Real-Time Implementation of Laguerre Neural Network-Based Adaptive Control of DC-DC Converter

Abstract

Applications of power electronic converters have increased invariably in fields of engineering such as robotics, e-mobility and smart grids. DC-DC converters are employed as a switching devices to obtain a required amount of DC voltage in various industrial applications. Under the class of non-isolated DC-DC power converters, the buck converters are of specific interest, as they provide lower DC output voltage than the source DC voltage. In order to obtain a faithful output voltage tracking despite disturbances affecting the system, the converter is connected in the closed feedback loop. In this respect, this paper presents the design, development and experimental findings of Laguerre neural network driven adaptive control of DC-DC buck power converter. The stability of the proposed controller is established through Lyapunov stability criterion. Further, the results are compared with adaptive backstepping control method, by subjecting the converter to start-up test, step changes in the load resistance, input voltage and reference voltage tests. Thereafter, the performance is evaluated on DSP-based dSPACE 1104 processor in the laboratory. Finally, the results are compared in terms of settling time of output voltage state. The results indicate an enhanced dynamic performance of both output voltage and inductor current with the action of proposed controller, thus making it suitable for fast practical applications.