Table 1. Analysis of Predicted Maintenance Requirements 



Complete Sample 







Random 



Sample 















Sample 



Size 



No. < D 

 (Sample; 



% < D 



X < D 

 Normal 



X < D 

 Weibulf 



Facility 

 Name 



No. of 

 Piles 



No. < 



a 

 D 

 cr 



D 

 cr 



No. of 

 Piles 



% of 

 Total 



cr 

 Sample 



Ratio 



Warren 



154 



35 



23 



24 



15 



7 



29 



23 



23 



Lobster 









21 



13 



6 



28.5 



23.5 



23.5 



House 





















Bldg 4 - 



347 



6 



1.7 



9 



2 











0.1 



0.5 



Fish Pier - 









9 



2 











0.6 



1.6 



Boston 









9 



2 











1.9 



3.2 



National 



647 



141 



21.8 



32 



5 



8 



25 



24.6 



25.7 



Park 









24 



3.7 



6 



25 



19 



19.5 



Service 









24 



3.7 



6 



25 



19 



19.5 



Boston Fish 



265 











12 



4.6 











0.2 



0.7 



Pier 









11 



4.2 











0.7 



1.6 



Castle 



2,782 



3 



0.1 



5 



0.01 











<0.01 



<0.01 



Island 









10 



0.03 











<0.01 



0.1 



Park 





















MBTA 



638 











10 



1.4 











<0.01 



<0.01 



All Piers 



4,833 



176 



3.6 



37 



0.07 



3 



12.3 



1.1 



2 











20 



0.04 



1 



5 



1 



2 



D = critical diameter of piling. 



predictions made from either the normal or Weibull distribution functions 

 were more accurate (within 5 percent of the actual requirement) than the 

 simple ratio method and should be used for estimation of maintenance and 

 repair projects. 



This study concluded that the distribution of damage due to natural 

 deterioration could be modeled using either the normal (Gaussian) dis- 

 tribution or a Weibull distribution. A sample randomly selected from 

 the population, using sampling theory equations to determine the required 

 sample size, could be used to accurately predict the average diameter, 

 distribution of deterioration, and maintenance requirements within the 

 accuracy and level of confidence selected prior to the selection of the 

 data. The next phase of the work required development of procedures to 

 implement the sampling criteria during actual inspections rather than 

 the selection of random data taken from previously inspected piers. 



19 



