Condition Monitoring: Need or Luxury?

Do I need Condition Monitoring? Is it a Need? Is it a Luxury?

To answer this question let us first go back to understanding “What is Condition Monitoring?”

Condition Monitoring is a process of scheduled data collection and evaluation to detect changes in performance or condition of a system or its components, so that a rectification action may be planned well in advance to reduce the maintenance cost and to maintain equipment/system reliability.

Let’s look at how failure patterns and statistics illustrate the necessity for condition monitoring and predictive maintenance strategies.

The traditional view of failure is that assets and their components wear out based on a predictable time interval.  Failure statistics, however, show that this is only the case for a relatively small percentage of the asset population (Fig-1).  This study, originating in the airline industry, shows that only 11% of failures exhibit wear-out zone patterns where time to failure based on run-time or calendar time can be predicted with reasonable certainty.  For these assets, a time-directed maintenance or repair strategy may be the most effective strategy.  

The majority of assets, 89% as shown in this study, show no predictable wear-out zone based on time in service.  A large percentage of these assets exhibited an infant mortality pattern, meaning many failures may have been introduced by performing unnecessary maintenance or an original design or manufacturing defect.  For these assets, a predictive maintenance strategy is the most effective.  A condition monitoring and predictive maintenance strategy should be therefore be selected as the first line of defense for these assets when designing and developing a maintenance and reliability program.

Once a failure mode initiates, the remaining time to failure usually follows a predictable pattern.  The P–F interval is the time interval between when a Potential functional failure mode can first be detected and when actual functional Failure occurs (Fig-2).  The P-F interval concept is used to set data sampling intervals so there is adequate time to plan and schedule corrective action before actual failure occurs.  Different types of failures have different P-F intervals.  For example, rolling element bearing failure may have a P-F interval of approximately 2 months (depending on the application).  A common rule of thumb is to set the sample interval at no more than half the estimated P-F interval.  This will allow the majority of failures to be detected early enough for corrective measures to be planned and scheduled while minimizing impact to the business.

On-line continuous and scanning monitoring systems eliminate possible error sources associated with identifying the best correct P-F and data sampling intervals (Fig 3). On average, continuous and scanning methods also enable earlier detection of potential failures than what can be provided with periodic sampling.  As shown above, failures are detected at point “P” with on-line systems, effectively eliminating the time lag associated with periodic data samples.  For critical and essential assets where the risk of a missed failure is not tolerable, on-line continuous or scanning condition monitoring methods and the additional diagnostic capabilities they provide are frequently justifiable.

Now let’s talk about the value of condition monitoring and predictive maintenance to an organization.  This graph describes several relationships between production, maintenance costs, and several types of maintenance and reliability strategies.  Many companies are faced with economic pressures that demand reductions in costs in order to remain competitive.  Unfortunately, a frequent course of action is to cut maintenance cost expenditures without any change in strategy.  Using the example in the graphic by moving from Point A to Point B (Fig 4), reducing preventative maintenance costs alone may achieve the desired results, but this achievement will be short-lived.  It won’t be long before equipment reliability and plant throughput will be negatively impacted. More importantly, the safety of plant personnel may be jeopardized due to unreliable equipment.

The challenge is to cut costs smartly with an approach that incorporates higher level maintenance and reliability strategies.  As shown in the curve defined by points A, B, C, and D, predictive and proactive maintenance strategies can help improve equipment reliability, availability, and plant production, while simultaneously reducing maintenance cost expenditures (Fig 5).  Case histories have shown that cost reductions up to 40% are typical, and sometimes even greater returns have been realized (Fig 6).  Condition monitoring and predictive maintenance programs are absolutely critical to enabling these higher level strategies and driving these results .

RTF –stands for Run to Failure and PCM – stands for Proactive Centered Maintenance in the EPRI study or “engineering out failure modes”

The first requirement in developing a condition monitoring program is a complete and accurate equipment data library.  This includes a master equipment list with the class and subclass of the equipment identified and supporting nameplate data.  Corresponding field tags on the equipment will also come in handy when configuring routes and sample points.

After the equipment data library has been developed, all the equipment must be ranked by relative criticality in terms of safety, the environment, production, maintenance costs, and product quality.  Weightings for each ranking criteria are determined based on input from team members representing key plant functions, and then the ranking of each asset is determined using consistent, balanced criteria.

The next step is to assign technologies to assets based on criticality and failure modes.  This will include development of an on-line condition monitoring plan.  Failure mode and effects analysis (FMEA) methods are used to determine what technologies and measurement points are needed to detect potential failure early.  For portable periodic sampling, PdM tasks are scheduled based on technology assignments and load-leveled resources allocations.

The next step is to address the human resources that will be required to support the technology plan.  Whether resources are in-house, outsourced, or a combination of the two, clear definition of roles and responsibilities must  be communicated, and training requirements must be met to support needed skills.  With new roles and responsibilities, and a new way of directing maintenance work, the importance of successfully engineering culture change cannot be overlooked.  Culture change methodology should be employed in parallel with the technical program to ensure top-down commitment is achieved and sustained.  

Once the technology and human resource plans have been developed, the methodology plan must be developed and implemented.  Policies and procedures that will direct the program need to be developed and put in place.  This will include new work and information flow processes that will reflect a revised organizational structure and the new roles and responsibilities.  Last but not least, key performance indicators must be developed and put in place to measure ongoing success of the program.  These KPI’s should include a mix of leading and lagging indicators.  Leading indicators, such as PdM task compliance can be achieved early in the program.  Lagging indicators, such as % of work resulting from predictive maintenance, will show results as the program begins to make an impact.