Just enter your email
It is not appropriate to compute power for the ANOVA example where the number of groups is increased. This is not a situation for ANOVA; it is a situation for a continuous dose-response analysis. You might derive the power for a 2 degree of freedom composite test for whether a quadratic effect is flat. ANOVA power surely goes down as number of groups expands unless the regression sum of squares doesn't also expand. But that is not relevant here.