Abstract：As the penetration of renewable energy in power systems increases，converter-based devices are

gradually becoming dominant in the grid. In terms of system frequency stability support，grid-forming converters，

which can provide both active inertia support and fast frequency response services，hold great potential for

application. However，due to power limitations on the DC side equipment，when there is a high demand for active

power from the grid-forming converter，the DC voltage may collapse. To address this，a nonlinear model predictive

control（NMPC）method was proposed to maintain DC voltage stability，prevent faults on the DC side when the

grid-forming converter provides frequency support to the grid. This control method senses the frequency of the bus

adjacent to the converter and tracks a predefined DC link voltage setpoint to determine the required level of active

power support，and ensures quantifiable frequency support following an emergency. The design of the nonlinear

MPC incorporates the current limiting and input control constraints of the grid-forming converter，enabling it to

deliver maximum output power in a short period following an incident，while also avoiding DC voltage collapse.

Simulations based on high-precision phasor models of Kundur 2-area 4-machine system and the IEEE 16-machine

68-node system were conducted using Matlab/Simulink. The results show that the proposed method can effectively

suppress the DC voltage collapse and provide power support for fast frequency adjustment.

      Key words：grid-forming converter（GFC）；droop control；nonlinear model predictive control（NMPC）；fast

frequency response（FFR）；sliding mode control