Improving PM's and Turning PM's into CM's (CM & Lube Newsletter Article 10/09)

From Imrtwiki

Return to PM & CM Optimization

Contents 1 Improving Preventive Maintenance Tasks (PM’s) and Turning PM’s into CM’s 2 Setting Up for ‘Find and Fix’ 3 System of PM Ownership 4 Using a Formal PM Improvement Process 5 Turning PM’s into CM’s

Improving Preventive Maintenance Tasks (PM’s) and Turning PM’s into CM’s

The top priority for maintenance people should be the quality of their equipment repair and servicing work. A very close second is the quality of their equipment inspections and condition monitoring activities. Preventive Maintenance tasks (PM’s) usually combine both equipment servicing and inspection activities and so it’s very important that they are done right. The last three IMRt NSW discussion meetings that the author has facilitated have been on 'Overhead Cranes', 'Conveyors' and 'Improving PM’s & CM’s'. Below are a number of common problems with PM’s that have been raised.

• They often are not done properly, left incomplete or not done at all
• People doing PM’s often don’t recognise the importance of the activity
• It is usually difficult to verify that PM’s are done properly
• Often not enough opportunity is taken to fix minor problems immediately
• The details of required tasks are often not well documented
• PM documentation is often ignored if the person has done the job before
• Often the tasks specified are not appropriate or not best practice
• Often PM documentation remains static and is not improved over time
• The information that comes back from PM’s is often not managed well

Some suggested solutions that might help address these problems is:-

• Setting up for ‘Find and Fix’
• Implementing a system of PM ownership
• Using a formal PM improvement process
• Turning PM’s into CM’s.

Setting Up for ‘Find and Fix’

The reason you do PM’s is to find and fix small issues before they become big issues. So if a maintenance person inspecting a piece of equipment finds a problem that would only take 5 minutes to fix, it is generally best if they fix it while they are there. The problem is that if they don’t have the right procedures, tools, and parts with them, a 5 minute job can easily turn into an hour job or impossible to complete. The example given in my overhead crane meeting was with inspection of lifting gear by a speciallist. If lifting gear is found defective and taken out of service, unless it is removed from the site there is always the risk that it will still be used. It is best if it is fixed immediately. You know in advance the likely items that will be found defective, that is chains and connectors. The inspecting person should have these parts and appropriate tools with them in their van or trolley so the problems found can be fixed immediately. Doing small quick repair work as Find and Fix is many times more efficient than the other option of find, report, plan, schedule and then fix.

The main problem with the Find and Fix approach is that often a particular PM may never be completed as there are always repairs initiated early in the inspection and there is then not enough time to complete it. Some tradespeople would much rather be doing repair work than inspection activities and so are more easily distracted away from PM tasks. If a safety or very urgent problem is found, then it is legitimate to complete the high priority activity in preference to the rest of the inspection. The argument against this is that sometimes some of the tasks in the inspection should actually have a higher priority than repair of a particular problem found. If tasks in a PM where categorised into 3 levels of priority, then it would be easy to set rules for when inspection tasks are more important than repair tasks. If a PM is being carried out is overdue, this should also be taken into account when assessing task priority. The priorities for the PM tasks might be:-

• High risk - Must be completed
• Medium Risk - Should be completed – only missed if there is an urgent or safety problem. If missed should break schedule or be rescheduled as soon as possible
• Low risk - Often PM tasks are only done because you are already at the equipment for an important task, so the additional cost of doing more tasks is small

My suggestion is to gear-up for as much Find and Fix as is practical, with inspection given increased time for additional 'Fix' work. If PM’s are not completed because of this, the priority of the missed tasks should be noted and the PM scheduled early if necessary. The best KPI I have seen used for managing PM’s and Inspections is recording the number of Finds and the number of Finds and Fixes per month from PM’s for a maintenance section. It does not matter how small the problems found are, down to small cleaning tasks. If you have lots of little problems being found and fixed efficiently then it gives you more confidence that larger problems will be controlled. If only a smaller number of larger problems are being addressed in an area, that that is an obvious area of risk.

System of PM Ownership

The most effective use of PM’s I have been associated with was when I was working in the Tin Mill (now shut down) in the Port Kembla Steelworks. Each operational line in the mill had a mechanical and electrical planner that controlled the section’s maintenance and each had an Inspector. The role of the Inspector was to do as much of the PM’s, CM’s and minor repairs for the section as was possible. Some of the inspectors doing this role were very knowledgeable, skilled and had a very high level of ownership for these tasks. They generally did a fantastic job. The only problem of having these top performers is that they documented nothing about what they did, how they did it or why they did it. There may have been some documentation for some of the inspection work they were doing but generally it was not used once they understood the job and the documentation probably had little resemblance to what they actually did. While they were doing the job it was great. When they left the job, even if there was a handover a large amount of knowledge and capability was lost to the section. So how can you encourage people to document what they do and keep improving the documentation as they learn more?

In the recent Improving PM’s & CM’s meeting Peter Morrison told the story of why their tradespeople started using the long text history fields in their CMMS to document problems that had occurred each shift. This information was being printed out as a report at the end of every shift. The tradesmen found using the report was the best way to communicate about problems to the following shifts. The learning for Peter was that a prime motivator for front line maintenance people to document thing is to help other front line people. At the front line in maintenance, there is often strong camaraderie and desire to support others, especially where people face similar difficult equipment problems.

So a suggestion to get ongoing improvement on PM documentation is to nominate an ‘Owner’ for each PM inspection and ideally have that person do that PM 60% of the time. If a person did not do the inspection at least 40% of the time, they would not feel like the inspection owner. If they did it more than 80% of the time, they would not feel they need to support others with improved documentation. The owner of the PM would have the major responsibility for front line improvement of the PM. The minimum requirement for the documentation should be to the point that others who have not done the PM before can do a competent job. Improving documentation is sometimes seen as an undesirable activity by many maintenance people so the responsibility has to be communicated properly and its importance must be emphasised. Where improving the documentation will be supporting others in their team, it is more likely to be given the required attention. In the recent NSW IMRt PM’s meeting the need to have an efficient formal process of approving and updating documentation was empasised. Improvement PM’s can be difficult and it needs someone with experience and maturity to review and approve proposed changes. This needs to be done in a way so as not to stifle innovation and people wanting to try different approaches.

The improving PM’s meeting suggested that documentation for PM’s should be at two levels. The first level is the Check Sheet, which identified the specific task but generally only contains abbreviated text & diagrams to prompt the information needed to be collected. There should also be a second more detailed PM Info Document that contains more details on the 'how and why' for the tasks in the PM.

Another great way to get more commitment to and ownership on PM's is with on-site auditing. Below are two links to auditing one page formats that might be useful.

PM AUDIT PROCESS AT TOMAGO ALUMINIUM (ONE PAGE SHEET) Tomago Aluminium One Page PM Audit Sheet

BLUESCOPE STEEL/ SILCAR RELIABILITY WALK (ONE PAGE AUDIT SHEET) Rod Bennett's Reliability Walk Proforma

Using a Formal PM Improvement Process

Documenting and improving PM’s is not a trivial task. The most technically correct method is a formal Reliability Centred Maintenance (RCM) approach, which is horrendously time consuming to do properly. RCM is only generally used on very critical equipment or where there are a large number of similar medium criticality equipment, where the one analysis can be used for many items. For most other equipment PM’s are developed by using OEM recommendation, using known standard practice, copying PM information from similar equipment and or a knowledgeable person writing their best guess of what PM tasks might be useful.

Making major improvement to PM’s and CM’s takes lots of time and effort so it is important to understand where the biggest benefits are likely to come from. There are things that can be done with little additional effort. The key one is using the people doing the PM’s to add regular small improvements as they identify issues. Below is a suggestion of some of the stages in the PM improvement process.

Health & Safety Issues – Does the documentation inform the PM user of all health and safety requirements for the plant area and the particular equipment?

Equipment Identification – Is there enough information in the PM documentation to clearly identify which equipment item the task will be carried out on, with no room for confusion?

Matching Documentation to Reality - Whatever the approach used to develop PM’s it is highly unlikely that the task specified in the PM will fully match the actual circumstances found by individual doing the PM task. If the PM task says ‘check the coupling’ and it is actually a V Belt drive, then this should be changed.

Purpose of the PM – Many PM tasks are written in very general terms such as ‘Check conveyor drive’. This leaves it completely up to the inspecting person of what is to be inspected. The RCM process looks at how the equipment might fail, what the causes might be, how this could be prevented and what symptoms might identify the problem. This information is very useful to understand the purpose of a PM task and to determine what PM task should be performed. Maintenance people often have a some knowledge of types of failures, causes, symptoms and preventive tasks and can update the PM’s with this information both on the Check Sheet and PM Info Document. Also for common equipment there is or should be best practice information available for the types of failures, causes, symptoms and preventive tasks that are likely. Symptoms and preventive tasks are what go into PM’s the check sheet activities and the failure types and causes should be updated into the PM Info Document.

Is the PM Task Useful? – Some questions that can be asked.

• Has the PM task detected or prevented a problem on this or similar machines and if not, how likely is it to give adequate warning of a problem and/or reduce the chance of it occurring?
• How difficult, time consuming and costly is it to do the PM task?
• How bad would be if it the failure or problem is not detected or prevented and how likely is it to occur if the PM task is not performed?

If a potential failure or problem has severe consequences you can justify a lot of effort, even if it seems to show little result. You continue to check the condition of fire extinguishers, even if you have never found a problem because the consequence of an extinguisher not working if a fire occurs can be huge.

How to do the Task – If the task is not trivial there should be a procedure which defines how to perform the PM task including details such as job steps, tools, etc and ideally has illustrative diagrams. This information should be in the PM Info Document. A number of times I have reviewed the history of measurements (eg. clearances), which have been collected in PM’s and found the information is useless as there has been no standardised measurement process used.

Setting of Repair Task Trigger Points – Where a measurement or qualitative assessment is being made to assess the condition on an equipment item or to identify a problem, there should be some alarm levels or set of rules to indicate when a repair or fix activity should be initiated. If this is not practical then at least the purpose of the PM task should be understood.

How Often To Do a Task

• Task defined by regulations are easiest as they are predefined.
• Where the PM task is for a servicing task like lubrication, the frequency should be related to the wear-out period for the item. There are various way to confirm the best change-out point, such as by post-mortem of removed components to determine the residual life.
• Where the PM task is measuring a wear-out deterioration, such as brake pad wear, then there should be at least 3 or 4 measurements within the expected life of the component.
• Where the PM task is monitoring a random type failure problem (eg. bearing failures), then the warning time given by the PM task, likelihood & consequence of the failure and the cost of doing the PM task are all important.
• Where the PM task is for Failure Finding, such as checking if standby equipment is working, then how often it is carried out is related to the likelihood of the failure occurring.

There are two levels of PM task optimisation. The first is optimising the frequency of the whole check sheet. The second is changing the particular tasks frequencies by moving tasks to different PM’s that are done at different frequency.

Turning PM’s into CM’s

My first experience working in maintenance was being assigned to work with a tradesman and going out to do a morning PM inspection of a large wharf ore unloading machine. The tradesman grabbed his tool bag and we climbed up to the unloaders motor room where he promptly sat down, pulled a newspaper out of his tool bag and read till ten minutes before morning tea. He then filled out the check sheet for the PM by ticking all the boxes as OK. This is always the main fear with PM’s that it is often difficult to verify that the tasks are actually being done. In comparison when I ran equivalent manual paper based condition monitoring systems, I never had this concern. If one of the guys collecting the mostly numerical data wanted to fudge the results, it would probably have been harder than doing the work properly. It is far easier verifying proper data collection when there is more than just a tick in check sheet tick box indicating completion.

What is the difference between a PM inspection task and a condition monitoring task? The main differences I suggest is in the level of discipline and precision used and the way the collected data or qualitative assessment output is managed. So what are some of the range of options from PM’s to CM’s that were discussed in the IMRt meetings.

1. An open undefined inspection task eg Check conveyor drive
2. A qualitative inspection task with no assessment method or alarm criteria eg check gearbox noise.
3. A qualitative inspection task with an assessment method but no alarm eg check gearbox noise using a listening rod placed on the output drive side bearing housing.
4. A qualitative inspection task with an assessment method and with a pick list of condition assessments (usually with a number code for each) eg check gearbox noise using a listening rod placed on the output drive side bearing housing. 1-OK/Normal, 2-Possible increase in gear noise, 3-Possible bearing noise, 4-Rubbing noise, 5-Significant increase in noise level.
5. A Semi-Qualitative inspection task using a Go/NoGo gauge assessment method eg check gearbox output shaft for drop using the provided Go/NoGo gauge between the marked point on the bracket above to the top of the shaft.
6. A simple measurement task with no condition assessment criteria with no alarm level eg measure EnV (enveloped acceleration) at gearbox marked point 8V using vibration pen meter.
7. A simple measurement task with an alarm criteria eg measure EnV (enveloped acceleration) at gearbox point 8V using vibration pen meter. If EnV is greater than 1.5, email Vibration Tech to check gearbox on next visit.
8. A simple measurement task with office based trending of reading and rule based alarm criteria eg measure EnV at gearbox point 8V using vibration pen meter. If EnV doubles from normal level or there is a definite increasing trend, email Vibration Tech to check gearbox on next visit.
9. A simple measurement task with on-site trending of reading, immediate investigation of cause and rule based alarm criteria eg measure EnV at gearbox point 8V using vibration pen meter. Please check location of maximum EnV, double check oil level, etc. If cause is not eliminated and EnV is doubled from normal level or there is an increasing trend, email Vibration Tech to check gearbox on next visit.
10. A complex measurement task with office based analysis and trending of data with rule based condition assessment criteria eg measure gearbox point 8V using vibration data collector using standard gearbox template setup with statistical alarms. If in alarm use standard gearbox condition assessment process.
11. A complex measurement task with immediate site based analysis, trending of data, cause investigation of alarms with rule based condition assessment criteria eg measure gearbox point 8V using vibration data collector using standard gearbox template setup with statistical alarms. If in alarm use standard gearbox condition assessment process.
12. Various online monitoring options.

Most of the problems with rigour and poor verification of PM’s are solved by using the point 4 above Pick List qualitative condition assessment. The Pick List approach gives guidance to the person doing the checks of what to focus on as well as adding some precision to the output options for the task. When a problem different to what is expected is found during an inspection, this becomes an option for an additional Pick List item. A few of the overhead crane service companies standardise on a 1 to 9 Pick List code for equipment condition with 1 being ‘good condition’ and 9 being ‘take out of service immediately for repair’. Crane companies also use the equipment condition code to reduce the precision of data given to customers by converting quantitative measures to Pick List codes.

For many condition measurements such as vibration, the level of the readings is important. But much more important to assess current condition is changes from previous measured levels and trends. The run chart or trend graph is a fantastic way to visualise information to make this assessment. It gives the ability see the variability in the measurement and to recognise even small changes outside this range of variation. Picking up problems very early often allows you to eliminate the cause of the problem before any damage is done.

The other often unconsidered advantage of trending is visualisation of condition history for equipment. The trend graph to the right shows the tale of woe from the view of the vibrations of a pump bearing. Someone decided to overhaul this pump, only because it had seen many years service. There was a bearing installation quality issue on the overhaul and after about 6 months the bearing had to be repaired. The vibration levels still ended up higher than they originals were. Another thing you see from trend graphs is the failure rate and warning time (PF Interval) given by a particular parameter. Snowy Hydro had trialled putting their PM qualitative assessment pick list numerical codes into their condition monitoring software to achieve this visual assessment capability for qualitative inspections.

The author has run a number of systems in the past to collect data on-site with return to office to graph the result. I have found two main problems with this process. The first is if there is a measurement error then you typically don’t find out till you return to the office the graph the data. The second is that you often don’t discover a real change in the data until you do the office based trending and you are not always in a position to go back out to site to further investigate the change. The obvious solution to this issue is to do the trending on-site directly after you complete the measurements. There are two systems approaches to manage this. The first is taking a laptop, PDA or one of the more sophisticated mobile phones to site to input the data so you can see a trend. This is the way of the future but the reality of most maintenance people’s current situation is that they will not be getting rid of their paper based system anytime soon.

The other system I have used is for on-site parameter trending is long life water proof type paper (DURACOPY) that can be taken on-site to manually trend quantitative data or Pick List codes (see article on Vibration Pen meters). Below is an example of some formats that can be used for Pick List code, numerical information and trending format using simple MS Word tables. If this type of format is used a significant amount of blank space should be left so additional information and new monitoring variables can be added by the PM users. This type of format deserves some consideration as a PM check sheet format option. With the appropriate amount of care being taken keep the checks sheets clean while on-site, all the advantages of condition monitoring can be achieved in PM activities. Some might see using paper based systems, no matter how sophisticated, as a step backwards but that has not been my experience. Good paper based systems tend to be people friendly solution.

Article by Peter Todd – Facilitator NSW Industrial Maintenance Roundtable