A multi-part answer to a multi-part question.
First, a caveat: much of the detailed domain knowledge in this field falls under export control laws, so publicly available citations are not available to support every statement made. This happens to be my particular line of work, so I will share as much as I feel comfortable.
On paint chips being more lethal than rifle rounds
The truth is more complicated.
It very much depends on the configuration of the shield. The penetrating capability of an incoming projectile does not vary monotonically with its velocity. For most spacecraft shield configurations, if you were to somehow create a graph of "lethality" vs. velocity for a given particle size, there would be very pronounced peaks at a number of different velocities due to changes in the physical interaction between the projectile and the shield. For a Whipple shield, a simple paint flake at any velocity is going to be far less lethal than a rifle round.
Further complicating this, the behavior of impact at interplanetary velocities is not well characterized. The best hypervelocity impact test labs in the world still cannot shoot projectiles at velocities anywhere near the expected encounter velocities for a spacecraft in interplanetary space. Extrapolations and engineering judgment are all that is left in these velocity regimes.
On repeat impacts on Whipple shields
The risk exists, but it is generally just accepted.
Space vehicle shielding design tends to look at the risk to broad areas of the spacecraft. Impacts into Whipple shields follow what I call the "Wile E. Coyote" effect: they tend to leave neat holes in the shield of similar-ish size to the incoming particle. The risk of impact more or less scales linearly with vulnerable area. If the risk of a lethal strike to a large vehicle is, say, 1 in 10, the risk of a lethal strike based on "threading the needle" through an existing hole is going to be orders of magnitude smaller -- probably tens of thousands of times less likely.
On replaceability of debris shields
If it is a design requirement, then it can be done.
For an interplanetary spacecraft, it comes with an added complication of needing to have spares or manufacturing capability on board, as resupply is out of the question, and it would require some way to install it, either through extravehicular activity or robotics.
On radiator design
All of these questions fall under "it depends."
If protecting for a meteoroid or debris strike is a design priority, then it can be designed as such. There are multiple design avenues available, but the selection depends on how the risk of a strike is weighed against other factors like heat transfer capability, mass, structural strength, etc.
On solar arrays
This is something I have direct test experience with: in general, individual strikes to solar cells are essentially undetectable at the system level. Instead, they are a contributing factor (among many) to a slow -- but steady -- loss of power production capability over time.
Where power production does present a vulnerability to strikes, it is overwhelmingly in "weak points" in the design: bus lines, power conditioning hardware, etc. All of these can be protected to keep the risk to an acceptable level. The degree to which risk exists, and the consequence of a strike to any specific piece of hardware is going to be a function of the architecture and specific detail design of that particular system.
The same statements from before apply to the repairability as well: if it's designed for, then yes.