Stephen Mitschke
August 12th, 2003, 11:37 AM
Everyone has seen the economic numbers in favour of Fieldbus, especiallyfrom suppliers. What we REALLY need are some firm numbers from actualprojects that have been completed. What I am asking for this time, is ifpeople will reply to me with answers to the following questions, I will thencompile it into a document and share it with the list. I will of courseprotect everyone's name in this document so each of you will remainanonymous to everyone else, (Unless you give me permission to do otherwise).Economic Benefit(s)/Disadvantages of FF:- ROI (% or $)- Commissioning (days, $)- Training ($)- Field Installation ($)- Host Installation ($)- Engineering ($)Other comments.Some uses for this information include:- Training courses- Foundation promotion- justification for YOUR project at your facilityThanks AGAIN, for your help folks.- - - - As a supporter and believer, I would like to add a few comments on thistopic:The savings that are enabled by FOUNDATION fieldbus are real and sometimestoo good to believe until experienced first hand. (I have won a few smallfriendly wagers here.) I would like to add a note of caution. The problemwith sharing case study savings numbers from other end-users is that themetrics are not always "portable". We ran into these issues when we startedreporting the actual, end-user validated project savings with our firstFOUNDATION fieldbus installation 4 years ago. Only a few Canadian end-usersreally believed them at first.There are many factors that determine the magnitude of savings: application,hazardous area requirements, distances, supported functionality in theproducts, site economics or cost structure, remoteness of the site, worldarea, how problematic the process is to maintain and keep at high rates,maintenance staffing, regulatory, quality and safety requirements, rootcauses of historical process interruptions, and on and on. All projects andsites are different. I would encourage you to ask your vendors about thesethings to validate claims when the numbers seem to high. Find out how muchthey really know and what they have already delivered.The key is in how the numbers are calculated. It takes some work,especially with estimating operational and maintenance savings. Acceptanceof the method of calculation and the base data is the key to convincingothers in your organization to accept the savings numbers. Face value justwon't cut it. It takes some proving to overcome the perceived riskassociated with the change.Therefore, the results don't have as much value without the data. The traphere is that sometimes the detailed data is proprietary! Some of thisinformation is held closely because of competitive advantage in low margincommodity markets. I would encourage those willing to share theirexperiences to provide as much background data as possible.Regards,Doyle Broom-------Ian,We have in SPDC not yet experience with fieldbus, not installations yet. Butthis is coming. The figures used here in SPDC (Shell Petroleum DevelopmentCompany in Nigeria)are from some vendors. I am not sure that these arecorrect, especially when you start for the first time an implementation. Ifyou are an experience user perhaps the figures going into that direction.RegardsJisk Kronemeijer.- - - - Dear sirs, I have a "story" comment on what Mr. Doyle Broom said in his email.Yes, I also believe you. I think that the fieldbus technology save money,particularly with FF.I work in a consultancy, basically for the petrochemical industry in theRotterdam harbour (Think about Shell and this sort of companies).I have gone to some seminars or presentations about FF (and otherfieldbuses) and the impressions that my colleagues and me have is always thesame. The things are not CLEAR enough to able to risk our reputation (It isjust a way to say it) saying to a client that the very best by far nowadaysin process control is FF. It goes about psychology, and psychology can beonly changed by means of really clear assortments. I think that the bestway for that is speaking in terms of "that is so BECAUSE... with examples".My colleagues and me have seen to many times the a list in a powerpointsheet titled "advantages of FF" and afterwards a list with more or less theitems Mr. Broom has mentioned like control to the devices, less wiring,higher availability,... The thing is that from my point of view there is akind of fog in all this issues, or at least this is the impression we have.So at the end it still remain the idea, "Yes, it is a really nice story, buttypical I/Ocables and devices are really reliable, and much more nowadays than this newFF".What I really miss is something like:1)we save money because we need lees cabling (cost estimate %) how manyintruments we use per segment, how long the average length... that meansbetween x1 and x2 % savings.2) control: Advantages of control in the devices, why is better, do wereally delivered what is expected?3) more functionality, why, how, cost saving that this suppose,...4) and like this with everythingI don't want to be really critic about that (I always send critics) but Ithink is the way to convince people.In the other hand, I have never heard what has to come yet, because I alwayssuspect from a product that is supposed to be already the best in the world,It can not be perfect, there could be limitations, not only physical (Exxareas, number of elements per segment,...) but qualitative (nice to have 30devices in a cable, but if the cable breaks, I don't have 30 devices in oneway).Well, I hope you all to see my critics from a constructive point of view.Kind regards to allJose M. Ferro- - - - Look in the April edition of InTech for a case history written by JohnRezabek for some guidance on justifying the implementation of Fieldbus.Although the hard numbers are not there in detail, John identifies a totalof ~$200,000. There is some excellent guidance on areas of savings toestimate. http://www.isa.orgRich Timoney-----Mr. Ferro,I understand what you mean about seeing through the fog. There is not muchmiddle ground between general claims or examples and specific applicationbenefits. I also agree with your request for examples backing claims. In away, that was the reason for my request that contributors provide backgroundinformation. Here are some quick responses to your questions. Don't expect too muchdetail. After all, my business is consulting.1) I have seen Total Installed Cost savings ranging from 15-40% (bothestimated and actual) with the result depending on the factors I mentioned.To get more specific, you have to look at a cost model of particular case.We use detailed costing as opposed to rules of thumb. 2) Why Control In the Field (CIF)? "Better" loop performance and a degree ofadditional redundancy with proper management of device LAS. Delivering onexpectations depends on what you expect or require (specifications).3) Functionality should be directly tied to a benefit. That is how youdecide what you need in your specs. Functionality and performance offieldbus devices and hosts vary quite a bit. Guess what? Since benefits gobeyond installation, grasping it all requires an understanding of operationsand maintenance (my former life as an end-user). 4) (cable breaks) How often does a cable break, when and where does thishappen? Analyze failure modes, criticality and current requirements for4-20. You have to arrive at your own conclusion on risk vs. benefit.Personally, I don't see any issues with cable breaks after discussing thisand implementing projects for several years. There are always limitations in the functionality-performance-cost orgood-fast-cheap tradeoff. A learning curve investment is required to getthrough all of the fog and still stay between the guardrails. The optionsare self learning or getting help from someone who really knows.Jump in. The water is fine.Best regards,Doyle Broom- - - - Well, here is my experience so far. I have made a decision to use FF for a very minor upgrade in my company'sfuel terminal. There are no control loops at all, so that is an irrelevant factor. In factit is just valve contol, open and close (not even variable flow required!)It is going to cost more to use FF than using standard wiring. There are only 6 valves, and if I used a standard system it would probablycost about A$8000 - A$10 000.(plus a bit of labour)Using FF it will cost A$8000 just for the smar 302 interface, plus ~ A$1500per valve for Smar positioner/Flowserve Buswitch, and maybe A$1000 forcable.total cost about A$ 20 000 with the labour. Why have I decided to do this? 1. Because within the company, sooner or later someone has to take astep/risk with the new technologies. If not me then who?2. How else are our electricians and techs and electrical engineer going tolearn about these things? 3. When we do a big (real) project where there maybe real savings, would wetake the risk on something that is new to us? Pretty unlikely.4. I also plan to install radars levels guages on some of our tanks, and ifwe use a bus system (FF or Profibus) there is not really any extra cost overstandard, for the two brands we are considering (Vega and Krohne) ~A$2500 -3000. By using FF (or Profibus for that matter) I do not need to run any newcables, or allocate i/os on a PLC etc. I just plug it in to the nearest busline (obviously a few metres of cable required, and I have made sure therewill not be too many instruments on the IS system)5 I feel (perhaps incorrectly, but it is not a deciding factor anyway) thatusing a FF system makes it easier for me to be able to access plantinformation (levels, flows etc) from my desktop pc. With a normal system thecontrols would need to go into a plc and need to be assigned etc etc. (Youcan tell I'm not an expert at plc stuff can't you!)With the FF HSE adaptor that goes onto the Smar 302, it should be very easyto interface with our existing ethernet system.I hope this helps. Russell Mackenzie (B.Eng- Mech) Tank Terminal Manager - Engineering Coogee Chemicals -------Just a note to let you know that even if FF does not save any wire, it savesat LEAST 50% on terminations versus a conventional analogue installation,even IS, which will present the 'worst case' scenario for number of deviceson the network/segment.less wiring, higher availability,... The thing is that from my point ofview there is a kind of fog in all this issues, or at least this is theimpression we have. So at the end it still remain the idea, "Yes, it is areally nice story, but typical I/O cables and devices are really reliable,and much more nowadays than this new FF".- - - - -Thanks Chuck, I still hold that FF requires fewer terminations, even if Iuse the same length of wire. I do agree with you that the cost of the cableis only a small part of the installed cost for the wiring infrastructure sothe 'wire savings' are to some degree a 'red herring' (if I mix metaphors.)IanIan -It was not long ago that we all "wired" our offices and added terminals oneveryone's desk! Why! Better communications? Sharing files? RemoteAccess? Improved information flow? NO!!!! It was to save paper!This is because we can count our paper cost but could not value the otherbenefits.Did we save Paper???? No way!!!!Will we remove the communication link (the wire) because we did not achieveour desired goal of a "Paperless Office"or lower paper cost? No!With Fieldbus, we can count the wire cost, so we do. In 2-3 years, we willlook back and laugh at the wire savings discussion. It needs to be this wayfor now but it's so funny how history can repeat itself!Each application has to evaluate FF technology or any other technology withthe specific application situation in mind. Critical factors that effectthe final decision vary depending on many issues - too many to list here.As with the introduction of any new technology, the benefits need to bemolded and modified to meet everyone's needs which takes time and fieldexperience. Do I need to remind you about the operator interface orfunctionality of the VIC 20 compared to today's PC / Software?????I just had to weigh in here with a few comments.Keep up the good work. The activity and questions are very exciting.Regards,Chuck Micallef- - - - - Chuck, Clever analogy! I agree that it seems that FF technology is slightly aheadof its time and that the real benefits will be realized only after it isimplemented.Chris George Technologist Southern Alberta Institute of Technology------Here is a paper Ian Verhappen prepared a few years ago, that will likelyhave better economics now, since there is more competition in avaialbledevices.http://www.isa.org/journals/intech/brief/1,1161,215,00.html?userid=&first=101&last=275- - - -- As responses to this thread have been indicating, this topic is to somedegree site specific. Well, here's another one in that category.As we all know, instrumentation and control are always at the back end ofprojects, especially the 'loop checks," calibration, and commissioning.Fieldbus, especially with the CFF file and off line configuration can reducethe "loop check" and commissioning phase by an order of magnitude or better.This means an earlier plant start up. An earlier start up means productionsooner. The question you need to answer is, "What is an extra day'sproduction worth to you?" On a larger project this changes to several daysand in our case with oil at its current price this can add up to a lot in ahurry.Other ideas folks? - Ian- - - - IanWhilst I agree with you that in theory this technology can reducecommissioning time, the actual situation can be very different. Thistechnology is still developing very fast and everyone is learning newthings. It is not something which can be included in a project andexpected to work perfectly first time ( such as a 4-20mA system ). Ourlimited experience is that commissioning actually takes longer as there aremore technical problems to resolve. As more people become experienced inthe application of the technology in the field ( Design, Construction andMaintenance personnel ), then commissioning times will be reduced and thepotentially large savings that you suggest achieved. Currently I suspectthat the HART based systems offer the greatest direct savings inCommissioning & Maintenance with minimal project impact, although over timeI would expect that to change.Laurence- - - -I would be interested to know if Lawrence's experience with "increasedcommissioning time" relative to HART was with a modern DCS versus a legacyDCS. I can see how someone might struggle more with a legacy DCS, whose FFinterface is poorly integrated and supported.The DCS we used in Lima was designed to support FF. Compared to thecommissioning of HART-smart 4-20 mA instruments that were connected to ourESD system, DCS-based FF instruments were much faster to commission, loopcheck, trouble shoot, etc. They were at least "as perfect" if not "moreperfect" than their HART counterparts.For one thing, the "visibility" of a device's configuration is so muchclearer in the FF world. I can get a lot of information out of a HARTdevice, provided I'm happy to drill down through numerous menus, andassuming you know where to look. If I ask a technician to "verify that allflowmeters are taking a square root where appropriate", he'll be putteringaround with a HART communicator a lot longer, versus the same task on awell-integrated FF DCS.A lot of DCS vendors would like us to keep buying their old legacy DCS's.They make more money for the vendor, and there's probably still a fairamount of inventory and manufacturing capability that's under-utilized.Likewise a lot of our own controls professionals hang on to the "old ways".We've gotten pretty good at making legacy DCS's do what we want over thepast 10 or 20 years. I like Dick Weismeijer's maxim, "Arrest Comfort Engineering". The sooner wemove away from legacy solutions, the less they'll be an albatross.John Rezabek- - - - - - I have inserted some initial comments below as well, hence my tardiness indistributing this note.Great discussion!I'm new to this, so please help me understand your (collective) perspectiveonthe system. I feel somewhat uncomfortable sharing my questions below, but Itrust that I am among a group of professionals who can help me honestlyevaluatethe strengths and weaknesses of this system. Many people have beenstrugglingwith how to present the savings potential of Fieldbus. Perhaps by yourresponseswe will all get a better feeling on how to justify these systems within ourrespective companies. If you could answer just one or some of thesequestionsyour input would be appreciated.1. TerminationsI don't understand the comment "FF requires fewer terminations."a. 4 to 20 mAI will use an example case with 6 instruments. With standard 4 to 20 mA wenormally use twisted pair wiring (not shielded), one cable directly from theinstrument all the way back to the I/O card in the DCS or PLC. Counting twoterminations at each instrument and two at the I/O for that instrument, Ihave 4terminations per instrument times 6 instruments = 24 terminations.If I use shielded pair this becomes 30 terminations (shield terminated atI/O,shield cut back at the instrument).b. FOUNDATION Fieldbus H1I will use shielded pair (per the specification). To go from the 6instrumentsback to their Fieldbus terminator will give me the same 30 terminations thatIhave above. In addition to this, I need a cable from that terminator to theterminator at the other end, and then from that terminator to the controlsystem. So it would appear that there are more terminations with Fieldbus.(If marshaling panels are used for the 4 to 20 wires, with multiconductorcablesbetween the marshaling panel/junction box and the control system, then 4 to20may have more terminations - but this is not the wiring method we typicallyuse.)[Verhappen, Ian] Terminators are normally / often incorporated in the hostsystem and FF terminals so these can be eliminated, saving a fewterminations. This does not significantly change your calculations but Ineed to clarify that the original statement assumed the use of marshallingcabinets AND I.S. barriers. (Not all of us are as fortunate as you to havesuch nice short home run lengths)2. Hardware CostsPurchased hardware prices appear to be substantially higher for Fieldbusoverconventional 4 to 20. In discussing prices with a vendor I found thefollowing:- FF pressure transmitter is $685 more than 4 to 20 PT of same type (ahigh-endPT in common use)- FF temperature transmitter is $407 more- FF control valve positioner is $281 morePlus FF usage license of about $20/instrument (from a DCS vendor).Cost for the fieldbus I/O is perhaps $50 more per instrument than 4 to 20 mAI/O(costs from one of our vendors) plus say $25 per instrument for terminationblocks (I am sure there are less expensive ones available).These costs seem to be excessive and will require substantial savingselsewhereto overcome.[Verhappen, Ian] Be sure to shop around. I'm sure you can do better thanthis.Hint, Hint vendors. This may be an opportunity to help FF become adoptedquicker if the price differentials are reduced. 3. Engineering - I am sure that there are different perspectives on this.a. Wiring DiagramsI have heard (but do not understand) that there are wiring drawing savings.Maybe someone could discuss specifically why - it would seem to me that thewiring diagrams for each type system would be roughly the same.b. Electrical Requirements - Fieldbus requires calculation of bus loading /update speed limitations, and power supply voltage drop issues. Signaldegradation / attenuation limitations and distortion effects on network sizeisalso discussed in the manuals. Many vendors have spreadsheets to help withthis,but it would still seem that the engineering contractor will charge for thisservice. Since I have not yet done this myself I do not know how easy ordifficult this is.c. Fieldbus requires additional training for engineers for wiring and DCSconfiguration, and for the plant maintenance people. One vendor suggestedoneweek of Fieldbus-specific training.[Verhappen, Ian] Experience is coming, and non-vendor specific training isalso becoming available, so the first job will have this steeper learningcurve but the future looks bright. Some of this engineering cost should beoffset by reduced configuration time, since it can be done in more of adatabase environment.4. Capital SavingsIt would seem that the primary capital savings is in the single cable "homerun"vs. 6 cables required in my earlier example. Most of our plants arephysicallysmall, and this length would typically be 200 feet. This would not seem tobe anappreciable savings. Are there other capital cost savings that I amoverlooking?5. Start-Up Time SavingsSome people have noted significant startup time savings. Discussing thiswithothers who have actual Fieldbus startup experience, and reading JohnRezabek'sexcellent article, it would seem as though the procedure for loop checkoutisvery similar for both systems (disconnect wires at the device and see if theDCSsystem reports the proper tag number dropped).[Verhappen, Ian] Do you 'ring out' each cable as well? with Fieldbus thisis not necessary.6. Device DownloadingIn theory I believe that Fieldbus should be similar to connecting a newprinterto your computer. You need to find the printer driver for your specificprinterand download it to your PC. (Of course none of this is required with 4 to20 mAsystems). Our (very limited) experience, where we have a DCS and fieldinstrumentation from different vendors, is that this is not so easy. Out oftheten or so instruments we have purchased, only one currently will workbecause ofvarious firmware/software incompatibilities. I know this is just a"start-up"problem, and as the Fieldbus specifications evolve this will go away.[Verhappen, Ian] The Capabilities File is intended to prevent this problem.Does your host support the CFF file? You should also check for HISTcompliance (host) and software revision information (Device) are the same.A couple of more significant concerns:- if an instrument needs to be replaced in the field with a new one, andthatnew one does not use exactly the same device drivers, the device will notwork.An instrument technician (if we even have one on staff at the time) willhave tofind the appropriate software file from the vendor's web site and downloadit tothe control system in order for the instrument to work. Some of our plantsmaynot have this capability on hand especially during off-hours.[Verhappen, Ian] Software revision control problem again.- Even if an instrument is certified to the latest version of ITK 4.01, theDCSvendors often still would like to test one of these devices with theirsystem.Their comment is that even ITK 4.0, although it goes a long way, is stillnot acompletely bulletproof interoperability specification. Having to insure thattheDCS vendor has tested and certified any given instrument for their systemcan bea real challenge.[Verhappen, Ian] INSIST that your host vendor uses the HIST program andthen this checking is done 'automatically'. Users asked for the HISTprogram, so now we should use it as intended.7. System RobustnessWith 4 to 20 systems, an open wire or a short typically will only take oneinstrument down. In Fieldbus there is a potential that a short will taketheentire segment down (reference "A New Hardware Concept for Fieldbus"). In myexperience shorts and open wires are not common, but they can happen. Thereferenced article pays some level of attention to this issue. In myexperienceI have short circuited many loops as I disconnected the instrument wiringwithout due care. The worst that happened to me was that I blew a smallfuse inthe I/O card. Taking an entire segment down would be more harmful.[Verhappen, Ian] This is the reason for special terminal blocks withinherent 'short circuit protection' for each spur. Check out Relcom Inc andWeidmueller for these devices.4 to 20 signals are inherently noise-immune. Fieldbus is obviously less so.Ihave heard people note that noise on the line has taken the entire H1segmentdown, but that these types of problems were generally due to impropergrounding.Once the grounding problem was identified and resolved (usually at start-up)thesignaling system proved to be reliable. What are other's experiences?8. Maintenance SavingsFirst, it is very hard to use projected maintenance savings to justifyadditional capital expenses. Second, as of today there appear not to be muchavailable in the way of diagnostics on Fieldbus that is not equallyavailablevia HART. There is little available in the way of predictive diagnosticsbeyondcontrol valves. I know this will be changing in the (near?) future asinstrumentvendors incorporate advanced diagnostics into their instruments.9. SpeedWe all have experience with the speed of 4 to 20 signaling technology,transmitters, etc. It is less clear how fast Fieldbus actually is indemandingreal-time control applications such as liquid pressure, flow, etc. Thingssuchas the number ofd instruments on the segment, type of instruments,configuration, where the controller is located all may have an impact on thespeed of a loop.[Verhappen, Ian] They do have an impact and hence the reason to engineerthe system, just like you need to do today. I just saw a paper on the DeltaVsite where FF is being used for compressor control.10. Batching/Dosing ApplicationsAll of the discussions I have had on Fieldbus have centered around its useincontinuous applications. How does Fieldbus fit into the world of high speedbatching or dosing applications, such as flow metering to fill a bottle inonesecond? Does Fieldbus have the capability to do this? If so, how? (Iwouldassume that Fieldbus would be used to download a target weight to theinstrument. All calculations, such as flow totalization and comparison tothedownloaded setpoint, would be accomplished in the instrument, and then acontactclosure output on the device would be used to trigger the cutoff valve.)11. Instrument Spare PartsToday pretty much any 4 to 20 mA device in a storeroom can be substitutedforany other, regardless of vendor, as long as the application and physicalrequirements are met.With Fieldbus devices, this may not be the case. If one instrument has adifferent revision of firmware from another, different software may need tobeobtained from the vendor's web site and downloaded to the control systembeforethat device will operate. I have heard that the way that some vendor'sdefinevariables (such as PID) are different from vendor to vendor, and that thismaycause problems or require reconfiguration of the point. Things such aswhetheror not negative numbers are allowed in a particular configuration field forthePID variables have been reported to cause problems.Others perspective and experience would be greatly appreciated!Best Regards,Jim Reizner- - - JohnThe experience was with a modern DCS. The extended commissioning time wasno great surprise as the object was to build up experience in this newtechnology, without excessive risk to plant production. The note from JimReizner describes some of the potential issues which need to be addressedwhen designing and installing a FF system. Since both suppliers and Userspersonnel are not yet familiar with all these issues they are not alwayscorrectly done. These faults then appear at commissioning and need to besolved before commissioning can be completed thus causing the delays.I have no evidence to suggest that a FF instrument itself is any more orless reliable than a standard HART 4-20mA model. In terms of 'visibility',clearly the use of a PC ( DCS ) based system is better than a HARTcommunicator. Such PC based systems are also available for HARTinstruments.I also favour moving to 'open' standards and I believe both HART and FFtechnologies are in theory 'open' standards. It is now up to the suppliers,working with the relevant User representatives to improve the practicalrobustness of these technologies in the field and to ensure that sufficientpeople with the relevant knowledge are available to implement them.Laurence