Abnormal Situation Prevention

You may recall Emerson's Bill Zhou from his demo video pieces here on the blog. I told him to be on the lookout for unscrupulous talent agents who'll want to whisk him away to Hollywood. Bill pointed me to a HART Communications Foundation announcement of Mitsubishi Chemical Corporation selection as 2009 Plant of the Year. Their use of HART communications helped deliver strong operational and financial results:

Diagnostic parameters that help detect signs of an abnormal situation or degrading performance are difficult to obtain with simple handheld devices because they require a time-consuming, manual, step-by-step approach," says Takayuki Aoyama, team leader, instrumentation group, Mitsubishi Chemical. "HART technology made it possible to access this data without manual operation. This made it much easier for us to gather data and detect abnormal situations from field devices and has reduced maintenance costs by 10 percent.

Bill and Takayuki presented at the 2009 Emerson Exchange with a session titled, Process Profiling: Investigation and Prediction of Process Upsets with Advanced Diagnostics. They share how the statistical process monitoring (SPM) technology found in the Rosemount 3051S measurement devices were used to measure flow through an orifice plate from a Naphtha tank to a series of furnaces. With the success of detecting plugs in their impulse lines, they expanded the SPM usage into other applications where the process was problematic.

The SPM diagnostics helped to quickly identify issues such as inadequate straight pipe length in DP flow, plugged manifold, compressor vibration issues, and even when strong winds affected the process.

Using the diagnostics to solve these conditions and spot problems early helped this award-winning plant to shift from time-based to condition-based maintenance and deliver quantified business results.

Update and bump: Tim just let me know that he's been elected to the position of 2^nd Vice Chair for the AIChE Fuels & Petrochemicals Division (FPD). As described below, he'll serve a four-year term including one year as division president. Congratulations, Tim!

Emerson's Tim Olsen, a refining industry performance consultant, recently presented to Argentinean refiners on advances in automation. Tim has 20 years of experience in refining starting as a technical advisor on refinery unit startups for UOP before joining Emerson as a performance consultant. He has been active in the AIChE Fuels & Petrochemicals Division (F&PD), is a 3-year, nationally elected director for the division and is the programming chair for Topical 7 on Refinery Processing.

In his presentation, he shared some statistics on the amazing march of technology over the last 40 years. Computer processing performance has increased 50% per year compounded over the 40 years. RAM memory and disk space storage have increased in a similar manner. Communications bandwidth for both wired and wireless have also experienced exponential speed increases.

The exception to this exponential technology performance increase is software productivity. It has increased only 4% annually from the 1970s to 2000 and is even less for the first decade of the 21^st century. It has become the dominant cost in system development and implementation. The upside is that software typically has a much longer life than hardware.

Tim highlighted some changes made possible by these trends including low-cost computing and communications, virtual organizations independent of distances, disappearing computational limits, and increased digitization since practical limits on storage space are disappearing.

Specific for refiners, the instruments that touch the process provide more than a process variable back to the operator. Build on digital bus networks, the operators and maintenance technicians receive the not only the process variables, but the goodness of the information. They also receive diagnostic and predictive alert information to help avoid abnormal situations. I recently highlighted some of these diagnostics accessible from handheld devices.

Refinery control rooms have also changed. Where once each unit had a separate control room with single loop analog control and panel board interfaces, there are now more site wide control rooms with coordinated multi-loop digital controls and high-resolution monitor-based operator interfaces.

The economic conditions in which refiners operate have also changed. Many may recall the days of cheap energy, cheap water, cheap waste disposal, less emission standards, predictable supply and price of raw materials, and a large talent pool. The controls and layered advanced software applications were the pricy component. These conditions have largely flipped over the past decade.

Tim offers many examples in the presentation. I'll highlight one--the hydrocracker quench control valve. The digital valve controller on this quench valve can identify low instrument air levels and alert maintenance techs and operators before poor control response occurs. Early recognition and reporting of this situation is critical since the reaction is exothermic. If the operations staff reacts and repairs the instrument air levels, emergency depressurization and shutdown can be avoided. See the presentation for other examples on predictive analytics, plant turnarounds, system modernization, key performance indicators (KPIs), advanced process controls, and more.

I'll also put in a plug for Tim having known him for many, many years. Tim has been active in the AIChE F&PD since 2002 and is seeking the 2nd Vice Chair position. When I asked how this works, he explained that you are elected as a 2^nd Vice Chair, become the 1^st Vice Chair the following year, assume the Chair/Division President the 3^rd year, and become the Past Chair in year 4 to advise the new Chair. It sounds like they have a solid, on-the-job training model to me!

Members in good standing with AIChE should have received an email to vote on-line between January 15^th and February 15^th. If you're a member, you can see the full list of candidates and positions and decide for whom you will cast your ballot.

MP3 | iTunes

Let's close this week with a new era for us in YouTube-based demonstration videos. The team has just added three new videos produced in high definition (HD) video format. The videos include:

If you click through and see the videos in HD on the YouTube site, what makes the HD quality so nice is that you can go full screen and really see the software screens in clear detail. Until now, videos created over the years and available on the DeltaVsystem YouTube channel did not have the resolution to view some of the software screens well.

In these videos, Emerson's Bill Zhou and Juan Gomez show the interaction of diagnostics available in the Rosemount 3051S integrated pressure, flow, and level multi-variable transmitter and how they make this data available to operators and maintenance personnel to avoid and resolve abnormal situations.

I know that some have written me to say that their IT department blocks these videos. I think the case must be made that there are significant business uses from training videos, to application videos, to even capabilities videos like these.

Perhaps this post is one more URL that you can share to help change the status quo.

Emerson's Bill Zhou is in town this week, escaping the cold north of Minnesota. You may recall Bill, from his Emerson Exchange plugged impulse line video demonstration. Actually, his purpose was not to see the Austin wildflowers in full bloom, but to work with the DeltaV product application specialists to share some of the capabilities in the Rosemount 3051S transmitter and what information it shares with systems like the DeltaV system. The team is also creating a video for us to add into the DeltaV YouTube site and other locations.

I've shared some of the statistical process monitoring (SPM) technology and how it helps spot abnormal situations by being closer to the process than the automation system is. These 3051S devices measure pressure, differential pressure (DP) level, and DP flow. They sample the process variable (PV) at 22 times per second compared with a typical 1-2 times per second from the automation system level.

It's this higher frequency data collection and applied statistical analysis to measure process variability that provides operations personnel more information about what's really happening in the process. Changes in this process variability can help uncover process- and equipment-related problems. Bill recommends using this high-resolution process variability as seen by the transmitters and following the four-step SPM model: Collect, Analyze, Decide, and Act.

The Collect step is about gathering as much process information as possible by trending the pressure and process variability inside the data historian application. With this collected information, step 2 is to Analyze the information and initially to establish a baseline. In subsequent times, process variability reductions, calculated by SPM can indicate process conditions like plugged impulse lines.

Step 3, the Decide step, determines the appropriate threshold limits to warn operators or maintenance personnel when actions need to be taken. The AMS Device Manager is used to set alerts based upon a change in process variability. Bill set a threshold of 30% process variability change in his demo example.

Step 4, the Act step, creates a notification for action once the threshold is crossed in the form of an alert on the DeltaV operator or maintenance station. Typically, specific operating procedures would be created for the critical measurement alerts. The actions might include checking the impulse line or other process/equipment condition based upon what the transmitter is measuring. These alerts are included with the historian process history to help achieve a wider view of what's happening in the process and identify future abnormal situations.

I'll update the post and embed the video once it has been produced and uploaded to YouTube.

MP3 | iTunes

Recently at the International Health, Safety, Environment, and Loss Prevention Conference in Kuwait, Emerson's David Walker had the opportunity to present on the topic of process safety. His presentation, Using a Smart Safety Instrumented System (SIS) to Make Better Operational Decisions, described the role of diagnostics in monitoring the health of the devices and surrounding process in safety instrumented functions.

David noted how accidents still manage to occur even with the safety standards and methodologies, which build layers of protection that guard against these incidents. The majority of these events occur because of the failure of plants to implement best practices in process safety and instill a pervasive safety culture. Also, safety implementations and standards are having difficulty keeping pace with the automation system alarm and display technologies. A few examples of these advances are alarm prioritization and narrowcasting by location and/or function. These capabilities often were not in existence when the plant process safety standards were developed.

David shared the data I mentioned in an earlier post where more that 90% of the failures in safety instrumented functions occur outside the logic solver--in the sensor or final control element (FCE). Traditionally, the focus of safety engineers has been on maintaining and supporting the logic solver, mainly because it could not provide diagnostics from the sensors and final control elements to identify potential problems. These missing diagnostics are even more critical as the components of the safety loop age and become less reliable. This equipment is more likely to cause spurious trips that unnecessarily shut down the process--decreasing overall availability.

Newer safety instrumented systems like DeltaV SIS communicate via HART digital communications to safety-certified or proven-in-use sensors and smart final control elements. The diagnostics from the sensors and FCEs help detect, report, and often give operators and maintenance technicians time to respond to the abnormal situation before a shutdown sequence is initiated. David listed some examples of these device diagnostics:

• Partial stroke testing of FCEs
• Low supply pressure to FCEs
• Temperature sensor failure and hot-backup capabilities
• Earth leakage detection
• Pressure transmitter impulse line plugging
• Flow transmitter two-phase flow
• Degraded voting logic upon transmitter failure
• Remote from hazardous location device testing by operations and maintenance staff.

David closed his presentation with the thought that the trend is to integrate this information from the SIS with the basic process control system (BPCS) instead of interfacing through a gateway because of the value of this diagnostic information. This helps operators and maintenance personnel make better operational decisions to avoid process shut downs and identify and address abnormal situations as quickly as possible.

MP3 | iTunes

You might recall Emerson's Bill Zhou from a quick, Rosemount transmitter demo video done at the Emerson Exchange a few weeks ago.

I asked Bill if I could get a copy of his and Andrew Klosinski's recent National Petrochemical & Refiners Association (NPRA) presentation, Advanced Diagnostics: 4 Steps to Better Decision Making.

The focus is on how advanced statistical process monitoring (SPM) technologies in intelligent field devices can help process manufacturers reduce maintenance costs, improve product quality and increase process uptime. All of that is easy to say, but the good thing is this presentation offers many case studies showing how.

First, from a technology standpoint, it's important to understand that a transmitter is much closer to where the action is, than the automation system. It touches the process as it measures temperature, level, flow, pressure, etc. Transmitters like the Rosemount 3051S, measure the process at 22 times per second instead of 1-2 times per second that is typical at the automation system level of the hierarchy. This higher resolution sampling is the basis for the statistical process monitoring to detect abnormal situations.

This statistical trending of process information is step one of the four steps to better decision making. It's followed by event correlation, then the creation of specific alerts to warn operators and/or maintenance folks, followed by actionable information to correct the situation before the unplanned shutdown, quality excursion, or asset failure occurs.

One example is a plugged impulse line. From a traditional view, an operator might see a quick drop in flow, with the valve position rapidly opening to try to compensate. It might take the operator quite a while to figure out why this occurred. During this troubleshooting period, process oscillations and shutdowns might occur. This same scenario seen from the transmitter's statistical perspective would show a sharp drop in the standard deviation. This indicates a plugged impulse line condition. In the real case study shared, Bill and Andrew show the dirt that had accumulated inside the pipe wall. Some dirt tore off the wall, which caused the plugging of the impulse line. Since the transmitter shared this insight, the problem was addressed far more quickly than with traditional troubleshooting methods.

Additional SPM-based advanced diagnosis and communication examples included furnace flame instability, DP level agitation loss, pump / valve cavitation, turbine blade wear, pressure transient detection, and distillation column flooding.

The common thread is the high-resolution, statistical monitoring of a process variable (PV) signal to identify and communicate the abnormal situation. In the case of burner flameout, flame instability shows a sharp increase in standard deviation of measured fuel gas pressure.

In the case of distillation column flooding, efficient separation stops, diagnosis is difficult, and repair is time consuming. Looking at differential pressure (DP) measurement across the packing from an SPM perspective shows an increase in standard deviation that correlates as a leading indicator of incipient flooding.

Make sure to view the presentation, if you have any of the other cases not highlighted in this post. Also, I'll keep working to try to get Bill to share some of these examples in video form, now that he's a YouTube star!

MP3 | iTunes

Update: I added a better link to the advanced diagnostics section of the 3051S and fixed the link to the NPRA.

Two very knowledgeable people in safety instrumented systems (SIS), Mike Boudreaux and Riyaz Ali, shared with me the story behind the recent news about the DVC6000 SIS digital valve controller (operated by 4-20mA) being certified to be compliant with IEC 61508 for use up to SIL 3 safety instrumented functions (SIF).

With this certification, DeltaV SIS logic solver's HART two-state, 4-20mA output and the Fisher DVC6000 SIS without any additional solenoids or other auxiliary devices can be used for SIL 3 applications. This configuration provides capturing trip events during safety demand, which provides crucial data for reliability and analysis by safety engineers of event. It's also helpful information for regulatory audits.

Now, I used my trusty friend Google to learn more about the HART two-state channel and found this page in DeltaV Books On-Line on the function block in the logic solver that helps make this happen. Basically:

...DeltaV SIS Logic Solver Digital Valve Controller (LSDVC) function block provides an interface to the DVC6000 SIS for safety shutdown and for partial stroke testing. The HART Two-state Output Channel provides the control signal and the HART communications path to the digital valve controller. You can configure the output channel to have an OFF_CURRENT of 0 mA or 4 mA. The control signal can command the valve controller to the tripped state regardless of the configured OFF_CURRENT value. Using an OFF_CURRENT value of 4 mA allows HART communication between the Logic Solver and the valve controller whether the valve controller is in the normal or the trip state. When the OFF_CURRENT is 0 mA, the power is removed entirely when the LSDVC function block drives the channel Off.

Mike noted that continuous diagnostics is possible because the valve closes when delivered a 4 mA signal. The DVC6000 SIS records the results of a demand event by logging all the results of travel and pressure data points in the microprocessor memory. This event log is critical for plant personnel, reliability engineers, and auditing authority to understand the final element status before and after the trip or demand event. Before the new certification was obtained, diagnostics would be lost on shutdown because the signal to the DVC would be 0 mA.

These on-line diagnostics coupled with partial stroke testing can be automatically initiated from the DeltaV SIS logic solver. This means that the final control element is periodically checked to help protect against spurious trips and to test for demand availability. The operator can also manually initiate these partial stroke tests from operator faceplates. The DVC6000 provides pass/fail status back via HART digital communications for alarming and historical event recording.

Riyaz pointed out that Type B devices (generally microprocessor-based) the IEC 61508 international safety standard (part 2, table 3) mandates redundancy in SIL 3 applications. This means the DVC6000 SIS connected to the DeltaV SIS HART two-state channel is suitable for SIL 1 and SIL 2 applications without redundancy, but for SIL 3 SIFs, IEC61508 mandates a full redundancy or hardware fault tolerance of one.

Achieving this certification helps reduce the components in these SIFs and increase the diagnostic coverage and capture of historical SIS information on demand.

I read Siemens' Charles Fialkowski's latest post, Introducing a non-redundant, redundant SIL 3 solution? about their SIL 3 HART I/O card. He discusses how technology has changed where newer SIL-3 rated safety instrumented systems (SIS):

...don't require redundancy to achieve high levels of safety. In the past, safety systems required dual, triple or even quadruple redundancy just to achieve high levels of safety.

He points out that advances in technology have allowed diagnostic coverage not possible in earlier SIS designs. He closes his post:

Another common misunderstanding is how these systems address field redundancy (sensors and final control elements). While I can't speak for the Emerson or Yokogawa system, I do know for a fact that the new Siemens HART analog input module handles redundant field devices just like any dual, triple or quadruple redundant system would.

I thought I'd give the Emerson perspective so I caught up with DeltaV SIS product manager Mike Boudreaux. He first pointed out that DeltaV SIS has HART I/O and the DeltaV SIS logic solvers are SIL3 certified in simplex (non-redundant) mode and have been since DeltaV SIS began shipping in 2005. Other safety instrumented systems also accept HART I/O, but only to pass-through the HART data to asset management systems. DeltaV SIS makes this HART status information available in the logic solver.

Mike noted that only the analog, 4-20mA process variable (PV) is used for the safety instrumented function (SIF). The digital HART PV's are not accessible for use in SIFs, but the device status provided by the HART digital communications protocol is passed along with an analog input in DeltaV SIS. If a HART transmitter detects a problem, the status for an analog input will become "Bad." Conditions for a Bad status include earth leakage detection, loss of HART communications, device malfunction and device fixed-loop current to name a few.

This Bad status can be used in the logic solver. For example, in a multi-transmitter SIF, a voter block can be configured to ignore an input value if it is Bad. In accordance with the international safety standard IEC 61511, this capability can be used to provide continued safe operation of the process while the faulty part is repaired. DeltaV SIS will alert operations of this problem so that the device can be maintained in the specified mean time to repair (MTTR). Alternatively, the voter block can be configured to treat a detected failure as a vote to trip, which provides increased safety.

When a HART device detects a problem, an alert is displayed on the DeltaV operator station. SIS faceplates and detail displays for HART devices help operators view and manage HART device alarms.

DeltaV SIS also uses the HART communications protocol to enhance partial stroke testing. It validates the operation of the final control element--the most critical and most likely to fail in a safety instrumented function. The logic solver can generate HART commands to initiate a partial stroke test in a digital valve controller. The operators can initiate partial stroke tests manually from their operator workstations or they can be scheduled to occur automatically based on the specified test interval. The results from these tests are captured and integrated with the system event history. An alarm can be generated if a partial stroke test fails, alerting maintenance that there is a potential problem with a valve.

This diagnostic coverage and information feedback to operations provide process manufacturers better tools for compliance with the IEC 61511 safety lifecycle compliance efforts.

Update: Welcome readers of Gary Mintchell's Feed Forward blog. Thanks for the shout out, Gary!

I was reading the great Control magazine article, Data Analytics in Batch Operations, by Lubrizol's Robert Wojewodka and Emerson's Terry Blevins. It describes the challenges in applying on-line analytics in batch production for early fault detection and prediction of end of batch quality parameters. If done right, this can help avoid out of spec batches, waste and rework and the opportunity cost of the lost batch. This can be critically important in specialty chemical and pharmaceutical manufacturing, which in the words of the authors, "depend heavily on batch processing to produce low-volume, high-value product."

The reward is great if you can make these on-line analytics work but the challenges are great. One of the biggest is the time differences between batches. Operators and events can halts and restarts to process. These may be for manual additive additions, waiting on common equipment to become available or abnormal conditions that may develop.

Another challenge is that online measurement of quality parameters "may not be technically feasible or economically justified." These lab samples must be available to online analytics toolset to perform these analytics.

Other challenges they noted included variations in feedstocks, varying operating conditions, concurrent batches and the assembly and organization of the data.

The concept of "Golden Batch" has been around a long time, which is the concept of comparing batches in progress, or just completed with ideal ones from the past. The authors point out two big weaknesses with this approach. The first is conditions indicated by each measurement may affect the product quality in different ways. Secondly, this is a univariate approach to a multivariate problem--no knowledge is gained of the relationships of process variations.

In a prior post, I've discussed some of the analytic tools, PCA and PLS. By applying these online, "...changes can be made in the batch to correct for detected faults or deviations in the predicted value of key quality parameters."

I mentioned to Terry that I gave this article a close read and was working on a blog post. He told me the really innovative thing they were able to do was to apply dynamic time warping (DTW). The article describes its use:

...allows such [batch time length] variations to be addressed by synchronizing batch data automatically using key characteristics of a reference trajectory.

This normalization process is covered in detail in the Greg McMillan and Mike Boudreau's book, New Directions in Bioprocess Modeling and Control and discussed in the article, PAT Tools for Accelerated Development and Improvement.

It's hard to give the article justice in a short blog post, so if you have 18 minutes and 30 seconds to spare, watch to the interview Walt Boyes did with Bob, Terry and Philippe Moro at last year's Emerson Exchange.

I use a service, WatchThatPage, to track changes to various pages around Emerson Process Management. It sends me an email when any page in a list of pages I have created has changed. I use this as one of my sources for the posts I create. This helps me keep track of changes in non-RSS enabled pages. For those who don't use RSS (really simple syndication), here's some resources on how it makes your information quest more efficient.

Late last week I received an email notifying me of a change to the Daniel liquid pipeline surge relief technical guide. I caught up with Dave Seiler to ask about this application and some of the challenges process manufacturers with high-pressure pipelines face. Pipeline operators and those with high-pressure pipelines are quite aware of the potential damage that can occur if a pressure surge occurs.

Dave noted that over-pressurization of a pipeline is commonly caused by sudden changes in liquid velocity. This may occur when a pump starts or stops or a valve opens or closes. When a pressure rise occurs above normal operating pressure, it's very important to analyze the rate of the pressure rise to determine the proper size and type of valve required.

Dave described line blockage as the most serious pipeline issue. To mitigate this condition, pipeline design includes valve interlocking logic and clear operating procedures. As noted in the technical guide:

...pressure is contained must have some form of pressure relief, which is often mandated and regulated by local authorities. The design of such systems is dependent on a complex range of factors including, but not limited to, the potential for pressure increases, the volumes which must be passed by the pressure relief equipment in operation and the capacity of the system to contain pressures.

This guide describes applications you may have in your facility. On application is a pilot operated pressure relief valve used for pump protection duty and for similar applications where pressure relief is required to maintain pressure at a given set point. Another application might have exceptionally fast response times that require gas-loaded systems. These are described:

The basic valve is the balanced piston design. Nitrogen gas is used to pressurize the valve piston to keep it in the closed position. The valve incorporates an integral oil reservoir mounted on the external surface of the cylinder head, which upon installation is partially filled with a light oil. Gas under pressure is applied to the reservoir.

Other applications described include surge relief valve closed position and open position. I found the pictures like this one help make the text easily understandable.

If you have high-pressure pipelines in your process, take a look at this guide and see how it might help you.

In the world of process safety, technology continues to advance to assist process manufacturers in their IEC 61511 safety compliance efforts. I saw a recent press release on enhancements to the Fisher DVC6000 digital valve controller. The news was:

...enhancements include manual reset, a stored safety demand event log, pass / fail status after a partial stroke test, and third party certification to SIL3, SIF loop.

I asked Riyaz Ali, whom you may recall from earlier posts, to simplify what this all means for me. The stored safety demand event log he likened to an airplane black box recorder. If a process upset condition triggers a safety demand on a valve controlled by the DVC6000 SIS (operated by 4-20mA input signal), it in turn automatically triggers an event log to capture the data into non-volatile memory locally in the digital valve controller.

This log keeps pre- and post-event data of the operating conditions surrounding the safety demand event. Examples of the type of data stored away in this event log include: travel, travel setpoint, output pressure with time in seconds, graphical representation of data points and date and time stamp of the trigger event for regulatory compliance.

Riyaz also described for me the partial stoke testing reporting. It now will provide pass/fail status and a signature curve of the valve stem movement. These partial stroke tests periodically diagnose the SIS valve to help ensure its availability. Also, a specially designed built-in relay provides protection against spurious trips which improves overall process availability. Other information provided back to the AMS ValveLink software includes diagnostics on stick slip, shaft integrity and maximum and minimum torque values.

For the DVC6000 SIS, the Fisher team achieved third-party certification for compliance to the IEC 61508 international safety standard for use in a SIL 3 safety instrumented function. This means that process manufacturers can use the DVC6000 SIS as part of the safety instrumented function in the SIL 3 loops they identify as part of their risk assessment and risk mitigation strategy.

Having all these digital valve controllers keeping logs of what's going on especially around upset conditions can greatly assist root cause investigations and help avoid future abnormal situations. And the diagnostics coming from the partial stroke tests can help process manufacturers avoid these abnormal conditions in the first place.

My RSS feeds pointed me to a great ChemicalProcessing.com article on batch manufacturing alarms. The article, Rethink batch-manufacturing alarm systems, was written by Joseph Alford.

He opens with provocative questions:

Do operators sing the praises of your plant's alarm system? No? Well, do they at least agree that generated alarms represent real abnormal situations requiring a response and that the automation/control system presents alarms in a timely, accurate and reliable way? No again? Well why not? Aren't operators the primary customers of your alarm system? Perhaps it's time for an alarm remediation project.

Many with continuous processes would agree that their alarm strategies implemented inside their automation systems need work. The complexity is amplified in batch processes because unique operating conditions are created within all of the numerous steps along the way.

The article boils down the crucial steps to take:

The key considerations in achieving effective alarm systems include defining objectives early in a project's life (i.e., in a plant's alarm philosophy or a system's functional requirements), adhering to the definition of an alarm, and implementing alarm-management best practices.

I ran this article by Todd Ham, a senior principal engineer in Emerson's Life Sciences industry organization. You may recall Todd from earlier posts.

Todd agrees completely that a successful alarm strategy begins in the functional requirements stage of a project. The project teams work with pharmaceutical and biotechnology manufacturers early on in a project to document their alarm requirements.

Todd stressed that a good strategy examines not only what conditions require an alarm--typically an adverse effect to personnel, product, or equipment--but also what is the desired response. Is alarm annunciation sufficient? Does this require a device interlock? Should this put the batch in hold? Does quality assurance (QA) need to be notified? This is called exception handling.

In a batch process, the requirements may differ based on the process step. For example, the alarm may be critical during processing, but not important during cleaning. Further, the alarm may only need to be monitored once the process is at steady state. In these scenarios, the control strategy developed by the project team will selectively enable/disable alarms at the appropriate point in the sequence.

Todd cautions that this is not a "one size fits all" exercise. The project team and manufacturer's staff must step through each process unit/system and ask a series of questions to arrive at a solution where alarms are both appropriate for a particular operating state and relevant for alerting operators to abnormal situations.

Todd agrees with the author that this work must be done up front, or the alarm flood, nuisance alarm scenario described in the article will be the result. You can't wait until the end of the project to think about alarm management. When you come right down to it, it's just as important as defining the control requirements for the project.

In the upcoming March issue of BioProcess International magazine, there is a great article by Emerson's Greg McMillan and Michael Boudreau, Broadley-James' Trish Benton, and the University of Texas at Austin's Yang Zang. The article, PAT Tools for Accelerated Process Development and Improvement, describes the collaborative effort between Emerson, Broadley-James, and UT, "...to examine and quantify the potential for faster optimization of batch operating points, process design, and cycle times." The specific objective of this collaboration:

...is to show that the impact of PAT can be maximized through the integration of dynamic simulation and multivariate analytics in a laboratory-optimized control system during product development.

Greg and Michael are putting many of the ideas they described in their book, New Directions in Bioprocess Modeling and Control: Maximizing Process Analytical Technology Benefits, into practice.

The authors outline the challenge for the 400 biotechnology medicines currently in development, which require overlapping and iterative stages for process development and commercialization. These stages include:

...cell line selection and development, media optimization, process conditions optimization and verification, scale-up, project definition, and plant design.

This team is working on beta tests using this new dynamic model and on-line data analytics and wants to make the results fully public to promote wide use and to advance these concepts and methodologies.

If you're like me and not in the biotechnology field, much of the article may get a little deep. I did glean a few tidbits you might find useful. By creating a dynamic model, one of the big benefits to the team is the ability to speed up the model by up to 1000 times real-time. Whether you're simulating the growth of mammalian cell lines or have another slow process, this can really help reduce trial and error time.

Another key is that the model, configuration, and tools can run in the "virtual plant" PC environment or can be downloaded to the automation system. With proper scale up factors:

...the embedded tools go readily from bench-top bioreactors to pilot plants and eventually industrial-scale bioreactors.

With the recognition by the FDA that quality cannot be tested into products, which led to the creation of the Process Analytical Technology (PAT) initiative, the authors discuss the role of analytics in their efforts.

Principal component analysis (PCA) and projection to latent structures (PLS) are two multivariate analysis techniques that can help analyze continuous and batch process operations. The authors' beta test is focusing on the on-line use of these analytical techniques where PLS detects deviations in quality parameters and PCA detects abnormal operations from measured and unmeasured disturbances.

Given the importance of new product development for pharmaceutical and biotechnology manufacturers, anything to reduce the overall development time and build in quality monitoring as prescribed in PAT should be a welcome addition.

Emerson's Pete Sharpe, an advanced automation consultant, was recently interviewed for a magazine article on the subject of process manufacturers' aging workforce. Manufacturers primarily in North America and Western Europe are feeling this loss of experience due to retirement.

Pete indicated that the reoccurring theme he hears is the loss of deep process knowledge. This knowledge, possessed by experienced operators, maintenance technicians and senior automation engineers is knowing when something in the process is not quite right, like when a measurement reading looks off, or when the process behavior is not the same. As seasoned operators retire, manufacturers feel this loss of experience most during non-normal conditions like startup, shutdown, or emergency situations.

Problems take longer to diagnose and resolve which can lead to less stable, more dangerous conditions. A recent refinery accident is one example where the problem diagnosis did not occur soon enough. You don't ever hear about most incidents but they certainly cost manufacturers money and often result in close calls that don't actually shut down a unit.

Pete discussed a couple of things automation suppliers are doing to address these issues. The first is improved operator training simulators, which I have written about in earlier blog posts. In this environment, less experienced operators can be challenged with operating problem situations so that they can improve their ability to diagnose the process and respond more quickly to abnormal situations.

As technologies continue to advance, more diagnostic capabilities are available in smart field devices and other plant assets. These devices can provide early warning about their own health and about the surrounding process. These predictive capabilities improve the ability of the maintenance organization to prioritize and respond to critical equipment alerts. Emerson's Abnormal Situation Prevention (ASP) algorithm uses process statistical signature data to give less experienced operators more time to react to abnormal situations and more diagnostic information to point to the root cause of the abnormal situation.

The final advancements that help to close the experience gap are advanced control technologies. As the technology has gotten increasingly scalable and easier to deploy in control systems like the DeltaV system, more and more processes can be operated as units and not as collections of loops. These APC technologies operate a process unit within its equipment constraints, at its most economical point. The operator's role changes from constantly adjusting individual loops to setting targets and constraint limits. APC applications are especially useful for process units that are tricky to run by less experienced operators--where many of the loops interact with one another or the process is highly constrained.

These advancements help ease the learning curve for future operators, maintenance technicians and automation engineers. On the positive side, today's engineers and young operators are nearly all computer-literate, so they can make good use of the modern tools and work processes that come with today's control infrastructure. This computer-savvy generation is more likely to adapt to computer-based control systems and modern fieldbus architectures. In addition, automation suppliers like Emerson are helping to ease this knowledge gap by having people like Pete and the other advanced automation consultants available to work with process manufacturers.

In our continuing series of screencasts, I'm trying to give examples of how advanced diagnostics in Foundation fieldbus devices can be used in control strategies to avoid abnormal situations and potential losses in production.

Emerson's Rune Reppenhagen shows in this quick 2 minute, 47 second screencast, how an advanced model predictive control strategy in a DeltaV controller automatically recognizes a failure diagnostic in a temperature transmitter and switches the mode of control over to a manual state.

At the same time, this diagnostic alerts the operator of the situation, and the AMS Device Manager software shows the condition of the transmitter so it can be quickly repaired.

By using the advanced diagnostics from these intelligent field devices in the control and advanced control strategies, conditions which impact the availability and quality of the process can be avoided.

Here is another in my series of screencasts, this time showing how an automation system uses predictive maintenance diagnostics to switchover a pump before it fails.

Emerson's DeltaV product manager, Randy Balentine, shows in this 2 minute, 43 second screencast a redundant pair of pump-motor trains. These pump-motor trains are being monitored with CSI 9210 Machinery Health Transmitters.

Randy shows a situation where one of the transmitters communicates excessive vibration via Foundation fieldbus digital communications to a DeltaV system. One of the DeltaV control modules receives the diagnostic alert, performs the logic to switchover to the backup pump-motor train, and notifies the operator of the problem so that it can be addressed.

By incorporating these predictive diagnostics into the control strategy, the switchover can happen before a failure causes a loss of production. Based on the severity of the diagnostic information reported by the smart Foundation fieldbus transmitter, the actions can range from notification of the operators to control actions performed by the control strategy.

I had mentioned in an earlier post that short screencasts are a great way to quickly convery ideas in lieu of hundreds of words. One of Emerson's product application specialists, Rune Reppenhagen, graciously agreed to demonstrate how advanced diagnostics can be used in automation system control strategies.

Today's example shows how air in a fluid can impact Coriolis flow measurement and cause the automation system control strategy to falsely assume it needs to increase the speed of a pump to try to raise compensate for the low flow measurement. This situation called entrained air or slug flow causes the measurement on the coriolis meter to go to zero. The actual flow is OK but the problem is with the measurement.

Rune demonstates in this screencast (runtime: 4:51) how advanced diagnostics like those found in Micro Motion Elite mass flow and density meters can be configured in systems like the DeltaV system to read these diagnostics and take action in the control strategy to turn the loop to manual control for the operator and notify him of the cause of the situation.

This immediate recognition of a process problem and operator notification of the situation is one example of how advanced diagnostics and digital communications protocols like Foundation Fieldbus provide ways for process manufacturers to avoid losses in production, quality excursions, and abnormal situations which can impact the efficiency of the production process.

Before the holidays, Dave Harrold wrote a post, A Wee Bit More About Safety Instrumented Systems, in his Dave @ AFAB Group blog. He describes his work with Dr. Angela Summers, founder/president of SIS-Tech Solutions on a guidelines book for the global IEC 61511 safety standards. Dave also referenced an SIS-related Q&A article Angela wrote for Flow Control magazine.

I forwarded the post and Flow Control article link to Riyaz Ali, whom you may recall from an earlier post. Riyaz wanted to add to the conversation and make three specific points in reference to the Flow Control article.

On the question regarding the use of digital valve positioners to perform partial testing and its relationship to the proof test interval, Riyaz agrees that the proof test is far more than a partial stroke test. The proof test can be performed on a final control element either on-line when a bypass valve exists or offline when the process is shutdown, such as during a plant turnaround. Many process manufacturers do not have large bypass valves and seek to extend the interval between plant turnarounds as long as possible. The on-line partial stroke testing provided by digital valve positioners can help extend the time between proof tests. They do not replace these tests. Riyaz points to a Control Engineering magazine article authored by Dr. Summers, Partial Stroke Testing of Safety Block Valves, in which she points out:

Also affecting the SIL is diagnostic coverage and testing intervals of partial-stroke testing to supplement full-stroke testing to reduce a block valve's PFD.

Riyaz did take exception to a statement in the article about throttling valves:

Positioner failures are the leading cause of control failure, so the positioner should not be used to actuate the valve in an SIS application when preventing events associated with a loss of control. Instead, a solenoid-operated valve should be used to independently close the control valve.

His final point is on the question regarding smart positioners for partial stroke testing of smart valves. Positioners operated by air have been used in process control industries for years to improve performance of control loop. It is becoming rarer to come across a process loop not without positioners, especially where the application improved process variability. Based on its usage and benefits in process control, process manufacturers have started using them for Safety Instrumented Systems also. Riyaz agrees with Dr. Summers comment that positioners have smaller orifice but any thing larger than 8"-12" size valve, even otherwise a Quick Exhaust Valve or similar mechanical device will be used, if fast stroking speed is desired. Len Laskowski adds that the driving factor is process safety time. Many times larger valves do not need to close in one or two seconds, and in fact require a more controlled closure to avoid negative effects on process and utility equipment. It all hinges on the process safety time for each application.

Positioners by design are to bleed very small air to keep the air flowing as well keep pressure higher than atmospheric so as avoid any external atmospheric corrosive gas getting inside the housing. Also during partial stroke test positioners exhaust and fill the air, which makes its mechanical parts moving and avoid any build up.

Digital valve positioners allows partial stroke testing, while process is running and provides date and time stamp of test with capability to store and compare test results. Also, being a microprocessor based, these positioners allow remote testing and retrieval of data remotely. The main advantage is predictive maintenance by providing valve degradation analysis, which is important to critical valves in safety related systems. If by any chance valve is stuck, digital valve positioners are capable of providing alerts to operators to fix the problem.

From my days as a young systems engineer working on offshore oil & gas platforms in the Gulf of Mexico, I know that abnormal situations in our processes are something we all wish to avoid. A 1999 study by the ASM consortium estimated $10 billion USD in losses for U.S. process manufacturers due to abnormal situations. The question is how best to prevent these abnormal situations from occurring in the first place.

Emerson's Ravi Kant and Roger Pihlaja recently presented a paper, "Abnormal Situation Prevention (ASP) in Complex Systems" at the recent NPRA Q&A and Technology Forum.

In their presentation they stress that the potential severity and cost of an incident increases if timely corrective action is not taken. An example cited from a refinery abnormal situation is the failure of a butterfly valve. After going several hours without detection by the automation system or operations personnel, it caused the Cat Cracker (FCCU) to shut down. In a matter of minutes this caused the refinery to shutdown, resulting in more that $1 million USD per day in lost revenue.

Ravi and Roger explained how abnormal situation prevention (ASP) technology embedded in the sensors, actuations, and machinery health are closest to the process and have access to better information. This ASP technology can predict root causes of abnormal situations through high-frequency spectral and statistical data analysis within these smart devices. The main reason for doing this analysis closest to the process is that the sampling frequency is greater--22 samples per second, instead of 1 sample per second to 1 sample per minute typical at the automation system level.

Data analysis at this higher frequency can uncover process anomalies including drift, bias, excessive noise, process spikes, and plugged conditions. Some of the detection and prediction algorithms and techniques which are employed include: polynomial extensible regression, principal component analysis, statistical process control, decision trees, fuzzy logic, and neural networks.

They cited some specific ASP applications in refineries including early detection of catalyst losses, catalyst circulation issues, afterburn conditions, column and heater coking, temperature runaway, and acid levels outside optimal or safe levels. The key to detecting these process conditions is sharing this data analysis at from the field device level, up through the equipment level, up through the process unit level to the operators and plant maintenance staff. Digital communications technologies like Foundation fieldbus and HART provide the information path.

Roger also shared with me other high-frequency data dependant ASP applications in the process including:

• Plugged impulse line detection for DP flow transmitters
• Flame instability
• Stick/slip in FCC solids transfer lines
• Stirred tank vessel agitator diagnostics
• Continuous rotary drum vacuum filter diagnostics
• Fouling & DP level transmitter plugging in evaporators
• Detection of developing ASP issues like arching, bridging, & rat-holing in bulk solids storage vessels
• In-situ proof testing of emergency relief systems

Work continues to refine and extend these predictive ASP technologies to more smart field devices to increase the "eyes and ears" on the process in order to avoid the costs and losses associated with abnormal situations.

Recently an email came in that said Refineries and Petrochemicals specialist, Ravi Kant, and ASP Validation and Verification Engineer, Ahmad Hamad, in our Performance Technologies division, won the Fuels & Petrochemical's Award for best paper (out of more than 80 papers) at the AIChE 2006 Spring National Meeting.

This was something I had to get my hands on and find out why, and extend hearty congratulations to Ahmad and Ravi. The predictive PlantWeb technologies developed by this team find their way into AMS Suite software products, Rosemount transmitters, and other Emerson smart field devices.

With many industries like refining and petrochemicals running near full capacity, abnormal situation prevention provides a method for early detection with problems in the process and provides an opportunity for timely corrective action--before down time, quality issues, or even safety issues occur.

The paper, Advances in Abnormal Situation Prevention in Refineries and Petrochemical Plants, looks at traditional ways of preventive maintenance and the drawbacks in performing unnecessary maintenance, sometimes requiring down time, and being unable to detect abnormal situations.

It also explores other techniques for abnormal situation management. These solutions use knowledge-based diagnostics with data drawn from the continuous historian to develop a multivariate model. The source data from the historian is typically very low frequency from once per second to once per minute. This approach fails to detect abnormal situation which can develop rapidly. It also often fails to find problems with machinery, devices, and transmitters in the process. An example might be a stuck valve.

Ahmad and Ravi describe how advances in microprocessor performance and digital communications like Foundation Fieldbus and HART make it possible to do high frequency diagnostics within smart field devices. Emerson Process Management has developed Abnormal Situation Prevention (ASP) blocks in smart field devices like Rosemount 3051s transmitter, which capture high frequency process data at 22 samples per second. The blocks perform statistical, frequency-based, auto-regression, wavelets and other diagnostic measures to try to discover problems in the process in their earliest stage. And automation systems like the DeltaV and Ovation systems can turn the most critical of these alerts from these ASP blocks into operator and maintenance alarms for corrective action to begin.

The paper describes for cases where this early detection can prevent abnormal situations from occurring. These include: coke detection in refineries, catalyst circulation in fluid catalytic cracking (FCC) units, maltrays detection in crude columns, and gas turbine abnormalities. These are but a few of the critical applications where abnormal situation prevention technology can be applied.

Like anything else, the closer you can get to the source of the abnormal situation, and the earlier you can identify it, the sooner you can mitigate or prevent the situation from occurring.