能源互联网重要基础支撑:分布式储能技术的探索与实践  

The cornerstone of energy internet:research and practice of distributed energy storage technology

作　　者：慈松[1,2] 李宏佳[1] 陈鑫[1] 王强文 

机构地区：[1]中国科学院声学研究所高性能网络实验室 [2]University of Nebraska-Lincoln, NE 68182,USA [3]Nova GreenTech Inc.,NE 68130,USA

出　　处：《中国科学：信息科学》2014年第6期762-773,共12页Scientia Sinica(Informationis)

基　　金：国家自然科学基金(批准号:60972083;61302108);美国国家自然科学基金(批准号:0801736;0830493)资助项目

摘　　要：可再生能源分布式发电的能量波动性以及用户驱动的能量需求的时空随机性,导致能源互联网中能量流本身具有先天的不确定性与无秩序性,因此,分布式储能技术由于可以有效消除能量流的不确定性,并使能量的时空转移和能量流的有序流动成为可能,成为能源互联网重要基础支撑.分布式电池储能作为一种重要的储能方式,电池单体本身非线性特性与电池成组或成网后单体间的差异性,使得电池储能系统"管理好"成为真正"使用好"的基础.但是,电池成组或成网后将构成复杂巨系统,这对"管理好"带来巨大的挑战.因此,本文基于作者多年理论与技术的成果积累,总结了一套适用于分布式储能的大规模电池网络优化管理的理论与方法,包括:单体和电池组或电池网络的建模与状态参数精确估算;基于模糊测度的电池网络特征提取及快速计算;基于自适应动态规划的电池网络优化管理.随后,本文将基于能源互联网思想的分布式储能架构及其管理优化方法与技术引入数据中心,并给出了其在数据中心中实际应用原型系统介绍.本文工作可以为能源互联网的重要基础支撑—分布式储能技术—的研究与应用提供方法指导与技术支撑.Due to the distribution and the fluctuation of energy generation, and random energy demands, energy flows in energy internet have inherent uncertainties and randomness. Therefore, distributed energy storage system is the fundamental component of energy internet, which can effectively smooth uncertainty of energy flow, and support spatial-temporal reallocation of originally distributed and fluctuant energy. Battery, as one of the most important energy storage carrier, has many non-linear physical effects, e.g., self-recovery effect, rate capacity effect, and individual physical differences among different cells. Therefore, effective battery network management determines whether distributed battery storage system could work well. However, from viewpoint of battery management, battery network, which contains many inter-dependent and non-linearly featured elements, is a complex system, which brings great challenges to battery network management. Therefore, based on our previous technical achievements, we propose a set of methods and techniques on battery network management： 1） modeling and state parameters estimation of battery network; 2） fuzzy measure based features extraction and fast calculation of battery network; 3） adaptive dynamic programming based battery network optimization. Finally, inspired by the core idea of energy internet, we introduce the distributed energy storage architecture, and its optimal management methods to the energy supply and backup system of data centers, and describe our developed prototype system. Our work could provide theoretical and technical guidelines to applications of energy internet and distributed energy storage system.

关 键 词：能源互联网 分布式储能 复杂系统 电池管理 数据中心 

分 类 号：TM910[电气工程—电力电子与电力传动]