[en] We present 1500 cycles of hydrogen shell flashes on a $1.38M_{\odot <!--\; ???\; -->}$ white dwarf (WD) for a mass accretion rate of $1.6\times <!--\; ??\; -->{10}^{-<!--\; ???\; -->7}{M}_{\odot <!--\; ???\; -->}$ yr⁻¹, the mass ejection of which is calculated consistently with the optically thick winds. This model mimics the 1 yr recurrence period nova M31N 2008-12a. Through these hydrogen flashes a helium ash layer grows in mass and eventually triggers a helium nova outburst. Each hydrogen flash is almost identical, and there is no precursor for the forthcoming He flash either in the outburst or in the quiescent until the next He flash suddenly occurs. Thus, M31N 2008-12a is a promising candidate of He novae, outbursting in any time within a millennium. The prompt X-ray flash of He nova lasts as short as 15 minutes, with the X-ray luminosity being about half of the Eddington luminosity, making the observation difficult. In the very early phase of a He flash, the uppermost H-rich layer is convectively mixed into the deep interior and most of the hydrogen is consumed by nuclear burning. In comparison with hydrogen shell flashes of M31N 2008-12a, we expect the forthcoming He nova to have a very short prompt X-ray flash (15 minutes), a very bright optical/near-IR peak (∼3.5 mag brighter than M31N 2008-12a), a much longer nova duration ($><!--\; >\; -->2$ yr), and a longer supersoft X-ray source phase (40–50 days or more).