Soft Sensing And Predictive Control Of Straw Fermentation Process For Ethanol Production

The suitable utilization of straw is beneficial to reduce the harmful gases and particles produced by straw burning,which plays an important role in controlling haze,protecting the environment and improving the comprehensive utilization of resources.With the expansion of the scale of straw fermentation engineering,the requirements of process control are also continuously improved.The internal mechanism of straw fermentation process is complex,which is affected by many environmental parameters,showing the characteristics of nonlinear,time delay,strong coupling and so on.Especially,some key variables that affect the fermentation efficiency and product yield cannot be detected online due to the limitation of hardware detecting instrument,which restricts the optimal control of straw fermentation and seriously affects the automation and industrialization of straw fermentation.Based on the technology of straw fermentation,the problems are studied that the important process variables on the fermentation(such as substrate concentration,fermentation concentration,ethanol concentration)are hard to realize the real-time detection and closed-loop optimal control.The soft sensor model of the important process variables is constructed.A generalized predictive control(GPC)method for feeding is proposed based on least squares support vector machine(LSSVM).The digital control system which can realize the real-time monitoring and control is developed based on the embedded Cortex A9 control unit and Visual C++ development platform.The main work is as follows:Firstly,in order to solve the problem that the important process variable(ethanol concentration)of straw fermentation is difficult to be detected in real time,a soft sensor modeling method based on LSSVM is proposed.Because the conventional method of selecting important parameters of LSSVM model is time-consuming and lack of direction,a method based on fruit fly optimization algorithm and chaos algorithm(CFOA)is proposed to select the important parameters.The experiment and simulation results show that the selected parameters by using CFOA have higher efficiency,faster speed and higher iteration precision than the conventional ways.Furthermore,the LSSVM soft sensor model based on the optimization parameters has shorter training time,higher prediction accuracy and better generalization ability.Secondly,a nonlinear predictive control model was constructed based on the CFOA-LSSVM model for the characteristics of multi-variable,time variant and nonlinear.Meanwhile,in order to avoid the complex calculation and the bad affection for real-time control when the nonlinear model is used directly,the linear prediction model is constructed by the linear operation in the sampling point.Based on this,a generalized predictive control method is proposed to control the important process variables of straw fed batch fermentation.The simulation results show that the prediction model has better control quality,higher measurement accuracy and stability.Thirdly,to reduce the disadvantages of a large number of operations and high complexity caused by the traditional Diophantine equation,artificial fish swarm algorithm(AFSA)is applied as a strategy of rolling optimization to realize the optimal control.Furthermore,the step length and range of vision in AFSA are optimized and the generalized predictive controller is designed.The experiment and simulation results show that with AFSA as the optimization strategy,GPC can optimize the control quality and has the characteristics of small overshooting and good stability,so as to optimize the production efficiency and increase yield of the product.Finally,based on embedded Cortex A9 control unit and Visual C++ platform,the control system of straw fermentation for ethanol is developed.And the soft sensor and the closed loop optimization control are transplanted into the system.The monitoring interface shows good performance during the operation of the system,which can realize the function of displaying,estimating and remote monitoring of the fermentation process.