Sometimes it seems universities and industry reside on planets that are light years apart. Too bad we don't have Star Ships with warp drive. Universities have leading edge research. Industry has "state of the art implementation."

Why are universities and industry "worlds apart?"

Engineers in industry don't seem to understand how to apply the research from universities. Professors don't appear to really know what is needed in industry. The tools are quite different. Engineers in chemical, pharmaceutical, and pulp & paper plants configure their control strategies in a distributed control system (DCS). Professors typically have their graduate students program their algorithms and test cases in Matlab.

One way to get industry and universities on the same page is to provide a DCS to the university with all the tools needed for research, such as a Matlab interface. In many cases the Matlab code can end up being configured in the DCS as part of the maturation of the innovation. The use of the DCS minimizes the reinvention of the wheel, such as the PID algorithm with all of its evolutionary enhancements. The setup facilitates the transfer of knowledge between the universities and industry. Being able to explore, prototype, and demo university innovations in a DCS makes it more real to industry and leads to rapid deployment and sharing of actual plant results.

If there is a unit operations lab, process control lab, or pilot plant, the DCS can be used to control the equipment used in the experiments. Students gain valuable experience in learning how to work with a toolset that is designed to meet industrial standards. Just learning the nomenclature and working with a DCS gives the student practical skills and confidence when as a new employee the student enters the control room. The window to see and affect the process is the DCS. Whether the student is going into automation or process design & technology, the student needs to be able to understand how to access and review modes, limits, options, and variables that determine how well a process runs. For example, the student gets to work in a university DCS on PID features commonly used in industry:

(1) PID limits (e.g. output, set point, and anti-reset windup limits)
(2) PID options (e.g. set point tracking of the process variable in manual, dynamic reset limiting, and nonlinear gain modification)
(3) PID form (series and standard)
(4) PID structure to determine whether each PID mode (proportional, integral and derivative) works on the process variable or the error (difference between the set point and the process variable)
.
The first semester I taught the Chemical Engineering course "Introduction to Process Dynamics and Control" at Washington University in Saint Louis as an adjunct professor, the students could not relate to my attempt to introduce practical plant applications and considerations in the normal course of Laplace transforms and bode plots. The second semester I added a virtual plant that consisted of a DeltaV DCS running in the Simulate mode integrated with HYSYS dynamic process simulations for each student. I later configured most of the process simulations directly in control studio. I was amazed how fast the students learned how to work in the graphical configuration environment and operator interface. All they needed was a few screen prints on navigation to get them started. Several of the students subsequently got intern or permanent positions doing configuration at the local DCS industry center. I had these students with experience in the automation industry come back to speak to the next class. The result was a dramatic turnaround in appreciation and understanding of what they would face in industry. The students decided on their own to go online to find and buy tee-shirts with Duncan, the DCS mascot, windsurfing. I ended up buying tee-shirts too and we all posed for a group photo by one of the students.

The main obstacle to the use of the DCS in the university is the initial installation and training. This is addressed by the support of industries with the same DCS who have a working relationship with the university and the local business partners of the DCS supplier. This method has enabled over 100 DeltaV DCS installations at educational institutions.

At the Automatic Control Conference in Saint Louis on June 11, I am co-chairing a session with Professor Tom Edgar from the University of Texas on "Bridging the Gap between Universities and Industry." The presentations are:

(1) "Bridging the Gap Between Universities and Industry"
(2) "Digital Process Control Lab at Washington University"
(3) "The Bioprocess Laboratory at Washington University"
(4) "Rose-Hulman Institute of Technology Unit Operations Laboratory"
(5) "Engineering Research Center for Structured Organic Particulate Synthesis (Rutgers, Purdue, New Jersey Institute of Technology, University of Puerto Rico at Mayaguez)"
(6) "Using a Distributed Control System (DCS) for Distillation Column Control in an Undergraduate Unit Operations Laboratory (University of Texas)"

My next blog will be June 22. In the mean time enjoy summertime.

I haven't had any special courses or training in writing and some may say it shows. I wouldn't advocate anyone following my style, especially if you trying to promote your products or ideas. I tend to lead the reader on path of discovery by laying out the situation and the problems and then some interesting ideas. Maybe it is the user in me (33 years at Monsanto and Solutia), the scientist in me (Physics), or my approach to writing that wants to leave it up to the reader to make the judgments and assessments.

However, people today want to know upfront the bottom line. They may not have time or the background to come to useful conclusions. Plus my emphasis on detailing problems can be formatted with a more positive approach of presenting opportunities and solutions. Next month I will offer a short introductory overview of my recent articles that emphasizes the scenario, essence, and value of the ideas developed.

The main point of this blog like all of my writing is to share what I have learned. My goal for next year is to help prevent significant expertise and knowledge in process automation from being lost forever. I would guess 100 or more automation professionals are retiring each year who have published at best an infinitesimally small portion of their expertise for posterity. Also, new engineers are facing special challenges. My sense is the new kid in the control room doesn't have the mentors or the internal technical training programs I took for granted. They may be thrown into the midst of a difficult problem with no guidance.

Beginning this Fall I will be making presentations at Local ISA sections and interviewing young and seasoned automation professionals recommended by the sections to get a better idea of new and lost process control expertise. The first stop may be the ISA Boston section. I lived in Cambridge when I was overseeing a project at Badger. I am looking forward to returning to Harvard Square and Legal Seafood Market. I can dream of returning to Fenway Park. The interviews will be published in my monthly Control Talk column in Control magazine. I enjoyed doing the 3-part series in Control Talk on "The Secret Life of pH Electrodes" based on interviews at Broadley-James and Rosemount Analytical in Irvine California (nice place to visit).

In the mean time here is what I have learned about writing in no particular order except what pops into my brain (kind of the way I do my first draft).

(1) An outline is a good idea but it is just a starting point. I don't truly know where the article or book will take me at the beginning. The value I get out of writing is the discovery process. I find concepts and ideas along the way. I gain knowledge besides sharing knowledge. In my next book, I realized after about 5 pages into the first chapter that an important message is the dramatic change in the performance of modern measurements. The installed accuracy of key measurements has improved by one to two orders of magnitude compared to my days in E&I design and construction. A smart closed coupled coplanar DP with static pressure and temperature compensation has 0.02% installed accuracy. A radar level gauge can detect changes as small as 0.04 inches in level. A Coriolis liquid flow meter can have an installed accuracy of 0.05% with a rangeability of 200:1. The bench top and installed accuracy of pressure, level, and flow measurements in 1970s and 1980s was typically 0.5% and 2%, respectively. I have an article from that era that quantifies the deterioration from uncompensated process and ambient operating conditions. Then there was the noise introduced by wiring problems (see May 19 entry). Smart wireless instrumentation offer a whole new ball game if you don't screw it up with a bad installation. The limit to control loop performance is not the measurement but more than ever is the control strategy, the final element, and how well you tune the controller.

(2) The most difficult thing is writing the first sentence. The second most difficult thing is writing the second sentence. The third most difficult thing is writing the first paragraph. The fourth most difficult thing is writing the first page. The lesson here is to just get started. Writing is an iterative process. My problem is that I get bored rehashing ideas I have unleashed and want to move on to the next page, article, column, or book. I don't iterate enough. Here is where a person who knows the subject can help by reading and commenting on your drafts. Beware of technical writers or copy editors who want to rewrite the whole thing because it often results in a loss or change in meaning and intent.

(3) Once you get going, don't stop. A flow is important. After about 10 pages, the thoughts start to flow fast and free. At this point, music (particularly the best of Concrete Blonde, The Goo Goo Dolls, Josh Groban, Don Henley, Meat Loaf, Matchbox 20, Bruce Springsteen, and U2) adds inspiration and makes writing more fun for me. I think this was the only way I was able to write a dozen serious technical books, a half dozen funny technical books, fifty articles and papers, and 8 years of Control Talk column. It also helps to have an understanding management and spouse. Can you envision getting approval of humorous books, columns, and "top ten lists" through the official channels of a big corporation? Can you imagine your spouse letting you spend 8 hours writing on a weekend? Lastly, I do the diagram and figures last. This is mind numbing work for me that would sap my creative energy and interrupt the flow.

(4) Use a lot of sub headings and bullet lists. This gets the reader interested and helps to cherry pick what is of greatest importance. My article "Maximizing PAT Benefits from Bioprocess Modeling and Control" is a good example of missing lists and sub headings. In my defense, the article was a last minute deal. I had just a couple of days and was just doing a core dump of what I thought was important.

(5) Go back and improve the introductory paragraph to each section and the introductory sentence to each paragraph. This is a good idea. I am going to try it when I have time.

(6) Develop the solution, include the assumptions, and detail the verification. I tend to do the first two parts of this suggestion and provide evidence of the last part. However, it would help the readers to make suggestions on how to prove out the idea for their application. No solution is universally true. There are always exceptions. I have benefited from the best minds in process control but I have noticed that the bigger the mind the bigger the ego. A blind spot is developed that makes experts unwilling to acknowledge when their solutions don't work in a particular application. Maybe it is my science background or an ego deficiency but I am always looking for exceptions and I never think any of my ideas as perfect. You learn more from things that don't work.

(7) Don't make statements that are to be accepted as fact. I am particularly sensitive to statements made to be accepted as true that are mostly untrue. For example, "Thermocouples (TCs) are faster than RTDs." This statement is generally accepted as true but is it useful and is it in fact misleading? What if someone chooses a TC instead of a more accurate RTD because he or she thinks the TC is faster? If you had a bare element there might be a difference of a couple of seconds in the response but the uncertainty in the time lags of most temperature systems are an order of magnitude larger. Once you put the element in a thermowell, the construction and fit and length of the thermowell determines a time lag for the assembly that is an order of magnitude larger as well (no pun intended). Also most controllers are tuned so slowly, you don't see the effect of small changes in time lags. The bottom line is that you will probably never see the difference in speed between a TC and an RTD. Having said that I can envision exceptions, such as the temperature control of inline mixing of streams with an aggressively tuned controller. I have just never seen this application in practice.

(8) Use short clear sentences and paragraphs that build on each other. Providing qualifications can result in long sentences. Also, one thought for me quickly leads to another and to another and to another, which leads to run-on sentences. I need to constantly go back and split my thoughts into separate sentences with a building block approach. I am currently trying in my first draft to break thoughts up. So far it doesn't seem to interrupt the flow, which was my original concern.

(9) Don't get hung up on perfect grammar or a perfect piece. If you don't give copy editors something to do, they will start some serious messing with the sentences. Everyone wants to feel like they are contributing and doing their job. Also, what copy editors are looking for in terms of commas and hyphens may change. Finally, obsession with sentence structure can lock up your mind. I talked to a copy editor who said he needs to find another job so he can write a book. As a copy editor, he thinks too much about the technical details of writing.

(10) Think of writing as if you were having a conversation with a good friend. This makes the whole writing process less intimidating and lets you be more frank and less formal.

(11) Use plenty of examples and illustrations. It takes time but sure drives the point home. I have seen a lot of misinterpretations of an idea or its intent. The fault is really my own and not the reader. For example, in my article "Is Wireless Process Control Ready for Prime Time" a reader thought there was a concern being expressed on noise from variable speed drives on wireless transmitters when what I meant was the opposite. Wireless transmitters should eliminate these and other noise problems associated with wiring. I no longer assume any concept is obvious.

(12) Use free association and both sides of your brain. This allows you to take creative leaps you probably didn't even realize before you started writing. In my case it also enables me to add humor such as "Top Ten Lists" and the cartoon descriptions for my illustrious friend Ted Williams.

While wiring problems may not be pervasive, when they do exist, the upset is significant and tends to persistently reoccur for months to years. The intermittent transient occurrence is difficult to diagnose. The short term steps and spikes cause kicks in the controller output from gain and rate action. In my experience, the spikes in pH transmitter outputs often go unresolved. The article by Fred Sanders titled "Watch Out for Instrument Errors" (Chemical Engineering, July 1995) gives examples of the insidious and disruptive nature of wiring problems.

The problems seemed to get worse in the 1980s and 1990s when new low cost invertors were installed for variable frequency drives (VFD). The VFD is also known as a "variable speed drive" (VSD) because the speed of the motor is regulated to be proportional to the controller output. In a recent conversation with Owen Campney, I got an idea of what happened to make the VFD inverter noise a bigger problem.

The switching in the new invertors was faster creating sharper edges. This reduced the heat, size, and cost of inverters making them more abundant. Instead of being in dedicated inverter rooms in the motor control center, they started to appear in instrument rooms among the interface panels. The faster switching created higher frequencies with higher energy. These invertors had an output choke to prevent damage to motor insulation but the input choke was optional and was often missing or insufficient. Eventually, the noise in instrument signals became bad enough that chokes were offered to meet the International Electrochemical Commission (IEC) standards. Alternatively, isolation transformers were located close to the inverter with the power wiring between the inverter and transformer in hard pipe conduit to minimize the noise from this section of wiring. You hear VFD war stories to this day.

One such war story was related to me by Owen. A large intermediate plant installed smart HART input cards on some critical HART transmitters to take advantage of the diagnostic information digitally superimposed on the analog signal. Unfortunately, 3 to 6 times each day the digital signal would be momentarily lost. It was suspected but never confirmed to be triggered by a VSD on coolant in a temperature loop when demand changed. The problem persisted for several years until an electrical engineer patiently tracked down the problem to a second ground hidden from view in the wall.

Whether the steps, spikes, and noise in an instrument signal is due to wiring or not, the wiring is always suspect particularly if the user has been burned by VFD incidents. Consequently more hours are wasted than is generally recognized on trying to track down real or imagined wiring problems. The WirelessHART network should be immune to the VFD interference and eliminate the wiring questions. The potential savings from WirelessHART in maintenance cost may be currently underestimated.

For more information on WirelessHART checkout the article "Is Wireless Process Control Ready for Prime Time?" in the May 2009 issue of Control magazine.

I need to minimize the time delay to dinner so I will minimize this discussion of how to minimize the oscillation from analyzer sample time delay. So many minimums and so little time.

Composition measurements with sample systems and cycle times termed "at-line analyzers" offer incredible opportunities for understanding and controlling what affects what you ultimately want to know for a process output stream - the composition. The sample time delay from the cyclic results from an at-line analyzer is more problematic than the transportation delay for a continuous measurement via a probe in a sample line termed "in-line analyzers". The "at-line analyzer" has a stepped response and sometimes spikes from bad readings with no intermediate values. The result is a propensity for oscillations when used for feedback control.

One might think a deadtime compensator would help the traditional PID deal with the deadtime from a cyclic time delay. However, these deadtime compensators are notoriously sensitive to a mismatch between the actual process deadtime and the estimated deadtime used in the compensator. The loop deadtime from unsynchronized digital devices and at-line analyzers is extremely variable and can at best be estimated after the fact.

It is interesting that the solution for suppressing oscillations from at-line analyzers resulted from improvements to the PID developed for variable updates from wireless devices (see February 9, entry on "Unexpected Wireless Benefits"). The control solution for WirelessHART requires no estimate of deadtime and is more robust than a traditional PID. The PID execution is kept relatively fast (once per second). The contribution of the proportional mode is computed every execution. The proportional action every scan provides a good set point response for a PID structure with proportional action on error. The contribution of the integral and derivative mode is only computed when the measurement has changed per the resolution setting of wireless device. Furthermore, the time used in the integral and derivative mode calculations is not the scan time but the elapsed time from the last measurement update.

The use of the elapsed time in the integral calculation and a reset time the same as the process time constant provides an integral correction that is equal to and opposite to the process response in the elapsed time. Even if the process time constant changes, making an integral correction only when there is update eliminates the extraneous ramp of the integral mode in the traditional PID acting on old information. The suspension of integral action until there is new information also helps the PID deal with a valve that is momentarily stuck provided position read back is used for dynamic reset limiting.

The use of elapsed time instead of PID execution time in the derivative calculation spreads the change in the process over the elapsed time rather than taking it to all occur in the single execution time. This more intelligent rate action eliminates spikes in the controller output that would occur in a traditional PID when there is an update.

The wireless PID greatly stabilized the glucose control of a bioreactor which had at-line analyzer sample time delays that varied from 6 to 12 hours. The improvement is greatest for self-regulating processes and controllers tuned for maximum performance. The suppression of oscillations can be seen on slides 29 - 33 of the Interphex 2009 Presentation "Advances in Bioreactor Modeling and Control."
Interphex2009_Advances_In_Bioreactor_Modeling_and_Control.pdf

If you want to know how to minimize oscillations from final elements and don't have time to read the supporting information you can use the following rules of thumb and move on to more important tasks like reading email. The final elements considered here are throttling control valves and variable speed drives (VSD) on pumps or fans.

• Use a sliding stem throttling valve with a properly tuned digital positioner (position feedback) or a VSD with a properly tuned speed controller (tachometer feedback) to minimize the amplitude of the limit cycle from a final element
• Make sure the DCS and final element I/O cards have at least 12 bits
• Enable "Dynamic Reset Limit" in PID block and use position or speed feedback as PV for BKCAL_OUT of AO block to prevent a burst of unstable oscillations when PID reset action is faster than valve or VSD response
• Set IDEADAND in the PID block equal to the limit cycle amplitude from the final element to kill the limit cycle during quiet periods of operation (e.g. periods when there are no disturbances or set point changes) for a self-regulating loop

Resolution is the minimum change in the element's output. Changes in the output smaller than the resolution cannot be made. For a control valve, the resolution limit is the result of friction in the packing, seat, and seal. For a VSD, the resolution limit is the result of an artificially imposed deadband, which is really a dead zone or from a speed sensing element resolution limit. Resolution can also result from a quantize limit from the number of bits in a microprocessor or I/O card. The number of bits in A/D and D/A cards for most DCS has increased from 12 bits to 16 bits. In both cases, the resolution limit from these I/O cards is negligible. However the standard input card of some VSD manufacturers is only 8 bit causing a significant resolution limit. The resolution in the stroke of a control valve or in the speed a variable speed drive will cause a limit cycle in any loop with integral (reset) action.

The term deadband is often used in automation systems to specify a dead zone (a bandwidth around a reference value where there is no response). Examples are deadband (dead zone) specifications in VSD configuration for noise rejection and in a PID configuration for integral action suspension.

For final elements, deadband has a significantly different definition. Here deadband is the change in signal required upon a reversal of direction to get a change in the element's output. Once the output reverse direction, deadband places no limit on how small a change can be made in the same direction. In reality, valve deadband is usually accompanied by a resolution limit. In the stroke of a control valve, deadband is the result of backlash from gaps or play in linkages and shaft or stem connections. Deadband normally doesn't exist in a VSD. Deadband will cause a limit cycle if there are two integrators in series in the control system. Multiple integrators in series can occur from a PID with integral action on a process with an integrating response such as level. Alternately, the limit cycle can occur if there is a cascade control loop where there is integral action in more than one controller. If both the temperature and flow PID blocks have integral (reset) action in a temperature to flow cascade control system, then deadband can cause a limit cycle. Most people forget that a positioner or digital valve controller creates a cascade loop where the positioner controller is the secondary loop. Positioners until recently were proportional only controllers.

The amplitude of the limit cycle is the smallest change in flow associated with the smallest possible change in valve position or speed multiplied by the process gain (change in process variable in engineering units divided by the change in flow). To get the smallest possible change in flow of a control valve, multiply the valve's resolution limit in % of stroke by the installed characteristic curve for the valve at its operating point. Note that valve stick-slip and the resolution gets worse near the seating or sealing surface. The manufacturer's quoted numbers are at a 50% throttle position. To get the smallest possible change in flow of a VSD multiply the resolution limit of the input card resolution of the tachometer sensing element, or noise deadband, whichever is largest, and convert to flow based on the interpolated shift in the installed characteristic curves with speed for the pump or fan. Be careful, many VSD have an adjustable deadband (dead zone) to prevent the VSD from responding to noise. This adjustment is often set with no regard to the effect on loop performance.

Resolution limits and deadband add dead time to the control loop for slow disturbances because it takes time for the PID output work through the zone of no final element response. The dead time is the resolution limit or deadband divided by the rate of change of the controller output. This additional deadtime increases the peak and integrated error for the upset. Note that step changes in the controller output larger than the resolution limit or deadband will not reveal the deadtime.

Control valves have an inherent velocity limit from the limitations imposed by actuator fill and exhaust rates. VSD have an application set velocity limit from the motor load limitations imposed by the impeller inertia. Make sure the valve actuator and VSD motor have enough muscle for the valve sticktion and pump inertia, respectively or you can get into poor valve position or speed control and hence even bigger loop problems.

Use the "dynamic reset limit" option of a PID block in a DCS, such as DeltaV, where the PID uses a positive feedback network for its integral action. The BKCAL_OUT for the AO block which in connected to the BKCAL_IN of the PID block should be actual valve position or VSD speed. Select the PV (position or speed) option in the AO block for the BKCAL_OUT. This feedback of actual position or speed to the PID enables the PID algorithm to curtail its integral contribution to the PID output so that the PID output from reset action does not change faster than the valve or drive can respond. If this protection is not in place, everything may look OK until the loop gets a disturbance large enough PID to cause the PID output to change faster than the final element. The mysterious bursts of instability for big load upsets often go unresolved.

Set the IDEADBAND option in the PID block to a value about equal to the limit cycle amplitude. IDEADBAND will suspend the integral action when the PID error is less the IDEADBAND. This suspension will stop limit cycles from a resolution limit or deadband for a self-regulating process at a steady state. It will not stop the limit cycle on a process with an integrating response because the process has no steady state and will continue to ramp until the process variable exceeds the IDEADBAND.

For more info on final element response, check out the "Deal or No Deal" Control Talk column in Control magazine, the article "What is your Valve Trying to Tell You" in Control Design magazine, and "Improve Control Loop Performance" in Chemical Processing magazine.

The loops with the most severe oscillations listed in order from biggest amplitude to smallest amplitude are pH loops, level loops, flow loops, pressure loops, batch temperature loops, heat exchanger temperature loops, and column temperature loops.

The following is a list of the sources of product quality oscillations in the approximate descending order of frequency of occurrence based on my experience. I have even offered my best guess in parentheses as to the percentage of applications that can be tracked to these root causes for chemical and biochemical products. You may wonder why pH loops didn't make the top of the list since it has the most severe oscillations. The main reason pH loops are down the list is that most pH loops are in waste treatment (WT). Also, the pH loops in reactors and bioreactors tend to have much lower process gains than WT pH loops and some process regulation from reagent consumption. Interacting temperature loops on furnaces, reformers, and reactors are severe problems but are near the bottom of the list for applications for specialty chemicals and biochemical products because multi-zone or profile temperature control are more prevalent in the petroleum, petrochemical, and bulk chemical industries. The following list is for normal operation of loops with good valves and does not consider oscillations that originate from the startup and shutdown and failure of equipment. Next week we will see the implications of "not so good" valves.

(1) Too much reset action in level loops on surge and feed tanks (40%)
(2) Discontinuities at split range point for pH, pressure, and temperature loops (20%)
(3) Interacting pressure and flow loops on headers (10%)
(4) Too much reset action in overhead pressure loops on columns and vessels (10%)
(5) Set point response of batch temperature loops (5%)
(6) Interacting temperature loops for 2 point composition control of columns (5%)
(7) Interacting temperature loops on furnaces and reactors (5%)
(8) Set point response of batch pH loops (5%)

All plants have oscillations. Process control improvement can reduce or eliminate these oscillations. In these days of tight budgets and resources, how do you justify cost and effort to fix the problem?

The "before" and "after" distribution and location of process variability depicted in slide 1 in ProcessControlBenefits.pdf is the classic presentation on how tighter control can result in benefits. If you can reduce the standard deviation (sigma), you can move the set point closer to the constraint without increasing the number of violations of the constraint. What I have added to the slide is the practical situation where operations give themselves a cushion or margin, particularly if there is no online monitoring system with data analytics that can provide the process knowledge and confidence needed to operate at edge of the product range to gain a competitive edge. In my experience the margin almost always exists, it is just a matter of how much. The margin is perhaps easiest to visualize in plastic sheet manufacturing. The greatest variability in sheet thickness and optical clarity occurs near the edge. An extra margin of sheet is trimmed off to make sure there are no off-spec sheets. Without doing anything to provide tighter thickness control, the trim width could be change if there was enough process knowledge and confidence. The benefit from less scrap can be taken as a decrease in raw material and utility cost to obtain the existing capacity or as an increase in capacity for the existing raw material and utility use as noted in the categorization of possible benefits on slide 2.

The key idea here is that most benefits are not achieved until we change a set point. We can find the existing margin by the intelligent use of an online data analytics system and we can create a new margin by tighter process control. Once we know the margin, we need to move the set point to eliminate the margin. A "good" process control engineer can draw straight lines. A "great" process control engineer can move the straight lines.

Often we are not so lucky to have an online measurement and closed loop control of the product quality or concentration that is the ultimate process output as implied by slide 1. What we have is lot of intermediate unit operations in a plant each with a multitude of process inputs and process outputs that can be oscillating. As a minimum many chemical and biochemical plants have a reaction unit operation followed by separation, purification, and formulation unit operations. For solid products, there is often additional equipment for crystallization, centrifuging, drying, and blending. Each of these unit operations has process inputs and outputs with a degree of variability.

So we have short term or long term oscillations at various points in the process and can reduce or eliminate these oscillations. How do we justify the cost and quantify the benefits of better process control?

In order to estimate what we can gain from process control improvement, we need to know process gains. A process gain is the change in a process output divided by the change in a process input. There is a steady state process gain which is the final change after all transients have dies out and the process has reached a new steady state. Steady state simulations can provide these process gains and through virtual experimentation quantify the changes in the product composition or quality for changes in an upstream process variable. For oscillations there is also a dynamic effect where the oscillations of a process variable are attenuated by downstream volumes. The attenuation is proportional to the period and inversely proportional to the residence time of volumes with back mixing from turbulence, recirculation, and agitation. The follow equation can be used to estimate the amplitude of oscillations in a process output (Ao) for oscillations in a process input (Ai), a steady state process gain (Kp) between the process output and input, a period of oscillation (Po), and for a residence time of a back mixed volume (Tm). The residence time is the volume divided by the total flow rate through the volume.

Ao = Ai * Kp * [Po / (6.28*Tm)]

We can compute the steady process gain from first principle equations as shown in Advanced Application Note 4 posted on March 25 or get it from a steady state simulation as long as we avoid a valve position as the process input. The installed characteristic of teh valve and hence the slope of this curve's contribution to the process gain is typically not simulated correctly. Dynamic simulations that have a flow-pressure solver should be able to predict the oscillation amplitude but in practice the results are poor because these simulations do not sufficiently model process and automation system dead times, valve backlash and sticktion, and control loop tuning that determines the period of the oscillations.

The best way to estimate the relationship is the find the process variable furthest upstream with the same dominant period of oscillations that are in the product. The ratio of the amplitudes (Ao/Ai) is the dynamic process gain. For a given reduction in the amplitude Ai, you can estimate the corresponding reduction in amplitude Ao. A power spectrum analysis of the process variables can be used to find the variables with the corresponding dominant frequencies. We then need to follow through and see how much of a margin we can create by a reduction in the product oscillation amplitude.

Once we have the margin, we need to work backwards (upstream) to get at what is the corresponding reduction in utility flow or feed flow. How do we do this? Again we need to use process gains. We divide the product margin by the process gain to get the change to be made in a key upstream loop set point once we have reduced the oscillations in the product. Consider the case where the key loop is a reaction or distillation temperature loop. We then divide the change in reactor or column temperature set point by the steady state process gain for the required change in coolant temperature and reflux flow, respectively. Next we divide this change in coolant temperature or reflux flow by the steady state process gain for the required change in coolant flow and steam flow, respectively. Finally, we multiply the required changes in utility flow by their cost per unit flow to get at the cost savings. Knowledge of the process and the gains are in the process gains and the periods of oscillation. Online data analytics can find the margins, power spectrum analyzers can find periods, and online controller tuning can find the process gains.

The leading cause of oscillations is a level loop with overly aggressive tuning and in some cases excessively sluggish tuning. Several very sophisticated process studies have come to down to this simple fix. Next week we will look in more detail at this culprit and explore the other causes of oscillations.