电压是电力系统的核心指标之一，是保障电网安全、优质与经济运行的关键。现代电网的电压主要依赖于自动电压控制（Automatic Voltage Control，AVC）系统的全局优化和闭环调节。随着新能源发电的大规模馈入，电网电压控制问题变得日益复杂，需要协调时间常数差异巨大且电气紧密耦合的众多特性各异的调节设备。本文针对自动电压控制中的多时间尺度协调问题，分别在三级电压控制（日前-小时级）、二级电压控制（分钟级）和一级电压控制（秒级）三个不同的时间尺度开展理论研究与工程实践，具体如下：在三级电压控制层面，由于大规模新能源集中馈入导致的电压快速波动，汇集区离散设备动作次数经常过早耗尽，其与连续调节设备的协调问题成为挑战。本文提出离散设备的日前鲁棒计划与日内动态调度方法，形成在日前确定动作时间区间、日内确定状态取值的两阶段控制模式。在日前阶段，提出鲁棒动态优化模型，兼顾了新能源不确定性和离散设备全日动作耦合，将复杂的离散设备动作次数和投切间隔约束转化为动作时间区间计划，降低了控制的保守性。在日内阶段，根据最新的预测信息进行滚动优化调度，利用日前确定的动作时间区间降低了问题规模，提高了控制的准确性和计算效率。在二级电压控制层面，电气耦合紧密但调节速度不同的连续设备之间存在无功竞争控制问题。本文构建了考虑下级控制器动态调节特性的二级电压控制模型，揭示了连续设备无功竞争的产生机理，并基于动态模型时间特性矩阵特征值推导了出现无功竞争的判据。提出了基于附加指令的动态二级电压控制方法，在控制过程中兼顾电压追踪和无功协调，有效消弭了无功竞争控制现象。在一级电压控制层面，新能源场站内部存在电压快速波动问题。本文提出了风电场自律电压控制方法，对风电场电气网络进行精细化建模，基于模型预测控制理论，将传统单时间点的反馈控制改变为未来时间窗内的动态轨迹最优控制，考虑了风机有功预测以及风机、SVG等各类设备的无功动态调节过程，在保证并网点电压动态效果的同时，实现了慢速风机无功的前瞻控制，保证了动态无功储备最大化。研发了风电场自律电压控制系统，在工程现场应用并取得了良好的控制效果。
Voltage is one of the core indicators of power system, and is key to the secure, qualified and econimical operation of the power grid. The voltage of modern power grid mainly depends on the global optimization and closed-loop control of the automatic voltage control (AVC) system. With the large-scale integration of renewable energy, it’s much more complex to regulate the grid voltage within expected ranges, so that the coordination between tight coupled reactive power resources with different time constants is quite necessary. Aiming at the multi-time-scale coordination problem in AVC system, this paper carries out theoretical research and engineering practice at three different time scales, which are tertiary voltage control (day-hour), secondary voltage control (minute) and primary voltage control (second). The main contributions of this paper are:With respect to tertiary voltage control, due to the rapid voltage fluctuation in the large-scale renewable energy integration area, the operation times of discrete devices is often exhausted prematurely, which exhibit extra burden on the continuous reactive power resources. In this paper, a two-stage control model is proposed to determine the operation time interval and the state value of the discrete devices. In the day-ahead stage, a robust dynamic optimization model is proposed, which considers both the uncertainty of renewable energy output and the times constraint of discrete device. In the intraday stage, according to the operation time interval determined at day-ahead stage, the rolling dispatch using the latest forecast information can reduce the scale of the optimization model and improve the accuracy and efficiency of the control.With respect to secondary voltage control, reactive power contention control exists between continuous devices with close electrical distance and different regulating speeds. This paper constructs a secondary voltage control model in which dynamic characteristic of lower-level controller is considerred, and reveals the mechanism of reactive power competition. According to the eingenvalue of the time characteristic matrix of the dynamic control model, the criterion of reactive power contention control is deduced. Based on this, a dynamic secondary voltage control method based on the additional command is proposed. Voltage tracking together with reactive power coordination is considered in the control process.With respect to primary voltage control, in order to eliminate the voltage fluctuation within a wind power station the traditional feedback control of single time point is substituted by the optimal control of dynamic trajectory in the future time window based on the model predictive control theory, which considers the active power prediction of wind turbine and reactive power regulation processes of devices such as wind turbine and SVG. The automatic voltage control system of wind farm is developed, applied in the practical engineering. Good field control performances are obtained.
