You cannot check an individual drive directly -- you have to break it first, as you guessed, but there is another way as I will show further down.
First run chkdsk on the mirrored drive, and that should really fix the problem. I would be surprised if one drive had file-system structural-problems needing chkdsk, and the other one didn't, as the disk volumes should be (literally) mirror copies of each other, but you can test that possibility out as follows next.
You can use Paragon Backup & Recovery, which has a USB-stick mini operating system that you can create which will allow you to load up the disks without affecting them. So, for each disk individually, you can backup the disk, restore that backup to an empty drive, and then bring up that restored copy in Windows and chkdsk it from there. I know Windows 10 recognizes either drive, but they are, in fact, different, and I have found that (with my older version of Paragon) Paragon recognizes one of the pair just fine, but the other appears to be completely empty (when it's not). If Paragon sees both mirrored disks at the same time, they are both properly recognized. But there must be some actual difference between the two drives, probably one is the working image, and the other is the backup image -- that's my guess.
If you have two extra empty drives, you can backup the pair (the whole mirror) and then bring the backed-up pair up in Windows, break the backed-up mirrored pair, and chkdsk individually (to prove to yourself whether one is messed-up, but the other is fine). If the Disk Management Mirroring is working properly, the disks should either both be good, or both need a chkdsk. If they are out of synchronization, please let us know, because that would be a serious reliability problem (and Thanks, if you do).
Finally, if getting a power outage before both mirrored volumes are synchronized, to me, that strongly argues for both running chkdsk automatically on a regular schedule, and also perhaps leaning towards using SSD's, as I have noticed that they take less time to update, generally speaking. I am arguing that If the disks update faster, there is more of a chance that they will finish writing what they have to write before power is lost. Still should regularly run chkdsk, though, IMO. Hope this helps.
