Nonlinear Bounded Error Compensation for Flux-Based Encoderless Controller of PMSMs

The encoderless control method based on a non-linear flux observer is considered effective in the medium-to-high speed range, but it fails in the low-speed high-torque operating region due to excessive angle errors, which forces the small drone to shut down when landing. To solve this problem, this paper proposes a nonlinear angle error compensator that acts directly on the angle estimation. The mathematical model of the motor flux observer and the flux compensation term is analyzed, which are proven to have corresponding error bounds. Experimental results demonstrate that the proposed method can operate effectively and stably in the low-speed and high-torque region where conventional methods fail, while also achieving good control performance across other speed ranges even under inaccurate motor parameters. Additionally, the theoretical relationship between observer parameters and the convergence speed after motor load disturbances has been verified in experiments. It is further confirmed that in scenarios including gradual loading, sudden heavy load at low speed, large external current disturbance, and speed ramp-down tests, the magnitude of the flux error does not exceed the error bound.