Comparing the Prediction Capabilities of Artificial Neural Network (ANN) and Nonlinear Regression Models in Pet- Poy Yarn Characteristics and Optimization of Yarn Production Conditions

In the manufacture of yarn, predicting the effect of changing production conditions is vital to reducing defects in the end product. This study compares, for the first time, non-linear regression and artificial neural network (ANN) models in predicting 10 yarn properties shaped by the influence of winding speed, quenching air temperature and/or quenching air speed during production. A multilayer perceptron ANN model was created by training 81 patterns using the Broyden-Fletcher-Goldfarb-Shanno (BFGS) algorithm. The hyperbolic tangent, or TanH, activation function and logistic activation functions were used for the hidden and output layers respectively. Results showed that the ANN approach exhibited a greater prediction capability over the nonlinear regression method. ANN simultaneously predicted all of the 10 final properties of a yarn; tensile strength, tensile strain, draw force, crystallinity ratio, dye uptake based on the colour strengths (K/S), brightness, boiling shrinkage and yarn evenness, more accurately than the non-linear regression model (R=0.97 vs. R=0.92). These results lend support to the idea that the ANN analysis combined with optimization can be used successfully to prevent production defects by fine tuning the production environment. INTRODUCTION Faulty fabrics may arise during production due to changes in spinning properties, even within a single lot (Ziabicki, 1967). Faults which occur during the conversion of yarns into fabrics by weaving or knitting are difficult to detect until the fabric is colored, at which stage reversal is impossible. Such faults slow production and lead to significant costs. The properties of polyester (PET) yarn are influenced by many interwoven factors during production including winding speed, mass transport, polymer melting temperature during extrusion and the quenching condition (quenching air temperature and speed), as well as other production parameters during the melt spinning process (Kim, 1986; Simmens, 1955; Stibal et al., 2005; K. Yildirim, 2007; Ziabicki, 1967). The physical properties are also considerably affected by the cooling process, during which disoriented molecules form chains and settle in a defined pattern. Filament properties are affected by the cooling rate, the feed fineness of the filament and stress induced during spinning. The crystallization ratio also plays a critical role in yarn production during spinning, second in importance only to the cooling rate (Simmens, 1955). Crystalline size and morphology are also affected during the spinning process. Small changes in the conditions of yarn production can lead to large variation in the properties of the final yarn produced, which may result in an unacceptable product and consequently increased production costs (Kothari, 2000). Management of even small changes in the conditions during spinning is critical to obtain an acceptable product at minimal cost. During production, the yarn characteristics can be obtained by testing. In real time, this information can be used to predict and control the characteristics of the yarn to be produced. If necessary, required changes can be made during production and as a result, faults may be minimized or prevented