Modeling and Control

pH control is an application where everything can be important. A less than right piping, equipment, measurement, control valve, and loop design and implementation can spell big time problems. It is a great training ground and the lessons learned are valuable for many other difficult and important process control applications. The understanding required to make a pH loop work contribute to a much more complete skill base.

I spent a lot of time on pH startups. I found most of the key design concepts needed for success where not discussed anywhere, For example, the normal dip tube design for reagent injection is disastrous and the mixing and valve resolution requirements are exceptional. I discovered how I could reduce the number of stages of neutralization, offer inexpensive alternatives to the classical neutralization vessel, and decide when signal characterization could help or hurt your control objectives.

In an intense ISA Live Web Seminar on May 16 at 2:00 pm EDT, I am summarizing the best of what I have learned. All registration fees go to ISA, which is the technical society most in tune with the automation engineer in the process industry. The link to the seminar registration page is

ISA Live Web Seminar on pH Control Solutions

The seminar is much more cost effective if the registrant connects in a conference room with a computer projector.

This seminar may help you avoid a tough pH startup.

Top Ten Signs of a Rough pH Startup

10. Food is burning in the operators’ kitchen
9. The only loop mode configured is manual
8. An operator puts his fist through the screen
7. You trip over a pile of used pH electrodes
6. The technicians ask: “what is a positioner?”
5. The technicians stick electrodes up your nose
4. The environmental engineer is wearing a mask
3. The plant manager leaves the country
2. Lawyers pull the plugs on the consoles
1. The president is on the phone holding for you

Packaged equipment is notorious for using the lowest priced instrumentation when customers base their selection solely on lowest bid. If the user was to consider the cost of maintenance and variability, the same type of consistency, intelligence, and performance demanded for their process installations would be required. When I worked on projects for Monsanto’s premier plants in Texas light years ago, bids for packaged equipment had to meet rigorous specifications and an acceptable vendor list for automation. These requirements were eventually made a corporate standard.

The benefit of Hart or Fieldbus smart instrumentation and digital valve positioners is becoming more commonly recognized as important everywhere. Still overlooked for skids is the core performance of the sensor or valve. Diagnostics provide a lot of data but may not give the most important variables of interest to reduce variability (repeatability for measurements and resolution for control valves). Particularly insidious is a common practice of using rotary on-off valves that employ actuator shaft rather than ball or disk position for feedback. In these cases, the digital positioner can be storing and sending gobs of data showing the valves resolution is 0.5% when in fact the resolution in terms of movement of the actual ball or disk is 8% (see April 9 entry “Control Valves that Can Turn on a Dime, or at Least a Quarter - Part 1”). Cost effective solutions exist. For example, a sliding stem valve designed for minimal seating friction and packing friction, with a diaphragm actuator and a smart positioner can provide exceptional resolution. The following test result shows that even operating near the seat a sliding stem control valve with a digital positioner can respond to changes as small as 0.1%. Note that the ISA technical report ISA-75.25.02-2000 (R2006) “Control Valve Response Measurement from Step Inputs” suggests that a resolution of 0.1% to 0.5% is a good valve. If you also consider that data on resolution supplied by valve manufacturers is for a mid throttle position where seating and sealing friction are minimal, the results shown here are particularly noteworthy. Besides reducing the limit cycle in the controller output from the resolution to the noise level in the process, such performance extends the rangeability and sensitivity of the valve, which for pH control can eliminate a stage of neutralization.

Valve with Exceptional Resolution Near Seat

Not commonly recognized is that resolution is more important than repeatability for the control valve, because the process control loop will correct for changes in the magnitude of the valve response. If the valve’s plug, ball, or disk doesn’t move, there is nothing the loop can do. Strangely enough, valve specifications do not normally require the valve actually to move in response to the size of the controller output changes seen during the execution of the loop.

The question for the day is where to locate measurements. My first choice would be a Caribbean island but if the plant is not there, the sensing or sample lines and the associated transportation delays would be quite long. The additional loop dead time would cause all sorts of performance problems that might take years to troubleshoot. I might even have to move to the island.

I have seen sensor locations chosen based purely on cost or ease of accessibility rather than performance. Technicians need to be able to safely remove and replace sensors but many of the accessibility issues can be reduced by better performance and by smart instruments, maintenance software systems, and ultimately wireless communication.

When considering the location in terms of loop performance the objectives are to:

1. Insure a representative measurement point
2. Decrease measurement noise
3. Increase sensor reliability
4. Reduce sensor fouling
5. Improve sensor response time
6. Decrease transportation delay

Maintenance and operations are often aware of items 2 through 4. Item 1 requires some process understanding, item 5 requires some testing or info from the vendor, and item 6 can be simply estimated as the length of the path to the sensor divided by the velocity. For liquid pressure and flow measurements, the delay is negligible except for very long distribution lines because the pertinent velocity is a pressure wave traveling at the speed of sound in the fluid. For composition and temperature measurements the transportation delay can be huge because the pertinent velocity is the fluid velocity (e.g. 0.5 to 5 fps).

For nearly all liquid process measurements, a partially full pipe line is bad news. Air pockets in the meter, sensing lines, and at the sensor can cause failure to meet all 6 of the performance objectives noted. Trapped bubbles can become pockets. Unless there are some extenuating circumstances, sensors should be located upstream of control valves but down stream of pumps to help insure the line is full and to minimize bubbles from flashing vapors. I prefer pump strainers rather than sensors to catch solids and wrenches.

To get the point, increase the signal to noise ratio and reliability, not get all fouled up, and improve the speed for making decisions on how to improve location, let’s consider briefly the measurement of flow and pressure this week. I will hold off on my personal favorites of composition and temperature till next week.

For flow measurement, the next most important consideration, particularly for vortex and differential head meters, is to make sure the velocity profile is uniform. Erratic and changing flow profiles cause poor measurement repeatability and noise. Flow meters must be located upstream of control valves with sufficiently long straight run upstream and downstream. The number of pipe diameters of straight run needed depends upon the meter and the piping details such as elbows, gate valves, and fittings. The number increases for flows at the low end of the meter capacity. A good but old resource for the straight run and instrumentation installation requirements in general is the Manual on Installation of Refinery Instruments and Control Systems - Part I (API RP 550).

Magmeters have a very minimal and Coriolis meters have essentially no straight run requirements although I wouldn’t recommend flanging any meter directly to a valve. A properly selected and installed Coriolis meter provides the most representative, accurate, and noise free flow measurement. It offers true mass flow independent of composition, something you cannot achieve even with even the fanciest pressure and temperature compensation.

The rate of coating build up tends to decreases as the fluid velocity increases. Since many meters have a rangeability based on minimum velocity, sizes that are too large will tend be get more fouled besides erratic at low flows. On the other hand there may be maximum velocity and meter sizes that are too small may have excessive pressure drop or cause excessive erosion when solids are present.

For pressure, the direct close coupled mounting of the sensor to the process eliminates the sensing lines and associated concerns about accumulation of solids and vapor pockets, freezing from inadequate winterization, and vaporization and cooking from excessive heat tracing.

The final point here is that a measurement location chosen purely based on accessibility is a self fulfilling prophecy in that maintenance will need more accessibility because of more performance problems. The more frequent service, removal, and handling of sensors have important safety implications.

I have about 30 minutes here before I head out for the Memorial Day weekend and you probably have less than 5 minutes to read this so it will be short and hopefully sweet.

By now I may have conveyed the sources of transportation delay. For conveyors and paper machines, the delay is the equipment length divided by its speed. For sample and process piping, the delay is the total length of piping including fittings divided by the fluid velocity. For extruders, heat exchangers, and static mixers, the delay is the equipment volume divided by the total volumetric flow. For agitated vessels, the mixing delay is the turnover time (liquid volume divided by the summation of the volumetric feed, recirculation flow, and agitator pumping rate). For dip tubes, the transportation delay is the dip tube volume divided by the volumetric dip tube drain flow. It gets worse. When the control valve closes to the dip tube, the stuff inside the dip tube slowly migrates into the vessel. For sensitive systems such as pH, there can be a noticeable drift for several hours.

Is this the whole story for process dead time?

Based on transportation delay and conventional wisdom, it would be best to locate an electrode or thermowell in the vessel rather than a recirculation line. Is this always the case? The electrode and thermowell time constant can be larger by 20 or more seconds in a vessel because the local fluid velocity is typically less than 0.5 fps compared to the 5 fps in a recirculation line. If the measurement location in the recirculation line is within 20 feet of the vessel, you are only talking about 4 seconds of delay. If you consider the electrode or thermowell tip may be too close to the wall and is much more likely to foul and develop a coating at low velocities, you have the scenario for sensor time constants greater than 60 seconds besides the need for more frequent cleaning. If there are not any safety issues, I prefer the faster and cleaner sensor you get in a recirculation line (e.g. elbow about 20 pipe diameters downstream of pump with sensor tip near the center of the line). For exothermic reactors, temperature sensors inside the reactor are normally required.

It is important to remember the loop dead time is not just process transportation and mixing delays. You need to add sensor time constants, final element dynamics (delays, lags, deadband, and resolution), signal filter time constants, and digital scan and execution times. Now without further delay, have a great holiday.