[FUN] Instrument maintenance

[Stephen Mitschke]

> Dear All,> > There prevails an assumption in the process industry that field devices > combined with their self diagnostics bring> savings to the predictive/preventive maintenance and one still gets the > same level of reliability of maintenance compared> to the era of 20 mA signaling. There is a lot of literature of control > valves and their diagnostics but other instrumentation> like pressure transmitters, differential pressure transmitters,> temperature > transmitters and flow meters are not as widely> discussed.> > My question is: How has better diagnostics of intelligent field devices > changed preventive/predictive maintenance schema> compared to the days of 20 mA signaling?> > I know that there can be a maintenance or check card for each device in> the > plant that tells how often and what kind> of maintenance the device needs. Do we need any maintenance cards now when> > device diagnostics can tell about device> problems before they actually occur? Or is it true that the maintenance is> > carried out when the device has failed (reactive maintenance)?> > > Jani Toro> VTT - Finland

[Stephen Mitschke]

> Jani,> As a strong supporter of getting fieldbus (and HSE based systems) into the> plants post haste, I cannot express how glad I am to see people finally> asking what will happen after the honeymoon. We recently began piloting> maintenance classes, which have brought up some very interesting points,> specific to maintenance workers. There seems to be a prevailing sense> that> maintenance people are largely unaffected by these systems or that there> job> will be simplified. My response is yes and no. Specifically, while most> devices do possess strong diagnostic and predictive capabilities, they> must> still be configured into the devices and configured in a way that is> useful> (useable?) by service people. I am not convinced this will be the case> for> many installations. Witness the capabilities of SMART valves installed> vs.> how many are truly configured to provide predictive information and I> think> the concern becomes clear.> > Another concern I have and we have been studying at Lee College is what> will> happen when the design and installation engineering people are done? The> college has been engaged in determining the overall maintenance impact> fieldbus will have on technicians, supervisors, planners, etc. The groups> involved have consisted of engineers, maintenance technicians and some> novices, all with instrument, DCS, and IT backgrounds. Again, when these> groups are singled out and there viewpoints considered, they see potential> challenges.> > As an example, consider the following hypothetical scenario:> > A fieldbus Level transmitter is installed as the primary of a cascade> system.> Along with standard AI block (which makes up the vast majority of device> requirements for the plant), this device also uses an Arith. block an> Alarm> block and a PID block.> It fails (sensors still go bad and cards still break).> It has been over one year since the device was put in.> In the interim, there have been several upgrades to the card (and upgrades> to DD or CFF files).> Due to min/max requirements, the storeroom only keeps one in stock, It is> the newest version and may or may not contain the functionality needed.> > The engineer is off at fieldbus school and on vacation after that.> > My concern is not that the presented scenario can be managed - it can. My> concern is that there is clearly a need to identify exactly who will be> responsible to maintain just which aspects of these systems after the glow> is gone and the unit has been operating smoothly for a time?> > > While I am very optimistic about the advantages of bringing fieldbus into> the plants, I also sense that there needs to be an initiative put into> place> which takes a cold look at the entirety of the issue FROM THE ASPECT OF> THE> DAY TO DAY MAINTENANCE PERSON who will be left to keep the unit running.> Anyone saying the diagnostic capabilities of fieldbus devices will> minimize> the maintenance requirements, in my estimation, has probably never been on> top of a tower in December or called out by operations at 2 in the morning> to make a valve work. Saying it is possible, and having it possible does> not always translate into it occurring.> > I am reminded by the engineering manager who told me his engineers> designed> systems which did not require unplanned maintenance.> > Chuck Carter> Center Director/Principle Investigator> Center for Digital and Fieldbus Maintenance Education> Lee College

[Stephen Mitschke]

> Dear Mr. Toro,> > First, this is a great question. Sorry about the long answer, I hope it> is> worth the long read. I have some examples of changes to maintenance> practices listed below, but first a bit of background. There are four> factors that work together to reduce, eliminate, or make reactive failure> maintenance into proactive predictive maintenance. They are:> > 1. Digital busses.> 2. Internal device diagnostics.> 3. Asset management software with record keeping and connections to> maintenance management systems.> 4. The ability to continuously monitor device health.> > 1. Digital busses. > > Some maintenance activities are eliminated because of the digital bus.> For> example, you no longer need to range the transmitter to the 4-20 mA> signal.> You also have the ability to send status information along with process> information. For example, a PV form fieldbus is flagged good, bad, or> uncertain. With this simple status information many if not most routine> device checks are eliminated.> > 2. Internal device diagnostics.> > There is a huge and rapidly expanding list of internal diagnostics. They> include diagnostics for the transmitter, for the process, and even for> equipment not directly a part of the transmitte. Some examples follow.> Many devices have a temperature sensor in the process sensor or> transmitter> electronics. Take a pressure transmitter for example. A temperature> measurement in the pressure sensor could alarm if the temperature is too> high. For example, insulation on a steam line could have been damaged,> and> the device is overheating. Or heat tracing could have failed, and the> device is about to freeze and burst, or the tracing is not turned off in> the> spring and the transmitter is overheating.> > Internal transmitter diagnostics can statistically monitor the performance> of the process to detect excess process variability, or other difficult to> detect conditions such as an evaporating wet leg in a level measurement.> Transmitters can also detect conditions such as plugged impulse legs.> Both> of these can eliminate routine maintenance activities, and alert the user> to> conditions either before they occur, or just after they occur and before> damage can extend to other equipment or effect product quality.> > 3. Asset management software with record keeping and connections to> maintenance management systems.> > Some maintenance activities are caused by systemic problems. For example,> if your asset management software keeps historical calibration records,> you> can check for calibration problems. If a transmitter is drifting out of> spec every 6 months, there is a problem with the device and it needs to be> replaced. If your calibration records show that the device has been in> specification and did not need calibration the last 3 times you checked> it,> that is an indiciation you can extend your tme interval between> calibration.> Another key is asset management software that eliminates manual data entry> as much as possible. A typical technician can spend up to 50% of their> time> filling in and handling paper records. If maintenance activities and> results can be automatically captured by the asset management system when> done, and communicatons with the work order system performed without the> need for the technician to fill in forms, most of the time spent in these> cativities is eliminated.> > 4. The ability to continuously monitor device health. > > With Foundation fieldbus a device can continuously send health and> diagnostic data. If the host can continuously recieve and present this> data> to both operators and maintenance in an easy to use way, problems are> caught> before the become serious. In addition, operator requests for maintenance> checks are reduced. About 30% of all field device maintenance requests> result in no problem found. Often things don't 'feel right' to the> operator, so maintenance is called. This is the MOST EXPENSIVE device> maintenance call because if there is no problem everything must be checked> and eliminated. If the operator can see correct operational status and> device health, eventually these calls reduce and stop, and significant> maintenance savings result.> > The combination of these four capabilities can reduce maintenance time per> device from 30% to over 80%.> > Again, sorry about the length of the reply, but this is a very important> question and deserved a more complete answer.> > Regards,> > Tom Wallace> Emerson Process Management

[Stephen Mitschke]

> Dear Mr. Toro,> > In addition to other points already mentioned on this topic I would like> to> add:> > 1. Just to make it clear, far from all network protocols provides the> framework for the interoperability required for efficient maintenance.> Only> HART, FOUNDATION(tm) Fieldbus and PROFIBUS PA do.> > 2. Another advantage of interoperability is that many devices are> configured> and operated in pretty much the same basic way. Configuration download of> a> replaced device can be done centrally.> > 3. Diagnostics can be achieved using a single device because the> diagnostics> is built into the device. In the past you needed two analog transmitters> and> check for signal discrepancy to find problems.> > 4. On top of maintenance, the FOUNDATION(tm) function block diagram> programming language has built in interlocks that shuts the loop down or> freezes it in case of failure.> > 5. The diagnostics is important because it lets you know about a> problem> immediately and where. In the past problems were only detected when the> product was out of spec and the QA department sounded the alarm, and then> you still did not know where exactly the problem was. Don't waste your> time> checking OK devices. With Fieldbus the problem is pinpointed faster so you> can get back online quicker. The more detail diagnostics tells let you> deduce what spares and tools to bring. You can find out quite a lot even> before going into the field. You may even be able to determine that there> is> indeed no instrument problem and that the upset is actually a process> problem.> > 6. You can get other info useful for maintenance other than> diagnostics.> E.g. you can see materials of construction, connection sizes etc. allowing> you to access the suitability of the device for the application and to get> spares. Identification information. Search for devices in your plant. All> of> this from your PC, before having to get out of your chair.> > 7. Check out this book:> http://www.isa.org/isastore/isastore.cfm?CFID=70190&CFTOKEN=56918965&SubSt> or> e=47&Store=11&ProductID=1816&ProductClass=P&ContentType=4> > Chapter 9 covers this topic.> > Jonas> Smar

[Stephen Mitschke]

> I have been a great believer in Asset Management, however on several> plants> have been unable to convince Operations of the benefits partly because of> the Plant Maintenance System.> > One of the main problems is that most plants have what they call the Plant> Maintenance System which is usually a Dinosaur which resides on the> Mainframe.> Most people who have to use this system (and choice is not an option)> loathe> it and find it inflexible.> Things can easily be scheduled as follows:> Daily> Weekly> Monthly> 3 Monthly> 6 Monthly> Yearly> 3 Yearly> > The Dinosaur does not like to provide too many options. It is usually tied> to Parent Equipment such as Vessels, Pumps and Relief Valves which fit in> well with the 3, 6, 9, 12 or 36 Monthly schedules. Instrumentation has to> fall in line with the above time schedule.> If your scheduled Shutdown is brought forward you need to fiddle with the> Dinosaur to come up with a semblance of what instruments need to be> maintained at short notice.> > When it comes to your control system and Asset Management System it is> obvious that all maintenence can be easily programed around the time the> plant is on line. If a plant is scheduled to be online for the next 99> days,> then should something start to play up tonight, then the Asset Management> System can provide a list of equipment that maybe maintened should the> plant> go down. It will have the calibration status of all the instruments at the> press of a button. It will also tell you wether these instruments have> been> performing within spec or not.> > The AMS can also provide details of equipment that may be worked on while> the plant is running without trying to check with the Dinosaur. It will> also> give running status (Run Hours etc) of running time of pumps and electric> drives which the Dinosaur cannot do.> > The AMS is easier to update than the Dinosaur when a transmitter or pump> has> been replaced as its data in uploaded when the instrument is put into> service. Usually the Dinosaur never gets to know about the change.> > When a plant has a problem with an instrument, it is not necessary to> replace it with the exact make and model. You may be able to review the> equipment already installed by reviewing the AMS database and use a> simillar> piece of equipment that is not currently in service. Easy with AMS> > A recent check of our store stock revealed a stock holding of 32 models of> 1151 Pressure and DP transmitters (probably more stock than than the local> representative). Sensible use of an AMS would reduce this to a few models.> > With the sophisticated instruments now available with a huge array of> diagnostics, and a three to five year calibration why bother to calibrate> except for custody transfer or statutory reasons. Even calibration of> instruments on a safety system would not need to be carried out as often.> > Finaly I have yet to see a Dinosaur connected to an Asset management> System> as it cannot cope with the real world and real time.> > Ambrose Hargan

[Stephen Mitschke]

> It is true that currently available Dinosaur Asset Managements that is> also> part of the whole ERP (Enterprise Resource Planning) are not the same and> integrated with those Asset Managements developed by some FF Equipment> Vendors. I would say that though it is called 'Asset Management' it only> covers those FF Devices (and its properties) within the system; I would> preferably call these as 'Instrument Asset Management' rather than 'Asset> Management' in general terms. We can see that the Dinosaur Asset> Management> Entities are mostly for Process Equipments with some sub-entities such as> pumps, valves, instruments, etc. which are defined as 'Parts' of the> Entity> itself.> > As a matter of fact, I haven't seen any 'Instrument Asset Management' that> is really integrated with that Dinosaur ERP even via their Maintenance> Suite> or MMS (Maintenance Management System). It would very ideal if we can> really> have integrated solution where diagnostic information from 'FF Instrument> Asset Management' could automatically trigger a job order to Maintenance> Suite / MMS which then also check Inventory and do automatic Purchase (if> stock inventory is not available) in the Dinosaur Asset Management and> ERP.> We still see a gap between 'Instrument Asset Management' and those> Dinosaur> Asset Management Systems. That many information we got from 'Instrument> Asset Management' Systems, must still be manually entered to the> Maintenance> Suite / MMS.> > Best regards,> Waskita Indrasutanta> > http://www.wifgas.com/

[Stephen Mitschke]

> Dear all,> > I would like to bring up few issues related to maintenance of smart> instruments> not been discussed yet.> > Yes, there is diagnostics built in smart devices but diagnostics> information is> device or communication protocol specific which means that device or> protocol> specific tools are needed which are pain to maintenance and even more pain> to> plant operators. There exist standard, protocol specific diagnostics> interfaces> in the devices, e.g. "device needs maintenance" status flag but the device> alone> is not able to say when it needs maintenance except in the obvious cases> where> it already has failed.> > Yes, the diagnostics can be achieved but not easily because process> control> system vendors provide limited access to device diagnostics information> which> means for example that no manufacturer specific diagnostics is supported,> existing PROFIBUS DP interfaces are limited concerning PA diagnostics,> HART I/Os> exist not yet too many in the market, HART 5 has very limited cyclical> diagnostics capabilities and data access for third party diagnostics> applications is very limited. OPC helps here but is not widely accepted by> DCS> vendors. Direct connection to field devices in the case of HART and> PROFIBUS is> not very economical from the installation point-of-view.> > The most powerful features from these viewpoints come with FOUNDATION> FIELDBUS> with its open HSE system bus but in practice such as 100% FF networked> plant (in> large scale) will not exist for a while and maintenance people have to> cope with> hybrid solutions meanwhile (do not ask me how long).> > After this pessimistic message I want to emphasize that the basis for the> more> efficient maintenance are the digital communication and microprocessor> based> instruments. To reach that level some other things need to be done too.> > Kari Hartikainen> Endress & Hauser