Research And Development Of Rice Wine Fermentation Process Monitoring System Based On WSNs And BP Neural Network

As a unique brewing wine in China,rice wine is gradually becoming popular with the concept of yellow wine health care,and the demand for rice wine will continue to grow.Under the background of the country's vigorous promotion of China's manufacturing 2025,the science and technology of various industries have developed rapidly,and the automation technology of the yellow wine industry,whether in technical research or technical application,is seriously lagging behind liquor,beer,wine,etc.Beverage industry.At present,most of the rice wine enterprises use the traditional rice wine brewing technology.Although the unique cultural heritage of rice wine is preserved to a certain extent,the quality stability of rice wine is poor,and the number of skilled workers is limited,and the output of yellow wine cannot be effectively improved.Therefore,under the premise of guaranteeing the quality of yellow wine,real-time wireless monitoring and accurate prediction of key influencing factors of key operations of rice wine should be carried out,and the informationization degree of yellow wine enterprises should be gradually improved,thereby improving production efficiency.In this paper,through the analysis of key processes and the analysis of influencing factors,the fermentation temperature and alcohol concentration of the rice wine fermentation process are finally studied.The main research contents include:(1)Based on Zigbee technology,develop Zigbee terminal equipment,routing equipment and coordination equipment using Arduino microcontroller,DS18B20 temperature sensor,MQ-3 alcohol concentration sensor,Zigbee module(CC2530+CC2591),and build wireless sensor network by using on-demand method(WSNs).In order to accurately calculate the alcohol concentration value,an alcohol concentration calculation model was established based on the interpolation method.By comparing the SSE,R-square,and RMSE of the exponential approximation,polynomial approximation,and power approximation interpolation methods,the constructive function between the alcohol concentration and the voltage was finally determined.(2)In order to improve the informationization degree of rice wine fermentation process,the rice wine fermentation temperature monitoring software was developed based on Python language,PyQt platform and SQLite embedded database.The software includes: order management,employee management,fermentation tank management,user management,workshop Kanban display,fermenter detail display,temperature prediction,and SMS notifications.The software can control the Zigbee coordination device through the serial port,send commands to the specified IP address Zigbee terminal device,and receive the temperature and alcohol concentration data returned by the Zigbee terminal device.(3)In order to realize the feedforward control of rice wine,it is necessary to predict the temperature value of the next stage of rice wine in advance.Therefore,based on BP neural network,a three-layer perceptron temperature prediction model of rice wine is established,and the temperature values of the previous five moments are used as nerves.The input of the network predicts the temperature value at the next moment.By comparing the mean square error(MSE)changes,it is verified that the BP neural network has good reliability in the prediction of yellow wine pre-fermentation temperature.The rice wine fermentation process monitoring system developed in this paper consists of hardware and software modules.The WSNs sensor network based on Zigbee communication module forms the hardware platform of the system.The software module is developed based on Python language and PyQt5.Taking key processes as monitoring targets,real-time temperature monitoring,temperature prediction,temperature warning and alcohol concentration monitoring of the whole process of rice wine fermentation process can be realized to maximize the informationization degree of the yellow wine industry,and the system is low in cost and stable in transmission.It is easy to maintain and has strong applicability.It is not only limited to the parameter monitoring in the production process of rice wine,but also suitable for most wireless monitoring fields.It has a good application prospect and is expected to be industrialized.