Research On The Fabric Moisture Content Of Nondestructive Measurement System Based On Artificial Neural Network

In the Textile Printing and dyeing industry, the textile fabric properties are closely relatedto the various physical parameters. It is very important to detect the moisture content, which isan important parameter of the printing and dyeing processing production process. Themoisture content will directly affect the structure performance and dielectric constantparameters of the printing and dyeing fabric. At the same time, the accurate detection of themoisture content can save printing paste, and improve the utilization rate of water cycle, andreduce the use amount of heating steam, etc. This is to save costs and reduce energyconsumption. In the national "the twelfth five-year plan", the textile industry is regarded as animportant direction of the manufacturing industry. Accelerate the high-end textile equipmentdemocracy and recycling the textile production raw material have caused the emphaticattention in the industry. Therefore, the designing of the textile printing and dyeing parameterdetection device with independent intellectual property rights has the great significance toimprove the automation process of the textile equipment.The traditional fabric moisture content detection method which used drying is very lowin efficiency, and is more used for the theoretical analysis. This thesis adopts the microwavetransmission method to detect fabric moisture content. Compared with the frequency method,the infrared method, and the resistance method, this method has a high accuracy, and lowinfluence on the fabric surface structure and color and the high anti-interference ability. Dueto the non-contact measurement, the influence of the analysis is almost zero. Through themeasured fabric, the energy of the microwave would change with the variation of the moisturecontent in the fabric. Measuring the variation of the microwave energy could contain themoisture content of the fabric. This is the theoretical basis of the microwave transmissionmethod to measure the moisture content of the fabric.Based on the theoretical basis of the microwave transmission method, this thesis designsa set of non-contact fabric moisture content measurement system. Firstly, this thesisintroduces the principles and methods of the microwave moisture technology. By analysingthe relationship between the microwave attenuation and the measuring voltage, the mathematical model of the measuring fabric moisture content is established.Secondly, this thesis designs a non-contact fabric moisture content measurementplatform, and respectively introduces the used microwave moisture sensor, data acquisitionsystem, the design of in-out cloth system, and the components and modules device. Throughsimulating the industrial printing and dyeing processing production process, the final fabricmoisture content measurement platform is set up to collect the experiment data.Finally, the experiment has respectively tested to three kinds data of the fine cotton,cotton fabric and denim in different moisture content condition. In the beginning, this thesisuses the regression fitting principle for the data analysis, and the error is so large. Then theartificial neural network method is used to deal with the data. At the same time, comparedwith the traditional drying method, the experimental results is ideal. The experimental resultsshow that the effective moisture content measurement range of fine cotton is35-75%, and themeasurement error is±1%; The effective moisture content measurement range of Jean is30-60%, and the measurement error is±0.2; The effective moisture content measurementrange of Cotton and linen is20-50%, and the measurement error is±1.5%. Based on theanalysis of the experimental error, the microwave transmission method for measuring fabricmoisture content has the very high measuring accuracy and feasibility which has the verygood guidance function for the localization development of the printing and dyeing parameterdetection equipment.