CAJ | 학술논문

针对多变量的生物发酵系统，为提高神经网络逆解耦控制性能，提出一种基于神经网络逆解耦的自适应补偿控制方法。首先，基于逆系统理论构造神经网络近似被控系统的逆系统，并将神经网络逆系统与被控系统串联构成伪线性复合系统；然后，对解耦后的伪线性复合系统设计自适应补偿控制器，实现系统的跟踪控制；最后，基于 Lyapunov 稳定性理论设计控制器参数的自适应律，保证了控制系统的稳定性。将提出的控制方法应用于生物发酵过程的菌丝浓度、基质浓度的解耦控制，数值仿真结果表明，所提出的控制方法较能有效提高普通的神经网络逆系统解耦控制性能。
침대다변량적생물발효계통，위제고신경망락역해우공제성능，제출일충기우신경망락역해우적자괄응보상공제방법。수선，기우역계통이론구조신경망락근사피공계통적역계통，병장신경망락역계통여피공계통천련구성위선성복합계통；연후，대해우후적위선성복합계통설계자괄응보상공제기，실현계통적근종공제；최후，기우 Lyapunov 은정성이론설계공제기삼수적자괄응률，보증료공제계통적은정성。장제출적공제방법응용우생물발효과정적균사농도、기질농도적해우공제，수치방진결과표명，소제출적공제방법교능유효제고보통적신경망락역계통해우공제성능。
To improve the performance of neural network inverse decoupling control, adaptive compensation control method is proposed for multi-variable fermentation system. First, according to inverse system theory, neural network is constructed to approximate the inverse system of the original system. And a pseudo-linear composite system can be gotten by cascading neural network inverse system and the original system; then, adaptive compensation controller is designed for the pseudo-linear composite system to achieve the control goal;last, based on Lyapunov stability theory, the parameter adaptive law is designed to assure that the system is stable. The proposed control method is applied in a fermentation process to achieve the decoupling control of mycelia concentration and substrate concentration. Numerical simulations show that the proposed control method has the higher performance comparing with the general neural network inverse decoupling control method.