Treating corrosion as a process variable

Corrosion, when examined as a process variable, can be both monitored and controlled in real time.

By Michael McElroy and Kristen Barbour, Pepperl+Fuchs, Inc. June 20, 2011

There are changes constantly occurring within any process system, both known and unknown. These changes can affect the corrosiveness of a fluid on the pipe wall surface. If these changes are known as they occur, then treatment methods or processing techniques can be designed to minimize the corrosive effect, minimize maintenance costs, and significantly reduce the potential for lost product and environmental damage. The difficulty has always been the inability to see these physical changes in real time so that process changes or preplanned maintenance can be effective.

Corrosion, when examined as a process variable, can be both monitored and controlled in real time. There are some techniques for monitoring corrosion, such as mass loss coupon analysis and smart-pigging, that will give you a measurement of how much material has been lost, but they only give you this information after the fact. When monitoring corrosion in a real-time environment, upsets are picked up instantly versus months later, thus reducing the over-maintenance cost as well as avoiding potential lost product. By reading back corrosion information in real time, the implementation of a control strategy becomes possible, allowing you to control your corrosion as it is happening by optimizing the use of neutralizing agents, for example, inhibitors.

Introduction

Monitoring for corrosion in the process industry is largely accomplished by using offline methods, such as mass loss coupons and smart-pigging. Analysis of the data collected from these methods provides the user with historical insight into what has already happened.

These techniques provide effective and accurate ways of determining how much corrosion has occurred over the sample period; however, they do not provide information as to when corrosion rates were high, low, or varying between the two. Most importantly, these techniques do not correlate a change in corrosion due to a changing process parameter.

Traditional corrosion monitoring techniques

There are four traditional corrosion monitoring techniques: mass loss measurements, ultrasonic examinations, resistance measurements, and polarization resistance determination. Mass loss, commonly referred to as coupons, are sacrificial pieces of metal that are inserted into the process. The coupons are made of the same properties as the pipe or vessel being monitored. Coupons are weighed prior to insertion and then weighed again after extraction.

Typically, coupons are supposed to be analyzed after a period of 90 days, although this does not always happen. Coupons are non-electrical and do not provide any means to collect data while in the pipe. Therefore, the corrosion rates provided by the analysis are purely historical and only represent the average corrosion rate.

Ultrasonic examinations use non-audible acoustic waves to determine metal wall thickness for pipes and vessels. By measuring how much time it takes for the signal to travel from the device to the pipe wall and then back to the device, a measurement of mass loss (corrosion) can be made. This measurement must always be in reference to a prior data point, which is initially done after the pipe is installed.

An example of a device that can use this technology is called a Smart-Pig. Often referred to as “pigging,” these robots travel through the pipe taking measurements at every desired increment.

Resistance measurements are similar to coupons in regard to interpretation. With resistance measurements, a wire or probe elements are exposed to the process. They start with a known resistance, and as they corrode, the resistance will increase. The change in resistance is measured and then interpreted as how much metal of the pipe is remaining. Electrical resistance probes rely on thin electrodes on the probe tips to increase the sensitivity of the technique.

As a result, the probe elements often have a very short life. This type of probe is not suitable for aqueous solutions as the result of a pitting attack could destroy the measuring element.

Linear polarization resistance measures the electrical resistance between the solution and the metal. To measure the electrical resistance, there must be at least two electrodes. When a piece of metal corrodes, oxidation occurs, which means that electrons are released by the corroding metal. By applying a controlled voltage, a measurable current will flow between the two electrodes.

This current is approximately linearly related to the potential difference (hence the term “linear polarization”). Through algorithms, these measurements become proportional to the corrosion current, and thus, a rate of general corrosion can be determined. This technique is used in an online environment where general corrosion rates can be determined continuously.

A deeper look

All of the techniques listed above have been successfully used in real-life applications. The question would then become “Why change what works?” Consider the scenario in Figure 1.

These data points were collected on a periodic basis, potentially by using mass loss coupons, ER probe, or pigging. Typically what happens is a decision is made on how much corrosion inhibitor is to be used based on the current data point with respect to the prior point. The goal is to find the acceptable balance of the rate of corrosion versus the amount of inhibitor used. Not only is this process time consuming, potentially taking a year, but what is happening in between the data points?

Figure 2 could represent the reality of the process. The cause for these swings could be virtually anything from changes in temperature, pressure, or flow to impurities in the product or an undetected upset. If Figure 2 is reality and corrosion was monitored per Figure 1, when the metal loss is above the data point, not enough inhibitor is being used, and when it is below the data point, inhibitor is being wasted.

On the other hand, if corrosion was monitored continuously like all other process variables, the corrosion engineer would have a wealth of data available. From this data, the correlation between process events and the rate of corrosion can be made. This gives the corrosion engineer the ability to inject the exact amount of inhibitor needed to keep the rate of corrosion at a defined value.

Real-time corrosion monitoring experiences

Linear polarization resistance technology was first applied in cooling water applications about four decades ago. For most of the ensuing years, it was used as a stand-alone monitoring technology particularly in the refining and petrochemical industries. As such, it was a very useful tool in detecting system upsets and pinpointing transient corrosion problems.

That was at a time when the HPI industry had dedicated corrosion specialists and supplier water treatment personnel on-site to gather and analyze the corrosion data. These data were keys to pinpointing transient corrosion problems. For example, LPR-based corrosion monitors are the only means by which one may detect the increased corrosion rate resulting from the overfeed of an oxidizing microbiocide, which may last only for a few minutes yet may cause the stripping of a corrosion inhibiting film from a heat transfer surface. It is a good way of detecting the loss of feed of a corrosion inhibitor before too much damage may occur.

There were a number of attempts to directly control the feed of water treatment chemistries to inhibit corrosion directly from LPR-based corrosion monitors. None of them were successful, for two reasons:

  • There is no direct relationship between the corrosion rate and the amount of corrosion inhibitor required in the system, and
  • There are too many other variables that affect the corrosion rate. Other variables must be controlled by other means.

Over the past 20 years, monitoring and control technology has rapidly advanced. Today, LPR-based corrosion monitors have been vastly improved and are capable of providing outputs that are being used as primary inputs for control logic. The chemical treatment of open-recirculating cooling water systems, the ones with cooling towers, is normally based upon feedback control of pH by the feed of an acid or an alkali and concentration of dissolved solids in the circulating water via conductivity by opening and closing or modulating a blowdown valve and control of chemical treatment materials by various means.

Systems that utilize oxidizing microbiocides may use an ORP controller to maintain the Red-Ox potential of the circulating water. To this combination, at least one water treatment service company has added both LPR-based corrosion monitoring and fouling monitoring as feed-forward inputs to the control system. Such systems are designated to be performance-based because they are actually controlling the efficiency of the cooling water system to reject heat and protecting the life of the system by minimizing corrosion.

One example of an LPR measurement used for control was in the cooling water system of a methanol unit in a petrochemical complex. This unit had experienced severe localized corrosion (pitting), while general corrosion remained within acceptable norms. Similar results appeared on both corrosion coupons and LPR electrodes.

At the same time, the ORP monitor detected an environment that was very conducive to the growth of microorganisms. The problem was traced to periodic and unpredictable process contamination. When the ORP was raised by increasing the feed of an oxidizing microbiocide, the localized corrosion rate decreased and at one point came to a near halt. Control logic was used to detect the localized corrosion with LPR and adjust the oxidizing microbiocide feed rate by monitoring ORP. The result was consistent good control and elimination of the pitting problem.

An example of performance-based control application was in a production unit of a large chemical complex in the southeastern U.S. Previously, the cooling system was controlled with a traditional pH/conductivity controller with separate chemical feed pumps feeding each of the treatment materials. There was no online corrosion monitoring. Only corrosion coupons were used.

Due to the production process, there was significant variability in the heat load to be rejected by the cooling water system. That variability affected the temperature of the circulating water, the rate at which heat was rejected in the cooling tower, and the corrosivity of the water. That in turn varied the demand for each of the treatment materials. The control system was inadequate for the application. In fact, the system suffered from excessive corrosion and fouling.

It was then replaced with a performance-based system that utilized the monitoring of five variables: pH, conductivity, ORP, corrosion (LPR), and fouling. In addition, the treatment program was changed by the new supplier. Following an initial cleanup and repassivation period, corrosion rates settled down well within the range of industry norms. General corrosion dropped to below 2 MPY on carbon steel and localized corrosion (pitting) decreased dramatically. Fouling also decreased to well below heat exchanger design allowances. The control system automatically adjusted the treatment chemistry to swings in heat loads.

Conclusion

Corrosion monitoring is slowly undergoing a change from manual methods to online methods. This could be compared to the changes that started about 25 years ago in pH monitoring, which went from manual to online with the development of new electrode technology and new microprocessor-based transmitters. This article described different types of corrosion and corrosion monitoring techniques with emphasis on online monitoring.

Traditional corrosion monitoring techniques provide proven ways to determine corrosion rates. Determining corrosion rates by only using these techniques does not allow processes to be controlled as a process variable. This information is static and cannot be correlated to events occurring in the process.

General corrosion or localized corrosion can be monitored online and in real time without additional hardware and software. The addition of HART protocol enhances the functionality of this standard 4-20mA signal. As a result of this innovation, corrosion can be monitored and controlled like other process variables, such as temperature, pressure, and flow. For the first time, corrosion monitoring can be achieved and can begin to control the huge annual cost associated with corrosion. Corrosion monitoring has now entered the world of process control and automation.

Michael McElroy is the business development manager for CorrTran and Kristen Barbour is product marketing manager for Pepperl+Fuchs, Inc., Twinsburg, Ohio. Their website is www.pepperl-fuchs.us.