基于耦合分析的热流系统协同设计

Co-design of thermal-fluid system with coupling analysis

中文关键词  热流系统  设计优化  整体设计  耦合分析

英文关键词  thermal-fluid system  design optimization  integrated design  coupling analysis

中文摘要

      作为综合能源系统的一部分, 热流系统通常在稳态假设下设计其物理系统, 在此基础上设计控制器. 这种 方法忽略了系统与控制器设计间的耦合关系, 导致系统暂态性能不佳. 为此, 本文提出一种协同设计方法, 以系统总 质量、熵产率以及控制效果为设计度量, 基于热流系统数学模型及其熵产率模型, 结合嵌套设计方法和同步设计方 法完成系统设计. 考虑到设计过程中设计参数多、变量范围大的问题, 引入时间序列模式距离, 构造相关度度量函 数来分析物理系统与控制器之间的耦合关系, 从而减少非耦合参数. 同时, 利用数据分析方法进一步缩小了待设计 参数范围. 为验证方法的有效性, 以一个理想热流系统为设计实例, 仿真结果表明, 通过相关度度量函数及其数据分 析, 可以将设计参数个数减少66%, 选取范围缩小26%. 相比于传统顺序设计方法, 经过协同设计方法设计的系统控 制误差可减少29%, 且设计控制器抗扰能力也优于频域方法和SIMC方法.

英文摘要

      As an indispensable component of the integrated energy system, the thermal-fluid systems are typically designed by using steady-state assumptions, and a design of controller is followed. However, this method ignores the coupling relationship between system and controller, resulting in poor transient performance of system. In this study, a co-design method involving the design metrics of total mass, entropy production rate and control effect is proposed. The system is designed by combining nested design method and simultaneous design method, along with the mathematical model of thermal-fluid system and entropy production rate model. The time series mode distance is introduced to construct correlation measure function to analyze the coupling relationship between physical system and controller, which reduces the number of non-coupling parameters. Meanwhile, the data analysis method is used to further decrease the range of design parameters. A design example of an ideal thermal-fluid system is given, and the simulation results show that the number of design parameters is reduced by 66% and the selection range is reduced by 26% measured by correlation measurement function and data analysis, respectively. In addition, the system designed by co-design method is superior to that by the traditional sequential design method, with a decrease of 29% in control error. The disturbance rejection ability of the designed controller is also better than that of the frequency domain method and the SIMC method.