Study On Multi-model Soft Measurement Method And Monitoring System Of Straw Fermentation Process

At present,the straw resources have reached 170 million tons in China,and the major crop straw like corn stalks,rice stalks,wheat stalks have reached 759 to 994 billion tons.Although there are many ways to deal with straw like returning to the field,feed,cultivation of edible mushrooms,biogas,etc,70% resources were still burned directly as fuel.This method not only destroyed the soil structure and effected soil moisture evaporation,but also resulted in waste of resources and the greenhouse effect.If using 10% of straw resources in producing fuel ethanol,we will 8 million tons of ethanol and save 3 million tons of grain.Therefore,the production of fuel ethanol from straw resources,while solving the waste of resources,protecting the environment and brings income to farmers.With the development of process in producing fuel ethanol from straw fermentation and the gradual expansion of production and scale,the control process has also been transformed into automation and intelligence.It is realized on-line monitoring of some parameters that show the quality of the fermentation process in the fermentation process.The premise and basis of automatic control.Due to the complex internal mechanism of fermentation,economic and technical reasons make it difficult to measure some parameters.At present,the traditional method is to remove the product from the fermentation process and carry out off-line testing.This method has a large measurement delay and error,which seriously affects the observation and control of the fermentation process in the later period.Based on the above problems,this project designs a "multi-model soft-sensing" method and matched it with an embedded-based monitoring system.Through the soft-sensing method,the key parameters(ethanol concentration,residual sugar content and bacterial activity)in the fermentation process are monitored online in real time.Firstly,the paper introduces the purpose and prospect of straw fermentation fuel ethanol,and pointed out the problems in the control of fermentation process,which lays a foundation for the research content of the study.Secondly,the theories of the soft-sensing algorithm and the system platform software and hardware design adopted in this paper are introduced in detail(Kernel-based fuzzy c-means algorithm,Least squares vector machine,Adaptive mutation different evolution).The soft-sensing theory includes the use of a multi-model soft-sensing modeling method for problems where a singleregression model does not adapt to new operating conditions,which covers data classification theory,modeling theory,and model optimization theory.The monitoring system design includes the system data acquisition module with the STM32F103ZET6 minimum system as the core to achieve the task of data collection and transmission from the fermentation site to the main control module.ARM11 is used as the core control module of the main processor that has completed the functions of data analysis,soft-sensing processing and display.Finally,the developed soft-sensing platform is applied to the actual process of straw fermentation,and the effectiveness of the platform is proved by the simulation and soft-sensing platform on the PC.The multi-model soft sensor monitoring system of straw fermentation studied in this paper has the function of real-time monitoring the key parameters of straw fermentation and displaying it through the display screen.It provides some reference for the monitoring and real-time display of the difficult measurement in the industrial production process in the future.