It is always difficult to advise if one does not have all the facts. On the basis of the few available facts I must, for the very first time, disagree largely with Jack's analysis. 95% coverage could well be the optimum for the part concerned. 95% has the advantage of being measurable! Even if you were required to reach 98% then adding 50% to the measured times for 95% would achieve that objective in a controllable manner.
Your estimates on peening time versus OD are mathematically correct - surface area of a cylindrical surface being proportional to square of its diameter.
Estimates for ID peening times depend to a large extent on the ID value. Jack is correct to point out the ricochet effect.This is particularly important for small-diameter holes.If, however, you are dealing with large-diameter thin-walled tubes then predictions based on diameter/peening time ratios would be a reasonable guide.
One thing we do agree upon is that predictions must be tested by actual measurements.
