A framework for environmental production of textile dyeing process using novel exhaustion-rate meter and multi-layer perceptron-based prediction model

Abstract

In textile industries, a lot of wastewater are discharged which are one of the major environmental pollution problems, because they release undesirable dye effluents. Owing to re-dyeing procedures performed to meet customized color specifications, environmental pollution is a serious problem because of the emission of large volumes of wastewater. To solve the environmental problems caused by re-dyeing, the right-first-time (RFT) %, which is the rate at which the target quality is obtained with just one dyeing, must be increased by considering the dyeing conditions that affect product quality. Here, this study suggests a framework for cleaner production of textile dyeing process using novel exhaustion-rate meter (NERM) and multi-layer perceptron-based prediction model to solve the environmental problems caused by re-dyeing procedure by controlling the exhaustion-rate outliers. The proposed NERM measures the exhaustion-rate based on absorbance of the dyeing solution and is composed of measuring and analysis section. The dyeing solution absorbance is metered in the measuring component through a detector, which performs high-resolution measurement (0.3–1.5 nm full width at half maximum) via a 25-μm slit in the 200–1100-nm wavelength range; the absorbance is then converted to the exhaustion-rate based on Beer's law in the analysis section. Using the NERM, an exhaustion rate dataset according to the Na2SO4 and Na2CO3 consumption is acquired and a surrogate model that augments the exhaustion rate data is developed. The MLP-based prediction model is then developed using the augmented data to control the real-time exhaustion-rate outliers. As a results, the model performance as regards Na2SO4 and Na2CO3 prediction is indicated by R2 values of approximately 0.985 and 0.998, respectively, and root mean squared errors (RMSE) of approximately 1.477 and 1.000, respectively. In addition, the effectiveness of the proposed framework is demonstrated through application to several scenarios in which the real-time exhaustion rate outliers are detected.