The reliability techniques used in the development of a fully solid-state control system for an off-shore oil-gas separator complex will be discussed. Three areas of reliability engineering will be considered:

  1. Component selection, test and derating

  2. Manufacturing techniques

  3. System design philosophy

Extensive data has been accumulated in the electronics industry in the area of component selection. Tables will be presented summarizing these data for electronic part failure rates.

Additional burn-in test, to eliminate "infant mortality" are discussed. Failure rates as a function of operating time have shown a very clear-cut indication that once electronics components pass a vigorous acceptance test and get beyond the infant mortality region as a result of burn-in or operating life, they seem destined for a relatively failure-free lifetime of service. in other words, most if not all of the reliability problem is, in fact, a qualityassurance problem. Failures experienced by the operational system, using burned-in devices, should be beyond the infant mortality region, and would be of the chance or random type. A table showing these failure rates will be presented. A characteristic of random failures is that their average rate of occurence remains constant with time; i.e., the mortality curve is a horizontal line as a function of time. As more operating time is accumulated, a point is reached when parts in the equipment start to fail from wear-out, and the failure rate curve then starts to rise. The portion of the life cycle that lies between the infant mortality and the wear-out regions is the "useful life" of the equipment.

The control system under consideration in this paper has been designed for a 20-year effective lifetime. Heavy derating of components, in addition to burn-in testing, enhances the effective Mean Time Between Failures (MTBF).

Manufacturing techniques used in production of this system were so exacting that the workmanship specifications and printed circuit fabrication specifications were actually the NASA Marshall Space Flight Center specifications for manned spaceflight hardware. These techniques will be discussed giving examples of elimination of potential failure modes inherent in off-shore applications.

The general system design philosophy is being presented in a joint paper. The reliability aspects of the design philosophy, are to be presented here. Significant features incorporated for the purpose of insuring system integrity include:

  1. All contact sense input lines have unique protection against inadvertent application of 440 volts, A.C.

  2. All outputs are short-circuit proof

  3. Logic Circuits have 15 volt noise margins

  4. Circuit design considering the mode of failure for circuit components

redicting the most probable mode of failure for a component can play a significant role in the design of a system, producing a virtually "Fail-Safe" design.

An example of this technique will be shown considering the use of resistors as circuit components. The carbon composition resistor and the metal-film resistor have, at appropiate quality levels, quite similar failure rates. The failure modes, however, are diametrically opposite; i.e., the carbon composition resistor typically shorts when it fails; the metal-film resistor typically open-circuits upon failure.

Armed with this knowledge, the designer used the appropriate device such that a device failure causes only false alarms.

This content is only available via PDF.
You can access this article if you purchase or spend a download.