Abstract:We propose a data-based method to solve a multi-stage stochastic optimal power flow (OPF) problem based on limited information about forecast error distributions. The framework explicitly combines multi-stage feedback policies with any forecasting method and historical forecast error data. The objective is to determine power scheduling policies for controllable devices in a power network to balance operational cost and conditional value-at-risk (CVaR) of device and network constraint violations. These decisions include both nominal power schedules and reserve policies, which specify planned reactions to forecast errors in order to accommodate fluctuating renewable energy sources. Instead of assuming the uncertainties across the networks follow prescribed probability distributions, we consider ambiguity sets of distributions centered around a finite training dataset. By utilizing the Wasserstein metric to quantify differences between the empirical data-based distribution and the real unknown data-generating distribution, we formulate a multi-stage distributionally robust OPF problem to compute optimal control policies that are robust to both forecast errors and sampling errors inherent in the dataset. Two specific data-based distributionally robust stochastic OPF problems are proposed for distribution networks and transmission systems.

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is how to balance economic efficiency and system reliability and security in the power system as the penetration of renewable energy sources (RESs) increases. Specifically, the paper focuses on how to use limited historical forecast error data to formulate control strategies in the multi - stage stochastic optimal power flow (OPF) problem. These strategies can deal with forecast errors and sampling errors, thereby reducing the system operation risks caused by the fluctuations of renewable energy. ### Main contributions of the paper 1. **Modeling of multi - stage distributionally robust optimal control problems**: - A data - based multi - stage distributionally robust optimal control problem for optimal power flow (OPF) is proposed. This model uses computationally feasible data - driven distributionally robust optimization techniques to solve the sub - problems at each stage. Different from the traditional single - stage distributionally robust optimization method, this method is extended to a multi - stage setting to obtain a closed - loop feedback control strategy. By using the Wasserstein ball to define the ambiguity set, this method can explicitly consider sampling errors, thereby improving the out - of - sample performance. 2. **Combination of linear approximation and AC OPF**: - Using the linear approximation of the AC power flow equations, computationally feasible, probabilistically - constrained AC OPF solutions are developed. These solutions are applicable not only to transmission systems but also to distribution networks. For distribution networks, the model includes inverter - connected renewable energy sources and energy storage systems, with a focus on voltage regulation problems. For transmission systems, the model includes power injections from synchronous generators and renewable energy sources, with a focus on the N - 1 security constraints of transmission lines. 3. **Numerical experiments and case studies**: - The effectiveness and flexibility of the proposed method are verified through extensive numerical experiments in a 37 - node distribution network and an 118 - node transmission system. These experiments show the inherent trade - off between economic efficiency and robustness to constraint violations and sampling errors. By explicitly incorporating forecast errors and sampling uncertainties, this method can help network operators better understand these risks and the inherent trade - offs, and design effective optimization and control strategies to appropriately balance efficiency goals and security requirements. ### Technical details - **Ambiguity sets and Wasserstein balls**: - The ambiguity set is defined using the Wasserstein distance, which contains all distributions within the Wasserstein ball. The Wasserstein radius can explicitly control the conservatism of the decision. A larger radius will reduce the dependence on specific data features and improve the robustness to sampling errors. - **Data - driven distributionally robust model predictive control**: - A distributionally robust model predictive control (MPC) algorithm is adopted, and only the control decision at the current time step is implemented each time. As time progresses, new forecast errors and states are realized, the optimization problem is re - solved, and the process is repeated. This method allows any forecasting method to be used to predict the uncertainties within the planning horizon, and the forecast error data set can be updated online. - **Network models and device dynamics**: - A symmetrically balanced power network model is considered, and the standard Pi model is used to describe the line characteristics. For grid - connected devices, the model includes traditional generators, inverter - connected renewable energy sources, fixed, dispatchable and curtailable loads, and energy storage devices. The dynamic characteristics of these devices are described by discrete - time linear dynamic systems. - **Network constraints**: - Network constraints include voltage magnitude limits, line power flow limits, and local constraints of devices. These constraints can be represented by linear inequalities, making the problem more computationally feasible. Through these techniques and methods, the paper provides a systematic framework that can effectively manage and optimize the operation of power systems in the context of high penetration of renewable energy.

Data-based Distributionally Robust Stochastic Optimal Power Flow, Part I: Methodologies

Stochastic Optimal Power Flow Based on Data-Driven Distributionally Robust Optimization

Data-based Distributionally Robust Stochastic Optimal Power Flow, Part II: Case studies

Data-Driven Stochastic Optimization for Power Grids Scheduling under High Wind Penetration

Distributed Linear Optimal Power Flow for Multi-area Unbalanced Distribution Systems

Data-Driven Nonparametric Probabilistic Optimal Power Flow: an Integrated Probabilistic Forecasting and Analysis Methodology

Two-stage Robust Optimal Power Flow for Distribution Network with Multiple Uncertainties

Toward More Accurate and Robust Optimal Power Flow for Distribution Systems

Optimal Power Flow Solutions Incorporating Stochastic Wind and Solar Power

Optimal Pump Control for Water Distribution Networks via Data-based Distributional Robustness

Data-Driven Probabilistic Optimal Power Flow with Nonparametric Bayesian Modeling and Inference

Interval Optimal Power Flow Applied to Distribution Networks under Uncertainty of Loads and Renewable Resources

A Data-Driven Stochastic Reactive Power Optimization Considering Uncertainties in Active Distribution Networks and Decomposition Method.

Distributionally Robust Multi-Energy Dynamic Optimal Power Flow Considering Water Spillage with Wasserstein Metric

A Hybrid Physical-Data Approach for Solving Dynamic Optimal Power Flow Considering Uncertainties and Different Topology Configurations

Distributionally Robust Chance-Constraint Optimal Power Flow Considering Uncertain Renewables with Wasserstein-Moment Metric

Parametric Distribution Optimal Power Flow with Variable Renewable Generation

Distributionally Robust Optimal Scheduling for Multi-Energy with Uncertainties

Analytical Uncertainty Propagation for Multi-Period Stochastic Optimal Power Flow

Dynamic Optimal Power Flow Model Incorporating Interval Uncertainty Applied to Distribution Network

A Novel Probabilistic Optimal Power Flow Method to Handle Large Fluctuations of Stochastic Variables