Overview
Research addressed the joint identification of permanent magnet synchronous machine (PMSM) flux maps and a simple inverter model. The method aimed to enhance offline flux identification from commanded voltage inputs, specifically by circumventing the need for prior identification of inverter parameters. The approach demonstrated improved accuracy for flux models by taking into account the non-linearity of the inverter, which typically disturbs the voltage output by the controller.
Research Context
Accurate prediction of torque, controller parameterization for torque tracking, and the creation of simulation models in electric drive modeling are dependent on the identification of magnetic flux maps. A common challenge in this process is the non-linearity inherent in the inverter, which can affect the commanded voltage. This non-linearity necessitates consideration during the identification of flux maps, particularly when relying on commanded voltage inputs.
Approach
The research proposed a novel approach for offline flux identification. This method specifically addressed the disturbance introduced by inverter non-linearity on the commanded voltage inputs. Instead of requiring prior identification of inverter parameters, the technique performed a joint identification of both the flux maps and a simplified inverter model.
In the dq reference frame, the flux maps were represented as static relationships between dq fluxes and dq currents. To capture saturation and cross-saturation effects accurately, a tensor product spline model was utilized. The study recognized that the effects of voltage disturbance on estimated flux maps were not negligible. The proposed joint identification strategy was expected to improve the flux model significantly.
Findings
- The effects of voltage disturbance on estimated flux maps were found to be non-negligible.
- Joint identification of flux maps and a simple inverter model significantly improved the accuracy of the flux model.
- The method was validated using both FEM simulation data and data from a test-bench.
Why This Matters
Accurate flux map identification is essential for predicting torque, parameterizing controllers for torque tracking, and developing reliable simulation models in electric drives. By providing a method to accurately identify flux maps while accounting for inverter non-linearity without requiring separate inverter parameter identification, this research contributes to improved electric drive modeling and control.