考虑分布式电源出力的波动性、负荷预测的误差以及各类调度资源的运行特性在时间尺度上存在差异等问题,提出一种在需求侧响应机制下主动配电网的多时间尺度优化调度模型。首先,负荷参与需求侧响应的形式分为价格型和激励型两类,分别对这两种需求侧响应的形式进行建模分析。其次,建立基于模型预测控制的主动配电网多时间尺度调度框架,根据模型预测控制方法分别建立日前调度、日内滚动和实时反馈三个阶段的配电网优化调度模型。考虑到配电网调度过程中的有功无功耦合特性,通过控制无功调度资源使得配电网各节点电压控制在允许范围。最后,通过对改进的31节点算例进行仿真,验证了所提模型可以有效地降低预测产生的误差并提高主动配电网运行的经济性和安全性。

Considering the fluctuation of distributed generation output, the error of load forecasting and the difference of operation characteristics of various scheduling resources in time scale, in this paper, a multi-time scale optimal dispatching model of active distribution network under demand-side response mechanism is proposed. Firstly, the forms of load participating in demand-side response are divided into price type and incentive type, and the two forms of demand-side response are modeled and analyzed separately. Secondly, a multi-time scale scheduling framework based on model predictive control for active distribution network is established. On the basis of the model predictive control method, three optimal dispatching models of daily distribution, intraday rolling and real-time feedback are established separately. Considering the coupling characteristics of the active and reactive powers in the process of distribution network dispatching, the voltage of each node of distribution network is controlled within the allowable range by controlling reactive power dispatching resources. Finally, through the simulation of the improved 31-node example, it is verified that the proposed model can effectively reduce the prediction error and improve the economy and safety of active distribution network operation.

激励型需求响应是一种通过补贴信号灵活调度需求侧能源的手段，对于维持电力系统供需平衡，提升经济效益有巨大潜力。在面向居民用户的激励型需求响应中，电力运营商选择削减电量潜力大的居民用户并向其提供补贴价格，以激励其削减电量。然而，站在电力运营商的角度，面对居民用户未知且不确定的用电行为，识别并选择削减电量潜力大的居民用户以及如何动态制定补贴价格是关键挑战。为了解决这一问题，本文提出基于在线学习的激励型需求响应算法(IDR-OL)，利用多臂赌博机框架在线学习居民用户削减电量潜力，建立电力运营商运营成本优化模型选择最优参与需求响应的居民用户并动态制定补贴价格。仿真结果表明，本文提出的IDR-OL算法能够在实现供需平衡的同时更大限度地降低电力运营商运营成本。

Incentive demand response is a means of flexibly dispatching demand-side energy through subsidy signals, which has great potential for maintaining the balance of supply and demand in the power system and improving economic benefits. In the incentive demand response for residential users, power operators choose residential users with high power reduction potential and provide them with subsidized prices to incentivize them to cut power to reduce electricity. However, from the perspective of power operators, in the face of unknown and uncertain electricity consumption behaviors of residential users, how to identify and select residential users with great power reduction potential and formulate the dynamic subsidy price are the key challenges. In order to solve this problem, this paper proposes an incentive demand response algorithm based on online learning. Based on the multi-armed bandits, it learns the load reduction potential of residential users online, and establishes an optimization model for the operation cost of electric power operators to select the optimal consumers participating in demand response and formulate subsidized prices dynamically. According to the simulation results, the algorithm can achieve a balance between supply and demand while reducing the operating cost of power operators to a greater extent.

为了减少海上风电场有功网损,提出一种基于二阶锥凸松弛的海上风电场无功出力优化模型。该模型考虑最优潮流过程中所涉及的电网稳定要求与安全裕度,在各风机最大化有功出力的前提下,优化各风机无功出力,以减少网损,提高风场整体出力。通过二阶锥凸松弛的数学优化方法,解决多种约束下的潮流优化问题,给出了最大化风电场整体有功出力下的各风机无功出力最优值。最后,以舟山普陀海上风电场为例,对各机组无功出力进行了优化,并与基于启发式的粒子群优化算法和经典的内点法优化算法进行了对比,结果表明所提算法能明显减少系统有功损耗,提升风电场整体有功出力。

This paper proposes a reactive power optimization model for offshore wind farms based on second-order cone convex relaxation for reducing active network loss. The model considers the grid stability requirements and safety margin involved in the optimal power flow process and optimizes the reactive power output of each turbine to maximize the active power output, reduce the network loss, and improve the overall power output of a wind farm. Through mathematical optimization of the second-order cone convex relaxation, the power flow is improved, the line active power loss is reduced, and the power-flow optimization problem under various constraints is solved. Finally, with the Zhoushan Putuo wind farm as an example, the reactive power output of each wind turbine is optimized and compared with the results of heuristic particle swarm optimization and the classical interior-point optimization algorithm. The comparison shows that the proposed method can significantly reduce active power loss and improve the overall active power output of the wind farm.

忽视配电网运行的负荷特点，会导致蓄电池状态以及功率交换情况的优化不足。为了解决配电网节点拓扑和负荷不确定性的问题，提出了一种考虑负荷不确定性的配电网分层协调控制方法。分析了包括负荷节点、变压器、断路器、隔离开关等节点在内的配电网拓扑结构，分别计算了并网模态、孤岛模态下的能量关系，建立源网荷储分层协调控制模型，并使用拉丁超立方抽样方法生成不确定性的场景，引入了惯性权重、学习因子等参数改进粒子群算法，设计多目标分层粒子群算法的调控流程，实现配电网源网荷储分层协调控制。实验结果表明：应用该方法后，蓄电池的能量状态更好，交换功率更高；分层协调控制效果较佳，并可以根据实际需求进行配电网负荷状态自适应调整。

Ignoring the load characteristics of distribution network operation will lead to insufficient optimization of battery state and power exchange. In order to solve the problems of distribution network node topology and load uncertainty, a hierarchical coordination control method for distribution network considering load uncertainty is proposed. The topological structure of distribution network including load node, transformer, circuit breaker, isolation switch and other nodes is analyzed, and the energy relationship under grid-connected mode and island mode is calculated respectively, and the hierarchical coordination control model of source, network, load and storage is established. In addition, Atin hypercube sampling method is used to generate uncertain scenes, and parameters such as inertia weight and learning factor are introduced to improve the particle swarm optimization algorithm, and the regulation process of multi-objective layered particle swarm optimization is designed to realize the layered coordinated control of load and storage in the source network of the distribution network. The experimental results show that the energy state of the battery is better and the exchange power is higher after the method is applied. The effect of hierarchical coordination control is better, and the load state of distribution network can be adjusted adaptively according to the actual demand.

考虑电源出力和负荷需求不确定性的影响，建立了新型电力系统“源-储”规划容量两阶段鲁棒优化模型。利用偶理论和卡罗需-库恩-塔克（Karush-Kuhn-Tucker conditions，KKT）条件，把具有min-max-min结构的鲁棒优化问题分解成主问题和子问题两阶段进行求解。主问题为最小化的优化问题结构，采用松弛形式，考虑了新型电力系统的碳排放要求以及弃风弃光的制约条件，确定各类电源的装机容量的最优值，得到原问题的最优解的上界。子问题为包含非线性变量的最大-最小化的优化问题结构，采用对偶理论和KKT条件进行建模，并利用Big-M法处理非线性约束条件，采用盒式不确定集描述风光出力的不确定场景，以第一阶段的优化结果为前提，确定各种最恶劣场景。算例中3个情景的仿真结果验证了所提模型和方法的可行性和适应性。

A two-stage robust optimization model for the planning capacity of the new power system "source-storage" of new power system is established, taking into account the uncertainty of power output and load demand. Using even theory and Karush Kuhn Tucker conditions (KKT), the robust optimization problem with min-max-min structure is decomposed into two stages: the main problem and sub problems for solution. The main problem is the minimization optimization problem structure, which adopts a relaxed form and considers the carbon emission requirements of new power system and the constraints of wind and solar power curtailment. The optimal values of the installed capacity of various power sources are determined, and the upper bound of the optimal solution of the original problem is obtained. The sub-problem is a maximum-minimum optimization problem structure containing nonlinear variables. It is modeled using dual theory and KKT conditions, and the Big-M method is used to handle nonlinear constraints. A box type uncertain set is used to describe the uncertain scenarios of wind and solar power output. Based on the optimization results of the first stage, various worst-case scenarios are determined. The simulation results of three scenarios in the example validate the feasibility and adaptability of the proposed model and method.