Reliability Growth Plan By André Vondran

A Reliability Growth Plan is a systematic development approach for achieving the product reliability requirements. Reliability Growth occurs by means of evolving and implementing product design improvements during the development phase. 

The plan for achieving and monitoring the growth is focused on first detecting design problems and then determining the root cause of the problem. Next, corrective actions to eliminate the root cause are developed. 

Then the design is revised to incorporate those corrective actions and new parts put on test to evaluate their effectiveness. 

This "design-test-fix-retest" cycle is continued until the reliability goals established for the program have been achieved.

Purpose: The Reliability Growth Plan is the basic structure under which the success in achieving the Reliability Goals is monitored and demonstrated. The reliability growth achieved at any point in the program provides a continuing measure of the product development throughout the CV-DS Gates of Planning, Development and Validation. The status of the reliability growth is one of the essential inputs to the Risk Assessment at the end of each Gate.

Requirements: This task must be applied to all programs using the CV-DS discipline. Its scope will be determined during the P6 Gate and will be subject to scalability considerations.

Prerequisites: The key inputs for this task are the Reliability Goals, the Reliability Allocation and the Design Verification and Product Validation Plans. Available resources, schedule requirements and CV-DS scalability levels are also a necessary consideration as inputs to establishing the Reliability Growth Plan.

Organization

Responsibility: The Reliability Engineer assigned to the CFT is responsible for leading the activity of creating the Reliability Growth Plan. 

The Project Manager is responsible for the performance of the Reliability Growth Program. 

The determination of the numbers of product to be tested and the schedule for performing this testing are developed with the participation of the Project Manager, Product Line General Manager and the appropriate members of the Cross-Functional Team (CFT). The Reliability Engineer is responsible for carrying out the monitoring of the reliability growth activity. 

The Reliability Engineer also performs a continuing process of coordination with the CFT to:

· review results,

· establish areas requiring corrective actions,

· spur the determination and implementation of necessary corrective actions,

· set criteria for evaluating the effectivity of the corrective actions.

Timing: This task begins during the Planning Stage as defined in CV-DS and continues through the Development and Validation Gates. It must be completed prior to the end of the Validation Gate so that its final conclusions can be provided for the Production Readiness Risk Assessment performed at the completion of the P2 Gate.

Procedure

Steps: A Reliability Growth program is used as the most effective approach to:

· Provide a management tool to identify the need for resources and to effectively allocate those resources.

· Provide a methodology for monitoring the reliability achieved at sequential design stages throughout the hardware development, both at a system and at a component level.

· Define a continuing set of targets and measure the progress being made toward their achievement.

· Provide detailed feedback as to where progress is unsatisfactory so corrective actions can be directed to deal with the problem area.

· Act as the framework for problem identification, for incorporation of design or process corrective actions and for assessing the effectiveness of those actions.

The reliability growth plan normally used at DDC is based upon Duane's Model [ 1 ]. For development systems being closely monitored during a pre-production test program, this model assumes the observed cumulative failure rate follows a consistent pattern when plotted on a log-log scale as a function of cumulative operating time. In general, Duane observed this growth curve to be a straight line in such a plot, indicating it can be expressed mathematically as:

Ln ( MTBF ) = b ( ln t ) - ln K,

where,

MTBF

=

mean time between failure of the system (which is the inverse of the cumulative failure rate at time "t" ).

t

=

cumulative test time

K

=

a constant which expresses the initial reliability of the system at the beginning of the development program.

b

=

rate of reliability growth ( i.e. the slope of the growth curve ).

Duane suggests this relationship will continue to apply to the system throughout its development, as long as active corrective action programs are in place to improve the reliability by analyzing failures, correcting the failure modes responsible and retrofitting the systems being tested with corrected hardware. In an early program [ 2 ], this model was shown to fit a typical automotive design and development cycle and it has been used successfully at DDC and elsewhere in the industry.

The first step in the development of the reliability growth model for a new program is to select the slope of the growth curve. The considerations used to carefully select this slope value "b" were developed in the referenced paper. Later, in the first program at DDC to actively use this method, the appropriate slope to optimize the test program was chosen to be in the 0.2 to 0.4 range. Following the selection of the appropriate slope for the system to be developed, the next step is to establish the time parameter. In DDC programs, this is normally test miles or test hours, depending on the application involved. Next, a figure for the initial reliability of the system is assessed. In most cases, the system to be developed is based upon a prior system that can be characterized by a demonstrated reliability. This can be used as a basis for the initial reliability of the new system and can be modified by the degree of difference between the new system and its predecessor. The final parameter that defines the growth curve is the reliability goal established for the new system.

These data are all incorporated into the log-log plot of MTBF vs. Cumulative Test Time. The sample size is then determined graphically from this plot as a total cumulative test time required to achieve the desired reliability growth. This in turn is converted into the number of individual test systems that must be committed to the reliability growth program. This is a function of the development program schedule and the rate at which each system on test can accumulate the necessary hours or miles per week.

Once the reliability growth program is underway, the reliability task moves to monitoring the progress of the reliability growth and working with the CFT to pursue necessary actions. This involves the collection of test times for each system involved and documenting each component failure. For purposes of tracking the reliability growth, the log-log curve is usually replotted as a linear plot relating test time and system reliability (R), rather than MTBF. This curve uses the mature system goal established earlier to define the reliability target that must be demonstrated at the start of production in order to achieve success in the development program. At any point in the test program, the reliability achieved for each component of the system, as well as for the system, can be compared to the target (expected) reliability for that component or the system. This will determine if the development is progressing satisfactorily or if additional work is needed in a given area.

Three primary areas of test experience are typically used in a formal reliability growth program. For a on-highway truck program, for example, the first of these is dedicated field tests. These are systems installed in a regular DDC customer truck and used by this customer in his normal business operation. The second of these are systems installed in DDC owned trucks and operated by a contractor over a selected route. The third test experience are systems that are mounted in instrumented test cells and operated over a simulated on-road operational cycle. A fourth area of test experience involved in some cases, are individual system components that are run on dedicated test beds to gain additional experience specific to that component. These may be DDC or supplier testing facilities.

The status of the reliability growth in a typical program is usually presented in chart form. This allows a comparison of the demonstrated values of reliability with the expected values at this point in the schedule. This comparison can then focus on those areas exhibiting a problem and can be used to direct program resources accordingly. The inputs to this assessment activity are detailed incident reports for every test system experience. Failures are documented by consistent with the Closed Loop Corrective Action system. Other incident reporting captures status information (accumulated hours or miles, routine maintenance actions, changes in hardware configuration, etc.) for each separate hardware system involved in the test program. In many cases, some components in an individual test system may be replaced during the test sequence by an updated part that incorporates changes made as a corrective action for a failure mode. Therefore, it is necessary to maintain the record of hours or miles down to a component level.

The final consideration in a Reliability Growth Program is to establish a basis for "purging of failures". This involves discounting past failure modes considered as addressed and therefore eliminated. One approach assigns partial or complete credit for the fix at the time the failure mode is identified and the corrective action determined. Another approach does this only when the corrective actions have been put into effect. The most meaningful approach takes credit for the fix only after the new hardware demonstrates the failure mode is eliminated, usually at a 90% confidence level. This requires that the systems in the Reliability Growth Program be retrofitted with the new hardware and sufficient test times accumulated after the retrofit to provide the demonstration. Calculating the test time required for the 90% confidence level demonstration is based on the failure rate of the test systems before the fix was applied. This is done for each failure mode for which a fix is being evaluated.

Final Output: The output of this task is the formalized, documented Reliability Growth Plan. It may be a separate publication or integrated into the Reliability Plan for future programs.

Conclusion

Reliability Growth is the method for evaluating and improving the reliability throughout the development of a new product and is the most essential tool for a successful program. The Reliability Growth Plan must accomplish this in a realistic, but uncompromising manner.

Best Regards · Mit freundlichem Gruß · Cordialmente · 

André Vondran

Think Green! Please don’t print if it’s not necessary.

Follow Me on Instagram