Stewart and Poudrier 



You may recognize the quotation from Act II, Scene II, Hamlet. In the 

 new and growing field of Reliability Engineering this is known as 

 Failure Mode and Effect Analysis. The Aero-space and Weapons industries 

 have developed these activities to a point far beyond that presently 

 practiced on oceanographic instrumentation, and have developed a body 

 of information, literature, and doctrine concerning the very real 

 necessity of an adequate feed-back of information during design, 

 development and field operation of their equipment. Lacking the 

 needed failure reporting system, we were forced to rely on interviews 

 witli operating personnel and the small amount of information obtained 

 from equipment survey reports. In many instances of equipment loss it 

 was not possible to determine either the Failure Mode (which is the 

 technical term used in Reliability Engineering corresponding approxi- 

 mately to "what happened" or the Failure Mechanism, which corresponds 

 to why it happened. The effect was usually more apparent, i.e., the 

 survey operation was delayed until new equipment was available, or was 

 discontinued altogether. In a sufficient number of instances to 

 warrant corrective action, it was determined that the Failure Mode was 

 a parted cable, and tliat the Failure Mechanism was either accidental 

 two blocking of an instrument while hoisting in or fatigue failure in 

 the cable or terminator. The term "two-blocking" as used here is 

 indicative of faulty semantics as well as seamanship. In its usual 

 Naval usage it means to hoist an item until it is against the upper 

 block. To our field survey personnel it means to hoist the item 

 unintentionally against the block with such force that the cable is 

 parted and hence the instrument lost. 



The second step in formulating the approach was to look at the losses 

 as a part of the total system cost, which is the cost of obtaining 

 certain specific data. Included in this total are program planning, 

 cost of procurement of necessary instrumentation, salaries of field and 

 supporting personnel, cost of ship time in implanting instrument or 

 taking data while underway, and possibly the cost of the time expended 

 in attempting to locate and recover lost instruments. This leads 

 naturally into a subdivision of instruments into two categories, those 

 from which the desired data is obtained even though the instrument it- 

 self may not be retrieved, and those from which no data is obtained 

 until the instrument is recovered. Examples of the first are instru- 

 ments designed to be expendable and also those telemetering data to ship 

 or shore stations; of the second, self contained recording current 

 meters, etc. If the instrument is of the first type the desired 

 mission may be completed without its recovery. The cost of recovery 

 then becomes a separate item which may be weighed against the cost of 

 replacement with a new instrument. If an expensive "Fail-Safe" device 

 would be required and/or ship time used in locating the instrument a 

 greater risk of loss can be accepted. 



In the second type, more attention must be given to loss prevention 

 and to recovery, since the total program costs may be lost unless the 

 instrument and data are retrieved. A further subdivision of instru- 

 ments would be according to method of deployment. Some are towed at 



295 



