I recently looked at some panels that list BFC in the range of 17k-22k. These panels had 10k rated breakers. In my non engineered head this is an under rated breaker problem. Am I looking at this correctly? Thanks in advance.

On the surface, it does appear to be an equipment short circuit adequacy problem. However, there is another method that might be used known as series ratings.

For a series rating, an underrated panel downstream (10 kA in this case) may be protected by an upstream device such as current limiting fuses or circuit breakers. Today, this type of rating is tested, listed and labeled by Underwriters' Laboratories but a long time ago (and on a few occasions today) let thru curves from the current limiting device manufacturers were used to "predict" the let thru current.

The concept is if a fault occurs downstream from the underrated equipment (17k-22k for this example) not only would the 10kA device try to interrupt but the device upstream would interrupt and reduce the current to a value that the underrated panel could safely interrupt. There are also non-current limiting series ratings which are also tested and labeled.

With all that said, my view is either they are using a series rating or the panel is inadequate.

It also depends on your methodology. ANSI short circuit calculation methods are frequently used because it's really easy. I can do it easily on paper with a calculator unlike doing arc flash calculations which are a royal pain to do by hand. However, the ANSI method is also extremely conservative and predicts higher currents than you see in reality. A more accurate calculation taking into account cable impedances, current limiting, and subtransients and transients from rotating loads provides a more accurate picture than the simple ANSI calculation which may make the resulting bolted fault current higher but in practice it is usually significantly lower.

The board construction, bus bar assembly etc may be type tested to 22kA for a period of time (1s, 5s etc), however your systems bolted fault level might be below 10kA in which case the breakers may not be underated.

As Jim correctly states, in the US there are two acceptable methods to determine if the panel is apropriately rated. Fully rated and series rated. Any method that uses an "engineering calculation" based on let-through or current limitation of the upstream device is not, to my knowledge, supported by any North American Circuit Breaker manufacturer. "Engineered" ratings are allowed by the NEC under some circumstances, but not supported by CB manufacturers. One respondent mentions that calculations can take into account current limitation by upstream devices. That is not a good practice. Unluckily many devices, not labeled as current limiting are, in fact, very fast regardless their label. And it is not unheard of that a 10kA CB with 20kA prospective fault current, will "try", to protect an upstream higher rated current limiting device. Unluckly, unless it is a tested series rating users are not provided with enough information to understand and predict the complex interation between two devices dynamically reacting to the prospective current and each other simultaneously.

In the situation described you either have an underated panel, or a series rated panel. If properly series rated the usptream protective device should be marked to indicate it is part of a series rating, and the downstream panel should have a label identifying the possible acceptable series ratings in that panel. Either printed on the label, or by reference to another document, the combination of dowstream CBs found in the panel, and the upstream overcurrent device should be described. Under some circumstances there could be third device in between.

That is not a good practice. Unluckily many devices, not labeled as current limiting are, in fact, very fast regardless their label. And it is not unheard of that a 10kA CB with 20kA prospective fault current, will "try", to protect an upstream higher rated current limiting device.