I would like to learn, if possible, an estimate of the total time-averaged power of gravitational wave emission at frequencies below GHz by the Milky Way. For example the value might be something like $$ (\mbox{number of mergers per unit time}) \times (\mbox{mean energy emitted per merger}) + (\mbox{number of binaries with long-lasting orbits}) \times (\mbox{mean power per binary}) $$ I tried to extract an estimate from the following figure (from "Upper Limits on the Stochastic Gravitational-Wave Background from Advanced LIGO’s First Observing Run", https://arxiv.org/pdf/1612.02029)
It looks like most of the energy is from BBH and BNS so I did a rough integral of the shown curves. However I'm not sure if I am doing it right. That integral gave an energy density about $2 \times 10^{-9}\,\rho_{\rm crit}$, but I'm not sure if I did that right, and I'm not sure how to convert it into a "total power per galaxy". (Here $\rho_{\rm crit} = 3 c^2 H_0^2/8\pi G$, the cosmological critical energy density).