An Extended Nonlinear Stability Assessment Methodology For Type-4 Wind Turbines via Time Reversal Trajectory

TL;DR

This paper develops an advanced nonlinear stability assessment method for Type-4 wind turbines, improving the accuracy and efficiency of estimating the system's stability boundary under weak grid conditions using time reversal and optimal sampling techniques.

Contribution

It extends previous work by validating the stability boundary with simulations, exploring parameter sensitivity, and incorporating nonlinear control elements like PLL saturation.

Findings

Enhanced estimation accuracy of the stability boundary.

Reduced computational time through optimal sampling.

Validated stability assessment with PSCAD simulations.

Abstract

As the integration of renewable energy generation increases and as conventional generation is phased out, there is a gradual decline in the grid's strength and resilience at the connection point of wind turbines (WTs). Previous studies have shown that traditional grid-following controlled converters exhibit deteriorating dynamic characteristics and may result in an unstable system when connected to a weak grid. Due to the limitations of linear analysis, transient stability investigations are necessary. However, existing methods, such as standalone time-domain simulations or analytical Lyapunov stability criteria, have drawbacks, including computational intensity or excessive conservatism. Our prior research proposed an innovative approach to estimate the system boundary - a time-limited region of attraction (TLRoA), using a hybrid linearised Lyapunov function-based method and the time-reversal technique to compensate for the known limitations. However, in that work, the accuracy of the estimated TLRoA was not investigated, i.e. the TLRoA was not compared against a forward simulated region of attraction, and the sensitivity of the system parameters on the TLRoA was not explored. Moreover, the framework did not consider nonlinear control elements such as PLL saturation. In this paper, we not only build upon our previous work and propose directions that address these gaps but also enhance its effectiveness by introducing optimal sampling to improve further the speed of estimating the TLRoA. Furthermore, the stability boundary is verified using time-domain simulation studies in PSCAD.