192 COLLAPSE OF TEXAS TOWER NO. 4 



and these — I think you can appreciate the fact that on this tower when 

 it is in motion, since you are on the ocean, you have ahnost no way of 

 determining how much it is moving, because you have no fixed ref- 

 erence point. 



rUNCTIOX AND USE OF ACCELEROMETER 



As a consequence, we decided to measure tlie motion, using acceler- 

 ometers and, Mr. French, you suggested that it might be well to ex- 

 plain very briefly what an accelerometer is. 



I think, perhaps, the simplest or, at least, the most direct analogy 

 would be to point out that the inertial navigation systems used in sub- 

 marines today, and also in our second line of missiles, employ accel- 

 erometers to determine the distance that, and the course that, the vessel 

 or missile has taken. 



This accelerometer is actually a weight, and in the case of these 

 accelerometers that we used, the weight is suspended by very fine 

 weblike wires, and these wires, suspending the weight, detect the 

 slightest motion of that weight by straining the wires, and by this 

 strain they change their electrical resistance, so it is possible for us 

 to measure very slight motions of the mass with respect to the earth 

 from these wires. 



Now, this is the same principle that is used in inertial navigation in 

 missiles and in submarines; that if you continuously measure the 

 acceleration, and then by mathematical means — double integration, it 

 is called, but it is a mathematical procedure — you can then determine 

 from the slight accelerations wliat the actual direction and distances 

 are that are traveled by the object. 



Dr. Draper, utilizing this device first put it in an airplane, and flew 

 directly by this integrating accelerometer to Los Angeles — this was, 

 I think, about 10 years ago — and it came within 10 miles of Los 

 Angeles. 



This device has been improved so that nuclear submarines can op- 

 erate for several weeks under water, and come up within a fraction of 

 a mile, where they are using techniques such as this one. 



We did the same thing. We put three on the deck of the Texas 

 tower, and by recording the electrical output of these accelerometers 

 we were then able to compute the motion in terms of inches. 



This was done by double integration, and we did it in this case just 

 by hand mathematics ; we did not use any computers. 



So that simultaneous with this, we measured bending stresses, and 

 the like, and by correlating the bending stresses in the leg with the 

 motions that we got by mathematical analyses of the accelerometers, 

 we were able to relate the two, so then we could go directly to the 

 strain data on the leg and get the motions. This saved all mathe- 

 matical analyses, but until you find the correlation between the mo- 

 tions in the leg, the stresses in the leg, and the actual motion in the 

 leg, you cannot tell anything from the stresses. 



So once there was established this relationship, we were then able 

 to go rather quickly as far as the analysis was concerned. 



Mr. French. Did any of the design engineers object at that time 

 or disagree with the method that you proposed to use? 



Mr. Bre^ver. No, they did not. I believe my memory is correct 

 that Mr. Kuss, of Moran, Proctor, Mueser & Eutledge felt that an 



