摘要
发酵过程是一个复杂的时变、非线性、强耦合过程。发酵过程中的关键参量菌体浓度通常难以用传统物理传感器实时在线检测。为了测量该参数,将CPSO算法与LSSVM相结合构建发酵过程软测量模型。模型采用CPSO算法优化LSSVM软测量模型参数,克服了常规交叉验证法选取参数的耗时和盲目性。仿真结果表明,CPSO-LSSVM软测量模型较LSSVM软测量模型更能在较短的时间内获得较高的收敛精度,其平均误差为2.05%,说明该软测量模型可用于发酵过程不可在线测量的菌体浓度的实时在线软测量,并且预测精度高,预测速度快,预测能力强。该软测量建模方法也为发酵过程其他关键参量的实时在线测量提供了新的途径。
Fermentation process is a complex time-varying, nonlinear and strong coupling process. It is very difficult to measure the cell concentration on line using conventional physical sensors. A soft sensor model based on the combination of CPSO and LSSVM is established to estimate the parameters. To overcome the time consuming and blindness in parameter selection of traditional LSSVM model using cross validation method, CPSO algorithm is applied to optimize the parameters of LSSVM soft sensor model. Simulation results indicate that CPSO-LSSVM soft sensor model performs better than the LSSVM model in terms of measurement accuracy and calculation speed. The average meas- urement error is 2.05% , which indicates that the model could be used in measuring the cell concentration that could not be measured online during the course of fermentation with high precision, rapid forecasting speed and strong forecasting capability. The proposed soft sensor modeling method also provides a new approach for measuring other key parameters in fermentation process on line.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2011年第9期2066-2070,共5页
Chinese Journal of Scientific Instrument
基金
江苏省农业科技支撑项目(No.BE2010354)
江苏高校优势学科建设工程项目
江苏省高校自然科学基金(No.09KJD510001)
江苏省高校优势学科建设工程资助项目(苏政办发(2011)6号)资助项目
江苏大学高级专业人才科研启动基金(No.10JDG086)
