Abstract: With the development of more electric aircraft and the widespread application of power electronic loads, aircraft high-voltage dc power generation systems are facing increasingly severe stability challenges. Conventional impedance-based analysis relies on continuous analytical expressions of generator inductance parameters, impedance, and stability regions. However, the inductance parameters of the doubly salient electro-magnetic generator (DSEG) exhibit strong nonlinearity and strong coupling, making it difficult to obtain accurate continuous analytical expressions and, consequently, to accurately characterize the system stability. A stability analysis method based on inductance–impedance–stability region discrete space mapping is proposed. First, a small-signal model of the DSEG power generation system is established, and the system output impedance model is derived. Then, a nonlinear inductance prediction model is developed, and a discrete inductance parameter space is constructed. The corresponding discrete inductance parameters are substituted into the output impedance model to obtain a discrete set of output impedances. Finally, according to the stability criterion, the discrete set of output impedances is mapped into a discrete loop-gain set, and stability is determined from its Nyquist trajectories, thereby completing the mapping from the discrete inductance parameter space to the stability region. Simulation and experimental results demonstrate that the proposed method can rapidly and effectively analyze the impedance characteristics and stability of the system.

Key words: high-voltage dc power generation system, doubly salient electro-magnetic generator, discrete space mapping, small-signal model, stability analysis, extreme learning machine