Combined MIMO Deep Learning Method for ACOPF with High Wind Power Integration

Altan Unlu,Malaquias Peña

DOI: https://doi.org/10.3390/en17040796

IF: 3.2

2024-02-07

Energies

Abstract:The higher penetration of renewable energy sources in current and future power grids requires effective optimization models to solve economic dispatch (ED) and optimal power flow (OPF) problems. Data-driven optimization models have shown promising results compared to classical algorithms because they can address complex and computationally demanding problems and obtain the most cost-effective solution for dispatching generators. This study compares the forecast performance of selected data-driven models using the modified IEEE 39 benchmark system with high penetration of wind power generation. The active and reactive power load data of each bus are generated using Monte Carlo simulations, and synthetic wind power data are generated by utilizing a physical wind turbine model and wind speed samples withdrawn from a Weibull distribution. The objective is to design and evaluate an enhanced deep learning approach for the nonlinear, nonconvex alternating current optimal power flow (ACOPF) problem. The study attempts to establish relationships between loads, generators, and bus outcomes, utilizing a multiple-input, multiple-output (MIMO) workflow. Specifically, the study compares the forecast error reduction of convolutional neural networks (CNNs), deep feed-forward neural networks (DFFNNs), combined/hybrid CNN-DFFNN models, and the transfer learning (TL) approach. The results indicate that the proposed combined model outperforms the CNN, hybrid CNN-DFFNN, and TL models by a small margin and the DFFNN by a large margin.

energy & fuels

What problem does this paper attempt to address?

The paper aims to address the following issues:

1. **AC Optimal Power Flow (ACOPF) Problem under High Wind Energy Penetration**: With the increasing penetration of renewable energy in modern power grids, effectively optimizing Economic Dispatch (ED) and Optimal Power Flow (OPF) problems becomes particularly important. The paper addresses the nonlinear, non-convex ACOPF problem by designing and evaluating an enhanced deep learning method.

2. **Prediction Error Reduction**: The study compares the performance of Convolutional Neural Networks (CNN), Deep Feedforward Neural Networks (DFFNN), combined/hybrid CNN-DFFNN models, and Transfer Learning (TL) methods in terms of prediction error. It finds that the proposed combined model outperforms other models to a certain extent.

3. **Multi-Input Multi-Output (MIMO) Workflow**: A MIMO workflow is utilized to establish the relationship between loads, generators, and node results to improve prediction accuracy. Specifically, this method aims to obtain the setpoints for all generators and the voltage/angle of each node in a single attempt, thereby simplifying the optimization output process.

4. **Wind Energy Integration**: The non-parametric nature of high-penetration wind energy is integrated into the IEEE 39-bus system to ensure accurate capture of wind energy variability and intermittency, reflecting real-world conditions.

In summary, the paper focuses on developing a deep learning technique that combines convolutional and deep feedforward neural networks to improve the prediction accuracy of data-driven ACOPF problems and validates its effectiveness through practical cases.