A Forecast Error Correction Method Based on Seq2Seq and Auto Encoder for Short-term Wind Power Forecast Enhancement

Accurate wind power forecast (WPF) is highly significant for the safe and economic operation of power systems, and implementing error correction is a common and effective manner to further enhance the forecasting performance. Nonetheless, existing error correction methods fail to consider temporal relationship and may inadvertently increase the error under certain circumstances, thereby compromising the local precision of forecast. In order to address these issues, this paper proposes an error correction model based on Seq2Seq and auto encoder (AE) to improve forecasting accuracy. Firstly, the seq2seq with attention mechanism is introduced to infer the probable error sequence through learning the temporal relationship between the numerical weather prediction sequence and the actual error sequence. Secondly, the samples that the seq2seq providing a positive error inference are collected and the AE model is employed to extract the meteorological features at these moments. Finally, on the basis of the above two models, the error correction method is constructed consisting of preliminary error inference and posterior correction determination. Using operational data from Ningxia province, the proposed method is validated for its remarkable superiority based on the evaluation metrics.

• Publication year

  2024

• Venue

  2024 IEEE/IAS Industrial and Commercial Power System Asia (I&CPS Asia)

• Publication date

  2024-07-09

• Fields of study

  Not labeled

• Identifiers
  DOI 10.1109/ICPSAsia61913.2024.10761861
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• DA-CBGRU-Seq2Seq Based Soft Sensor for Mechanical Properties of Hot Rolling Process

  2023cited by this paper
• Short-term Wind Power Probabilistic Forecasting Based on Attention Mechanism and Bidirectional LSTM

  2023cited by this paper
• Multi-Floor Indoor Localization Scheme Using a Seq2Seq-Based Floor Detection and Particle Filter With Clustering

  2023cited by this paper
• A wind power forecasting method based on optimized decomposition prediction and error correction

  2022cited by this paper
• Combined Approach for Short-Term Wind Power Forecasting Based on Wave Division and Seq2Seq Model Using Deep Learning

  2022cited by this paper
• Predicting short wind speed with a hybrid model based on a piecewise error correction method and Elman neural network

  2021cited by this paper
• A Short-Term Regional Wind Power Prediction Method Based on XGBoost and Multi-stage Features Selection

  2020cited by this paper
• Probabilistic Prediction of Regional Wind Power Based on Spatiotemporal Quantile Regression

  2019cited by this paper
• Very short-term spatial and temporal wind power forecasting: A deep learning approach

  2019cited by this paper
• A Model Combining Stacked Auto Encoder and Back Propagation Algorithm for Short-Term Wind Power Forecasting

  2018cited by this paper
• Direct Interval Forecast of Uncertain Wind Power Based on Recurrent Neural Networks

  2018cited by this paper
• Identification and Correction of Outliers in Wind Farm Time Series Power Data

  2016cited by this paper
• Wind Power Forecasting Using Neural Network Ensembles With Feature Selection

  2015cited by this paper
• A Short-Term Wind Power Forecasting Approach With Adjustment of Numerical Weather Prediction Input by Data Mining

  2015cited by this paper
• Hybrid Forecasting Model for Very-Short Term Wind Power Forecasting Based on Grey Relational Analysis and Wind Speed Distribution Features

  2014cited by this paper
• Wind Power Forecasts Using Gaussian Processes and Numerical Weather Prediction

  2014cited by this paper
• Very short-term probabilistic wind power forecasting based on Markov chain models

  2010cited by this paper
• Study on the Physical Approach to Wind Power Prediction

  2010cited by this paper

• No citing papers are available for this paper.