动力锂电池最优等效电路模型研究

可靠的等效电路模型能够精确的模拟动力电池的充放电特性,是建立各种状态估计方法的基础。首先建立了锂电池Thevenin模型、DP模型,并采用带遗忘因子最小二乘法进行在线参数辨识,利用电路原理进行离线参数辨识。其次,在MATLAB/Simulink中结合实验数据对模型辨识精度和运算速度进行仿真验证。最后,基于辨识精度和运算速度建立模型评价方法,得出在线辨识过程DP模型相比于Thevenin模型能够在精度和速度方面取得更好的平衡;离线辨识过程,Thevenin模型能够取得更好的平衡。     

数据驱动的变频空调负荷模型参数在线辨识方法

准确辨识空调负荷模型的参数是挖掘其节能及需求响应潜力的重要基础,当前研究大多采用精度较差的离线辨识方法。为此,基于数据驱动思想,提出一种变频空调模型参数在线辨识方法。首先,建立了数据驱动的空调负荷模型参数在线辨识架构。然后,基于空调负荷模型提出数据驱动的在线辨识机制和方法。其中,数据驱动的在线辨识机制设计为基于参数显著变化事件驱动的参数更新判别机制和基于历史参数波动范围的参数动态阈值设定机制,在该机制下通过粒子群优化算法建立了快速在线辨识方法。最后,通过实测环境,验证了所提在线辨识方法的有效性,与离线辨识方法相比,所提方法极大地提高了计算速度及准确度,可满足在线应用需要。     

最小二乘算法优化及其在锂离子电池参数辨识中的应用

传统最小二乘法（LS）用于锂离子电池模型在线参数辨识精度低,通过带遗忘因子递推最小二乘算法能够有效地提高辨识精度,但固定的遗忘因子影响模型动态特性。遗忘因子的自适应处理能提高算法对动态系统的参数辨识能力,而目前的自适应方法容易忽略模型参数的稳定性,同时方法待定系数范围较大且难以确认。为了得到高精度且稳定性良好的模型参数,该文设计了一种精度和稳定性兼优且更简单的自适应遗忘因子递推最小二乘（AFFRLS）改进方法,并与其他AFFRLS、可变遗忘因子递推最小二乘（VFFRLS）进行仿真对比分析。结果表明,改进的AFFRLS能够在模型精度和参数稳定性取得更好的平衡,且对不同的在线工况具有良好的适用性。     

基于电动汽车锂电池的SOC动态估算策略

针对实际运行中电池参数的变化,建立了基于Thevenin模型的锂离子动力电池状态空间模型,采用递推最小二乘法进行模型参数在线辨识,对参数做出实时修正,同时克服广义卡尔曼滤波(EKF)估算的不足,提出了基于无色卡尔曼滤波(UKF)估算锂电池SOC估算的新方法。实验结果验证了在同等条件下,UKF比EKF具有更好的滤波估算精度,提高了系统的适应性。     

锂离子电池模型极化误差的修正方法

以锂电池为研究对象,采用电流脉冲,对Thevenin模型参数进行辨识,考察了不同工况下该模型的仿真精度。发现特定工况下,模型极化误差较大。采用充放电实验数据对模型参数进行修正,并建立新的状态方程。改进后的Thevenin模型在恒流和变电流工况下均能达到较高仿真精度,电压误差维持在0.016 V以内。     

锂电池动态系统Thevenin模型研究

建立高精度的电池模型对于电动汽车动力锂电池的应用研究有重大意义。锂电池在使用过程中,其系统参数会跟随外界环境及荷电状态变化而改变,选用固定参数的电池模型会导致模型精度差。提出一种动态系统Thevenin模型。结合影响锂电池特性的荷电状态和环境温度因素,将经典Thevenin模型中的欧姆内阻、极化内阻、极化电容等固定参数,在动态系统Thevenin模型中描述为随荷电状态与温度动态变化的变量。最后选取单体锂电池为实验对象,采用HPPC实验辨识模型参数,对经典模型与动态系统模型分别进行仿真分析,结果表明,动态系统Thevenin模型能更准确描述锂电池性能。     

基于模糊建模技术的PEMFC非线性控制

针对PEMFC等一类具有严重非线性的复杂被控对象,提出一种基于模糊模型的非线性预测控制算法对PEM-FC系统进行建模与控制。在建模与控制过程中,采用离线学习和在线修正辨识出对象的模糊模型;其中,模型的参数通过模糊聚类初始化和离线反向传播算法进行学习,必要时,可通过在线调整后件参数,使模型的预测精度能满足实时控制的需要。线性辨识方法在线建立PEMFC控制系统的T-S模糊预测模型,然后基于分支定界法的基本原理对控制量进行离散寻优,从而实现PEMFC的非线性预测控制。仿真和实验结果证明了该算法的有效性和优越性。     

基于天牛须搜索算法的电池模型参数辨识

为实现高精度的动力电池SOC估计,需要建立准确的电池模型.针对磷酸铁锂电池平台特性,采用二阶RC等效电路建立电池的等效电路模型;用双指数函数拟合OCV与SOC的函数关系;采用双线性变换法将电池的数学模型离散化,得到可辨识的电池数学模型.为保证电池模型的精度,要对电池模型的参数进行辨识.首次提出一种基于天牛须搜索算法的电池模型参数辨识方法.用该方法对建立的电池模型进行辨识后,用Simulink对模型进行了脉冲放电与恒流放电仿真测试,并与离线测量数据进行对比,结果表明模型的最大动态误差不超过30 mV,说明该辨识方法有较高的辨识精度,建立的模型能够准确模拟磷酸铁锂电池的动静态特性.     

基于MRAS的改进永磁同步电机参数辨识方法

针对传统参数辨识方法收敛速度较慢的问题,提出了一种基于MRAS的改进参数辨识方法。该方法在MRAS参数辨识方法中引入了离线辨识技术,将离线辨识所得结果作为在线辨识算法的初值,提高了辨识结果的收敛速度。建立了基于永磁同步电机矢量控制系统的参数在线辨识模型,对所提方法进行了仿真验证,结果表明该方法可有效提高辨识结果的收敛速度。     

模型参数优化的卡尔曼滤波SOC估计

荷电状态(SOC)表征电动汽车动力电池剩余电量。以精确在线估计电动汽车动力电池SOC为目的,基于Thevenin等效电路模型,结合电路原理对模型参数进行辨识,对辨识后的模型参数进行合理优化。利用基于优化后模型参数的扩展卡尔曼滤波对SOC进行在线估计,对比基于优化前模型参数的扩展卡尔曼滤波SOC在线估计结果,仿真结果表明,利用优化后的模型参数,在不影响扩展卡尔曼滤波初值误差收敛性的前提下,估算误差比优化前减小了43%左右,具有更高的SOC在线估计精度。     

网络分析应用中的公用信息模型

从能量管理系统应用的一个重要组成部分——网络分析应用的角度出发,介绍了网络分析应用的数据需求,讨论了公用信息模型(commoninformationmodel,CIM)对上述数据需求的支持,并进一步明确了3个问题:CIM模型库对于网络分析应用提供了哪些信息模型;这些模型应该如何使用;模型库在网络分析应用方面存在哪些不足。围绕上述问题介绍了5类模型(包括量测模型、保护模型、控制模型、拓扑模型和设备模型)及其在网络分析中的应用,展望了CIM在网络分析中的应用前景。     

Analysis of nonequilibrium phenomena on electric characteristics of advanced MOSFETs by using a quantum energy transport model

In this paper, we have investigated nonequilibrium effects for advanced MOSFETs by using a device simulator with quantum energy transport (QET) model. The QET model allows to simulate nonequilibrium carrier transport as well as quantum confinement. The QET model includes the mobility model as a function of carrier temperature in order to consider the nonlocal effects. We have simulated advanced MOSFETs down to 20 nm gate length using the QET model. The QET model is compared with the quantum drift diffusion (QDD) model which includes a mobility model with local assumptions. It is found that the nonlocal mobility model is needed to simulate the advanced MOSFETs with less than 40 nm.     

Research on virtual actor action editing and movement control

To directly use a virtual surface model for action editing and movement control, a general method for creating virtual actor skeleton models and controlling movement is presented. The method includes judging borderlines of the block virtual surface model, calculating the joints, confirming the above block, and using the block hierarchical layout to create the skeleton model. Then, according to the virtual actor model and movement restriction, the study focuses on the generation of movement animation using the key frame technique and smoothing movement technique by automatically adding animation and adjusting the actor’s pose by different weights on movement amplitude. Finally, movement control of the actor in the virtual environment is implemented by real-time control and path point control, which achieve a good result. __________ Translated from Journal of Southeast University (Natural Science Edition), 2006, 36(5): 720–724 [译自: 东南大学学报 (自然科学版)]     

半干法脱硫塔内SO2脱除过程数值模拟

建立了2段含湿颗粒干燥模型,将该干燥模型与分段SO2传质模型及气固双流体模型耦合,编制了求解程序。应用该程序通过UDF接口,对FLUENT计算软件进行二次开发。模拟结果表明:沿塔体高度方向SO2浓度逐渐降低;脱硫反应主要发生在喷口后1.5 m距离内;从喷口开始塔内脱硫产物浓度沿高度方向逐渐升高,并呈现中心浓度低、边壁浓度高的趋势。     

Small-signal Modeling and Modal Analysis of Wind Turbine Based on Three-mass Shaft Model

Abstract—One critical task in wind turbine shaft torsional vibration study involves the modeling of a wind turbine and power grid. With a focus on the mechanical rotational system of wind turbine, this article provides a three-mass shaft model upon which one wind turbine to infinite bus model can be developed. The model, based on small-signal stability analysis, is used to study the wind turbine shaft torsional vibration. For this reason, this article concentrates on the union model of stall wind turbine and power grid. The small-signal stability model includes the mechanical system and electrical system. Data for the turbine's blades and shaft as well as electrical generator are given to allow replication of the model in its entirety. Each of the component blocks of the wind turbine and power grid is modeled separately so that one can easily expand and modify the model to simulate variable-speed wind turbines or multi-turbines system to suit their needs. This is then followed by a case study that explains how the small-signal stability model can be used to study wind turbine shaft torsional vibration issues. Results obtained from the case study show that the model performs as expected.     

Accurate and efficient RTL power estimation based on power contour model

This paper proposes a new RTL power estimation method based on a power contour model. The model achieves accurate and efficient power estimation. The experimental results show that the maximum error is reduced by 56.19% and the RMS error is also reduced by 3.41% compared to the conventional table‐based RTL power estimation. © 2012 Wiley Periodicals, Inc. Electr Eng Jpn, 182(3): 48–56, 2013; Published online in Wiley Online Library ( wileyonlinelibrary.com ). DOI 10.1002/eej.22340     

电力系统自组织临界特性分析与仿真模型

电力系统自组织临界性的研究尚处于初级阶段,其概念并不十分清晰,需要进一步研究,其还没有公认的可以作为研究电力系统自组织临界性的标准模型出现。针对这种情况,首先重新对比了沙堆模型与电力系统,对电力系统的自组织临界现象给出了新的分析和解释;建立了一个用于仿真电力系统自组织临界性的模型;最后利用这个模型对我国的一个实际电网进行了仿真计算,得到了表征电力系统自组织临界性的幂律特性曲线,同时分析了决定电网自组织临界性的参数。     

Electronic model of a Ferroelectric Field Effect transistor

Abstract

A pair of electronic models has been developed of a Ferroelectric Field Effect transistor. These models can be used in standard electrical circuit simulation programs to simulate the main characteristics of the FFET. The models use the Schmitt trigger circuit as a basis for their design. One model uses bipolar junction transistors and one uses MOSFET's. Each model has the main characteristics of the FFET, which are the current hysterisis with different gate voltages and decay of the drain current when the gate voltage is off. The drain current from each model has similar values to an actual FFET that was measured experimentally. The input and output resistance in the models are also similar to that of the FFET. The models are valid for all frequencies below RF levels. Each model can be used to design circuits using FFET's with standard electrical simulation packages. These circuits can be used in designing non-volatile memory circuits and logic circuits and are compatible with all SPICE based circuit analysis programs. The models consist of only standard electrical components, such as BJT's, MOSFET's, diodes, resistors, and capacitors. Each model is compared to the experimental data measured from an actual FFET.