Consider fieldbus for retrofits Oil Industry Publication: Recent spate of oil industry accidents "traceable to instrument repair or malfunction"

Benefits result from reduced maintenance, greater safety and increased performance

R. Dunlap, Fisher-Rosemount, Austin, Texas

Eonomic benefits of the digital fieldbus instrument communication standard have been developed and studied for new construction cases. Versus a traditional analog installation, fieldbus achieves cost reduction through reduced footprint, shortened configuration and commissioning times, and reduced wiring and labor costs. These are easily tracked, tangible savings. Fieldbus implementation can also be economically beneficial in the case of retrofits and revamps of existing hydrocarbon processing plants. When calculating the life cycle cost and benefit of a plant, stages beyond construction must be considered.

Although the economic discussion applies to all industries, it is especially important in hydrocarbon processing. Hydrocarbon processing is unique in terms of public scrutiny, industry scale, financial pressures, safety risk and constant revamp activity.

For the purposes of this article, benefits of fieldbus in retrofits will be split into the following categories: maintenance, safety, performance and profitability. We will consider a hypothetical refinery with 200,000 bpsd crude capacity and 100,000 bpsd FCC capacity.

Maintenance. Maintenance is necessary and expensive. Predictive maintenance is much more effective than planned or reactive maintenance. Maintenance is one of the largest controlled expenditures in any industrial concern and can soak up over 10% of the cost of goods sold.

Traditional 4  20 mA transmitters only report process variables. In contrast, much more than just process variable information flows over the fieldbus network. Fieldbus instruments create networked intelligence.

Now information such as time since last calibration and total valve stem travel can move over the network. Maintenance data can then be collected and stored for each instrument.

Fieldbus-compliant instruments can also warn when they are approaching failure. Experienced personnel will be more able to accurately predict how different field devices behave in different services and adjust the repair schedule for devices in particular services. They can also identify device-to-device idiosyncrasies.

Present technology allows a fieldbus-enabled control system to transfer instrument information to plant LANs, maintenance programs and engineers workstations. Data can then be used for vendor selection and purchasing decisions, evaluating metallurgy and enhancing scheduling considerations.

Regulatory and internal corporate regulations depend on instrument performance. A fieldbus-compliant system can keep its own audit trail.

As an example, during a refinery turnaround, many valves will be removed, checked, inspected and have new parts installed.

Assume the 200,000 bpsd refinery has 10 process units with 40 control valves each. Further assume that half of these will be repaired, inspected or have some effort spent on them. Under present methods this repair process would cost approximately $175,000 once labor of operators, pipefitters, machinists and technicians is included. Transportation, packing and gaskets would also be required.

Process industry studies show that only a fraction of the valves typically pulled for repair actually need to be repaired. The rate of unnecessary repair can be 50% to 70% or more. Fieldbus allows tracking the condition of each valve and determining when it needs repair. If the example refinery cut its valve repair rate to 30%, this would save $122,500 in labor and material costs. Determining the optimum interval for preventive maintenance will only occur with accurate and complete instrument data. Fieldbus-enabled devices take the trial-and-error out of determining the optimum PM interval.

Safety. Although as a whole the hydrocarbon industry is very safe, it has suffered a recent spate of accidents. Accidents may be occurring more frequently because fewer dollars can be allocated to staff, training and maintenance. Commodity price shifts eat at profits, resulting in capital and operating spending cuts. These shifts have put unprecedented strains on operators and machinery. Equipment and personnel run closer to physical and safety limits than ever before. Concomitant with throughput increases, accident size and severity have risen.

Thus, importance of safety and shutdown equipment is growing rapidly. Protection system complexity is rising with the complexity and throughput increases of the processes being protected. Safety system integrity depends on the integrity of its constituent field devices.

A large portion of these accidents are traceable to instrument repair or malfunction. For example, every time a manifold is opened to check a transmitter or a valve is removed to the shop for repacking, there are several points of contact between atmosphere and chemical, and between human and chemical.

Each contact point is a possible fire, explosion, exposure or illness.

Minimizing single points of failure has traditionally extended from the marshalling rack upward. In contrast, fieldbus can maximize safety by making an entire plant self-monitoring.

In a typical facility, 20% of the loops can be considered critical. Let us consider the definition of failure. If a transmitter or field device in a critical loop goes out of calibration, this can be a more severe problem than if the device becomes totally unavailable. A totally failed transmitter will make itself known by causing a shutdown or showing an out-of-range value. A transmitter in a critical loop with traditional field wiring architecture can prevent a necessary shutdown or cause one with no indication of an impending problem. Let us assume that a transmitter experiences a calibration failure every five years. Taking the reciprocal, every transmitter has 0.2 failures per year. Referring to our hypothetical plant, it will have on the order of 10,000 I/O points, 25% of which are transmitters and 20% of those are considered critical. Simple multiplication reveals that there will be 100 failures per year of critical transmitters. None of these failures will be reported over a 4  20 mA wiring scheme.

Total transmitter failures are much less frequent. At an assumed MTBF of 100 years the number of failures is still significant. Using the same calculation method as before and the same estimation for the number of critical loops means there will be five total failures per year of critical transmitters. This probably understates the actual case: as a plant ages, the probability of device failure rises and the number of failures will not be constant or predictable.

Considering the need to ensure both safety and plant uptime, the frequency of both of these types of failures is not acceptable. The safety, insurance and avoided shutdowns pay off the fieldbus implementation instantaneously. Only fieldbus-compliant devices can report their calibration status and warn of impending failures.

Performance. In many services, valve variability cost has not been well determined. However, research indicates that a small amount of variability results in large hidden costs. For every 0.5% reduction in valve variability, a 1% production benefit can be achieved. At this rate, an increase of just a few percent of variability in a critical service results in an enormous production cost.

It is not unheard of to find valves with 5% variability. This has a significant impact on refinery profitability. For example, cost of a single hour of 5% variability in an FCC unit feed valve is over $200 at present margins. This is equivalent to $1,600,000 per year. This variability cannot be determined by an outside operator and cannot be detected over traditional analog signal lines. However, a fieldbus-equipped valve could report its stem friction, position anomalies, time since last calibration and even give insight into wear on the internals.

Blending is another area in which every instrument is critical. Depending on which components are used to make up octane, 0.1 RON giveaway costs $3,000,000/yr/100,000 bpsd crude capacity. Rvp giveaway is also significant.

Tied in with the maintenance cost, instrument performance directly affects equipment reliability upstream and downstream of it. Cycling valves cause extra wear on pumps; wandering transmitters can cause fired heater damage. Detuned controllers, frustrated or apathetic operations staff, operating in a wider tolerance from setpoint, crowding the specifications, inability to engage APC and accidental trips decrease profitability. Fieldbus can cure many of these problems at the source.

Profitability. The information gathering devices in a plant form a pyramid:  Enterprise resource planning (ERP) links a site to a business strategy  Optimization schemes run over advanced process control (APC)  APC feeds new setpoints to the DCS  DCS systems process and display field data  Field devices collect data and provide control.

Profits are frequently increased through sophisticated software at the ERP level. However, this scheme cannot function without a solid base. Effectiveness of each successive layer depends on the integrity of the layer beneath.

Many of the benefits accrued to higher-level implementations are due to careful study of the layer beneath. For example, benefits of implementing APC largely result from tuning controllers internal to the DCS and repair of field devices below that. Before an APC project can be executed, a study of field devices and controller tuning commences. A study is required because the devices connected to 4  20 mA systems cannot inform engineers or staff of their health.

Let us assume a 5% benefit from APC, a 5% benefit from optimization and a 5% benefit from ERP. If the hypothetical refinery has profits of $4,300,000/day, then $4,300,000 x 1.05 x 1.05 x 1.05 is $4,900,000/day. The difference is $600,000/day. Let us also assume that the on-stream factor for APC is 90%. Raising the on-stream factor to 95% will increase profits by over $1,000,000/year. At this rate, payout for a fieldbus retrofit project is less than one year from APC alone.

The author Rob Dunlap is a business development specialist. His areas of expertise are in hydrocarbon processing, petrochemicals and refining. Rob provides applications knowledge as well as sales and customer support to the Performance Solutions group. He has over eight years of experience in the hydrocarbon processing industries. His experience includes: catalytic reforming, alkylation, simulated moving-bed separation, isomerization, hydrotreating, distillation, catalytic polymerization, phenol and other process units. Rob was employed by UOP, a process licensing company, prior to joining Fisher-Rosemount. He was responsible for inspecting, commissioning and troubleshooting instrumentation, distributed control systems and electrical systems. Rob also served as a technical advisor for process matters and plant startups. Rob also holds a BS in chemical engineering from Cornell University, New York, and is a member of the Instrument Society of America.

http://www.hydrocarbonprocessing.com/archive/archive_00-09/00-09_consider-dunlap.htm

-- Carl Jenkins (Somewherepress@aol.com), September 22, 2000

Answers

An excellent article Mr. Jenkins

"As for me...I shall finish the Game"!

~~~~~~~~~~~~~~~~~~~~Shakey~~~~~~~~~~~~~~~~~~~

-- Shakey (in_a_bunker@forty.feet), September 23, 2000.

I crossposted this interesting item on EZBoard and added the following comments of my own: ....After much sleuthing at the Department of Energy in March of 2000, including talking with a group of the team that oversaw DOE's Y2K efforts, I discovered that no one there had the embedded systems expertise necessary to identify, let alone track and assess oil sector-related embedded systems and complex integrated systems problems. Sad, but true. (That is why the connection between refinery disruptions and embedded systems has not been made by DOE.) No one on John Koskinen's staff had that expertise. That will help explain why the Information Coordination Center failed to track refinery disruptions and pipeline explosions, and other high hazard sector problems that have occurred and that are occurring at record levels since the first of the year. Y2K has been and remains an incredible example of flawed policymaking and implementation. An unexpected set of factors converged to bring us to this point in time with no high impact, high visibility catastrophes or disasters involving the chemical industry, nuclear power industry, the oil and gas sector, etc. For the most part, only fragments of the story concerning how we got this far so relatively unscathed are known and so far as I can determine, they are known by only a very few people. They tend to be individuals who feel themselves to be free from political, professional, or organizational constraints and who have sufficient time and interest to continue to track these concerns. Y2K is an issue area that involved and continues to involve technological complexities. With only occasional exceptions that I know of, Federal efforts, not to mention United Nations-sponsored efforts, failed to include individuals who possessed the requisite technological expertise in the most complex aspects of Y2K: embedded systems and complex integrated systems. By one software engineer's estimate, there are only around 104 firms and 2000 people in the world possessing such expertise. I am aware of only a few individuals with such expertise who have thus far indicated a willingness to come forward to share what they know or to alert policymakers and others in roles of responsibility concerning problems that are continuing to occur. This is for a variety of reasons, many of which are quite understandable. These reasons include liability concerns, fear of reprisals, threats of being fired, or fear of jeopardizing their ability to earn a living, etc. The failure to pay adequate attention to the most complex aspects of Y2K can be found in the fact that John Koskinen had no technical experts on his small staff of 11. In fact, he said in March in response to a question that I posed to him concerning the omission of technical experts from his staff, that he would not include a technical expert on his staff if he had it to do over. (You can link to that exchange at http://www.gwu.edu/~y2k/keypeople/gordon) As with other parts of the Federal Government's efforts, including those of Congress, the more complex and daunting aspects of the problem were not understood and there is certainly no evidence that there is any greater understanding of them now. In fact, to my knowledge, the more daunting problems that have continued to occur are not even being acknowledged, tracked or assessed, let alone acted upon. (This includes the many different types of malfunctions and failures that were predicted.) Other less daunting IT system-related problems are, in some quarters, not even being referred to as being "Y2K-related" and are simply being dealt with as any technical problem would be. In parts of the public sector that I am aware of, it is even politically incorrect to mention the fact that Y2K-related problems are occurring and are having to be addressed. Strange, but true. I think that there is some chance that the full story may come out one day. That day may well be years away; then again it could be months away. Right now, most people have little interest in the matter. They have no reason to believe that the government's January "declaration of total victory" was indeed premature. Of course, to many, the sole concern was always about worst case scenarios. While worst case scenarios of converging disruptions and disasters have been averted, low and mid-range impact scenarios are still within the realm of possibility. If we end up with a full scale energy crisis by the winter, the impact scale by the first quarter of next year could be a level 3 or more. (See Part 1 of my White Paper on my website regarding what such impacts could entail.) In some ways, it may well be a moot point whether or not Y2K played any role whatsoever even as regards the energy sector. In the following two ways, however, the role of Y2K is not a moot point: ~ It is impossible to develop sound policy if you do not understand the causes and the nature and scope of the problems being addressed. Repairing something that has malfunctioned or failed requires an understanding of what caused the malfunction or failure. Preventing a problem also requires understanding what might go wrong. ~ It does not bode well for the future if policymakers and those in roles of responsibility fail to understand what has happened and what is happening regarding Y2K and embedded systems. It does not bode well for the future if they approach other complex technological challenges, now and in the future, with equal disregard for technological expertise and understanding. It does not bode well if they fail to apply the lessons that should have been learned and that have yet to be learned from the Y2K and embedded systems problems that have occurred and that are continuing to occur.

-- Paula Gordon (pgordon@erols.com), September 23, 2000.

The crossposting on EZ Board can be found at http://pub5.ezboard.com/fyourdontimebomb2000.showMessage? topicID=11603.topic&index=1

-- Paula Gordon (pgordon@erols.com), September 23, 2000.

By the way, Carl, thanks for much for this find.

-- Paula Gordon (pgordon@erols.com), September 23, 2000.

Paula, it would seem to me that IEEE should be continuing to monitor the situation, given that they were first and most in-depth in their analyses of embedded sytems and their potential problems. Haven't been at their site for a while; think I'll toodle on over there....

-- Rachel Gibson (rgibson@hotmail.com), September 23, 2000.

I'm back. First, I should have said IEE (British). This particular page has a great deal of information that will take me a while to sift through. Any help would be appreciated.

-- Rachel Gibson (rgibson@hotmail.com), September 23, 2000.

Here's a link to the EZBoard Article
A recent spate of oil industry accidents

-- spider (spider0@usa.net), September 23, 2000.

Rachel,

What a good idea! Perhaps a way to begin would be to post a half dozen or so of the most interesting pages or items that you have found at the IEE site. A number of people will doubtlessly want to explore them further. You could put out a "plate of appetizers"....

Sorry, I am not able to pitch in at the moment.

Regards,

-- Paula Gordon (pgordon@erols.com), September 23, 2000.

Thanks, Spider, for the EZ Board link.

-- Paula Gordon (pgordon@erols.com), September 23, 2000.

Thanks for the link, spider.

Paula, this ap petizer appears to be most appropriate presently:

"L2.3 SLOW DRAG: Just as problems with the year 2000 appeared years before, in this scenario, problems will appear over time after 2000. As daily, weekly, monthly, quarterly, yearly programs encounter the problem, there is a constant but only slowly realised drag on all activities. In this scenario, everything done for the first time after 2000 will be problematic, and delays, errors and decreased productivity will diffuse through the economy, not always attributable to the year 2000 problem. The drag could be as significant as an increase in tax rates or energy prices. The year 2K could result in a recession, but the connection might not be obvious."

The above was written in 1998.

-- Rachel Gibson (rgibson@hotmail.com), September 23, 2000.

Extraordinary quote...

Thanks for the insurance link as well.

-- Paula Gordon (pgordon@erols.com), September 23, 2000.

The title of this thread is taken totally out of context. If you are still looking for Y2K evidence, there isn't any in this article. The author indicates that he thinks the recent spate of accidents is due to inadequate staffing, training, and maintenance exacerbated by the recent price increases which are causing "Equipment and personnel run closer to physical and safety limits than ever before. Concomitant with throughput increases, accident size and severity have risen." So, he is saying that accidents have increased because production demands have increased. Nowhere does he allude to digital equipment causing accidents directly, due to Y2K or otherwise.

Get your facts straight.

By the way, I'd say the author of this article would be a good one to contact to discuss the minute possiblility of Y2K errors causing any of these accidents. I'd bet he could set you straight in 5 minutes of conversation.

"Safety. Although as a whole the hydrocarbon industry is very safe, it has suffered a recent spate of accidents. Accidents may be occurring more frequently because fewer dollars can be allocated to staff, training and maintenance. Commodity price shifts eat at profits, resulting in capital and operating spending cuts. These shifts have put unprecedented strains on operators and machinery. Equipment and personnel run closer to physical and safety limits than ever before. Concomitant with throughput increases, accident size and severity have risen.

Thus, importance of safety and shutdown equipment is growing rapidly. Protection system complexity is rising with the complexity and throughput increases of the processes being protected. Safety system integrity depends on the integrity of its constituent field devices.

A large portion of these accidents are traceable to instrument repair or malfunction. For example, every time a manifold is opened to check a transmitter or a valve is removed to the shop for repacking, there are several points of contact between atmosphere and chemical, and between human and chemical.

Each contact point is a possible fire, explosion, exposure or illness. "

-- Buddy (buddydc@go.com), September 26, 2000.

Buddy, this is a stand alone sentence from the report in question:

"A large portion of these accidents are traceable to instrument repair or malfunction."

I think I have my facts perfectly straight. This is probably the key sentence in the entire report. In the journalism trade, this is called burying your lead. A good editor doesn't allow it.

My questions for you are:

what is causing the spate of instrument failures?

why are you knee jerk reacting to the possibility there may be y2k related instrument/embedded chip failures?

-- Carl Jenkins (Somewherepress@aol.com), September 26, 2000.

I dont understand this.

How can Paula state that: After much sleuthing at the Department of Energy in March of 2000, including talking with a group of the team that oversaw DOE's Y2K efforts, I discovered that no one there had the embedded systems expertise necessary to identify, let alone track and assess oil sector-related embedded systems and complex integrated systems problems.  when she herself has no expertise in the matter? She cant seem to tell the difference between normal  failures and ones that might be an exception. The refineries are said to be working at 96% to 98% of capacity. Anytime you run flat out the chances of human mistakes or equipment failing greatly increases.

One question that has to be asked is about the statement:  (That is why the connection between refinery disruptions and embedded systems has not been made by DOE.) is: What connection? The logic here seems to be from Alice in Wonderland via the Red Queen first we have the conclusion and the proof comes later. The idea that maybe, just maybe, the failures that are occurring have nothing to do with Y2K or even embedded chips/systems seems to be not even considered. After all if you know all the answers from the start why bother with questions?

For the most part, only fragments of the story concerning how we got this far so relatively unscathed are known and so far as I can determine, they are known by only a very few people.

No, actually they are known to a lot of people. You just dont want to believe them.

By one software engineer's estimate, there are only around 104 firms and 2000 people in the world possessing such expertise.

I cant help wonder if that is Ed? (LOL) I suspect there are a lot more people in who have that the expertise. And how do we know THAT engineer has the expertise to know how many people there are?

As for John Koskinen: He and his staff had a lot of experts feeding them information. There are a lot of experts who work for General Electric (as an example) but Jack Welch doesnt have them all in his office. He can contact any of them anytime he wants. The same thing with Koskinen. If he needed the expertise all he had to do was pick up the phone.

-- The Engineer (spcengineer@yahoo.com), September 26, 2000.

Sorry, Carl Jenkins, but it is not a stand-alone sentence. The preceding paragraph gives it context, and the sentences following clarify it.

"A large portion of these accidents are traceable to instrument repair or malfunction. For example, every time a manifold is opened to check a transmitter or a valve is removed to the shop for repacking, there are several points of contact between atmosphere and chemical, and between human and chemical.

Each contact point is a possible fire, explosion, exposure or illness. "

As for knee-jerk reactions, I'll fight fire with fire. The Y2K-doom movement has been one big knee-jerk reaction from the get-go.

-- Buddy (buddydc@go.com), September 27, 2000.

Buddy, you failed to answer the most interesting question:

what is causing the spate of instrument failures that have lead to the accidents?

Though the report in question doesn't provide sufficient information to decide, these are the types of failures that were predicted as an outcome of y2k.

I have not said that these failures are y2k related. Frankly I don't know. This report is merely a good investigative lead.

Also, far from being a "doomer" I am merely investigating and posting incidents that are ignored by the global media. I do find it very intriguing that we are facing the worst energy crisis since the 1970s, a situation which was predicted to be one of the primary indicators of the severity of y2k related disruptions.....

I find it interesting that there are many people like yourself who don't even want the questions posed......

-- Carl Jenkins (Somewherepress@aol.com), September 27, 2000.

"a situation which was predicted to be one of the primary indicators of the severity of y2k related disruptions..... "

This makes me laugh every time I hear it. The present "oil crisis" has nothing to do with equipment breakdowns. OPEC cut back their output on purpose to make the price rise. They said they would do it and they did it. Where is the mystery?

"what is causing the spate of instrument failures that have lead to the accidents? "

According to the guy who wrote the article you posted here: "Accidents may be occurring more frequently because fewer dollars can be allocated to staff, training and maintenance. Commodity price shifts eat at profits, resulting in capital and operating spending cuts. These shifts have put unprecedented strains on operators and machinery. Equipment and personnel run closer to physical and safety limits than ever before. Concomitant with throughput increases, accident size and severity have risen. "

"these are the types of failures that were predicted as an outcome of y2k."

So were nuclear plant problems, problems with trains, planes, ships, military equipment, you name it. Why would Y2K-related problems be happening only in the oil industry? Many industries use the same type s of digital equipment. Also, those predictions were made by people who we now know didn't know what they were talking about (I knew before, but it's been proven beyond a doubt now).

I never said you were a "doomer". I don't know if you were/are or not. I'm merely pointing out that the Y2K-doom movement was led by folks who didn't know what they were talking about, didn't have any Y2K project experience, and didn't fix any Y2K problems. That's where your predictions came from.

-- Buddy (buddydc@go.com), September 27, 2000.