And the news this time is better.
After the initial boot, I simply open a terminal window to confirm connectivity. I didn't, so I (*)corrected that so I could ping bbc, and rebooted. All OK, still on the original kernel.
I then selected 'System Tools'/'Software Update', and was offered 196 updates including the same kernel that caused grief last night:
3.10.0-229.14.1.el7.x86-84
I allowed the updates to run and then checked whether the initramfs for the new kernel
was installed in /boot. Unlike last night it was indeed present , the subsequent reboot worked fine, uname confirms the new kernel is now running.
Having seen how long the updates took to install I wonder, with hindsight, whether I may have rebooted the system before they were done? Quite apart from the silly power-off, that may be another factor. In my view, neither of these 'human error' explanations is a terribly good explanation, but better than none. None of the other instances that I found from web searches had confessed to any human error.
FWIW, all file systems (except swap) are xfs, in 'traditional', non-LVM, partitions.
Anyway I will now proceed with a slightly dented confidence, but at least with fresh enthusiasm.
* I have not detailed the issue with networking after today's install as I am sure it was not relevant. Merely something I'd got wrong in the install sequence.