Probabilistic wind power forecasting resilient to missing values: an adaptive quantile regression approach

Honglin Wen

2024-04-24

Abstract:Probabilistic wind power forecasting approaches have significantly advanced in recent decades. However, forecasters often assume data completeness and overlook the challenge of missing values resulting from sensor failures, network congestion, etc. Traditionally, this issue is addressed during the data preprocessing procedure using methods such as deletion and imputation. Nevertheless, these ad-hoc methods pose challenges to probabilistic wind power forecasting at both parameter estimation and operational forecasting stages. In this paper, we propose a resilient probabilistic forecasting approach that smoothly adapts to missingness patterns without requiring preprocessing or retraining. Specifically, we design an adaptive quantile regression model with parameters capable of adapting to missing patterns, comprising two modules. The first is a feature extraction module where weights are kept static and biases are designed as a function of missingness patterns. The second is a non-crossing quantile neural network module, ensuring monotonicity of quantiles, with higher quantiles derived by adding non-negative amounts to lower quantiles. The proposed approach is applicable to cases under all missingness mechanisms including missing-not-at-random cases. Case studies demonstrate that our proposed approach achieves state-of-the-art results in terms of the continuous ranked probability score, with acceptable computational cost.

Applications,Systems and Control

What problem does this paper attempt to address?

### Problems the Paper Aims to Solve

This paper aims to address the issue of missing values in probabilistic wind power forecasting. Specifically:

1. **Background and Motivation**:
- Wind power generation poses challenges to power system operations and electricity markets due to its uncertainty. Probabilistic wind power forecasting can address this inherent uncertainty, but existing forecasting methods often assume complete data, ignoring issues like sensor failures and network congestion that lead to missing data.
- Current methods for handling missing values (such as deletion or imputation) have problems in both parameter estimation and actual forecasting stages, potentially leading to information loss or reduced prediction quality.

2. **Research Objectives**:
- Propose a new probabilistic forecasting method that maintains prediction effectiveness even in the presence of missing values. This method does not require data preprocessing or retraining and can adapt to various missing patterns, including Missing Completely at Random (MCAR) and Missing Not at Random (MNAR).

3. **Main Contributions**:
- Unlike traditional "impute-then-predict" strategies, this method employs an adaptive quantile regression model whose parameters can adjust according to the missing patterns, thus avoiding manual parameter tuning and preprocessing steps.
- The method is applicable to all types of missing mechanisms and demonstrates superior performance in various experiments, particularly excelling in Continuous Ranked Probability Score (CRPS).

Through this approach, the paper aims to provide a more robust and less labor-intensive solution for probabilistic wind power forecasting in industrial applications.