We present a scenario for the variability of the microquasar GRS 1915+105. This starts from previous works, leading to the tentative identification of the Accretion-Ejection Instability as the source of the low-frequency Quasi-Periodic Oscillation of microquasars and other accreting sources. We follow the physics of this instability: its conditions (the magnetic field and geometry adapted to MHD jet models), its instability criterion, and its consequences (cooling down of the disk, heating and excitation of the corona). Comparing them to the observed properties of the source, in particular the detailed properties of its spectral states, we first derive a model for the $\\sim$ 30 minutes cycles often exhibited by GRS 1915+105. In our model this is a limit cycle determined by the advection of poloidal magnetic flux to the inner region of the disk, and its destruction by reconnection (leading to relativistic ejections) with the magnetic flux trapped in the vicinity of the central source. We show how it leads to natural explanations for observed behaviors of GRS 1915+105, including the three basic states of Belloni et al (2000). We then discuss how this could be extrapolated further to understand the longer-term variability of this and other microquasars. |
