微电网作为提高可再生能源利用率的主要手段，需要保证其可以稳定运行。其中，孤岛微电网没有主网支撑调节，为了维持正常运行，一般采用分层控制模式。分层控制模式中的一次控制通常采用下垂控制，但是传统的下垂控制属于有差调节，无法将孤岛微电网系统的运行频率及电压恢复至额定值，因此必须进行二次控制。一方面，某些工程应用对微电网运行稳定性要求很高，必须使二次控制目标在有限时间内完成；另一方面，分布式控制依赖通信，容易受到网络攻击的影响，所以需要设计合理的控制策略来抵御网络攻击。本文针对上述问题，设计了新的控制策略，使得孤岛微电网可以快速且安全稳定地完成二次控制。

首先，推导了分布式电源的数学模型，在此基础上，分析了逆变器的三个控制策略和微电网的三个控制模式。本文重点研究了其中的下垂控制和分层控制，使微电网系统通过下垂控制来完成一次控制，实现输出频率、电压的稳定和功率均分。介绍了三种稳定性理论，为之后证明所设计控制策略的稳定性打下理论基础。

其次，将多智能体理论的分布式一致性控制策略与微电网二次控制结合起来，将分布式电源看作智能体，再利用一致性算法初步完成二次控制。由于微电网对收敛速度有着较高要求，因此本文在传统有限时间算法的基础上进行改进，设计出自适应有限时间分布式二次控制策略，根据系统自身运行状态合理选取收敛系数，达到提高收敛速度的目的，并对所提控制策略进行了仿真验证。

然后，建立了二次控制器受到虚假数据注入攻击的模型，分析得到传统分布式二次控制无法抵御虚假数据注入攻击。对此本文引入滑模控制，通过设计合理的滑模面以及趋近律，提出了基于滑模自适应有限时间的分布式二次控制策略，使得二次控制在受到常值虚假数据注入攻击的情况下仍然可以在有限时间内实现控制目标，并通过仿真实验验证了所提方法的有效性。

最后，对不同种类的虚假数据注入攻击进行建模，包括常值、时变以及状态变量虚假数据注入攻击，并提出了可以抵御多种形式虚假数据注入攻击的分布式二次控制策略。通过所提出的控制策略，使系统可以在很短时间内抵御虚假数据注入攻击，将输出频率恢复到额定频率，同时有功功率也保持均分，极大地减小了虚假数据注入攻击对微电网系统产生的影响，并进行了仿真实验。

Microgrids as the main means to improve the utilization of renewable energy, need to ensure that they can operate stably. Among them, the islanded microgrid does not have the main grid to support the regulation, in order to maintain its normal operation, the hierarchical control mode is generally used. The primary control of hierarchical control is usually used droop control, but the traditional droop control is a differential regulation, cannot be restored to the rated value of the frequency and voltage of islanded microgrid system, and therefore must be carried out to restore the secondary control. On the one hand, certain engineering applications have high requirements for microgrid operation stability, it is necessary to make the secondary control objective complete in a limited time. On the other hand, distributed control relies on communication and is susceptible to cyber-attacks, so it is necessary to design reasonable control strategies to resist cyber-attacks. In this thesis, a new control strategy is designed to address the above problems, so that the isolated microgrid can complete the secondary control quickly, safely and stably.

Firstly, the mathematical model of distributed power is derived, based on which, three control strategies of inverter and three control modes of microgrid are analyzed. This thesis focuses on droop control and hierarchical control, so that the microgrid system can complete the primary control through the droop control to achieve the stabilization of the output frequency, voltage and the power equalization. Three stability theories are introduced to lay a theoretical foundation for later proving the stability of the designed controller.

Secondly, the distributed consistency control strategy of multi-agent theory is combined with the secondary control of microgrid, and each distributed power source is regarded as an agent, and then the consistency algorithm is utilized to complete the secondary control initially. Since the microgrid has high requirements on convergence speed, this thesis improves on the basis of the traditional finite time algorithm to design a self-adaptive finite time distributed secondary control strategy. The convergence coefficients are reasonably selected according to the system's own operating state to achieve the purpose of improving the convergence speed. And the proposed control strategy is simulated and verified.

Then, the model of secondary controller subjected to spurious data injection attack is established, and the analysis obtains that the traditional distributed secondary control cannot resist the spurious data injection attack. In this thesis, sliding mode control is introduced, and a distributed secondary control strategy based on sliding mode adaptive finite time is proposed by designing a reasonable sliding mode surface and a convergence law. So that the secondary control can still achieve the control objective in finite time under the attack of constant-valued false data injection. And the effectiveness of the proposed method is verified by simulation experiments.

Finally, different kinds of false data injection attacks are modeled, including constant-value, time-varying, and state-variable false data injection attacks. A distributed secondary control strategy that can resist multiple forms of false data injection attacks is proposed. The proposed control strategy enables the system to resist the false data injection attacks in a very short time and restore the output frequency to the rated frequency while the active power is kept equally divided. The impact of the false data injection attack on the microgrid system is greatly minimized. And simulation experiments are carried out.