The Reliability Development Test (RDT) is a planned Test, Analyze, and Fix (TAAF) process in which development items are tested under actual or simulated mission profile environments to disclose design deficiencies and to provide engineering information on failure modes and mechanisms. The purpose of RDT is to provide a basis for early incorporation of corrective actions and verification of their effectiveness in improving the reliability of equipment.
The RDT by itself, however, is not the most efficient or economical means of achieving acceptable reliability. Proper emphasis must be placed on design fundamentals such as derating, stress analysis, thermal analysis, and failure mode and effects analyses so that a potential for reliability is designed into the equipment prior to the start of RDT. Reliability growths will then result when positive changes are made to the design to correct problems identified during test.
The RDT is conducted under controlled conditions with simulated operational mission and environmental profiles to determine design and manufacturing process weaknesses. The RDT emphasizes reliability growth rather than a numerical measurement. Unlike the RDT, the Reliability Qualification (demonstration) Test (RQT) is not designed for reliability growth and has proven to be ineffective for improving equipment reliability. The RQT normally is performed too late to provide major impact on the design effort, and if corrective actions are required, they seldom are incorporated and verified before production because of the need to meet delivery schedule commitments.
The RDT, using the TAAF process, is a key requirement to achieving acceptable system reliability. The RDT, however, must be tailored to the needs of the specific project and must be integrated with other development test activities to provide for the more efficient use of test resources. For example, the temperature and vibration portions of the design limit qualification tests could be used for the initial portion of the RDT. Corrective actions developed as a result of other tests should be incorporated in the RDT test units to verify their effectiveness and to prevent unnecessary duplication of failure analyses and corrective action efforts.
The efficient use of test resources also requires that reliability growth test emphasis be placed on those equipment that will have the most impact on system and mission reliability. Selection of equipment for RDT should be based on consideration of reliability allocation and prediction, state-of-the-art, equipment similarities, and complexity. RDTs should be performed on subsystems of low (predicted) reliability. High (predicted) reliability subsystems not selected for RDT should be evaluated during the system-level development testing. If any of these subsystems exhibit problems during the system development tests, then suitable corrective action can be identified and incorporated to preclude the recurrence of the problems, or additional RDTs can be considered.
The RDT must be monitored and kept flexible to allow for changes as the reliability database grows. When reliability data indicates that further testing will produce only insignificant changes in reliability, the RDT should be terminated. Early termination of one subsystem RDT will permit test resources to be applied to other subsystems where additional testing is expected to provide significant reliability improvement.
Reliability growth during RDT is the result of an iterative design process. Equipment is tested to identify failure sources and further design effort is spent to correct the identified problems. The rate at which reliability will grow during this process is dependent on how rapidly the failure sources are detected and how well the redesign effort solves the identified problems without introducing new problems. It is essential, therefore, that periodic reliability growth assessments be made and compared with the planned reliability growth values. These periodic assessments will provide visibility of achievements and will identify deficiencies in time to affect system design.
Prior to initiation of an RDT, the design reliability
should have advanced to such a stage that the predicted MTBF is at least 1.25
times the required MTBF. To estimate the amount of time for conducting the
RDT, a plot of MTBF versus time can be constructed on log-log paper. An
initial starting MTBF estimate for a low risk project of 30 percent of the
predicated MTBF may be used with lower values for higher risk projects (as low
as ten percent in some cases). A growth slope of 0.5 or less should be used,
with lower slopes for less aggressive reliability projects. Figure 1 shows
examples for both low risk and high risk projects. To monitor satisfactory
progress, the actual growth curve should be compared to the ideal 0.5 slope
and additional emphasis placed on failure analysis, corrective actions, etc.,
for significantly lower-growth slopes.