转换器
控制理论(社会学)
被动性
整流器(神经网络)
PID控制器
功率(物理)
控制器(灌溉)
计算机科学
仿射变换
数学
控制工程
工程类
人工神经网络
控制(管理)
物理
随机神经网络
生物
农学
电气工程
循环神经网络
温度控制
量子力学
人工智能
纯数学
机器学习
作者
M. Hernández-Gómez,Romeo S. Ortega,Françoise Lamnabhi‐Lagarrigue,G. Escobar
标识
DOI:10.1109/tcst.2009.2023669
摘要
Linear proportional-integral (PI) controllers are widely used in power converter applications. In a recent paper, a methodology to design such controllers ensuring asymptotic stability was proposed. The technique relied on the basic fact that if an affine system can be rendered passive with a constant control, then it is stabilizable with a PI. A structural condition was imposed then on the power converter to satisfy the former property with a passive output generated as alinearcombination of the states. This condition is technical and has no clear physical interpretation. In this brief this result is extended in three directions: first, the aforementioned condition is removed; second, a larger class of converters (with switching external sources) is considered; third, the load resistance is assumed unknown and an adaptive PI controller (with three different estimators) is proposed. Instrumental to establish the result is the proof that the nonlinear incremental model of power converters defines a passive map-a property first observed by Sanders and Verghese in the early 1990's. The methodology is applied to the problem of power factor compensation of a 3-phase voltage source rectifier, already considered in our previous paper, which is revisited from the incremental passivity perspective yielding simpler proofs. Also, a stable adaptive PI is designed for the output voltage regulation of a quadratic boost converter. Simulations complete the brief.
科研通智能强力驱动
Strongly Powered by AbleSci AI