New method simplifies modeling of power converters
A new framework called DQsym uses smaller matrices to simulate complex electrical systems.
Engineers have developed a new framework called DQsym for simulating power converters. This method uses a smaller solver state matrix than previous approaches.
Modern power systems use modular multilevel converters (MMCs). These converters show periodic dynamics at many different frequencies.
Engineers usually use harmonic state-space (HSS) or dynamic phasor (DP) models to study these systems. Both methods use state matrices that grow larger as more harmonics are included.
The size of these matrices depends on the truncation order . This growth affects the sparsity and the coupling of the models.
The DQsym framework represents harmonics in the signals instead of the states. This makes the size of the solver state matrix independent of the truncation order .
In DQsym, harmonic coupling happens in nonlinear signal blocks outside the state matrix. This shifts the computational cost to the width of the signals.
Testing showed that for , the DQsym solver has 45 states. In comparison, HSS and DP both have 135 states.
The DQsym model stays within about 4% of the reference model. Once the coupling blocks are closed, it yields eigenvalues identical to HSS and DP.
A harmonic vector is a set of complex Fourier coefficients used in HSS modelling. A dynamic phasor is a coefficient computed over a sliding window of one fundamental period.
Key facts
- The DQsym solver state matrix size is independent of the truncation order .
- For , DQsym uses 45 states while HSS and DP use 135 states.
- DQsym achieves accuracy within about 4% of the reference model.