase in yarn fineness, while the sizing rate has a relatively small impact on it. When different types of sizing agents with the same sizing process are used, it is found that the quality of sizing yarn with PVA sizing agents is better than that of sophora bean powder and starch. However, the modified sophora bean powder sizing agent is a green one that can be used as a substitute for PVA and starch. This comprehensive evaluation method based on equal sizing rate technology and entropy weight TOPSIS theory provides a new approach for comprehensive quality evaluation and the development of sizing agents.

Key words: equal sizing rate technology, capacitive sizing rate detection system, modified sophora bean sizing, entropy weight TOPSIS model, Maxwell Garnett model

摘要： 为了构建一个能够综合评估浆纱质量的评价体系，选用淀粉、聚乙烯醇（PVA)和改性槐豆粉3种浆料，使用电容式上浆率检测系统来实现等上浆率上浆，在单纱上浆机上对棉纱上浆；将熵权TOPSIS模型引入浆纱综合质量评价中，用熵权法计算各评价指标的权重，利用TOPSIS方法计算各方案与最优理想点之间的贴近度，并据此对各个方案进行排序。结果表明：采用细度为38.6 tex的棉纱、含固率为3%的PVA、上浆速度为100 m/min、上浆率为4.9%的上浆方案时，浆纱综合质量最优；使用同种浆料、不同细度的棉纱上浆时，粗支的浆纱质量更优；等上浆率条件下，采用PVA制成的浆纱质量优于槐豆粉浆料和淀粉浆料。等上浆率上浆技术和熵权TOPSIS模型为浆纱综合质量评价、新型浆料开发开辟了新途径。

关键词: 等上浆率上浆, 电容式上浆率检测系统, 改性槐豆粉浆料, 熵权TOPSIS模型, Maxwell Garnett模型