the piston separates the piston from the fluke and initiates keying. 

 The fluke was designed to provide a reasonable keying distance while 

 maintaining structural integrity. 



An optimum fluke design was attained by an empirical and theoreti- 

 cal analysis and some model testing. Using anticipated seafloor soil 

 types, a trade-off study was made between fluke sizes and shapes to 

 determine which fluke (s) could be driven into the seafloor with a mini- 

 mum of energy while attaining the required long-term holding capacity 

 of 20,000 pounds. This allowed for a small, light gun system to be 

 chosen to embed the flukes because the required energy had been 

 minimized. Appendix A outlines the trade-off study. 



Figure 5 presents the results of a series of laboratory and small- 

 scale field tests to evaluate keying distance as a function of keying 

 arm length. These tests were the basis of selecting the eccentric 

 keying distance of .3 L (L = fluke length) to cause keying in a distance 

 equal to about 1 3/4 L measured from the fluke tip. The top of the 

 fluke, therefore, moves less than the fluke length before keying. Full- 

 scale at-sea tests indicated that keying actually occurs somewhat 

 quicker than shown in Figure 5. Rapid keying ensures that the anchor, 

 once established, will function as a "deep" anchor, an anchor whose 

 capacity is not significantly affected by small upward movement. Taylor 

 and Lee (1972) have given a more complete description of this 

 phenomenon . 



Rock Fluke . There is little information available to design pro- 

 jectiles that will penetrate rock and then resist a specified pullout 

 load. As a result, the design of the rock projectile was accomplished 

 using considerable engineering judgement based on the results of full- 

 scale anchor tests of others (Smith, 1971), results of Sandia Labora- 

 tory's investigations of the penetration phenomenon (Young, 1970; 

 Feltz, 1972), and results of model penetration tests. 



A model test program (True, 1972) was performed to determine the 

 penetration characteristics of projectiles of various shapes in simu- 

 lated rock. The results provided information useful in the rock fluke 

 design. A three-fin fluke was chosen because model tests indicated 

 that this configuration developed, for a given amount of penetration 

 energy, about the same holding capacity as a plate or long solid shaft 

 of equal mass. Since penetration may not be sufficient to completely 

 bury the fluke, a three-fin fluke is more desirable than other shapes 

 because it affords increased moment resistance to randomly directly 

 loads. Model test data also indicated that large fluke serrations are 

 unnecessary; a fluke with a roughened surface is as effective. 



Full-scale tests (Smith, 1971) showed that a three-fin fluke was 

 very effective in coral and partially effective in basalt. In one 

 basalt test the fluke cracked on impact and failed prematurely, while 

 in another the fluke exceeded design goals in holding capacity. It 

 was apparent from the tests that the basic configuration was satisfactory. 

 However, a fluke with stronger and tougher steel, free of residual 

 stresses, and free of points of stress concentration was needed. 



