Abstract:Renewable energies and their operation are becoming increasingly vital for the stability of electrical power grids since conventional power plants are progressively being displaced, and their contribution to redispatch interventions is thereby diminishing. In order to consider renewable energies like Wind Power (WP) for such interventions as a substitute, day-ahead forecasts are necessary to communicate their availability for redispatch planning. In this context, automated and scalable forecasting models are required for the deployment to thousands of locally-distributed onshore WP turbines. Furthermore, the irregular interventions into the WP generation capabilities due to redispatch shutdowns pose challenges in the design and operation of WP forecasting models. Since state-of-the-art forecasting methods consider past WP generation values alongside day-ahead weather forecasts, redispatch shutdowns may impact the forecast. Therefore, the present paper highlights these challenges and analyzes state-of-the-art forecasting methods on data sets with both regular and irregular shutdowns. Specifically, we compare the forecasting accuracy of three autoregressive Deep Learning (DL) methods to methods based on WP curve modeling. Interestingly, the latter achieve lower forecasting errors, have fewer requirements for data cleaning during modeling and operation while being computationally more efficient, suggesting their advantages in practical applications.

What problem does this paper attempt to address?

### Problems the paper attempts to solve This paper attempts to solve two key challenges in wind power (WP) prediction: 1. **Scalable and low - error wind power prediction models**: - As the wind power capacity increases, there is a need for automated design and operation of wind power prediction models to keep up with the growth rate of wind power capacity. - Offshore wind farms, due to their uniform terrain, are suitable for holistic modeling approaches, and computationally - intensive methods such as autoregressive deep - learning models can be considered. - Onshore wind turbines are distributed dispersedly and have diverse terrains, resulting in different characteristics of wind turbines, which makes the cost - effectiveness of using computationally - intensive deep - learning models in prediction a problem. 2. **The impact of regular and irregular interventions on wind power capacity**: - Regular shutdowns include time - controlled shutdowns, such as those for bat protection and maintenance. - Irregular shutdowns include rescheduling shutdowns that change the operation plan to prevent line overloading. - For planning rescheduling interventions, the prediction needs to convey the available wind power generation amount and should not include predictions of future rescheduling interventions. - Autoregressive methods rely on past wind power values and exogenous variables (such as wind speed and wind direction forecasts), but rescheduling - related shutdowns also appear in these past values, which may cause the model to wrongly predict future shutdowns, thus affecting prediction - based planning. ### Solutions The paper solves the above problems through the following methods: - **Comparing autoregressive deep - learning methods with methods based on wind power curve modeling**: - Three autoregressive deep - learning methods are selected for comparison: DeepAR, N - HiTS, and Temporal Fusion Transformer (TFT). - Methods based on wind power curve modeling only rely on weather forecasts and do not consider past values, including three machine - learning methods: XGBoost (XGB), Support Vector Regression (SVR), and Multi - Layer Perceptron (MLP), as well as the Original Equipment Manufacturer (OEM)'s wind power curve and an automated machine - learning method AutoWP. - **Handling shutdown data**: - For autoregressive methods, five shutdown handling methods are considered: no handling (none), explanatory variables, drop, imputation, and drop - imputation. - For methods based on wind power curve modeling, only the drop method is used because these models do not require a time context during operation. ### Main findings - **Methods based on wind power curve modeling**: - Have lower prediction errors. - Require less data cleaning. - Are more computationally efficient and more suitable for practical applications. Through the comparison and analysis of these methods, the paper improves the understanding of the challenges of shutdown handling in wind power prediction and provides a valuable reference for practical applications.

On autoregressive deep learning models for day-ahead wind power forecasting with irregular shutdowns due to redispatching

A Comprehensive Framework of the Decomposition-Based Hybrid Method for Ultra-Short-Term Wind Power Forecasting with On-Site Application

Short-term Wind Power Ramp Forecasting with Empirical Mode Decomposition Based Ensemble Learning Techniques

Wind Power Short-Term Forecasting Model Based on the Hierarchical Output Power and Poisson Re-Sampling Random Forest Algorithm

Prediction of Electricity Generation Using Onshore Wind and Solar Energy in Germany

Advanced Ensemble Methods Using Machine Learning and Deep Learning for One-Day-Ahead Forecasts of Electric Energy Production in Wind Farms

Design and Development of a Short-Term Photovoltaic Power Output Forecasting Method Based on Random Forest, Deep Neural Network and LSTM Using Readily Available Weather Features

Forecasting model for short-term wind speed using robust local mean decomposition, deep neural networks, intelligent algorithm, and error correction

Machine Learning and Deep Learning for Wind Power Forecasting

DeepFore: A Deep Reinforcement Learning Approach for Power Forecasting in Renewable Energy Systems

Design of a Wind Power Forecasting System Based on Deep Learning

Application of autoregressive dynamic adaptive (ARDA) model in real-time wind power forecasting

A novel deep learning-based evolutionary model with potential attention and memory decay-enhancement strategy for short-term wind power point-interval forecasting

Forecasting enhancement of Wind Power Generation using Adversarial Networks: A Data Driven Approach

Combating Uncertainties in Wind and Distributed PV Energy Sources Using Integrated Reinforcement Learning and Time-Series Forecasting

Demystifying the use of ERA5‐land and machine learning for wind power forecasting

Wind power predictions from nowcasts to 4-hour forecasts: A learning approach with variable selection

WindDragon: Enhancing wind power forecasting with Automated Deep Learning

Short-term offshore wind power forecasting - A hybrid model based on Discrete Wavelet Transform (DWT), Seasonal Autoregressive Integrated Moving Average (SARIMA), and deep-learning-based Long Short-Term Memory (LSTM)

A novel automatic wind power prediction framework based on multi-time scale and temporal attention mechanisms

Short-Term Wind Power Prediction Based on LightGBM and Meteorological Reanalysis