控制理论(社会学)
前馈
转换器
自适应控制
计算机科学
控制器(灌溉)
带宽(计算)
数字控制
查阅表格
电压
电子工程
工程类
电气工程
控制工程
控制(管理)
电信
人工智能
程序设计语言
农学
生物
作者
Davide Cittanti,Matteo Gregorio,Enrico Vico,Fabio Mandrile,Eric Armando,Radu Bojoi
标识
DOI:10.1109/tia.2022.3178394
摘要
The LLC resonant converter is typically adopted in battery charging applications due to its excellent performance in terms of efficiency, power density, and wide input/output voltage regulation. However, this converter is a complex high-order system characterized by a strong nonlinear behavior, featuring large variations of the small-signal gain/phase and pole location depending on the operating point. Consequently, these features pose substantial challenges in designing a closed-loop controller and providing constant dynamical performance over a wide operating range. Therefore, this article proposes a digital multiloop control strategy for LLC resonant converters ensuring constant closed-loop bandwidth and excellent disturbance rejection performance across the complete converter operating region. The control scheme consists of two cascaded voltage and current loops. To design and tune these controllers, a simplified LLC dual first-order small-signal model is proposed. The system nonlinear behavior affecting the current control loop is counteracted by a real-time controller gain adaptation process, which ensures constant control bandwidth. In particular, the adaptive gain values are provided by a static switching frequency lookup table obtained experimentally. Moreover, the steady-state switching frequency value is fed forward at the output of the current loop regulator, providing a further dynamical performance enhancement. The proposed control strategy and the controller design procedure are verified both in simulation and experimentally on a 15-kW LLC converter prototype. The results demonstrate the superior reference tracking and disturbance rejection performance of the proposed control strategy with respect to a state-of-the-art solution based on a proportional–integral regulator.
科研通智能强力驱动
Strongly Powered by AbleSci AI