电工技术学报  2023, Vol. 38 Issue (12): 3264-3276    DOI: 10.19595/j.cnki.1000-6753.tces.220691
电力系统与综合能源 |
基于模型预测控制的直流微电网虚拟惯性优化方法
赵书强1, 王慧1, 田娜2, 孟建辉1, 王琛1, 田艳军1
1.华北电力大学分布式储能与微网河北省重点实验室 保定 071003;
2.国网河北省电力有限公司经济技术研究院 石家庄 050024
Model Predictive Control Based DC Microgrid Virtual Inertia Optimal Method
Zhao Shuqiang1, Wang Hui1, Tian Na2, Meng Jianhui1, Wang Chen1, Tian Yanjun1
1. Key Laboratory of Distributed Energy Storage and Microgrid of Hebei Province North China Electric Power University Baoding 071003 China;
2. Economic Research Institute of State Grid Hebei Electric Power Company Shijiazhuang 050024 China
全文: PDF (3303 KB)   HTML
输出: BibTeX | EndNote (RIS)      
摘要 为了提高直流电压的稳定性、动态特性以及安全性,提出一种基于预测模型控制的直流微电网虚拟惯性滚动优化方法。首先,建立含虚拟惯性控制单元的直流微电网线性离散模型,用以预测系统输出量的未来趋势。其次,设计模型预测控制器:以直流电压跟踪误差、虚拟惯性系数两者的加权二次方和最小为目标,希望直流电压维持稳定并且惯性响应的强度不要太大;以直流电压及其变化率限制为约束,防止因动态电压超过安全阈值而导致切机、切负荷甚至系统崩溃问题;在每个采样时刻,求解各控制单元的虚拟惯性系数并将其输出,从而实现虚拟惯性的优化控制。同时,通过控制性能分析,给出控制器主要参数的选取原则。最后,通过负荷突变、风速随机波动等典型工况下的硬件在环测试,验证了所提方法的可行性和有效性。
服务
把本文推荐给朋友
加入我的书架
加入引用管理器
E-mail Alert
RSS
作者相关文章
赵书强
王慧
田娜
孟建辉
王琛
田艳军
关键词 直流微电网虚拟惯性优化控制线性离散模型模型预测控制    
Abstract:To improve the stability and dynamic characteristics of microgrid DC bus voltage, researchers have proposed adaptive virtual inertial control methods. These methods have clear analytical formulas for virtual inertia, but providing the design details for the “optimal” parameters is challenging. Moreover, the operation state determines the virtual inertia, and the controller acts in a delayed manner with the system variations. As a result, emergency preventive control cannot be realized due to dynamic voltage constraints. Therefore, a model predictive controller has been proposed to realize the virtual inertia optimization of the DC microgrid by dynamically adjusting the units’ virtual inertia coefficients.
Firstly, the linear discrete model of a DC microgrid with virtual inertial control units was established to predict the future trend of system output. Secondly, the model predictive controller was designed to minimize the weighted sum of DC voltage tracking and virtual inertia, maintain the DC voltage stability, and make the intensity of the inertial response acceptable. Taking the DC voltage and its climbing rate as the constraints, preventing the excess safety threshold for the dynamic voltage is also crucial in case of machine cut-off, load cut-off, or even system collapse. At each sampling period, the individual virtual inertia coefficient is solved and updated to realize the optimization of virtual inertia. Finally, the selection principles of the main parameters are given through theoretical analysis.
Hardware-in-the-loop simulation experiments have been built to verify the proposed method. Compared with the adaptive virtual inertial control, the proposed method provides a stronger inertial support for the system, increases rapidly in the initial stage of the sudden load increase, and reduces the drop of DC bus voltage. In reverse recovery, the virtual inertia coefficients were reduced, so the DC voltage recovery is accelerated in the beginning, and the recovery speed is reduced in the following time to eliminate the overshoot. In addition, the test result shows that the proposed method can maintain a steady state and its climbing rate within the desired range. Finally, the model mismatch scenarios have also been tested. The result shows that, under the line model mismatch, the performance of DC bus voltage and virtual inertia coefficients do not degrade substantially by compensating for the optimized rolling, thus guaranteeing the controller’s adaptive performance.
Conclusions can be drawn: (1) The proposed method can improve the stability, dynamic characteristics, and safe operation of microgrid DC bus voltage. Aiming at minimizing DC voltage tracking error and virtual inertia coefficients, the proposed control has advantages in improving DC bus voltage stability, accelerating recovery speed, and reducing overshoot. It can also prevent the voltage and its climbing rate out of the thresholds. (2) The proposed method can reduce the optimization size problem, which is beneficial for optimizing the computational burden of the MPC controller. The prediction model can be effectively simplified and well controlled by simplifying the DC microgrid line model and converter control loops, thus reducing the optimization scale and contributing to accelerated calculation speed. (3) The proposed method provides certain tolerance to line model mismatch. MPC uses a rolling optimization mechanism to update the optimization problem with updated measured values, which mitigates the model uncertainty caused by the mismatch.
Key wordsDC microgrid    virtual inertia optimal control    linear discrete model    model prediction control   
收稿日期: 2022-04-27     
PACS: TM712  
基金资助:国家自然科学基金资助项目(51807064)
通讯作者: 王慧 男,1982年生,博士研究生,高级工程师,研究方向为新能源电力系统稳定控制与经济调度。E-mail:wanghui@ncepu.edu.cn   
作者简介: 赵书强 男,1964年生,教授,博士生导师,研究方向为电力系统分析与控制。E-mail: zsqdl@163.com
引用本文:   
赵书强, 王慧, 田娜, 孟建辉, 王琛, 田艳军. 基于模型预测控制的直流微电网虚拟惯性优化方法[J]. 电工技术学报, 2023, 38(12): 3264-3276. Zhao Shuqiang, Wang Hui, Tian Na, Meng Jianhui, Wang Chen, Tian Yanjun. Model Predictive Control Based DC Microgrid Virtual Inertia Optimal Method. Transactions of China Electrotechnical Society, 2023, 38(12): 3264-3276.
链接本文:  
https://dgjsxb.ces-transaction.com/CN/10.19595/j.cnki.1000-6753.tces.220691          https://dgjsxb.ces-transaction.com/CN/Y2023/V38/I12/3264