Smart grid inverter control: integrating RNN, model predictive, and adaptive sliding mode controller for optimal...

Area of Science:

• Electrical Engineering
• Control Systems
• Power Electronics

Background:

• Grid-connected voltage source inverters (GC-VSIs) face challenges like harmonic distortion and grid instability.
• Traditional Model Predictive Control (MPC) is accurate but computationally intensive; Recurrent Neural Networks (RNNs) are fast but lack formal control guarantees.

Purpose of the Study:

• To develop a robust and computationally efficient hybrid control strategy for GC-VSIs.
• To address issues of harmonic distortion, grid fluctuations, and external disturbances in smart grids.

Main Methods:

• A hybrid control combining offline MPC trajectory optimization, real-time RNN implementation, and an Adaptive Barrier-Condition Super-Twisting Sliding Mode Controller (ABC-STSMC).
• RNNs trained with MPC data for reduced online computational load.
• Lyapunov analysis for stability and error bounds.
• Improved Grey Wolf Optimization (IGWO) for parameter tuning.

Main Results:

• The hybrid ABC-STSMC demonstrated superior harmonic mitigation and dynamic response compared to standalone MPC and RNN controllers.
• Achieved minimized Total Harmonic Distortion (THD) under weak-grid, unbalanced load, and distorted voltage conditions.
• Validated through extensive simulations and Hardware-in-the-Loop experiments.

Conclusions:

• The proposed hybrid control system offers a computationally efficient and robust solution for advanced GC-VSI control in smart grids.
• Effectively enhances stability and performance in nonlinear and uncertain grid environments.
• Presents a significant advancement in managing complex inverter control challenges.