Soft Sensor Model for Straw Fermentation Process Based on Least Squares Support Vector Machine Optimized by Chaos Fruit Fly Algorithm

It is difficult to directly measure the product concentration by using traditional physical sensors during the straw fermentation process, which makes the monitoring and real-time control impossible. To resolve this problem, the chaos fruit fly optimization algorithm (CFOA) is introduced to least square support vector machine (LSSVM) to optimize some key parameters, which overcomes some shortcomings of the cross validation method such as time consuming and blindness in parameter selection. Using this way, the CFOA-LSSVM soft sensor model is built for the straw fermentation process, which realizes the real-time measure of product concentration in this process. The simulation shows that the average measurement error of the proposed CFOA-LSSVM soft sensor is 4.55%, which is smaller than the traditional LSSVM model. The proposed CFOA-LSSVM soft sensor model has strong forecasting capability and high accuracy.