关键词: 纱线回潮率, 纱线线密度, 环境相对湿度, 覆盖系数, 数学模型

Abstract: The sizing moisture regain rate is one of the three major indicators of sizing performance. It is the ratio of the weight of moisture in the sized yarn to the dry weight of the sized yarn, expressed as a percentage, reflecting the drying level of the sized yarn. The drying level of sized yarn is not only related to the energy consumption of sizing, but also affects the properties (elasticity, softness, strength, re-viscosity, etc.) of the sizing film, thus affecting the mechanical properties of the sized yarn. To realize the accurate control of sizing moisture regain, modern sizing machines are mostly equipped with a resistance type moisture regain tester to monitor sizing moisture regain online. However, current research mainly focuses on the relationship between yarn raw materials and sizing moisture regain. Other key factors in actual sizing production, such as yarn linear density, coverage coefficient, environmental temperature and humidity, size type and sizing rate, are rarely involved, resulting in some errors in the actual measurement of sizing moisture regain. 
To improve the testing accuracy of moisture regain, on the basis of the existing resistance method, the effects of yarn linear density, coverage coefficient and environmental relative humidity on moisture regain were studied. First of all, a set of sheet yarn moisture regain testing device was established in this paper based on the operation simulation of the sizing machine to realize the convenient adjustment of experimental parameters and the accurate collection of test data. Based on the relationship between yarn linear density, coverage coefficient, environmental relative humidity, test moisture regain and real moisture regain, a mechanism model was established, a deviation correction model between moisture regain and various influencing factors was established, and an objective function was established. The deviation correction model was used to improve the detection accuracy of sizing moisture regain. The levels of the three influencing factors were set at 4, 4, and 10, respectively, and a comprehensive experiment was carried out on all the combinations of the three parameters. The collected data were fitted by using the least squares method with the particle swarm optimization algorithm, and the model fitting coefficient was obtained. The goodness-of-fit R² value of the model was 0.8827, and the mean squared error (MSE) was 0.0273, which showed that the mathematical model established in this paper had a good ability to express the experimental data.
To validate the aforementioned mathematical model, another batch of samples were prepared and subjected to testing with the same four kinds of yarn linear density, 10 kinds of coverage coefficient and four kinds of environmental relative humidity. The different yarn linear density t, coverage coefficient f, environmental relative humidity h and test moisture regain value W were input into the model to obtain the corrected test moisture regain W' after correction. The average absolute error (MAE) between the uncorrected test moisture regain W and the true moisture regain W1 was 0.3086, while the MAE between the true moisture regain W1 and the corrected test moisture regain W' using the model constructed in this paper was 0.2321, and the error was reduced by 24.8%. Therefore, it can be concluded that the correction effect of the mathematical model constructed in this paper is satisfactory, and the testing accuracy of moisture regain is effectively improved.

Key words: yarn moisture regain rate, yarn linear density, environmental relative humidity, coverage coefficient, mathematical model