If that's a regulated supply, and judging by the inductors, it is, then the cause is probably excess voltage applied to the caps. The 63V caps are probably seeing something like 80V momentarily while the voltage is being integrated down to the averaged PWM value.
This is common too in PC, off-line SMPS, and power factor correction circuits where only the average regulated supply voltage is considered in cap selection (and for obvious financial reasons) and not the peak, pre-integrated transformer secondary voltage.
Remove the caps, clean the board thoroughly, repair any damage, and replace the caps with 80V, 105C caps and it should last much longer.
Also check the 16V input filter cap - it may have opened up allowing excess ripple at the primary and causing the PWM regulation to do weird things.
This is common too in PC, off-line SMPS, and power factor correction circuits where only the average regulated supply voltage is considered in cap selection (and for obvious financial reasons) and not the peak, pre-integrated transformer secondary voltage.
Remove the caps, clean the board thoroughly, repair any damage, and replace the caps with 80V, 105C caps and it should last much longer.
Also check the 16V input filter cap - it may have opened up allowing excess ripple at the primary and causing the PWM regulation to do weird things.