Andromeda is our nearest neighbouring disk galaxy and a prime target for detailed modelling of the evolutionary processes that shape galaxies. We analyse the nature of M31s triaxial bulge with an extensive set of N-body models, which include Box/Peanut (B/P) bulges as well as initial classical bulges (ICBs). Comparing with IRAC 3.6$mu m$ data, only one model matches simultaneously all the morphological properties of M31s bulge, and requires an ICB and a B/P bulge with 1/3 and 2/3 of the total bulge mass respectively. We find that our pure B/P bulge models do not show concentrations high enough to match the Sersic index ($n$) and the effective radius of M31s bulge. Instead, the best model requires an ICB component with mass $M^{rm ICB}=1.1times10^{10}{rm M_{odot}}$ and three-dimensional half-mass radius $r_{rm half}^{rm ICB}$=0.53 kpc (140 arcsec). The B/P bulge component has a mass of $M^{rm B/P}=2.2times10^{10}{rm M_{odot}}$ and a half-mass radius of $r_{rm half}^{rm B/P}$=1.3 kpc (340 arcsec). The models B/P bulge extends to $r^{rm B/P}$=3.2 kpc (840 arcsec) in the plane of the disk, as does M31s bulge. In this composite bulge model, the ICB component explains the velocity dispersion drop observed in the centre within $R<$190 pc (50 arcsec), while the B/P bulge component reproduces the observed rapid rotation and the kinematic twist of the observed zero velocity line. This models pattern speed is $Omega_p$=38 km/s/kpc, placing corotation at $r_{rm cor}$=5.8 kpc (1500 arcsec). The outer Lindblad resonance (OLR) is then at $r_{rm OLR}$=10.4kpc, near the 10kpc-ring of M31, suggesting that this structure may be related to the bars OLR. By comparison with an earlier snapshot, we estimate that M31s thin bar extends to $r_{rm bar}^{rm thin}sim$4.0 kpc (1000 arcsec) in the disk plane, and in projection extends to $R_{rm bar}^{rm thin}sim$2.3 kpc (600 arcsec).
Download