Scheduled Maintenance
This topic covers:
The criteria for determining scheduled maintenance requirements for an item relate to the decision on what maintenance tasks and processes will be performed and influence the decision of when they will be performed. A range of information must be established about each item or system under consideration. This information must be systematically evaluated using specific decision criteria combined with the knowledge and experience of the analyst.
The reliability characteristics of an item are potentially the most significant criterion for determining scheduled maintenance requirements, since they form the basis for determining whether scheduled maintenance is potentially effective for the item. If the characteristics of an item are such that scheduled maintenance cannot be effective, a task should not be included in the scheduled maintenance program.
The second major criterion for determining scheduled maintenance requirements is the effect of failure of an item on safety, environment, function, cost and/or secondary damage. This criterion can only be considered after the existence of an effective task has been established through an evaluation of the failure characteristics of an item. If the failure of an item does affect safety, environment, function and/or cost, but no effective maintenance task can be performed to prevent or delay the failure, inclusion of a task in the scheduled maintenance program serves no useful purpose. Under these circumstances the answer does not lie with maintenance but in design changes or modification action to alleviate or eliminate the undesirable failure effect.
These basic principles govern the sequence of application of the criteria for determining the scheduled maintenance requirements. The intent is to formalize the essential considerations so that the criteria for determining scheduled maintenance requirements are applied consistently thereby ensuring consistent conclusions.
In general, the following items will be analyzed:
- All items for which scheduled maintenance has been defined in existing schedules
- Any other item not included in the existing maintenance program but which warrants consideration because of its complexity, criticality, failure rate or other reason. In many cases, these items will be identified during detailed examination of other items/systems and should be added to the list of items for analysis as they are identified.
The first form used in the analysis process is the System Criticality Analysis Form. The form serves as a permanent record of the plant breakdown into functional systems and sets the priority, degree of effort, and details for all subsequent analysis.
An item is a specific functional unit, which may incur costs for maintenance, restoration or replacement and which may be distinguished in the asset register. (This could be an area, a part of an area, or specific piece of plant or equipment).
The second form in the analysis process is the Item Criticality/Failure/Maintenance Analysis Form. A separate analysis form is raised for each item within the system being analyzed, with the exception of grouped items. A number of situations arise where several identical items require analysis. These items may be in the same system, may perform similar functions in different systems or may be part of multiple installations. Examples of this situation are:
- Multiple circuit breakers in an electrical power distribution system
- Wiring and hydraulic plumbing
- Multiple pumps in a fluid transfer system
- Identical fans in a ventilation system.
In such cases, it is sometimes both convenient and efficient to analyze a group of items, rather than individual items. However, the use of this technique is dependant on whether the items concerned can be regarded as the same for maintenance analysis purposes, bearing in mind their actual function, the consequences of failure of each function and the operating conditions for each item. In some cases, variations of these factors between systems may generate the need for separate analysis.
The Item Criticality/Failure/Maintenance Analysis Form comprises three distinct sections:
- Identification and Description
- Failure Modes and Effects
- Maintenance Analysis.
Each section is further subdivided. All sections are to have information recorded in them. Where sections have alternative selections, one or more of the selections are to be completed as appropriate.
The reliability characteristics of each "item" selected for analysis (i.e. an item, a group of items or a system) are determined from:
- Failure data
- Design characteristics
- Maintenance documentation held at the site
- Breakdown reports
- The experience of technicians
- Any other reliable information such as manufacturer's data, similar items used elsewhere
- The analyst's experience with the type of item being analyzed.
The first step in determining the reliability characteristics is the establishment of the number of item failures over a selected period. If the failure rate is low then this fact should be recorded for the task evaluation stage. On the other hand, if the failure rate is substantial, the various likely failure modes must be identified (i.e. the ways in which the item is likely to fail) to enable the effectiveness of each maintenance action on each failure mode to be evaluated.
The failure modes should be described briefly in terms of the actual fault that is likely to occur (e.g. worn brushes, hydraulic oil leaks, corrosion, particular part failures). In addition to these failure modes, which are derived from the item's failure history, possible failure modes typical of the item being considered should also be identified for consideration during subsequent task evaluations.
For each failure mode any failure patterns and/or failure progression rates must be identified. The available data should be examined to determine whether:
- The failures occur at random, i.e. different parts fail at various times, or the failures depend on operating/elapsed time or after a number of events such as starts or liters pumped, i.e. a predictable wear out pattern exists
- The item starts to deteriorate immediately it is put in use or after some maintenance process, or after some initial delay (e.g. corrosion)
- The deterioration progresses, slowly, fast or the rate changes during the item's useful life.
Scheduled maintenance could be effective if the item under consideration has likely failure modes but if likely failure modes do not exist then scheduled maintenance would be ineffective. However, any existing scheduled maintenance tasks must still be evaluated to determine whether some of these tasks are actually preventing failures and hence causing the low failure rate. Obviously, these effective preventive tasks, or more efficient alternatives, would be retained in the maintenance program.
This section covers the examination of each existing maintenance task to assess its effectiveness in preventing, delaying or detecting one or more of the identified failure modes. Failure modes that are not covered by the existing maintenance actions are identified and the appropriate additional maintenance tasks defined. The evaluation process leads to a set of recommendations regarding the tasks which should be performed.
The final step in the evaluation of maintenance requirements is the determining of the maintenance interval at which an effective task should be performed. The process of determining the correct maintenance interval is imprecise because of the differences in equipment reliability performance and a lack of complete knowledge about the item in some instances. In general, the assessment of the interval will be of a subjective nature based on the analysis of the failure data and the characteristics of the item. However, the interval for each task must be consistent with the cause of the deterioration (i.e. operating hours or calendar time) and the rate at which the deterioration occurs.
When very few failures are reported, the possibility that the item is inherently reliable should be also considered.
The main criterion for this assessment is a comparison of the rate of unscheduled occurrences and scheduled discoveries. No strict guidelines can be set concerning the proportion of unscheduled to scheduled occurrences that must be experienced before a task may be classified ineffective. Generally, a task is considered ineffective if the number of unscheduled failures exceeds the number of failures discovered during scheduled maintenance. However, the following possibilities should also be considered:
- The frequency of the task may not be aligned with the parameter causing the failure, i.e. operating hours or calendar time
- The task may be detecting a particular failure mode but failures due to other causes may far exceed those failures that are being detected
- The intent of the task may be such that a high rate of failure detection is not expected. This would occur in the case of:
- General condition or system monitoring
- Confidence checks/tests of hidden function
- Specific condition monitoring tasks to detect a failure appearance and then the associated failure progression rate for high availability equipment.
- In these cases the task cannot be considered ineffective simply because the rate of detection is low.
The condition monitoring process consists of finding a failure appearance and then by means of additional surveillance, predicting when failure is likely to occur so that preventive or corrective maintenance can be applied before total failure takes place.
The effectiveness of a specific condition monitoring process may be difficult to determine from failure data as:
- The failure appearance could be reported as discovered during scheduled maintenance
- The additional surveillance tasks for determining the failure progression rate may be attributed to a special maintenance task
- The preventive/corrective action would be reported as an unscheduled occurrence although it would have been pre-planned.
The failure data, therefore, should be examined carefully for related events and the maintenance staff consulted to determine whether the process has been effective in predicting likely failures.
Evaluation of the demonstrated effectiveness of current tasks should provide an indication of the worth of the task. However, in all cases the analyst must consider whether there is an alternative, more efficient means of accomplishing the intent of the task, i.e. of achieving the same results. In addition, the possible need for an additional task to prevent or detect failure modes assessed as likely, but not covered by an existing task, must be considered. In either case a new task my need to be identified on the analysis sheet.
In many instances when considering what maintenance is most appropriate a number of options will be available. Generally it is preferable to prevent a failure than to delay it or predict it. Conversely, prediction may be a more cost effective action. The analyst still quickly realizes which is the best option with each specific failure.
The basic question to be answered in evaluating the potential effectiveness of a task is: could this task provide an effective means of preventing or detecting one or more of the likely failure modes identified for the item? The following factors must be considered in making the evaluation:
- Preventive Tasks. As the intent of the selected task is prevention, does the item exhibit a failure mode that is capable of being prevented e.g. does it exhibit some wear/out-of-tolerance characteristic or require lubrication? In many cases, the answer to this question will be NO since the majority of items are not responsive to this form of maintenance. However, the fact that a failure mode does not appear in failure history should not be regarded as final evidence that it will not occur. The possibility that the current task is highly effective, and is preventing failures, must always be considered.
- Monitoring Tasks. When the task is a monitoring type, two basic questions must be answered. They are:
- Could the completion of the task lead to the detection of one or more of the specified failure modes? In many cases, the task may be simply incapable of detecting the type of failures that are likely to occur, e.g. an external examination of electronics. Alternatively, the methods applied may not be sufficiently sensitive to detect an impending failure or actual failure.
- Could relatively infrequent application of the task be effective in finding a fault? In many cases the failure progresses rapidly from the point where it is just noticeable to the time when a complete failure occurs. Therefore, most impending failures would not be detected by infrequently scheduled tasks and the operator would have reported the complete failure unless a hidden function was involved.
The consistency with which the tasks/processes may be expected to produce results must also be considered. A task or a process, which does not have a reasonable chance or success, i.e., one that might work, must be considered very carefully before it is assessed as effective and consequently considered for inclusion in the scheduled maintenance program. The consequences of failure may be taken into account where the task appears to be marginally effective for a safety or environmentally critical item.
Increasingly modern equipment is equipped with built-in monitoring systems that provide general monitoring of the equipment. For maintenance analysis purposes this monitoring may be considered as “alternative tasks”. When evaluating the potential effectiveness of system monitoring against specific monitoring tasks the following should be considered:
Scope of Monitoring
The monitoring should have the same intent and cover the same parameters, items or systems as the proposed scheduled tasks. If, for example, the intent of the task is to detect impending failures then a monitoring system that indicates an occurrence of failures would be inadequate. Similarly, if the failure mode in question related to overheating, then the monitoring system should measure the coolant temperature or flow rate.
The Standard of Monitoring
The accuracy and reliability of the monitoring system should be considered to ensure it is adequate. In some instances, the monitoring system itself may require preventive maintenance before it can be relied on to replace scheduled specific monitoring tasks in the operating equipment. In that situation the relative efficiency of the two scheduled maintenance requirements should be considered. In either case, the effect on the availability of the operating items or systems should also be evaluated.
Automatic Condition Monitoring
A built-in condition monitoring system (i.e. as a maintenance process) will determine when preventive or corrective maintenance is required. Scheduled maintenance would be redundant in these circumstances.
Even though a task may be effective, a final recommendation covering inclusion of the task in the program is dependant on an evaluation of task efficiency. The prime question to be answered is: does the benefit to be derived from completion of the task justify the cost of completion?
Evaluation of task efficiency is mainly subjective, in that neither the cost of completion nor the value of the benefit can be accurately assessed. The following general guidelines apply:
- If the item involved is Safety or Environment critical, the action should be retained (but not necessarily at the same frequency)
- If the item involved is Function or Cost critical, the action should contribute to the preservation of function/cost reliability
- If the item has No criticality, the decision may be considered from the viewpoint of what would be the likely consequences of not doing the task. Specific considerations are:
- Is the item subject to secondary damage and would the omission of this task be likely to increase the number of failures or the severity of the failures? If the answers to these questions are YES, the task should be retained.
- Is the workload associated with this task unusually high, or is it a simple task that can be performed quickly and perhaps in conjunction with other tasks (i.e., the cost is low).
- Is this an item which is normally monitored by the equipment operator and for which normal operator reporting action could be used as a means of determining when maintenance is due?
In general, for items which are not safety, environment or function critical and for which the probability of secondary damage after failure is low, it will be necessary to identify some clear benefit which will accrue if a specific maintenance task is performed.
Essentially this decision can only be made subjectively after consideration of all the factors outlined above. The essential consideration is, “Is there some evident advantage to be gained from performing this task”? If a benefit cannot be identified, the task should be omitted and the item monitored by general surveillance.
Effective tasks or adequate system monitoring may not exist for some failure modes. In these cases, the functions associated with each of these failure modes must be isolated and whether each function is hidden and whether safety, environment or function critical determined for further consideration.
In most cases, the criteria for determining the effect of failure of an item rely on the fact that the state of the item (i.e. operating correctly or failed) will be evident whenever the equipment is operated. However, some items and systems have a hidden function, i.e. their function is not evident during normal operation. A hidden function exists where either:
- An item has a function which is normally active whenever the equipment is used, but there is no indication to the operators when the function ceases to perform
- The item has a function that is normally inactive and there is no prior indication to the operators that the function will not perform when required (e.g. emergency stop equipment).
In many cases, those systems or items having a hidden function are intended for emergency use and their failure has direct safety or environmental implications. In other cases, secondary damage could result if a failure remains undetected.
Consequently, for items having a hidden function, a maintenance task will normally be included in the scheduled maintenance program. Such tasks are usually in the form of an operational check or functional test and serve either to detect failures that have occurred or to establish that the item or system was capable of operating at the time of the check.
A confidence check may be a task, an operational check or a functional test. In all cases, however, the confidence check must have the scope and sensitivity to ensure that the state of readiness of an item or system will be provided. If an effective confidence check is not possible and the item or system provides safety or environment critical function(s), redesign or modification action is required.
The converse to a hidden function is operator monitoring. In this situation, the operator during normal duties can consistently detect the loss of a function or the impending loss of a function (e.g. noticeable performance deterioration) and therefore indicate when maintenance is required. However, the following two exceptions should be considered:
- Safety/Environment Critical Item/System. If an item/system is safety or environment critical, a Before-use or After-use servicing may be required.
- Function Critical Items/Systems. Function critical items/systems, which are used infrequently, may require a after-use servicing to detect failures which could have occurred during the period of use.
Before-use and After-use maintenance tasks are performed immediately before or after use of an item or system.
Pre-use Maintenance Tasks
- Checking designated safety and environment critical items and functions that may have been damaged or deteriorated since the equipment was last used.
- For mobile equipment or motion capable equipment, visually checking the exterior of the equipment for damage that may have occurred since the equipment was last used.
- Replenishment of consumables to the levels required.
- Procedures for activating the equipment which are prescribed for the purpose of preventing unsafe conditions or damage to the equipment.
After-use Maintenance Tasks
- Identifying any maintenance required to restore the equipment to a fully functional condition.
- Replenishment of consumables.
A preventive task would have been assessed as effective or potentially effective only because some wear or drift pattern had been identified. This failure pattern (i.e. the failure distribution with respect to time) would also indicate the appropriate interval for the preventive task. However, in specifying the interval consideration must be given to the proportion of failures that should be prevented. A shorter interval may prevent more failures but the maintenance effort and costs will increase accordingly, and the equipment availability may suffer due to the greater down time for maintenance. Furthermore, some degree of random failures will always occur, therefore, 100% prevention cannot be achieved.
When an existing preventive task appears to be very effective, the small number of unscheduled occurrences may be merely random failures. In that case, the rate of unscheduled arising may not change substantially if the interval was increased. An extension of up to 25% may be adopted.
When the existing task has been assessed as effective but the number of unscheduled occurrences is high, the interval should be reduced (i.e. the preventive effect may "wear off" before the next performance of the task). However, if the preventive task involves the replacement of components or even the whole item, the possibility that the failures may be due to the "infant mortality rate" of these replacements should be considered. In that situation a reduced interval will not produce any improvement and in fact, an increase in the interval may be possible.
Further guidance for the selection of a suitable interval is the condition of the item or its parts when the preventive task was carried out, if such information is available. If the condition is such that the task appears unnecessary (e.g. brushes hardly worn when replaced or very little adjustment required) then the task interval may be increased. Conversely, if the item has consistently deteriorated to a point where a much deeper level of maintenance than that covered by the scheduled preventive task is required, the task or some part thereof should be performed more often.
Finally, if the spread of failures is such that a suitable interval is difficult to find, or the frequency of the task becomes uneconomical, then the task should be reassessed. In these situations, a specific surveillance task with a back-up preventive task may be more appropriate.
The proportion of faults found during scheduled maintenance to those discovered on an unscheduled basis provides an indication of the effectiveness of performing the task at the existing interval. If a specific condition monitoring task consistently leads to the detection of either a fault or a significant level of deterioration, the existing interval is approximately correct. Alternatively, the following general guidelines are to be followed in making the evaluation:
- If the proportion of unscheduled faults is high, (arbitrarily more than 50% of the total), the maintenance interval may be too long. Alternatively the scheduled task may be ineffective in either preventing or detecting failures, and should be either omitted or a more effective task developed, in which case a new interval must be established
- If specific condition monitoring performed during scheduled maintenance detects few faults, and the rate of unscheduled occurrences are also few, extension of the maintenance interval may be warranted. Exceptions to this general rule are tasks performed on safety or environment critical items, items having a hidden function and examinations performed to detect structural deterioration. However, as a general principle any scheduled specific condition monitoring task that does not lead to the detection of failures, is either being performed at the wrong interval or is ineffective.
Another intention of specific condition monitoring is to predict when failure will occur. This requires discovery of the start of failure and the subsequent determination of its progression rate. If the interval for determining the start of failure is too short, normal operating fluctuations in the monitored parameter may lead to incorrect presumption of failure start. Generally, the task interval should be set to at least 25% of the expected period between start of failure and total failure.
When the start of a failure has been detected, the interval of the monitoring task should be decreased to establish the time to total failure. The decrease in task interval will depend on the condition at failure discovery. The monitoring task should be performed several times before total failure time is calculated.
Areas and systems requiring general monitoring are subject to general monitoring as part of a scheduled program. As a general rule, area examinations and system monitoring are carried out periodically. However, if general condition monitoring results in the discovery of a significant number of faults in a particular area or system then the interval should be reduced. Conversely, extending the general condition monitoring interval beyond a year is also feasible, but this decision should be considered carefully as it would be contrary to the intent of general surveillance.
The period at which the proposed maintenance should take place is established from the criteria above. It is then qualified using manufacturers recommendations, clients' established maintenance programs, and published failure data. These should always be reviewed by the analyst.
Care should be exercised when selecting a period in balancing the requirements for optimum operating time against minimum maintenance down time.
In some cases the nature of the failure mode is such that the period will be as a result of some event, but only where the condition can be established.
When the period is chosen, it must be logically related to the failure mode. There is little point in nominating an operating based time period for a failure mode which is not related to operating time.