An optimization method for flexible interconnection planning based on improved CNN-LSTM prediction and tunable...

Area of Science:

• Electrical Engineering
• Artificial Intelligence
• Optimization Theory

Background:

• Power systems are increasingly complex, requiring advanced solutions for reliability and economic efficiency.
• Flexible interconnection technology is crucial for managing evolving grid demands and enhancing operational performance.

Purpose of the Study:

• To develop a data-model dual-driven planning methodology for flexible interconnection systems.
• To integrate advanced forecasting and optimization techniques for improved power system design.

Main Methods:

• Utilized a multi-scale spatio-temporal cross-enhanced CNN-LSTM model for accurate load forecasting.
• Introduced a Tunable Relative Entropy (TRE) metric for quantifying spatio-temporal complementarity.
• Employed a Chaotic Evolutionary Optimization (CEO) algorithm to optimize system configuration and operational parameters.

Main Results:

• The CNN-LSTM forecasting module achieved high accuracy (MSE: 0.000368, MAE: 0.006334).
• The TRE index improved complementarity efficiency by 3.8%.
• The methodology effectively reduced load fluctuation indices and enhanced planning efficiency.

Conclusions:

• The proposed data-model dual-driven approach offers a viable pathway for intelligent power system development.
• The integration of CNN-LSTM and CEO algorithms optimizes system design, reliability, and economic performance.