Maintenance

What is maintenance

Maintenance is actions to retain a component in or restore a component to a state where it can perform a required function.

Why

Maintenance will increase safety and reliability of components and thus reduce operational cost if the maintenance activities are properly planned and performed.

How

Maintenance is planned during the development phase by working out a maintenance program and plan. Maintenance is executed and improved during the operational phase. The maintenance work process is:

Figure 1: The main phases of the maintenance work process

Information related to each phase follow. This is not a guideline for maintenance. A guideline describes the following content for each activity in the work process: What, why, deliverables, recommendations and responsibilities. The descriptions below gives only short introductions and practical advices to each phase in the working process.

Maintenance program

Preventive maintenance is used to avoid critical events. Corrective maintenance can be performed on non-critical components. The maintenance program describes type of maintenance for components. An optimal preventive maintenance program can be developed by executing a RCM analysis. This analysis will find the right preventive maintenance principle and optimal maintenance intervals for each component. It will also find the optimal relation between preventive and corrective maintenance. The maintenance program shall also be in accordance with laws, regulations and standards.

• Corrective maintenance
• Preventive maintenance
• Time related terms
• RCM analysis

Maintenance plan

The maintenance plan is a detailed plan specifying what to do, why, when, where, support, what if, costs and responsibilities. The maintenance plan should be formulated during the detail design phase and be modified during the operational phase when real experiences are available.

More about details follow:

What

Corrective or preventive maintenance, preventive maintenance principle, preventive maintenance interval, procedures for work execution, required maintenance environment and repair policy.

• Repair policy

Why

Reason for the described maintenance task (motivation).

When

Time schedule (integrated with operational activities), grouping of maintenance activities, and time for renewal.

Where

First line maintenance (on the site) or depot maintenance.

Support

Type and number of spare parts, test and support equipment, the overall distribution and flow of materials. It could be wise to have the same electric motor on more pumps, even if some of the pumps require smaller motor. The purchase cost is slightly higher, but the cost of spare parts could be considerable reduced. Type of spare parts should be modified during operation based on user’s experiences.

What if

What to do if abnormal situations occur during the task.

Costs

Cost estimates for each maintenance task.

Responsibilities

The prime responsibilities for maintenance between producer, consumer and supplier should be described.

Personnel quantities

Number of persons for executing the job should be described.

Skill and knowledge

The maintenance personnel should have the skill and knowledge required to execute the maintenance job but also to improve the maintenance plan. External assistance should be used if required.

The maintenance plan and revision stop should be worked out together with the production department. This will result in less disturbance of the ongoing production. The maintenance plan should always be improved based on practical experiences from operation. That means collection of data and performance of analysis. Backlog is maintenance hours not executed in accordance to the maintenance plan. There will normally be a delay between increasing backlog and reduced availability. When maintenance comes out of control (large backlog) it is important to initiate operational restrictions (closing, reducing performance etc). Work orders with highest criticality are not accepted in the backlog.

Maintenance execution

This is the phase when maintenance is executed in practice. Different execution tasks are:

Replace

Replacement of an item with a new or refurbished of the same type and make. An example is replacement of a bearing.

Repair

Action to restore an item to its original state. Examples are weld, plug, reconnect, and remake.

Modify

Change the item (type, make, material and design). Examples are a filter with smaller mesh and replacement of an oil pump with another type.

Adjust

Bring any out-of-tolerance condition into tolerance. Examples are alignment, set and reset, calibrate and balance.

Refit

Minor repair service to bring back an item to an acceptable condition. Examples are polish, clean, grind, paint, coat, lube and oil change.

Check

Regain a functional failure by restarting or resetting.

Service

Periodic service normally without dismantling of an item. Examples are cleaning, replenishment, adjustments and calibrations.

Tests

Functional tests are performed manually at defined time intervals. Safety systems in the oil industry normally have 1, 3, 6 or 12 months intervals. The advantage is increased availability. Disadvantages are additional wear, human induced errors (forgets to reset by-pass of component, damages), downtime and costs. The test will not detect all kind of failures.

There has been an increased use of automatic self-test. An example is partial stroke test (PST) device on the emergency shutdown valve (ESV) actuator. Such test will have no influence on production and thus no downtime. Not all failures are detected. The diagnostic coverage expresses the efficiency of the test. Also not that the automatic self test will increase the spurious failure rate.

Inspections

Periodic inspections with and without dismantling normally by use of senses. An operator may detect failures between tests. For instance the panel operator may detect a transmitter that is stuck. He may also detect a sensor left in by-pass.

Overhaul

Comprehensive overhaul with extensive disassembling and replacement of items as specified.

The practical part of the maintenance work process is:

Figure 2: The practical part of the maintenance work process

A notification is made when after a fault detection and maintenance is needed. All maintenance jobs shall be given as a work order. Permit-to-work system is introduced to avoid maintenance work on equipment that is energised (e.g. high pressure gas pipeline), to ensure that work is performed at the right place and time, for identification and implementation of necessary safety precautions, to ensure the equipment is brought back to operational status before start-up and to get overview of active work permits to person responsible for the operation.

Maintenance reporting

Experiences from a maintenance execution should be reported to improve the maintenance program and the maintenance plan. The following should be reported:

• Installation time of a new unit
• Time to first failure
• Time between failure (from second failure etc)
• Maintenance actions
• Maintenance quality
• Maintenance downtime
• Maintenance man-hours
• Spare part consumption
• Failure causes (normally bad reporting)
• Environment
• Other systems that are disabled
• Fault that is not repaired
• Technical condition

Note in respect to data analysis that it may be important to differ between “first failure”, “second failure”, "third failure" etc. for repairable components. A repairable component will often not be “as good as new” but “as bad as old” after repair. Thus, the  time-to-failure distribution for the first failure is not always equal to that of the second failure etc. and the data analysis approach needs to adopt to this type situation. Most databases do not include this type of failure history, which makes it difficult to find the most realistic model that describe the compont reliability.

Maintenance analysis

Analysis of maintenance data is an important part of maintenance management. As a minimum the analysis should include:

• Top-ten list for cost drivers
• Top-ten list for bad MTBF
• Failure cause analysis
• Identify common cause problems (CCF)
• Update reliability parameter data
• Reliability trends
• Failure propagation speed (P-F interval)
• New failure modes that has not been considered in the maintenance plan
• New critical components that was not identified in the initial analysis
• Indication of wrong use
• Indication of wrong component design specification
• Key performance indicator (KPI) trends

The analysis should also identify the need and relevance for feedback to the manufactures and vendors. The result from the analyses is used to suggest improvement measures, but also for updating of risk analysis and maintenance intervals.

By proper collection and analysis of failure modes and failure causes, it is possible to eliminate failure mechanisms by implementing measures against these failure causes. On the other hand, if this systematic approach fails it is likely that reliability performance is impaired to an unacceptable level.

For a new installation only generic reliability parameters are available (WellMaster, SubseaMaster, OREDA, from manufactures, etc.). By proper collection and analysis of the reliability data for a given site, it is possible to establish site specific reliability parameters, and thus establish a maintenance program that is adapted to the actual reliability performance of that site. Further collection and analysis is an important means to ensure reliability growth during the lifecycle of an installation.

When the initial maintenance program is set out there is a large uncertainty whether the generic data will be relevant for the actual installation. Therefore conservative estimates for the initial failure rate should be used. However, if a detailed FMECA analysis is performed, it may be possible to argue that the new installation will have a lower failure rate and hence maintenance intervals could be increased.

Typical reliability parameters are:

• Failure rate λ
• Shape factor α
• Mean time to failure MTTF
• Rate of occurrence of failures ROCOF
• Mean time between failures MTBF
• Mean time to repair MTTR
• Mean downtime MDT
• Average availability

Recommended frequency of analyses is:

• Tasks should be evaluated daily
• Backlog should be evaluated weekly/monthly
• Cause analysis and common cause analyses should be analysed for critical failures

Issues to be analysed from time to time are:

• Is the rate of failures as expected
• Is there a negative development in the rate of failures
• Can the data tell about P-F intervals
• Is there detected other failures than expected in the action plan
• Are new component with high criticality discovered
• Is there indications on “wrong use”

More information of the P-F interval follows:

• PF interval

Improvements

Make improvements when required.