Related Articles:
ROI on PHM


Prognostic Terms

Glossary

 

Adjusting "Time at Failure" to "Time to Failure"
for
Critical Failure Modes Defined in the Integrated Systems Diagnostics Design Process

As previously stated, effective PHM cannot rely on prognostics alone but must be designed to a well integrated prognostics and diagnostics system.  Integrated Systems Diagnostics Design (ISDD) will provide the design capability to detect a failure at any level of a system with high confidence.  The issue is understanding what can be allowed to run to failure and what needs to have a maintenance action performed prior to failure.  Those failure modes that cannot be allowed to run to failure are typically those with high criticality as determined through a Failure Mode Effects and Criticality Analysis (FMECA).

Those of you who work with a FMEA, leaving out the criticality, are missing the point of even performing a failure mode analysis in the first place.

The ISDD is based on the diagnostics analysis using the eXpress diagnostics engineering design tool.  The primary purpose of the diagnostics analysis is to design and develop a PHM system that will support a platform that provides the required Operational Effectiveness along with high Availability and a cost effective Logistics Support system. As anyone who has been through this process can tell you, this can only be accomplished through a well planned and implemented Systems Engineering Process, and the core of this process in the ISDD.

Without a thorough understanding of the systems diagnostics capabilities, an effective PHM system cannot be developed.  This has been experienced too many times in recent attempts at PHM.  If the ability to detect a failure when it occurs, or is about to occur, it not understood with high confidence, the PHM system is operating with blinders on.  The ISDD provides the detection knowledge, and many more parameters, needed for a PHM system that will meet today’s requirements.


The emerging technology that can enhance the ISDD is Prognostics. It is well understood that the ISDD provides the design knowledge to detect a failure when it occurs, and also the ability to isolate to the lowest defined unit causing the failure.  We have always looked for means to predict a failure before it occurs. For many years now failure trend information has been collected to try to predict this incipient failure.  This reliability based trend analysis has been very ineffective in predicting a failure within any usable time frame.  It has helped position spares and other maintenance resources but that is about the limit of this prediction’s effectiveness.

With the development of the prognostics analysis, a selective failure mode can be monitored, and based on physics of failure algorithms, can be identified as going into an abnormal operation / wear pattern.  Based on the algorithms and information being received from this failure mode, a prediction can be made as to the remaining useful life of this component.  This is an important technology that will allow maintenance actions to be preformed prior to failure of selected components.  A point that needs to be made here is that, with todays, and even envisioned, technology, you cannot prognose a full system.  In

fact, due to technology limitations and high cost, prognostics needs to be limited to selected components with the primary focus on critical failure modes. The selection of the prognostics candidates needs to come from the ISDD process.  A good example of a program that jumped in too deep and far too early on “prognostics PHM” is a new technology fighter air craft. It was not until well into the program that it was understood that the air craft’s system as a whole could not be prognosed, or even diagnosed.  This could have been prevented with a well planned ISDD process in place.

The important point here is to understand the integration of prognostics with ISDD.  Figure 1 shows a typical reliability Bathtub curve showing the probability a unit’s failure over time.  Based on a diagnostics analysis alone, the probability of detection for a failure weighted with this curve would be based on a failure occurring at Tf.  If this is a critical failure as determined by the FMECA, operating to Tf could result in an operational / mission failure.  The typical severity parameters for of this type of failure are:
Probability of Loss of Operation / Mission (LOM)
Probability of Loss of Platform / Vehicle (LOV)
Probability of Loss of Life (LOL)
It can be readily seen that running a critical failure mode to Tf is not acceptable.  There are maintenance actions in place today to prevent a critical failure from failing a system: 

These “work around” actions have been used in the past because nothing else was available.  With prognostics coming into the picture, some critical failures can be prevented.  This prevention is accomplished by prediction a pending failure in a selected failure mode and performing a maintenance action prior to failure.  This prognostics action needs to be accomplished with the full understanding of the systems ISDD.  The bottom line is, if you cannot observe or classify a failure, you cannot effectively prognose it.

The ISDD process is established to integrate the Reliability Centered Maintenance based diagnostics design with the Condition Based Maintenance based prognostics analysis.  By using the results of the prognostics analysis, the ISDD can move the failure detection point from Tf to the Time of Anomaly Detection, Tad.  Tad now becomes the failure point where the diagnostics analysis provides the probability of detection parameter.  This new detection point is important when analyzing the diagnostics capabilities at the systems level. The results of the new systems analysis are then used to support the maintenance process in determining the time period where a maintenance decision must be made to prevent a unit from running to failure. This is critical when defining the capability of a PHM system.

image

Figure 1, ISDD Integration with Prognostics

The ISDD as integrated with the prognostics development provides the proactive PHM design that must replace the existing reactive PHM system development, otherwise modern systems will not meet their operations and support requirements. Working with new and future technologies is important to keep ahead of the technology curve, but this cannot replace a tried and proven technology that continues to advance within logical boundaries.  The ISDD process that supports the Operational and Maintenance requirements is shown in Figure 2 and is defined in the steps below this figure.