Why pH Control is a Problem

Systems pH control systems are known by their extreme rangeability and sensitivity, but contact between the measurement electrodes and hostile fluids can cause issues. Case studies of representative installations show that success in implementing these pH controllers depend not only on investigating the complexity of the loop and selecting a control strategy, but also on realizing and avoiding problems while sourcing and installing instrumentation, equipment, and piping.

I have helped to rescue more than 50 pH control systems over the past 5 years and I am still rational enough to tell you about it. Most people don’t realize that I twitch at the mention of hydrogen ion concentration.

Why it's a Problem

Why is pH control a problem? After all, it is a simple scale of measurement from 0 to 14 dimensionless units. Measurement electrodes have been around for long enough to be well understood and easily applied. Instrument vendors must have become acquainted with every possible application of pH meters by now.

The Real World

There are other factors that add to the difficulties of pH control. These range from the requirement of wetting the electrodes - with a constant potential of leakage and attack by the fluid, to long timescales from the need to mix large amounts of process material with small concentrations of reagent. Even with a fundamental understanding of measurement and control concepts, these real-world effects introduce a new element of mystery to pH.

Choose the Right Digital pH Controller

pH Panel Controller

Its small size allows for a simple installation and will accommodate any pH electrode with a BNC connector. Temperature compensation can either be manual or automatic when used with a 1000O platinum RTD.

Multi-Parameter Input Transmitter for pH/ORP

These devices provide a single channel interface for many different parameters including flow, pH/ORP, conductivity/resistivity, salinity and temperature.

Some Typical Problems

There are no typical pH control complications. The systems that are easy to install don't get referred back to those of us whom InTech refers to as the noodnicks from Central Engineering. But the installations I will subsequently detail are typical of those I have encountered recently and demonstrate some of the expected problems.

Where’s the Tank?

An application involved a strong acid waste flow, to be neutralized by a strong basic reagent. It was sent in because the pH was fluctuating from 0 to 14 despite efforts to optimize the controllers, manually manipulate the reagent, and regulate the influent flow. When I arrived at the plant, I looked across the space and didn't see any tanks. I suddenly realized that there was a major problem.

Figure 1a shows the original pH control system. This used a ratio controller to distribute reagent to an acid waste flow upstream of an in-line mixer. A separate pH controller was used in a loop on a sump. The system designers did not understand that the flow measurement error and the flow control valve hysteresis need to be less than 0.00005% to remain within 1 pH of the 7 pH setpoint. The designers assumed that disturbances would be minimal since the change in waste composition was slow and its flow was designated by a controller. The design team did not know Fact #1.

A system containing a strong acid and a strong base usually need three control stages to keep a solution within 1 pH of 7 pH (Ref 1). Since cost was a major factor, I kept the existing mixer and sump as a single stage and added an extra two vertical well-mixed tanks downstream for the second and third stages. I also agreed to not install controls on the third stage until there was requirement to do so. The third stage volume therefore acted as a filter for the oscillation from the second stage.

For the first control stage, we replaced the ratio flow system with a fast inline pH loop. A remote setpoint was received from a second pH controller on the sump. The fast in-line loop would initiate the correction and depend on the sump volume to average out any deviations in the hydrogen ion concentration. Using a linear control system analysis, it was anticipated that this combination would be as effective as an individual well-mixed vertical tank. It didn't work. Dynamic simulations showed that the in-line loop would oscillate at a range between 0 and 14 pH for all controller settings. The result was confirmed by a plant test.

I first thought the sump was not providing the predicted filtering for some reason. Then I remembered Fact #2. The filter was reliant on hydrogen ion concentration, not pH. Concentration oscillations from the sump were attenuating by a factor of 100, but this only affected the pH by 2 units. The attenuation was improved by minimizing the distance between the mixer, control valve and electrodes, so that the oscillation was quicker.

The second state possessed an output in the form of a notch-gain pH controller which provided a pulse frequency proportional to an analog signal. Above 25% controller output, the valve was throttled normally; below 25%, the rangeability of the valve was increased using pulse frequency or interval control.

Figure 1b shows the upgraded installation. This system could keep the pH within the desired offset band at the outlet of the third stage. However, the sump controller was difficult to tune and there was a slow recovery change from startup or waste flow controller setpoint.

Figure 1. Where's the tank? (a)-unsuccessful and (b)-successful pH control systems for a continuous neutralization process initially having no mixing tank.

If I designed this system today, I would put a feedforward loop on the sump and install controls on the third stage. I would also characterize the feedforward and feedback signals. The characterization would involve determining the reagent demand from the pH measurements using a titration curve and programme the control command using the result. This would reduce nonlinearity, recovery time, sensitivity, and tuning difficulty. Microprocessor-based controllers can provide the required calculation accuracy and ease of implementation.

As with any novel system, the startup was not without teething problems. Some were of the common issues - like transposed wires and incorrectly calibrated positioners.

E.g. At high pH levels, the measurement reduced as the strong base reagent flow increased. As you can imagine, this drove us all mad – including the control system. There was difficulty in that the measuring electrodes in the in-line loop were not specified with high-pH glass. This would normally cause the measurement to read under by about 1 pH at the upper end of the scale. In our case, it caused an opposite response. This performance was confirmed by the vendor and was remedied by switching the electrodes with low sodium ion error devices.

Another mystery effect was that the electrode response became erratic for the well mixed tank. It was found that there was water on the terminals inside the submersion assembly. The vendor informed us that the leakage would cease if we bought an assembly that cost double the amount of the existing one. We did; it didn't. The vendor then tried to persuade us to buy a newly developed assembly, at four times the price of the original, and then the leakage would definitely stop. Rather than make the same mistake again, I shopped around and found a throwaway electrode assembly completely protected by plastic - at half the price of the original. It worked fantastic. A similar experience with a submersion assembly from a different vendor led me to Fact #3.

Where's the Valve?

Another application required a small amount of a highly concentrated viscous reagent for continuous neutralization of a waste stream. The control system was so behind speed, that disturbances passed through the plant well before any corrective measures took effect; further, the pH trend recording possessed a noise band that far outstripped the allowable setpoint offset. When I looked over the system, I stood near the injection point at the pipeline mixer inlet, scanned the space and didn't see one reagent control valve. I quickly realized that I had a major problem. Figure 2a shows what I found.

Can you see a pH control issue exclusive of the pH loop in this figure? The sump level controller sets the flow in the upper outlet branch. The valve is simultaneously manipulated in the lower branch by the mixer flow controller to keep a constant flow out of the sump. The system is obviously overcontrolled. We got around this issue by cascading the level controller output to flow controller setpoint.

Onto the pH loop. The reagent was being fed into the pipeline through a positive displacement metering pump. The pump was about 300 feet away from the mixer. When the pump was activated, this distance caused a delay because the process fluid would backfill into the injection tubing and had to be forced out of the line before any reagent could be administered. It doesn't take much complex mathematics to work out that it takes an hour to push a gallon through a pipe when the speed is one gallon per hour. This led to Fact #4. We also encountered a delay when the pump speed was changed, but the cause was never isolated. We would have attributed it to air pockets, if there were any. The answer probably lies in the ketchup bottle - related to low flow of viscous fluids.

Moving on, we reduced the delays and noise band by an order of magnitude when the remote metering pump was replaced with a close-coupled control valve. The valve was operated using a wireless pH controller to control the reagent flow to the sump discharge flow, correcting the ratio using the in-line pH loop.

Some noise was still present due to a poor distribution of the injected reagent in the pipeline. This was impossible to eliminate, as it required making the injection port smaller to enable the reagent velocity to be increased. However, a hole small enough to do the job was too small to prevent blockage. The noise was more of an inconvenience on the trend chart than in the system itself, so the record was tidied up by filtering the measurement signal through an electronic filter.

Just when we thought our problems were finally over, another mystery reared its ugly head. When the miniature reagent valve was changed from closed to open, the reagent flow measurement spontaneously increased and then changed to zero. The magnetic flowmeter was the immediate suspect - but came through unscathed; we checked the wiring and found it to be the issue; the vendor inspected and confirmed the integrity of the electronics; we tested the meter on water and found that it responded correctly. We then tried changing the valve trim, but several tests reached the same conclusion.

I was on the verge of discarding the tiny but costly trims away, leave the engineering profession, and enter a seminary. During this time of contemplated thought, I suddenly spotted something that looked to be a reverse taper on the trims. It was hard to know for certain, because the parts were small, but I confirmed the observation with a micrometer. In a desperate attempt to get away from this startup, I calculated the contour of the plug for a linear characteristic, drew a sketch, and had the parts made up.

The valve worked perfectly with the homemade trim. The reverse taper had made the flow decrease as the stroke increased. The momentary surge inflow at the start of the stroke was caused by the plug lifting off the seat as to provide a large enough annular clearance. How did the reverse taper get there in the first place? I never found out the reason, but I did learn that the trims were too small to be of standard size and were specially produced by the vendor for the order. As far as I was concerned, they were too special. Without a reagent flow meter, you can probably imagine how difficult it would have been to diagnose this valve problem. This leads to Fact of Life #5.

There was another instrumentation issue later on, when one of the design engineers decided to modify the system and recover some panel space. In place of the ratio station and pH-based flow controller, he decided to install a feedforward controller. The device added a flow feedforward signal to the flow command from the pH controller. The vendor, wanting to sell a feedforward element, thought that this was a great idea. In operation, as you will have probably guessed, the flow controller adjusted its output to negate the effect of the feedforward signal and maintain a flow designated by its setpoint. To work as anticipated, the feedforward action would have to be located on the flow controller setpoint - multiplied by and not summed with the pH controller output. Multiplication will force the reagent flow to zero if there is zero process fluid flow, or the fluid is at the designated setpoint. Also, for you control jocks out there, multiplication cancels the composition loop gain - a term inversely proportional to sump flow. This leads to Fact #6.

All of these corrections are reflected in Figure 2b above. The system, as shown, has been controlling well since startup.

Where's the Agitator?

A process incorporated a vertical tank for neutralization. It exhibited a poor performance because the response was slow and the effluent was not efficiently mixed. I viewed the drawings and noticed that the vertical unit was too tall for its diameter. I enquired as to how high it was, and the designer replied, "50 feet," I gasped, "It's not nice to kid an old engineer." He responded, "Who's kidding?" He replied, "You're the only agitator on this project." I instantly knew I had a major issue.

Figure 3. Where's the agitator? (a)-unsuccessful and (b)-successful pH control systems for a process involving an extremely tall mixing tank without an agitator

Figure 3a shows how the pH was initially being controlled. The difficulties could have probably been resolved with some axial agitation, but could not be provided from an economic standpoint because the tank was too tall. A shorter tank would have resolved the issue - again at a higher price than the plant wanted to shell out for. I decided that the best method to deal with the tank would be to use its volume as a filter, deducing that it would attenuate the hydrogen ion concentration oscillations of an in-line loop by a factor of 10,000 - 4 pH units. We installed a circulation pump to act as a low-deadtime in-line mixer. The Influent and reagent were added to the new suction and an injector probe was installed on the pump discharge. The new system is shown in Figure 3b.

Issues still occurred, mainly attributed to the quick opening characteristic and plug valve’s large positioner hysteresis on the influent. However, the in-line pH controller loop quickly reverted to its setpoint after a disturbance. Furthermore, after flowing through the tank volume, the pH created the straightest line I have ever seen; for a moment, we thought somebody had tied the pointer down. The performance was so high that the plant mentioned that we should standardize this type of system for pH control throughout. I cautioned them that the setpoint of this system was quite a few pH units below the neutral zone, on a fairly flat position of the titration curve. On a steep part of the curve, Fact #2 would prevail and there would be a large number of oscillations.

Figure 4. Where's the electrode? (a)-unsuccessful and (b)-successful pH control systems for a process in which electrodes have to be installed in inconvenient locations.

Where's the Electrode?

I was asked to troubleshoot the pH control system shown in Figure 4a. This simple configuration should have worked with fail, but it was plagued by an unusually wide control band around the setpoint. I investigated the exit nozzle of the vessel and couldn't locate the pH electrodes. I quickly realized that I had a major problem on my hands. In this case, the source of the difficulty was political. The maintenance department in charge of this instrument had specified that the electrodes be placed in the analyzer house, to avoid servicing them outside during the winter months. Unfortunately, this locale introduced a major deadtime in the loop. To aid in the avoidance of this problem in the future, I feel compelled to state Fact #7.

I managed to convince them that the electrodes needed moving by arguing about the extreme safety hazards and quality control problems that accompanied large pH excursions. The change, indicated in Figure 4b, minimized the control band to around 0.1 pH.

We opted for electrodes in this application. Experience has shown that these provide a greater performance and require a lower amount of maintenance than sample chamber electrode holders. These benefits are particularly obvious when the electrodes are positioned in the discharge nozzle piping where there is a high fluid velocity – because the high flow guarantees a swift response by reducing the boundary layer thickness and preventing the electrode from being coated by impurities in the stream.

Compared to sample chamber elements, injection electrodes seem to be less prone to leakage. By inspecting 30 installations of injection devices from one manufacturer, I found no occurrence of leakage; in fairness, when we received products from an alternative source, there was some leakage. However, every sample chamber electrode holder I have ever come across has eventually leaked. Also, unlike sample chambers, leakage is visible with injector assemblies. For dangerous fluids, you don't want any nasty surprises when you open the top cover of the electrode holder. This leads me to Fact #8.

Is the Size of the Tank Important?

A plant used the system depicted Figure 5a for waste a neutralization process. The eductor shown in the figure was factored in because the mixing deadtime was too long. But even with this device, the deadtime was still over 40 minutes. The subsequent natural period of the pH loop was 160 minutes, meaning that the maximum reset should have been fewer than 0.01 repeats per minute. As this was below the controllers lowest setting, the loop adopted a continuous reset cycle; furthermore, the integrated error - which is proportional to the square of the deadtime - was way out. I gazed at the engineering flow diagram and saw the largest storage tank that I had ever seen. I enquired to the process engineer where the neutralization tank was located, and he pointed me to the elephant I just thought was for storage. I immediately understood that there was major concerns.

The intent of the large tank was feasible. It would serve to mix acidic and basic waste streams from multiple sources and minimize the reagent demand. So long as there is no requirement to put control loops on them, large tanks are useful. A large tank can filter out disturbances and reduce the reagent requirements when positioned upstream of control loop. Downstream, it can filter out loop oscillations - which is advantageous because these fluctuations are normally quicker than the variations in influent concentration. Therefore, they are more effectively attenuated. This reminds me of Fact #9.

Figure 5. Is bigger better? (a)-unsuccessful and (b)-successful pH control systems for a process in which an extremely large tank was initially employed for mixing.

The new control system is shown in Figure 5b. The large tank was switched with two smaller tanks positioned in series. A pulse frequency controller was implemented to avoid plugging the valve at low reagent flows and to meet the large rangeability requirements imposed by the wide variations in influent flow and pH. To counteract the steep slope of the titration curve at the setpoint, signal characterization was employed.

Startups are no fun without some mystery. In this case, we observed that there was erratic pH measurement on the first tank. When we removed the electrodes and inserted them directly into the buffer solution or connected them to measurement system located at the second tank, the problem was not duplicated. We individually replaced the pH transmitter, preamplifier, cable, and electrodes but the erratic behavior continued. Eventually, someone remembered that field maintenance department had replaced the fiberglass preamplifier enclosure with metal housing. The enclosure mounting plate was grounded, thus creating a second ground point in the circuit, causing a large current to flow through the circuit. The problem was not present in the second tank as the preamplifier housing was not mounted on a conductive structure. Similarly, the erratic behavior was not observed during the buffering process because the bottle was made of plastic. We solved the problem by isolating the preamplifier enclosure from ground with a plastic mounting plate.

Other than some periodic pluggage of the electrodes in the overflow sample line, the control system has performed well from startup. The liquid head was too low to achieve a sufficient velocity to clean the electrodes. A new electrode holder with a large flat electrode surface will be trialed. If that doesn't work, we may have to find enough money to install a sample pump and an injector electrode assembly.

This information has been sourced, reviewed and adapted from materials provided by OMEGA Engineering Ltd.

For more information on this source, please visit OMEGA Engineering Ltd.