3. Regression analysis shows promise as a tool for determining 

 which parameters are important and what the magnitude of the relevant 

 coefficients should be. However, any regression analysis should be 

 formulated carefully to include the physical characteristics of the 

 problem. 



4. A well-founded physical analysis based carefully upon engineering 

 judgment appears to offer the most promise for immediate use in 

 predicting penetration. In the analysis presented above, the use of 



what is basically a static pile bearing capacity equation in conjunction 

 with a Newtonian equation of motion appeared to offer the most nearly 

 accurate results. 



5. The approach presented in the SUGGESTED PREDICTION TECHNIQUE 

 section appears to be the most reasonable practical technique for use 

 in the interim period before additional research is performed. 



6. The effects of soil viscosity and penetrator velocity on 

 object penetration are not well understood. The present investigation 

 yielded conflicting results in this regard. 



RECOMMENDATIONS 



1. Additional complete investigations such as that reported 

 herein are required for a better understanding of the penetration 

 phenomenon. 



2. This requirement can best be satisfied by the performance 

 of a large number of laboratory penetration tests in which a variety 

 of small penetrators are allowed to penetrate various simulated soils. 

 It is imperative that each experiment be totally controlled. The 

 motion of the penetrators as they penetrate should be accurately 

 recorded either by means of an accelerometer or through photography. 

 Soil properties should be accurately measured and controlled. The 

 experiments should be designed statistically so that sahpe, velocity, 

 size, and soil effects can be isolated. 



3. The laboratory experiments should be supplemented by a 

 limited series of in-situ tests of the nature of the tests described 

 above only involving a greater variety of objects and entry velocities. 



4. Analyses of the data obtained should include a regression 

 analysis in conjunction with an improved physical analysis based on 

 dynamic concepts. 



27 



