-
Our new ticketing site is now live! Using either this or the original site (both powered by TrainSplit) helps support the running of the forum with every ticket purchase! Find out more and ask any questions/give us feedback in this thread!
Reliability centred maintenance
- Thread starter taepn
- Start date
- Status
Sponsor Post - registered members do not see these adverts; click here to register, or click here to log in
R
RailUK Forums
100andthirty
Member
The idea of RCM is to do the maintenance that will deliver the required performance of trains. Safe, reliable, clean and preserves asset condition. The rigour of an RCM programme will address the gaps between what is required and what is currently being achieved.
Thank you, and it does not involve the punctuality of train?
D365
Veteran Member
- Joined
- 29 Jun 2012
- Messages
- 13,208
"Just in time" maintenance - replacing components right before they fail, therefore achieving the maximum possible lifespan.
The punctuality of a train is only affected when a failure occurs.
The punctuality of a train is only affected when a failure occurs.
Thank you, and if I said that the RCM can reduce reliability issue 25%, it can be implied that PPM level is improved up to 25% ?
edwin_m
Veteran Member
No. There are lots of reasons why a train can be delayed which aren't related to train faults. Examples are infrastructure faults, traincrews not available, problems caused by passengers. These will account for the huge majority of "primary" delay. Primary delay in turn leads to reactive delay (trains delaying other trains) but it is reasonable to assume that the reactive delay will be caused roughly in proportion to the primary delay that gave rise to it. PPM is influenced by all these sources of delay so even the elimination of all train faults would not make a huge difference to PPM in most cases. Hull Trains may be an exception at present, as it seems all their trains are out of action currently!
Your statement "PPM improved up to 25%" is also meaningless. Most TOCs have PPM of over 75% so either adding 25% or uplifting it by 25% will take it over 100% which is impossible.
Thank you and if the currently service delay 20% , 30% of the cause of delay due to the failure of rolling stock, and establish RCM can reduce reliability issue 25% , how much of PPM increase? I think that is 0.25*0.3*20 = 1.5% is correct?
edwin_m
Veteran Member
That would probably be approximately true.
100andthirty
Member
I don't believe there is an easy mathematical calculation to show that "this" much improvement in rolling stock reliability leads to "that" much improvement in PPM.
PPM is largely discredited as a measure. it has led to perverse behaviours - such as padding of the public timetables - leading to trains which are delayed along the route being on time at its destination. The industry is moving to "right time throughout the journey".
PPM is largely discredited as a measure. it has led to perverse behaviours - such as padding of the public timetables - leading to trains which are delayed along the route being on time at its destination. The industry is moving to "right time throughout the journey".
BestWestern
Established Member
- Joined
- 6 Feb 2011
- Messages
- 6,736
'Just in time' is surely entirely related to cost minimalisation and not performance enhancement? Performance is clearly at greater risk by 'sweating the assets' in this way as opposed to routine earlier replacement of possible failure items.
100andthirty
Member
Determining when things are going to fail is, in the rail industry, not a precise science. Often performance falls away before failure. RCM amongst other aspects seeks to identify the ideal intervention point which is usually long before ultimate failure.
Sometimes condition monitoring techniques are identified to detect impending failure earlier. For example hot axlebox detectors identify failing axle bearings just before they fail. They cause delays but prevent serious safety issues (eg derailment). If the vibration from bearings is monitored, signs of impending failure can be identified earlier and therefore the repair can be planned with no impact on service.
More generally, it is really difficult to determine when things will fail which makes any attempt to develop "just in time" pointless. The output of RCM is the "appropriate maintenance", not too often and not too infrequent. Note, the appropriate mIntenance for a train on a hilly route with frequent stops will be different from the requirement for the same train type on flat infrequent stop routes.
Sometimes condition monitoring techniques are identified to detect impending failure earlier. For example hot axlebox detectors identify failing axle bearings just before they fail. They cause delays but prevent serious safety issues (eg derailment). If the vibration from bearings is monitored, signs of impending failure can be identified earlier and therefore the repair can be planned with no impact on service.
More generally, it is really difficult to determine when things will fail which makes any attempt to develop "just in time" pointless. The output of RCM is the "appropriate maintenance", not too often and not too infrequent. Note, the appropriate mIntenance for a train on a hilly route with frequent stops will be different from the requirement for the same train type on flat infrequent stop routes.
dp21
Member
- Joined
- 10 May 2017
- Messages
- 358
I may be repeating a few things from above but:
From an operator's perspective, one of the core concepts of RCM is to ensure that the right things are being maintained to achieve a high level of reliability. It therefore means targeting maintenance to those components which require the most attention and, based on data analysis, are more statistically likely to fail. If you're examining something on every A exam (say, ever 28 days or so) and it has never EVER failed, why do you continue to do that exam when the man-power could be better used elsewhere and technology (e.g. remote condition monitoring) used to undertake monitoring of the system in service.
As has been said above, it's nigh on impossible to predict an exact PPM improvement unless you are consistently having failures which cause a predictable amount of delay which, in this industry, is extremely unlikely. You could do some delay minute reduction projections which are probably going to be fairly accurate but with PPM being affected by almost everything it's hard to predict.
Just a few things to consider.
From an operator's perspective, one of the core concepts of RCM is to ensure that the right things are being maintained to achieve a high level of reliability. It therefore means targeting maintenance to those components which require the most attention and, based on data analysis, are more statistically likely to fail. If you're examining something on every A exam (say, ever 28 days or so) and it has never EVER failed, why do you continue to do that exam when the man-power could be better used elsewhere and technology (e.g. remote condition monitoring) used to undertake monitoring of the system in service.
As has been said above, it's nigh on impossible to predict an exact PPM improvement unless you are consistently having failures which cause a predictable amount of delay which, in this industry, is extremely unlikely. You could do some delay minute reduction projections which are probably going to be fairly accurate but with PPM being affected by almost everything it's hard to predict.
Just a few things to consider.
randyrippley
Established Member
- Joined
- 21 Feb 2016
- Messages
- 5,382
The basic premise of the question is wrong: RCM is a cost saving measure, not a reliability measure.
If you assume that current maintenance schedules are based on proven need, then changing to RCM will delay maintenance and so run the risk of unscheduled failure due to some components prematurely wearing outside of the expected schedule - or at the extremes of the expected schedule
millemille
Member
RCM is a concept is fantastic, but falls own on execution when it meets the real world needs of a running railway.
As a maintainer you've got a set of limitations imposed upon you by the need to run trains. The operational side of the business - which is what running a railway is all about - gives you, the maintainer, your availability requirements for the fleet. So, for example, if you've got a fleet of 100 trains then 95 of them will be needed for service every week day. Therefore you can only have 5 out of service. You'll need a heavy maintenance allowance (Bogie overhaul, wheelset changes etc), deep cleaning, modification, defects... and pretty soon you've got nothing left for running maintenance. So your maintenance has to be done when the trains aren't running (during the day between peaks if you're running a metro service or at night for all kinds of operation), which means your down time for an exam is limited to a real world, hands being laid on the train after you've allowed for shunting into and out of the shed, of about 5 or 6 hours.
And then there's a limited number of fitters/technicians available to do the work - limited both by budget and the sheer number of people you can get working on the train at the same time.
And then there's depot capacity, shunting capacity, diagram arrivals and departures (can your first arrival be your last departure for example?) and so which all combine to give a very constrained window of opportunity and the resources that can be deployed within that window.
And then you've got the train and you need to work out, in a structured manner which will stand up to scrutiny as being compliant with all applicable legislation, guidance and best practise, what maintenance it actually needs vs. what's nice to have.
So you break the train down into systems; brakes, traction, air, bogies, propulsion, doors, comms etc. Then you break each system down into components and you determine all of the failure modes each component can have and what the impact of each of those failure modes would have upon the operation of the train.
And you take of each of these impacts and ask a series of questions; Does the failure directly affect safety? Does the failure indirectly, another safety system also needs to fail, affect safety? Is the failure evident to the train crew? Does the failure affect train reliability? etc.
This allows you to categorise each component failure mode's impact as either "Hidden Safety", "Evident Safety", "Performance", "Availability", "Economic", "Reputational" and so on.
You can then prioritise the component failure modes that you are attempting to use maintenance to prevent, with Hidden Safety being the first priority and you come up with, if possible, maintenance tasks - checking, testing, replacement etc. - could be done. And then you work out the frequency, man hours, down time, work flow for these tasks and subtract the sum of these from the pot you've got as constrained by unit availability, people available etc.
Then repeat this for Evident Safety. And then for Performance. And I absolutely guarantee at this point you're running out of down time and/or man hours before you can do all of the Performance related maintenance tasks....
As a maintainer you've got a set of limitations imposed upon you by the need to run trains. The operational side of the business - which is what running a railway is all about - gives you, the maintainer, your availability requirements for the fleet. So, for example, if you've got a fleet of 100 trains then 95 of them will be needed for service every week day. Therefore you can only have 5 out of service. You'll need a heavy maintenance allowance (Bogie overhaul, wheelset changes etc), deep cleaning, modification, defects... and pretty soon you've got nothing left for running maintenance. So your maintenance has to be done when the trains aren't running (during the day between peaks if you're running a metro service or at night for all kinds of operation), which means your down time for an exam is limited to a real world, hands being laid on the train after you've allowed for shunting into and out of the shed, of about 5 or 6 hours.
And then there's a limited number of fitters/technicians available to do the work - limited both by budget and the sheer number of people you can get working on the train at the same time.
And then there's depot capacity, shunting capacity, diagram arrivals and departures (can your first arrival be your last departure for example?) and so which all combine to give a very constrained window of opportunity and the resources that can be deployed within that window.
And then you've got the train and you need to work out, in a structured manner which will stand up to scrutiny as being compliant with all applicable legislation, guidance and best practise, what maintenance it actually needs vs. what's nice to have.
So you break the train down into systems; brakes, traction, air, bogies, propulsion, doors, comms etc. Then you break each system down into components and you determine all of the failure modes each component can have and what the impact of each of those failure modes would have upon the operation of the train.
And you take of each of these impacts and ask a series of questions; Does the failure directly affect safety? Does the failure indirectly, another safety system also needs to fail, affect safety? Is the failure evident to the train crew? Does the failure affect train reliability? etc.
This allows you to categorise each component failure mode's impact as either "Hidden Safety", "Evident Safety", "Performance", "Availability", "Economic", "Reputational" and so on.
You can then prioritise the component failure modes that you are attempting to use maintenance to prevent, with Hidden Safety being the first priority and you come up with, if possible, maintenance tasks - checking, testing, replacement etc. - could be done. And then you work out the frequency, man hours, down time, work flow for these tasks and subtract the sum of these from the pot you've got as constrained by unit availability, people available etc.
Then repeat this for Evident Safety. And then for Performance. And I absolutely guarantee at this point you're running out of down time and/or man hours before you can do all of the Performance related maintenance tasks....
- Status