Since this will be my last blog post as Society President, I would like to thank you for all of your contributions to our profession and to ISA as well as for your support over the past year.

My experience as 2018 ISA president has made me deeply aware of the importance and relevance of the Society—particularly in light of the many changes and challenges that we faced over the last 18 months—and of the outstanding contributions and efforts that have been asked of ISA leadership, members, and staff during this time.

It is immensely gratifying and inspiring to see how much has been accomplished over these many months and recognize how these accomplishments have positioned the Society for much greater success in the months and years ahead. ISA has put in place the strategic foundation and plans needed to sustain a bright and prosperous future for the automation profession. We must never lose sight of the fact that automation makes our world a better and safer place for everyone.

In the New Year and beyond, ISA must look both inward and outward to: develop new growth opportunities, increase awareness of its value proposition, strengthen its brand, tap into new revenue streams, boost membership, and develop a proactive plan for engaging the next generation of leaders. To do so, ISA must be successful in three vital areas: operations, collaboration, and innovation.

We need to closely examine the operations at all levels of the Society to optimize systems and resources. Secondly, we need to seek out opportunities to collaborate, both within the Society (across geographic, technical, and operational boundaries) and with external entities (including government, academia, the Automation Federation, other professional organizations and members of the automation industry) where synergy either already exists or can be created. Strategic partnerships are key to collaboration and operational excellence.  

We entered 2018 with new and emerging technologies that allowed us to better engage with automation professionals in new and exciting ways and across the global playing field. New and emerging technologies are creating new opportunities for automation around the world, and are changing the roles, responsibilities, and needs of automation professionals. All of these developments are impacting ISA and its products, services, and customers.

ISA’s success depends on its ability to seize these opportunities while continuing to deliver value—both to individual members and to the global automation community.

In responding to its challenges, ISA has clearly defined its mission, vision, and goals in 2018, and the products and services to be delivered, the partnerships to be secured, and marketplace opportunities to be explored. We will continue to engage members, volunteers, and staff in the ongoing conversations in the weeks and months ahead.

ISA needs your help; you are already contributing by virtue of your ISA membership. In 2019, I challenge you to do even more. Seek out ways in which you can use your experience and expertise to serve. Introduce your colleagues and company to ISA. Make them aware of what ISA offers. Get more involved in your ISA section or division. Join an ISA LinkedIn group. Reach out to local, national, or global ISA leadership to inquire how you can help. Get involved!

Every generation of ISA membership has the opportunity and, I believe, the responsibility to move the Society forward in the world of automation. This is our time. The automation profession continues to make the world a better place. We should all be proud to be a part of the positive change automation has created. ISA plays a pivotal role in this process—helping members and other automation professionals improve their technical skills and knowledge, and enabling companies increase throughput, reduce waste, and improve productivity and profitability—both safely and securely. 

ISA has an obligation to help industry leaders as well as the public better recognize the value and benefits of automation. Through our collaboration with the Automation Federation, we’re just beginning to scratch the surface in leveraging our capabilities in industrial cybersecurity. While there is growing awareness among industry leaders of the risks of industrial cyberattack, we need to work harder to foster recognition in the marketplace that ISA offers real solutions to mitigate these risks. It’s also important to note that conversations about cybersecurity can serve as the door opener to educate those about other important ISA offerings and capabilities.

ISA also has benefited greatly over these past many months from the support and involvement of our corporate partners and sponsors. These relationships inspire new and collaborative ways of solving common challenges, make our members and customers aware of additional resources and capabilities that could benefit them, and foster best-practice approaches that advance the automation and control profession.

Section Engagement

Given their geographically based structure, ISA sections offer a convenient way for members to take part in ISA initiatives and events. Here are just a few ways you can get involved and contribute at the section level:

• Team up with other ISA members to explore common professional interests.
• Invite guest speakers to section meetings, creating a learning environment.
• Arrange section tours of local plants and facilities.
• Develop new networking, social and recreational events.
• Speak at local technical colleges and universities to generate student interest in automation careers.
• Encourage local students to attend section events and become ISA student members.

Division Involvement

As an ISA member, you should take full advantage of your two free technical division memberships: one from the Automation and Technology Department and one from the Industries and Sciences Department. Division memberships enable automation professionals the opportunity to:

• Attend, help plan, and conduct technical division symposia and events.
• Stay up to date on current technical trends and news by reading division newsletters and web sites.
• Write, review, or present technical papers for ISA publications.
• Network with colleagues across the globe.
• Explore professional development opportunities and gain leadership skills.
• Develop workshops and short courses for division members.
• Exchange ideas and insights through email discussions.

ISA Training Programs

ISA, through its leading automation and control training programs, can better prepare technicians and engineers—from those new to the job market to the highly experienced—for the workplace demands and advanced manufacturing jobs of the future.

ISA’s worldwide leadership in automation and control training begins with its subject matter experts. ISA’s instructors and consultants are at the forefront in their field; their unrivaled knowledge and marketplace experience provide practical, real-world solutions. Why don’t you organize a training course for your section members or discuss ISA training courses with your employer?

As I reach the twilight of my year as 2018 Society President, I am gladly handing over the helm to President-Elect Paul Gruhn, who will steer the Society along the next leg of its voyage. I have full confidence that Paul—with the assistance of ISA Executive Director Mary Ramsey and ISA staff manning the engines—will guide ISA through the dawn of a new day and onto a brighter and prosperous future in 2019.

In closing, I want to express how grateful I am to have learned so much from previous ISA leaders with whom I have served, and to have gained and benefited from the experience of our current ISA leadership team that are now so many valuable friends. These relationships have been vital to me during my year as President and instrumental to my success on a professional and personal level.

My sincere thanks to all and best wishes for the New Year.

A version of this article also has been published at ISA Insights.

Dead time is the source of the ultimate limit to control loop performance. The peak error is proportional to the dead time and the integrated error is dead time squared for load disturbances. If there was no dead time and no noise or interaction, perfect control would be theoretically possible. When the total loop dead time is larger than the open loop time constant, the loop is said to be dead time dominant and solutions are sought to deal with the problem.

Anuj’s Question

Is there any other control algorithm available to improve loop performance for dead time dominant systems other than using Smith predictor or model predictive control (MPC), both of which requires identification of process model?

Greg McMillan’s Answer

The solution cited for deadtime dominant loops is often a Smith predictor deadtime compensator (DTC) or model predictive control. There are many counter-intuitive aspects in these solutions. Not realized is that the improvement by the DTC or MPC is less for deadtime dominant systems than for lag dominant systems. Much more problematic is that both DTC and MPC are extremely sensitive to a mismatch between the compensator and model deadtime versus the actual total loop deadtime for a decrease besides an increase in the deadtime. Surprisingly, the consequences for the DTC and MPC are much greater for a decrease in plant dead time. For a conventional PID, a decrease in the deadtime just results in more robustness and slower control. For a DTC and MPC, a decrease in plant deadtime by as little as 25 percent can cause a big increase in integrated error and an erratic response.

Of course, the best solution is to decrease the many source of dead time in the process and automation system (e.g., reduce transportation and mixing delays and use online analyzers with probes in the process rather than at-line analyzers with a sample transportation delay and an analysis delay that is 1.5 times the cycle time). An algorithmic mitigation of consequences of dead time first advocated by Shinskey and now particularly by me is to simply insert a deadtime block in the PID external-reset feedback path (BKCAL) with the deadtime updated to be always be slightly less than the actual total loop deadtime. Turning on external-reset feedback (e.g., dynamic reset limit) on and off enables and disables the deadtime compensation. Note that for transportation delays, this means updating the deadtime as the total feed rate or volume changes. This PID+TD implementation does not require the identification of the open loop gain and open loop time constant for inclusion as is required for a DTC or MPC. Please note that the external-reset feedback should be the result of a positive feedback implementation of integral action as described in the ISA Mentor Program webinar PID Options and Solutions – Part 3.

There will be no improvement from a deadtime compensator if the PID tuning settings are left the same as they were before the DTC or by a deadtime block in external-reset feedback (PID+TD).  In fact the performance can be slightly worse for even an accurate deadtime. You need to greatly decrease the PID integral time toward a limit of the execution time plus any error in deadtime. The PID gain should also be increased. The equation for predicting integrated error as a function of PID gain and reset time settings is no longer applicable because it predicts an error less than the ultimate limit that is not possible. The integrated error cannot be less than the peak error multiplied by the deadtime. The ultimate limit is still present because we are not making deadtime disappear.

If the deadtime is due to analyzer cycle time or wireless update rate, we can use an enhanced PID (e.g., PIDPlus) to effectively prevent the PID from responding between updates. If the open loop response is deadtime dominant mostly due to the analyzer or wireless device, the effect of a new error upon update results in a correction proportional to the PID gain multiplied by the open loop error. If the PID gain is set equal to the inverse of the open loop gain for a self-regulating process, the correction is perfect and takes care of the step disturbance in a single execution after an update in the PID process variable.

The integral time should be set smaller than expected (about equal to the total loop deadtime that ends up being the PID execution time interval) and the positive feedback implementation of integral action must be used with external reset feedback enabled. The enhanced PID greatly simplifies tuning besides putting the integrated error close to its ultimate limit. Note that you do not see the true error that could’/ have started at any time in between updates but only see the error measured after the update.

For more on the sensitivity to both increases and decrease in the total loop deadtime and open loop time constant, see the ISA books Models Unleashed: A Virtual Plant and Predictive Control Applications (pages 56-70 for MPC) and Good Tuning: A Pocket Guide 4th Edition (pages 118-122 for DTC). For more on the enhanced PID, see the ISA blog post How to Overcome Challenges of PID Control and Analyzer Applications via Wireless Measurements and the Control Talk blog post, Batch and Continuous Control with At-Line and Offline Analyzers Tips.

The following figures from Models Unleashed shows how a MPC with two controlled variables (CV1 and CV2)  and two manipulated variables for a matrix with condition number three (CN = 3) responds to a doubling and a halving of the plant dead time (delay) when the total loop dead time is greater than the open loop time constant.

 

This automation industry quiz question comes from the ISA Certified Control Systems Technician (CCST) program. Certified Control System Technicians calibrate, document, troubleshoot, and repair/replace instrumentation for systems that measure and control level, temperature, pressure, flow, and other process variables. Click this link for more information about the CCST program.

When bundling and running cables in conduit and duct, the listed conductors are normally separated in different duct or conduit runs based on signal type. Which type would normally be shielded?

a) power wires
b) signal wires
c) ground wires
d) control wires
e) none of the above

When I was much younger I spent a lot of time driving the complex freeway network around San Francisco. Often there were several engineers in a car together, all motivated to understand how to get from where we were to where we were going more quickly. Eventually, the discussion turned to how similar the traffic was to the flow of water through pipes and conduits. We wondered if fluid flow science could be used to predict or even model traffic flow. After a few mornings of this the smartest guy in the car offered to buy lunch for the first guy that could verbalize the answer.

Confronted with a deafening silence he said that each car-particle in the “flowing” stream was different from fluid flow in at least one profound way (beyond its modularity). In the case of a car, where it was going to go, or what it was going to do, next depended on what happened in the car-module itself; while in a flowing stream, where any element of the flowing stream went was the result of a vector field of circumstances surrounding it. If course, our late friend Chuck was observing that the science we used in fluid mechanics was based on the Navier-Stokes analysis which was, in turn based on the Cauchy momentum equation, all of which predicted fluid motion based on fields of parameters (e.g., pressure, energy, temperature) over the field of flow. It also depends on conditions in the fluid itself as it encounters that Cauchy field.

Fluids have a lot of condition-dependent properties. They consist of solids, liquids, gasses, or combinations thereof. At any region, their flow-state condition impacts their density and viscosity, both of which have a lot to do with how they respond to the field of parameters around them. While most of elementary fluid mechanics is based on the idea of “homogeneous flow” it is very uncommon in the petroleum world to find one. Vapor pressure, for example, is the surrounding pressure required to maintain a liquid in a vapor-free state. If the surrounding pressure is below the vapor pressure some portion of the liquid will “flash” into vapor.

When the fluid temperature increases so does the vapor pressure. In any region of the pipeline, of course, the density of the fluid found there depends on how much of it is gas and how much is liquid. All parameters dependent on density are affected by these changes. Sometimes these changes impact the ability of process equipment to operate as intended. For example, a meter designed for gas flow measurement may mismeasure a gas-liquid stream of the same chemical composition.

A lot of process design goes into anticipating and mitigating such problems. The more precise understanding of the fluid field needs to be the more important these issues are. Of course accuracy and understanding of these issues can have a lot to do with process measurements and characterization of the impacts that may stem from them.

In leak detection there are a lot of fluid state issues, beginning with changes in density appearing as changes in velocity or changes in mass that adversely affect calculations based on unwarranted assumptions.

Observation of these conditions, or even the parameters that influence them, are exacerbated by the generally substantial distances pipelines traverse. Further, some parts of the pipeline may be buried, some parts may be exposed to sun and weather, and some may be buried in soil with weather-dependent thermal conductivity.

Observation may be difficult; and without it, precise parameter assessment and anticipation of the resulting fluid characteristics may not be possible. It all sounds impossibly complex but with some good design decisions and some anticipation of the likely effects, success is possible albeit with accuracy and sensitivity dependent on the quality of what can be known about the situation.

Can these issues be resolved by modeling? Generally yes, to a degree; but it remains impossible to manufacture data. One can draw a straight line between two points. One can draw a parabola or a hyperbola through a couple of points, and estimate a cubic with three. Beyond the simple things, either more data or more assumptions are required. Once upon a time I was assisting a researcher on a process enhancement issue.

During the discussion I observed that for a curve as complex as what he was expecting we needed more that the two points we could put a simple curve through. He then pronounced in jest that he could present a family of curves using but a single point. This guy was very smart and a pioneer in process control, and perhaps a lot of his ability came from knowing more about these process dynamics that the rest of us. In any case, he went on to do some really great things, but…

This post is an excerpt from the journal ISA Transactions. All ISA Transactions articles are free to ISA members, or can be purchased from Elsevier Press.

Abstract: An improved proportional integral derivative (PID) controller based on predictive functional control (PFC) is proposed and tested on the chamber pressure in an industrial coke furnace. The proposed design is motivated by the fact that PID controllers for industrial processes with time delay may not achieve the desired control performance because of the unavoidable model/plant mismatches, while model predictive control (MPC) is suitable for such situations. In this paper, PID control and PFC algorithm are combined to form a new PID controller that has the basic characteristic of PFC algorithm and at the same time, the simple structure of traditional PID controller. The proposed controller was tested in terms of set-point tracking and disturbance rejection, where the obtained results showed that the proposed controller had the better ensemble performance compared with traditional PID controllers.

Free Bonus! To read the full version of this ISA Transactions article, click here.

2006-2018 Elsevier Science Ltd. All rights reserved.

This automation industry quiz question comes from the ISA Certified Automation Professional (CAP) certification program. ISA CAP certification provides a non-biased, third-party, objective assessment and confirmation of an automation professional’s skills. The CAP exam is focused on direction, definition, design, development/application, deployment, documentation, and support of systems, software, and equipment used in control systems, manufacturing information systems, systems integration, and operational consulting. Click this link for more information about the CAP program.

Common mode noise can be best defined as:

a) noise that appears equally and in phase from each current carrying conductor to ground
b) noise that appears equally and 90 degrees out-of-phase between each current carrying conductor to ground
c) noise that appears between the phase or signal and its return
d) noise that cannot be measured by conventional means
e) none of the above

Normal

0

false

false

false

EN-US

X-NONE

X-NONE

/* Style Definitions */

table.MsoNormalTable

{mso-style-name:”Table Normal”;

mso-tstyle-rowband-size:0;

mso-tstyle-colband-size:0;

mso-style-noshow:yes;

mso-style-priority:99;

mso-style-parent:””;

mso-padding-alt:0in 5.4pt 0in 5.4pt;

mso-para-margin-top:0in;

mso-para-margin-right:0in;

mso-para-margin-bottom:8.0pt;

mso-para-margin-left:0in;

line-height:107%;

mso-pagination:widow-orphan;

font-size:11.0pt;

font-family:”Calibri”,sans-serif;

mso-ascii-font-family:Calibri;

mso-ascii-theme-font:minor-latin;

mso-hansi-font-family:Calibri;

mso-hansi-theme-font:minor-latin;

mso-bidi-font-family:”Times New Roman”;

mso-bidi-theme-font:minor-bidi;}

Our annual meetings aren’t just about passing the gavel to the next president or attending a few sessions and dinners.

Of course, outgoing ISA President Brian Curtis did pass the gavel to Paul Gruhn, who will take over as president in 2019. The most recent Annual Leadership Conference (ALC) in Montreal, Canada was more of a time for automation professionals and members to collaborate, brainstorm, and make decisions about future direction.

Even in today’s highly connected culture, there is no substitute for face-to-face networking. “This was a great reminder that in-person events are still very relevant in a digital world,” said Mary Ramsey, ISA executive vice president.

Longstanding traditions—such as the Honors and Awards Gala—melded with designating new section leaders and talking with students about how ISA can become a more positive force in the rapidly changing automation field. A celebration of success and achievements was as much a part of the event as reviewing international standards.

As Brian Curtis noted, all members need to come together to keep focused on growth and critical objectives. “If we work as a team and stay actively engaged, we’re sure to keep the positive energy going,” he said. You can read more in this previous post about our outgoing president’s conference review, and why we need you to help lead our continued march to the automation future.

The 2019 ALC will be held in beautiful San Diego, Calif., USA. If you’d like to become more involved with ISA, visit our website.

The two most common categories of process responses in industrial manufacturing processes are self-regulating and integrating. A self-regulating process response to a step input change is characterized by a change of the process variable, which moves to and stabilizes (or self-regulates) at a new value. An integrating process response to a step input change is characterized by a change in the slope of the process variable. From the standpoint of a proportional, integral, derivative (PID) process controller, the output of the PID controller is an input to the process. The output of the process, the process variable (PV), is the input to the PID controller.

In Part 1 and Part 2 of this series, I presented tuning a PID controller for the basic “integrating” and “first-order, self-regulating” process responses. These responses are shown in figure 1. However, in some cases, the process response has more complex dynamics than those discussed in these two articles. The following complex process responses have been observed in the processes of many industries, including chemical, refining, petrochemical, pulp and paper, oil and gas, and pharmaceutical.

• self-regulating, second order, overdamped
• self-regulating, second order, underdamped
• self-regulating, second order plus lead
• self-regulating, second order plus lead with overshoot
• self-regulating, second order, nonminimum phase
• integrator plus first-order lag
• integrator plus first-order lead
• integrator plus nonminimum phase

The two most common of these complex dynamic responses are the “self-regulating, second-order, overdamped” and the “integrator plus first order lag” responses.

Another topic that is often overlooked when tuning PID controllers is the concept of resonance. When a PID controller is used on a process that has dead time (and all have some!), the closed-loop response has a resonant frequency at which it will amplify variability with frequency components at or near the resonant frequency. The more aggressive the tuning is, the higher the amplification.

This article covers the tuning of a PID controller for two of the most common complex dynamic responses—“self-regulating, second order, overdamped” and “integrator plus first order lag”—and the concept of resonance.

Challenges

It is rare to have a perfect model fit to an actual process response. Thus, the goal is to obtain a reasonably “good” model fit. This involves selecting a model type that is “complex” enough to be a good fit to the process response.

The two most common of the complex dynamics are the self-regulating, second-order, overdamped (hereinafter called “self-regulating second order”) and the integrator plus first order lag (hereinafter called “integrator plus lag”). Figure 1 contrasts the response of a self-regulating, first order to a self-regulating, second order and an integrator to an integrator plus lag.

The closed-loop response of a tuning method should be selectable in order to optimize the process performance that may or may not require optimized control loop performance. Optimizing the process performance may require that the loops in the unit have a coordinated response: either a very slow response to use process capacity to absorb variability or a very fast response to maximize regulation of load disturbances.

Tuning for a second-order, self-regulating process

A tuning methodology called lambda tuning addresses these challenges. The lambda tuning method allows the user to choose the closed-loop response time, called lambda, and calculate the corresponding tuning to achieve the desired response time. The lambda time is chosen to achieve the desired process goals and stability criteria. This could result in choosing a small lambda for good load regulation, a large lambda to minimize changes in the controller output and manipulated variable by allowing the PV to deviate from the set point, or somewhere in between these two extremes. More importantly, the lambda of the loop can be used to coordinate the responses of many loops to reduce interaction and variability.

Lambda tuning for self-regulating processes can cause a closed-loop response that is slower or faster than the open-loop response time of the process. Though lambda is defined as the closed-loop time constant of the process response to a step change of the controller set point, the load regulation capability is also a function of the lambda of the loop. The response to a step set point change and a step load change for a self-regulating process response with lambda tuning is shown in figure 2.

Self-regulating process responses typically include dead time and can usually be approximated by a “first-order plus dead time” or “second-order plus dead time, overdamped” response. This article describes the lambda tuning procedure when the process response can be best approximated by the latter. The lambda tuning for a second-order plus dead time response can be approximated with manual analysis and calculations. A more rigorous analysis is required for more exact tuning.

Procedure

The lambda tuning method for self-regulating processes has three steps:

1. Identify the process dynamics.
2. Choose the desired closed-loop speed of response, lambda.
3. Calculate the required PID tuning constants.

Figure 3 shows the dynamic parameters of a self-regulating, second-order process, which include dead time (Td), in units of time; primary time constant (tau 1), in units of time; secondary time constant (tau 2), in units of time; and the process gain (Kp), in units of percent controller PV/percent controller output. Typically several step tests are performed; the results are reviewed for consistency; and the average process dynamics are calculated and used for the tuning parameter calculations. If the controller output goes directly to a control valve, any significant dead band in the valve will reduce process gain if the output step was a reversal in direction. If the controller output cascades to the set point of a “slave” loop, the slave loop should be tuned first.

The next step is to choose the lambda to achieve the desired process control goal for the loop within the allowable stability margin for the expected changes in process dynamics. A shorter lambda produces more aggressive tuning with less stability margin. More aggressive tuning also has a larger amplification of disturbances with a period of oscillation near the resonant period of the loop. A longer lambda produces less aggressive tuning and more stability margin. It is not uncommon for the process dynamics, particularly the process gain, to vary by a factor of 0.5 to 2. If testing during different conditions reveals that the process dynamics change significantly, then an additional margin of stability is required. The process response could also be “linearized,” or adaptive tuning could be used.

If the potential change in process dynamics is unknown, starting with lambda equal to three times the larger of the dead time or time constant provides stability even if the dead time doubles and the process gain doubles. If it is desirable to coordinate the response of loops to avoid significant interaction, the lambda of the interacting loops can be chosen to differ by a factor of five or more. For cascade loops, the lambda can be chosen to ensure the slave loop of the cascade pair has a lambda 1/5 or less of the master control loop.

If derivative is used, the lowest recommended lambda for a self-regulating, second-order process is approximately equal to the larger of the dead time or lag time constant. If derivative is not used, then the minimum lambda varies depending on the amount of the secondary time constant, but is in the range of two times (dead time + tau 2). Both of these lower limits on lambda result in aggressive tuning with a very low gain and phase margin. Thus, moderate increases in the dead time or process gain can cause instability of the loop. Using properly set derivative action actually provides more stability for these complex dynamics.

From a stability standpoint, there is no upper limit on the lambda. If the lambda is not chosen based on a coordinated response, a good starting point for stability for PI or PID tuning is:

Equation 1. Lambda = 3 × (larger of dead time or time constant)

The tuning performance can be monitored for a time period and adjusted to be a shorter or longer lambda as needed.

The final step is to calculate the tuning parameters from the process dynamics. Care should be taken to use consistent units of time for the dead time and the lambda. For a self-regulating, second-order process response, the controller gain, integral time, and derivative time are calculated with the following equations. These equations are valid for the series (sometimes called classical or interactive) form of the PID implementation. Note that only the controller gain changes as lambda (λ) changes. The integral time remains equal to the time constant regardless of the lambda chosen. Conversion of the tuning from series PID form to standard form (sometimes called ideal, noninteractive) is provided in the last section.

Example

Consider the simulated temperature controller shown in figure 4. The temperature controller, TIC-302, manipulates a properly selected control valve that has a high-performance digital positioner.

Figure 5 shows a step test of the temperature controller to identify the process dynamics. Several such steps were analyzed, and the average process dynamics were calculated. The process gain is 1.0%PV/%OUT; the dead time is 20 seconds; the primary time constant is 80 seconds; and the secondary time constant is 60 seconds.

In this example, there are no “loop response coordination” requirements, so the initial lambda is chosen to be 3 × (larger of dead time or primary time constant) = 3×80 seconds = 240 seconds.

Now, the tuning for a series PID form can be calculated with the lambda tuning rules.

If the process dynamics have been tested or proven over time to be consistent in the overall operating range, more aggressive tuning may be desirable. The following table shows the tuning for different values of lambda. Note that the integral and derivative time remains the same for all choices of lambda.

Lambda (seconds)	Gain	Integral time (seconds)	Derivative time (seconds)	Amplitude ratio
240	0.31	80	60	1.09
120	0.57	80	60	1.12
80	0.8	80	60
20 (minimum value = dead time)	2.0	80	60

Figure 6 shows the response to a step set point and a step load change for each of the lambda values in the table. Note that the tuning is stable for shorter lambda values than the starting point of 3 × (larger of dead time or time constant). However, this is with unchanging process dynamics in a simulator. Additional tests on a real process, in different operating conditions, will help determine how consistent the process dynamics are and whether more aggressive tuning is stable and provides the desired process performance.

The next step is to choose the lambda to achieve the desired process control goal for the loop within the allowable stability margin for the expected changes in process dynamics. The same warnings about variations in process dynamics, control valve performance issues, and the need to tune the slave loop first in a cascade arrangement apply to this process response. Without using derivative, the minimum lambda considering stability and resonance requirements is approximately 3 × (dead time plus lag time constant). Using derivative, the minimum lambda considering stability and resonance requirements is approximately the larger of the dead time or the lag time constant. If a lambda is chosen that is near these limits, the resulting tuning should be tested on a simulator to verify its stability and resonance. Analysis tools can provide a more accurate calculation of lambda tuning values and the limits of lambda based on stability and resonance.

The final step is to calculate the tuning parameters from the lambda and process dynamics. Care should be taken to use consistent units of time for the dead time, lag time constant, and the lambda. For an integrator plus first-order lag process response, the controller gain, integral, and derivative times are calculated with the following equations. These equations are valid for the series (sometimes called classical, interactive) form of the PID implementation. Note that both the PID gain and integral time change as lambda changes. The derivative time remains the same for all choices of lambda. Conversion of the tuning from series PID form to a standard form (sometimes called ideal or noninteractive) is provided in the last section.

Example

The process shown in figure 8 will be used as an example. The level controller, LIC-102, manipulates a properly tuned flow controller, FIC-102. The level process on these types of applications typically has an integrator plus lag response.

Figure 9 shows a step test of the level controller, LIC-102, to identify the process dynamics. Several such steps were analyzed, and the average process dynamics were calculated. The process gain is 0.005%PV/second/%OUT; the dead time is 20 seconds; the first-order lag time constant is 60 seconds.

The next step is to choose the lambda. One method to choose the lambda is the “allowable percent PV deviation” method. If it is desired to keep the %PV within an “allowable deviation” (AD) from its previous value due to step change load disturbance (MLD, in units of the %OUT of the controller), then the required lambda can be calculated from equation below.

Equation 8.   

For this example, AD = 20%PV, MLD = 50%OUT. From above, Kpi = 0.005 %PV/sec/%OUT.

Thus, Lambda = (2 × 20%PV)/(0.005%PV/sec/%OUT × 50%OUT) = 160 seconds.

Now, the tuning for a series PID form can be calculated using equations 5–7.

Integral time =   = 2 × 160 seconds + 20 seconds = 340 seconds

Controller gain == (340 sec)/((0.005 %PV/sec/%OUT) × (160 sec+20 sec)2)

Derivative time = Lag time constant = 60 seconds

The following table shows the tuning for several different values of lambda. Note that both the PID gain and integral time change as lambda changes. The derivative time remains the same for all choices of lambda.

Lambda (seconds)	Gain	Integral time (seconds)	Derivative time (seconds)
240	1.48	500	60
160	2.1	340	60
60 (minimum value ~ larger of lag or dead time)	4.4	140	60

Figure 10 shows the response to a step set point and a step load change for each of the lambda values in the table. Note that the tuning is stable for much shorter lambda values than the starting point of 3 × (larger of dead time or time constant). However, this is with constant process dynamics in a simulator. Additional tests on a real process, at different operating conditions, will help determine how consistent the process dynamics are.

Meeting process goals

Most published PID controller tuning methods are designed for optimum loop performance, not necessarily optimum process performance. Optimizing the process performance may require that the loops in the unit have a coordinated response, a very slow response to utilize process capacity to absorb variability, or a very fast response to maximize regulation of load disturbances. The lambda tuning method provides the ability to tune the PID controllers in a process system to achieve process performance goals regardless of the loop requirements. The lambda tuning method can be used for all of the common complex dynamics that were listed in the introduction.

Loop resonance and variability amplification

The combination of a PID controller (with P, PI, or PID tuning) is used on a process that has dead time; the closed-loop response will have resonant frequency at which it will amplify variability with frequency components at or near the resonant frequency. And, the more aggressive the tuning is, the higher the amplification. Figure 11 shows a frequency response plot (Bode plot) of a PV disturbance for the temperature loop, TIC-301, used in the example for a self-regulating, second-order process response. The aggressive tuning of lambda = dead time = 20 seconds results in an 82 percent amplification of any variability with a period near 138 seconds. In other words, if there is variability that has a period near 138 seconds with the controller in manual, putting the controller in automatic will increase the variability by 82 percent! Interestingly the limit cycle from a poorly performing control valve might have a period around this value! Figure 12 shows a time series presentation of the data to illustrate this concept. Figure 13 shows how the aggressiveness of the tuning affects the amplification ratio. The concept of resonance and variability amplification should be considered in the controller tuning process.

Conversion of the tuning from to series PID form to a standard form

Equations 9–11 below can be used to convert tuning from a series PID form to a standard PID form. The equations are based on series PID tuning parameters with the following units.

• controller gain: %OUT/%PV (often said to have “no units”)
• integral time: seconds
• derivative time: seconds

Equation 9. Standard PID gain = Gain × (integral time + derivative time)/(integral time)

Equation 10. Standard PID integral time = (integral time + derivative time)

Equation 11. Standard PID derivative time = (integral time × derivative time)/(integral time + derivative time)

Useful models

It has been said that “All process models are wrong, but some are useful.” Of the more than 4,000 process responses that I have analyzed in the past 15 years, I have never measured the same exact model parameters for different tests of the same process! Thus, the goal for process model–based tuning is to identify a reasonable fit to a particular process model. The better the model fit, the more accurate the tuning and results will be. In some cases, this requires choosing a more complex process model. The consistency (linearity) of the process response will affect the margin of stability that is required when tuning the loop. In some cases, techniques to linearize the process response from the perspective of the controller may be required. During the controller tuning process, the process objectives should considered when choosing the control loop objective. Once the control loop objective is identified, a tuning methodology that can achieve the desired control loop object will help achieve the process objectives.

A version of this article also was published at InTech magazine. 

Pressure transmitters used in the process industries are very durable and reliable instruments. Even so, they still require periodic maintenance and calibration to ensure optimal performance. This is an area of confusion for many. Typical questions include:

• Are we calibrating our transmitters too often, resulting in excessive downtime and unnecessary maintenance expenses?
• Are we calibrating our transmitters too infrequently, resulting in quality issues and possible loss of product?
• Are we calibrating our transmitters correctly?

As with most things in life, there is no “one size fits all” answer. However, there are simple best-practice guidelines, which can be modified to fit specific applications. This article helps answer the basic questions facing process plant personnel with regard to calibration.

How often?

Each process plant has to determine correct calibration intervals based upon historical performance and process-related requirements. Factors you need to consider that may influence this decision are:

• Are there any local, national, safety, or environmental regulations that must be observed?
• What is your reason for requiring calibration: quality, safety, or standard maintenance?

Process conditions

• Is there a homogeneous process fluid with a stable pressure/temperature?
• Will the process conditions fluctuate significantly?
• Is there risk of buildup, corrosion, or abrasion to the pressure transmitter?
• Will heavy vibration be present?

Ambient conditions

• Will the pressure transmitter be installed in a well-controlled environment with low humidity, normal or stable temperatures, and few contaminants, such as dust or dirt?
• Is an outdoor transmitter exposed to widely varying weather conditions or high humidity?

If you have no significant history or regulatory requirements to guide you in developing your calibration procedures, a good place to start is with the following general guidelines.

• Direct-mounted pressure transmitters installed inside in a controlled environment on a process with stable conditions should be calibrated every four to six years.
• Direct-mounted pressure transmitters installed outside on a process with stable conditions should be calibrated every one to four years, depending upon ambient conditions.

If a remote diaphragm seal is employed on a pressure transmitter, the calibration interval should be reduced by a factor of two (i.e., a four-to-six year interval is reduced to two to three years). This is because a remote diaphragm seal employs more fill fluid than a direct-mounted configuration. Consequently, it will experience more mechanical stress from process or ambient temperature fluctuations. Most remote diaphragms are flush faced where the diaphragm/membrane is susceptible to physical damage (dents or abrasions) that can cause offset or linearity issues. If the process regularly experiences significant pressure swings or overpressurization events, reducing the calibration interval by a factor of two is a good rule of thumb.

How accurate?

How good is good enough? In other words, what is the maximum permissible error (MPE) for your calibration? Many make the mistake of adopting the manufacturer’s reference accuracy as their calibration target. Unfortunately, this means they will have an MPE that is too tight, with a high rate of nonconformance in their calibration process. In the worst case, with a very tight tolerance MPE, it may not be possible for their field or lab test equipment to calibrate some of their transmitters.

A manufacturer’s reference accuracy is based upon tightly controlled environmental conditions that are seldom, if ever, duplicated in a plant environment. Using that reference accuracy for a calibration target also fails to take into account the long-term stability of the instrument.

Over time, all instruments experience slight accuracy degradation due to aging and simple wear and tear on mechanical components. This needs to be considered when establishing the MPE. In general, unless there are mitigating circumstances, it is better to set a reasonable MPE that is achievable with standard field and lab test equipment.

Test equipment starts with an accurate pressure source to simulate the transmitter input. The corresponding output is measured with a multimeter for a 4–20 mA transmitter, or with a specialized device for smart transmitters with digital outputs, such as HART, FOUNDATION Fieldbus, Profibus, or EtherNet/IP.

The test equipment you intend to use should be traceable to the National Institute of Standards and Technology. As a general recommendation, your reference equipment should be at least three times more accurate than the pressure transmitter being calibrated.

 

 

Performing the calibration

Once you have established the calibration interval and MPE, you are ready to perform the actual calibration procedure on your pressure transmitter. The best-practice recommendation is:

1. Mount the transmitter in a stable fixture free from vibration or movement.
2. Exercise the sensor or membrane before performing the calibration. This means applying pressure and raising the level to approximately 90 percent of the maximum range. For a 150 psi cell that means pressurizing it to 130–135 psig. Hold this pressure for 30 seconds, and then vent. Your overall results will be much better than if you calibrate “cold.”
3. Perform a position zero adjustment (zero the transmitter). This is important because the orientation of the fixture used for calibration may be different than the way the transmitter is mounted in the process. Failing to correct for this by skipping this step can result in nonconformance.
4. Begin the calibration procedure. Typically this means three points up (0 percent/50 percent/100 percent) and then three points down. The 4–20 mA output should be 4 mA, 12 mA, and 20 mA at the three points (or the correct digital values for a smart transmitter). Each test point should be held and allowed to stabilize before proceeding to the next. Normally that should take no more than 30 seconds. You can use more points if you require a higher confidence in the performance of the instrument.
5. Compare the results of your pressure transmitter to your reference device.
6. Document the results for your records.

The calibration should be performed in as stable an environment as possible, because temperature and humidity can influence the pressure transmitter being tested as well as the pressure reference. If the results of your calibration are within the MPE, do not attempt to improve the performance of the transmitter.

One mistake many end users make is to regularly perform a sensor trim adjustment of their pressure transmitter—even on new units. A sensor trim corrects the digital reading from the sensor after the analog/digital conversion. Performing a sensor trim on a new transmitter is essentially a single-point calibration under the current plant environment conditions, as opposed to sticking with the original factory calibration.

Factory calibrations of pressure transmitters are performed in a tightly controlled environment and incorporate as many as 100 test points. Performing a sensor trim on a new pressure transmitter under field conditions will cause the unit to operate at less than optimal capacity. A sensor trim should only be performed by a qualified technician under the manufacturer’s guidance.

Who should perform the calibrations?

Even with the sophisticated calibration and reference equipment currently available, there is no substitute for a properly trained technician when it comes to calibrating pressure transmitters. Not only does the technician need to be trained on the mechanics of the calibration process, he or she also must be equally qualified in completing and maintaining the documentation. Repeatability is the key. In the world of calibration, if it is not properly documented, it did not happen.

 

 

Occasionally there are calibrations that cannot be performed in a standard maintenance shop by maintenance technicians. For these cases, an ISO 17025–accredited organization is required. Not only can an ISO 17025–accredited organization perform more stringent calibrations, it provides other value as well:

• Accredited labs can simplify the calibration audit process.
• The process and methodology used by an accredited lab is extremely repeatable, thus there is a high level of confidence in the results from an auditor’s perspective.
• Annual audits of the accredited lab ensure it is consistently performing at a high level for its registered scope of work.

The “correct” calibration cycle for a pressure transmitter depends on the purpose of the calibration and the application. The same pressure transmitters employed in different operating units or processes at the same plant may require different calibration intervals.

Even more important than the calibration interval of the instrument are:

• establishing correct and realistic MPEs
• following correct calibration procedures
• properly training the person performing the calibration
• properly documenting calibration results

Following these guidelines and using judgment based on actual plant operational conditions will help establish proper calibration practices, saving money while maintaining acceptable performance.

A version of this article also was published at InTech magazine. 

This automation industry quiz question comes from the ISA Certified Control Systems Technician (CCST) program. Certified Control System Technicians calibrate, document, troubleshoot, and repair/replace instrumentation for systems that measure and control level, temperature, pressure, flow, and other process variables. Click this link for more information about the CCST program.

When a controller that possesses integral action receives an error signal for significant periods of time, the integral term of the controller will increase at a rate governed by the _______(1)_______. This will eventually cause the manipulated variable to reach 100% (or 0%) of its scale, i.e., its maximum or minimum limits. This is known as ______(2)_______.

The best answer set to fill in the two blanks labeled (1) and (2) above is:

a) (1) offset; (2) integral wind-up
b) (1) integral time of the controller; (2) integral windup
c) (1) sample time of the controller; (2) integral start
d) (1) integral time of controller; (2) derivative time
e) none of the above