Stephen Mitschke
August 12th, 2003, 10:47 AM
Ferril Ford and I have exchanged the attached a few times as possible waysto represent multivariable transmitters as is now available due to FF onP&ID's. We would like to seek your input as well. Once we have reached someconsensus on this list, we will then approach ISA and SP5.1 about preparinga Technical Report along the same line. After the Technical Report isapproved, we will likely approach all of you again to help make a completedocument for publishing.The difference between the 2 drawings is that the second sketch shows howall the devices on 1 segment 'hooked together' or connected, thus beingcloser to the real world installation and duplicating to some extent theinformation on a loop diagram.Ferril and I would like to know which version you folks prefer and why.Thanks.IanSyncrude Canada Ltd.PO Bag 4009, MD 0032Fort McMurray, AB T9H 3L1P 780 790-4079, Cell -799-6017F 780 799-5190verhappen.ian@syncrude.com<<Sec_4_9.ppt>> - - - IanP&IDs are designed for use by all plant personnel ( operators, ProcessEngineers, Plant Managers etc. ). For these people clarity over thefunction of the control scheme is more important than how it is implementedin the field. In other words they need clarity over which measurementinfluences which valve ( assuming the electronics all works ). TheInstrument Technician and Instrument Engineer on the other hand need tounderstand how the failure of a given component or fieldbus segment willeffect other components. They can use the loop diagram for this as opposedto the P&ID.I would therefore suggest sketch one which shows more clearly the controlscheme.Laurence Thring- - - - I have a concern when instrument bubbles start crowing out the processequipment on a P&ID. We have tried to simplify the main drawing, and justshow the functions or tags that the operator must deal with. We show thebehind the scenes details in a references detail on a P&ID detail page or inmargin notes. If the loop is really complex, we will have a complex loopwrite-up. I think we will need to continue that practice for complex loopsin Fieldbus.So, I would show alarms in a Transmitter AI block, and PID placement in avalve. We will show the operator where his interfaces are located. We willnot tell him what segment, controller / rack, ... the device is connected toon a P&ID. That will be on a segment drawing.I do not want FF by my Fieldbus devices as I have seen on some drawings. Iwould rather use a different connection line if anything. I liked your ideaof shading the open circles to make them fat dots.I am still a bit unclear on what to do with the 8-point blocks where I have8 measurements that come in a single block read. If they are alltemperatures, I am going to use a T prefix rather than a U. After that, Idon't have a good plan.I'm sure this will be like herding cats. Good luck.Herman Storey- - - - The symbol of the circle for an "instrument" is kind of universal and is OK,but the original ISA standards did not include a horizontal line inside thecircle unless the item was meant for a panel mounted display. It seems thatall of your circles have horizontal lines. The SAMA symbols I never liked,but they tended to show the computations at one location by shape andputting them inside one box. That might be more useful than drawing a boxaround circles.Anyway, just my thoughts.Dick Caro- - - - Ian:I hate to be a wet blanket, but I feel that we should stay with traditionand use the PID drawings and instrument symbols simply to represent thecontrol scheme.What I am getting at is that trying to show the FF segment connectivity onthe PID will make the PID more complicated than we can stand during thelimited time we have during project execution and the number of iterationsthat PID go through.My opinion is that we need to restrict PID representation to showinginstruments connected to the pipe or equipment and the SIGNAL FLOW requiredfor the control scheme to work plus any intervening control functions suchPID, low select relays, computational blocks etc. Since the host equipmentcomes with built in alarming, trending, etc, I question even showing alarmsin an explicit manner. In this age, alarming and trending requirements maybest be handled in a data list format maybe as an extension to instrumentindex.Regards,Charles- - - -I agree entirely with this sentiment. I want to treat Fieldbus as a nuanceof our control implementation, like I/O card loading was in traditional DCS.Experienced engineers were / are judicious about I/O card loading just as wewould be about segment loading. The added information on the P&ID wouldconfuse more than clarify.How many EPC firms use P&ID nomenclature to guide I&E designers? In theproject I just did, if a balloon on a P&ID didn't have a dashed line goingto it, it didn't get any wire or conduit routed to it. The more oneoptimizes the capabilities of Fieldbus (e.g., using hybrid multidrop /chickenfoot topologies) the more difficult it's going to become for EPCfirms to "turn the crank" and generate drawings.We may not be ready to make decisions about where a PID block resides at thefirst iteration of P&ID's. Just the same, I like the idea of having somemethod for flagging that PID is being solved in a field device, so long asit can be done without excessive clutter.Contiguous balloons is as far as I would go to show multivariable, and thenonly if I was actually using the secondary variables for control / processpurposes. I would also be satisfied if only a single field device was shown,with "signal flow" to multiple "shared display" (i.e. DCS screen) balloons.John Rezabek- - - -I prefer the version that shows what is on the segment.I would want to keep the field devices without the squares around them toshowthey are in the field vs. control system.I would keep the multivariable flow transmitter tagged as FT and eliminatetheUT. The lines thru the bubbles should only be for control house devices.For the future, multivariables (multiplexers) of the same variable, e.g.temperature, I would go back to the old ways, TJI-1 is the multiplexer andTJI-1-1 is the first indicator point and TJI-1-X is the last point.Regards, Ron Szanyi- - - - Ian, I tend to agree with Charles. In terms of P&ID representation itshouldnot be necessary to show that the device is part of a multivariable.Conventional symbols should be used along with the signal representing afieldbus loop.Kind Regards, Jim- - - - IanI support the idea of not using additional symbols to reflect the differentvariables of a multivariable device, ie not to use DT-XXX and FT-XXX inaddition to UT-XXX or UYT-XXX(Y for computing function) . This willhopefully reduce clutter on the P&ID. I would like to have the info on P&IDon what multivariable measurements are available from the device. This Isuggest can be in the form of texts alongside the UT symbol. See attachments(don't mind if I made modification to your original sketch!). Also, I would like to see functionality (control, alarm, indication...etc)of the loop originated from the multivariable field device represented onthe P&ID using conventional standard tags and symbols, examples FIC,DIC,...etc. This ensure information expected of a conventional instrumentP&ID is retained for other users (operators, process engineers..) of P&ID.And lastly I would note the data link for fieldbus with the letter F in thelittle circle to differentiate from other non fieldbus data links. Thiswould be useful when you have a mixed of different data link types coexistin a conventional instrument plant and fieldbus is used in the expansion ofthe facility.RegardsLim<<Sec_4_9_nic.ppt>> <<Sec_4_9_nic.ppt>>