Previous
abstract 
Next
abstract
Luminous blue variables (LBVs) are known to vary on many-decade timescales like $\eta$ Car and P Cyg, on many-year timescales as their usual observed outbursts, and with many-day period microvariations that could be stellar pulsations. These pulsations, observed only when the stars are not in outburst, may be intrinsic, just as for the less luminous $\beta$ Cephei and related variables. This investigation studies whether the new Livermore OPAL opacities can allow theoretical predictions of nonradial g-modes. A model with initial mass of 50 $M_\odot$ and current mass of 38.85 $M_\odot$ at a luminosity of $5.36$x$10^5$ $L_\odot$ and surface effective temperature of $24,430K$ has unstable few day nonradial modes near $g_5$ with degree $\ell$=1. An approximate allowance for the effects of time-dependent convection in the driving layers between $10^{-6}$ and $10^{-5}$ of the mass into the model can nullify the radiation kappa effect and stabilize predicted pulsations. Thus it seems necessary to suppress convection so that the driving is not reduced to a value less than the radiation damping occurring in deeper lying layers. A likely convection suppression cause is a composition $\mu$ gradient produced by rapid helium diffusive settling in layers that originally were not convective. When the temperature gradient becomes superadiabatic in the gradually deeping layers as mass loss proceeds off the surface, a helium $\mu$ gradient can prevent convection and allow pulsations. Possibly outbursts are caused by pulsations driven so strongly that they get out of hand. They then remove any inhomogeneous composition layers. As the outburst recovers, helium settling below any new surface convection zone can reestablish the $\mu$ gradient that can again suppress convection when the deepening convection zone arrives there. Cycles of pulsations and an outburst followed by no pulsations can occur in this very thin surface layer as the star evolves to cooler F star regions with greatly deepening convection and no pulsational instabilities. The hundreds of days periods observed are probably the result of beating between our predicted adjacent few day period g-modes that could be observed individually.
