Dynamic Modelling of Coordinated UFLS and DER Shedding in Wide Area RMS Studies of Non-credible Events

Under-Frequency Load Shedding (UFLS) schemes face critical effectiveness challenges. High

DPV generation reduces the net load on UFLS circuits, and during periods of low synchronous generation, Rate of Change of Frequency (RoCoF) is exacerbated due to low inertia.

Furthermore, "reverse-flow" conditions can cause UFLS to inadvertently disconnect active power generation provided by DER, potentially worsening a frequency decline. Whilst the rollout of “dynamic arming” relays with the ability to disarm during periods of reverse power flow has been commenced in the NEM [1], current RMS simulation models lack the scalability required for detailed wide-area analysis without detrimental increases in runtime.

This paper introduces an augmented UFLS relay model designed for computational efficiency in wide-area studies. Improvements include:









 Expanded number of Under-Frequency Stages (from three to seven)

Additional RoCoF trip function

Partial shedding of DER generation in proportion to load

Dynamic arming mode, which disables the UFLS relay in response to reverse-flows.

Compatibility with dynamic variations in DER model behaviour, such as fluctuations in active power and inverter protection trips (DPV shake-off). These upgrades provide an industry-leading capability to simulate the loss of load and DER generation due to underfrequency and RoCoF events. This paper benchmarks the model’s performance against a historical power system event involving the trip of multiple synchronous generators in rapid sequence, resulting in the loss of over 2.7 GW, and the subsequent shedding of 1.3 GW of load and 400 MW of DER [2]. A total of 367 instances of the model were mapped on an RMS model containing a 3,500-bus representation of the NEM. The quantity of Total

Load and DER shed accurately, with errors less than 4% compared to reported data.

Benchmarking the dynamic arming functionality of the model was undertaken in a modified case with higher DER penetration (offset by reducing large-scale asynchronous generators).

The model demonstrated statistically significant improvements to load and DER ride-through with the preferred Scenario 2 (dynamic arming threshold of net flows at 0 MW). This reduced the amount of load shed by 388 MW (23%) and DER shed by 426 MW (48%) respectively, with faster frequency recovery and higher frequency nadir were also observed. This performance indicates that dynamic arming can significantly improve the retention of DER generation to assist in frequency stability and provides a detailed and scalable framework for assessing UFLS responses in broader applications.