Integrating RCM with Effective Planning and Scheduling - Part 1
This paper was presented to the West Australian Optimising Maintenance Conference, Perth, Australia on 14 May 1999
Author : Sandy Dunn
Go to Part 2
Plant Maintenance Resource Center Home
For many organisations, there is a collective sigh of relief once they have completed their RCM analysis - at last this job is over! However, the reality is that the job is, at best, half complete. A lot of work remains to be done to successfully implement the RCM decisions, and many RCM projects fail because they do not successfully complete the required implementation tasks. This paper outlines the key activities that must be performed in order to successfully implement the RCM decisions following completion of the RCM analysis, and integrate these with an effective planning and scheduling system. Some of these activities are technical in nature, but the most important ones deal with people and administrative issues.
The outputs of RCM
The first thing to note about an RCM analysis is that the key output is a collection of recommended actions, each of which is focused on a specific failure mode. These actions can be classified into three groups.
- Routine things we want maintenance people to do - these could include routine inspections, functional tests, overhauls, services - indeed anything that needs to be done on a routine basis, which requires trades skill in order to be successfully performed.
- Routine things we want production people to do - these could include routine inspections, minor maintenance tasks or functional tests - things that need to be done on a routine basis, but are tasks which the production people are in the best position to conduct, and have the necessary skills to successfully perform.
- One-off actions - these could be performed by a wide range of people, and could include such actions as redesigning or modifying an item of equipment, modifying a standard operating procedure, modifying a standard maintenance job procedure, retraining operators or maintainers etc.
Successful RCM implementation requires that all of these recommended actions are completed.
The key thing to note is that the first two types of actions are routine - they are required to be performed at a regular, pre-determined frequency. However, in arriving at the proposed frequency of the task for each failure mode, we have considered the failure mode in isolation. (Or at least we should have, if we have truly performed our RCM analysis from a "zero base"). Before we can implement the proposed routine tasks, we must first group and optimise these tasks, taking into consideration their interaction with each other, and the impact of the tasks on Maintenance Labour productivity and equipment downtime.
For example, our "zero-based" RCM analysis, may have suggested that, for a gas turbine, that an internal inspection be carried out for blade cracking every 10,000 operating hours. We may also have another task which is to drain and replace the turbine's lubricating oil every 8,000 hours. Both of these tasks require the turbine to be shut down and isolated. The question is, should we perform both of these tasks at the same time? And if so, what should the frequency be - 8,000 hours, 10,000 hours, or something in between?
Let's examine the steps involved in creating maintenance schedules from the resulting RCM decisions in more detail.
Creating Maintenance Schedules
The objectives of compiling Maintenance schedules are:
- To "faithfully" translate and package RCM tasks into Maintenance schedules
- To minimise the impact of Routine maintenance tasks on equipment availability
- To maximise labour resource productivity and utilisation, and
- To provide documented Equipment Maintenance Strategies, with an audit trail.
The first point cannot be overemphasised, and raises questions about the type of person, and the skills required, to compile the maintenance schedules resulting from an RCM analysis. On some occasions, RCM implementations have failed as a result of a person being given the task to compile the maintenance schedules from an RCM analysis, when that person:
- Has no knowledge of RCM principles
- Has not been involved in the original RCM analysis, but
- Was responsible for setting up the maintenance schedules currently in place.
The result is generally that the maintenance schedules do not faithfully reflect the intentions of the original RCM decision team, but do reflect the particular biases and opinions of the individual involved (which generally closely resembles the status quo).
The steps involved in compiling the maintenance schedules are:
1. Identify constraints
This includes identifying any operating cycles and "windows of opportunity". If, for example, for a shiploading facility, you know that you can rely on having at least one 12 hour period per week where no shiploading will be taking place, based on forecast shipping schedules, then this creates a "window of opportunity: to do some maintenance. Other opportunities may be presented through the seasonality of production requirements, or through the nature of the production process (batch vs continuous process, for example).
Other constraints that you may need to consider include limitations in the availability of people, skills, facilities and workshops, equipment, and the capabilities of your CMMS. If, for example, your CMMS cannot handle schedules of longer, or shorter, than a particular frequency, then you may either need to make some changes to your CMMS, or alter the way in which you manage and control those schedules.
2. Sort and Assess tasks
In this step, you should initially sort the tasks from your RCM analysis:
- by trade type (mechanical fitter, electrician etc.)
- by equipment operating requirement (the equipment must be shut down to do the task, or must be running, or can be done with the equipment either running or shutdown)
- by RCM task frequency
This is the first step in optimising the schedules. Computer tools can assist with this step, and many RCM software packages, including RCM Toolkit from Aladon, perform this task.
3. Assess tasks for "rolling up"
After the initial sort, some tasks will become obvious candidates for combination. For example, if, for one failure mode, a task was identified to be done on a daily basis which was to:
"Check that No1 turbine oil temperature does not exceed 130 degrees C"
and another task was identified for another failure mode which was also to be done weekly which was to:
"Check that No1 turbine oil temperature is not less than 110 degrees C"
then the obvious combination task would be to:
"Check that No1 turbine oil temperature is between 110 and 130 degrees C"
4. Establish Schedule Intervals and Cycles
You can now group tasks together, and establish the appropriate frequency for each group of tasks. It is preferable that the schedule frequencies be multiples of one another (eg. 250 hours, 500 hours, 1000 hours, 2000 hours etc.), however often there will be "rogue" tasks that don't fall neatly into the appropriate grouping. For example, see the chart below.
In this example, you can see that the tasks do not all neatly fall into frequencies which are multiples of one another. How do we resolve this situation?
In general, you should:
- Shorten the task frequency for those tasks which have safety or environmental consequences (including those hidden failures where the ultimate consequences of the multiple failure are an impact on safety or the environment. This will ensure that the specified minimum safety standard is met.
- Conduct a cost-benefit analysis for those tasks which have operational and non-operational consequences. This is a bit more involved, but in general, the following rules-of-thumb apply:
- Never extend the task frequency for scheduled restoration and scheduled discard tasks - a requirement for the selection of these tasks is that the probability of failure rapidly increases as the task frequency is extended. As a result, extending the task frequency will rapidly increase the failure costs.
- For On-Condition tasks, extending the task frequency increases the risk that the impending failure will not be detected prior to the failure. This risk increases proportionally with the task frequency. For example, increasing the task frequency to twice the PF interval reduces the probability of the failure being detected to 50%. Increasing the task frequency to 3 times the PF interval reduces the probability of the failure being detected to 33%. The cost benefit analysis must take this into account.
- For Failure Finding tasks where the ultimate consequences of the multiple failure are economic in nature, the standard Failure Finding Frequency calculation can be used to optimise the task frequency - all that needs to be modified is the cost of performing the Failure Finding Task.
Once again, there are computer packages that will assist you to make these optimisations, Typically, they will require information regarding:
- The Beta factor for the failure mode (it is worth considering where you will get this from, and how accurate it is likely to be!)
- The downtime associated with the failure being prevented
- The cost of downtime associated with the failure being prevented
- The frequency with which the failure being prevented would occur, if it were not being prevented
- The downtime associated with performing the scheduled task
- The cost of performing the scheduled task
- The frequency of performing the scheduled task
Typically, these packages cannot take into account:
- The potential to take advantage of "windows of opportunity"
- Any constraints that may exist when performing tasks simultaneously (such as access to work areas etc.)
These tools are helpful in giving guidance to the optimisation process, but have functional limitations (as outlined above), and also rely heavily on accurate quantitative data (such as Beta factors and MTBF's) that may not exist.
On occasions, tasks may still remain that do not neatly fit into the ideal packages of work - these will need to be managed as "special" schedules.
5. Sequence Tasks in each package
Once the groupings of tasks have been established, they should be sequenced so as to make the most productive use of a tradesman's time. This involves:
- "Flowing" tasks in a sequence to minimise travel time
- Maximising the use of available time
- Avoiding conflicts with other schedules (for example, when a mechanic and an electrician may both need access to the same item of equipment)
6. Write the Work Instructions
This is a fairly standard step - however, if the full benefits of the RCM analysis are to be obtained, attention must be paid to the following points:
- Make sure the task descriptions are "faithful" to the RCM decisions
- Spell out exactly what is to be done
- Where the task is an on-condition task, ensure that the standard associated with the task is clearly specified, and describe the action to be taken if the item fails the inspection or test.
- Use consistent terminology and terms and phrases
- Make sure task descriptions are concise and readable
- Use detail that is consistent with the skill level of the person performing the task.
- Embellish tasks with additional actions or information
- Add tasks of opportunity
Typical mistakes made with task descriptions are as tabulated below:
Examine feedscrew coupling rubber for deterioration. Replace if deterioration evident.
Fit 0-20 bar test gauge to compressed air system test point 3 and check reading on Operator's Panel Gauge PI204. If reading variation is greater than 5%, replace gauge and repeat check.
7. Upload the tasks, frequencies and descriptions to your CMMS
This may require some significant data entry work.
Much of this discussion has concentrated on the technical tasks to be performed to compile and optimise the schedules, but, as we will see later, this still does not ensure that the new schedules actually get performed.
Go to Part 2
Copyright 1996-2009, The Plant Maintenance Resource Center . All Rights Reserved.
Revised: Thursday, 08-Oct-2015 11:54:32 AEDT