Scholarlink

Volatility and randomness of wind speed and solar irradiance can make it difficult to predict the power of wind and photovoltaic in short time, which reduces power grid’s scheduling ability and affects the stability of grid-connected operation of new energy power generation. In order to address this problem, the Quantum Particle Swarm Optimization (QPSO) algorithm is adopted to optimize the network structure and parameters of Long-Short Term Memory (LSTM) network. Through theoretical analysis of working principle and process of LSTM network and optimization algorithm, an optimized LSTM for predicting wind and photovoltaic power in short time is established. The accuracy of the proposed model is proved by using actual data, then compared with unoptimized model and particle swarm optimization (PSO) optimized model. The results of experimental indicate that the LSTM model optimized by QPSO has better prediction precision and demonstrates good prediction results in both wind and photovoltaic power generation in short time, with a wider applicability.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Short-Term Wind and Photovoltaic Power Prediction Based on Optimized LSTM Model

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Yuhui Li,
  • Jiandong Duan,
  • Pengfei Zhang,
  • Bo Shao,
  • Luxiao Wang,
  • Yiming Xu

摘要

Volatility and randomness of wind speed and solar irradiance can make it difficult to predict the power of wind and photovoltaic in short time, which reduces power grid’s scheduling ability and affects the stability of grid-connected operation of new energy power generation. In order to address this problem, the Quantum Particle Swarm Optimization (QPSO) algorithm is adopted to optimize the network structure and parameters of Long-Short Term Memory (LSTM) network. Through theoretical analysis of working principle and process of LSTM network and optimization algorithm, an optimized LSTM for predicting wind and photovoltaic power in short time is established. The accuracy of the proposed model is proved by using actual data, then compared with unoptimized model and particle swarm optimization (PSO) optimized model. The results of experimental indicate that the LSTM model optimized by QPSO has better prediction precision and demonstrates good prediction results in both wind and photovoltaic power generation in short time, with a wider applicability.