Maintenance

Insidious corrosion of fixed equipment detected via predictive maintenance

Pipes and vessels monitored to detect and mitigate corrosion using wireless instrumentation.
By Jake Davies, Ph.D. February 4, 2020
Courtesy: Emerson Automation Solutions

Discussions of predictive maintenance in the oil & gas industry usually focus on rotating equipment such as pumps and turbines. Those are certainly valid areas of concern, but the result of a failure is usually limited to the equipment itself. On the other hand, static equipment such as piping, vessels and similar equipment is not as maintenance intensive, but a failure can be catastrophic. This equipment should also receive predictive maintenance attention.

Major loss of hydrocarbon containment can result in fatalities while damaging equipment, the environment and a company’s reputation. The economic impact will often be felt long after the initial clean-up and repair is completed. Two well-documented events tell the tale: BP Cherry Point in February of 2012 and Chevron Richmond in August of 2012 both experienced containment losses caused by corrosion damage from inside piping, and major fires ensued.

The threat of internal corrosion and erosion exists throughout the hydrocarbon production chain, from the wellhead, through midstream, to refinery and distribution. To avoid unplanned outages or disaster, equipment must be repaired or replaced before the metal thickness reaches a critical minimum limit. But how can operators determine when that point is approaching?

Monitoring metal loss

Traditional approaches to this challenge send technicians into the plant to take manual thickness measurements. Since sections can’t be cut open, the most common method is ultrasonic thickness measurement. This technique is non-intrusive, so it can be performed with the plant in operation, with no effect on the process or safety risk.

Naturally, manual inspections incur costs for the technician to gain access to the desired measurement location, which may involve erecting scaffolding, removing insulation and so forth. Despite these costs, a refinery will typically have several thousand locations scheduled for inspection at periodic intervals ranging from every few weeks in high-risk locations to once every five years in other less-critical areas. These inspections produce volumes of data, manually entered into inspection management software for analysis.

Figure 1: Manual ultrasonic thickness measurements from a single point over 30-year period illustrate the type of variability often found when monitoring corrosion in piping and vessels. Courtesy: Chevron

Figure 1: Manual ultrasonic thickness measurements from a single point over 30-year period illustrate the type of variability often found when monitoring corrosion in piping and vessels. Courtesy: Chevron

Even with thousands of measurements generated each year, this inspection data is inadequate to understand plant health in real time. Plant personnel can’t see how the plant is coping with the ever-changing corrosion and erosion sources or use data to predict when metal thickness will reach its retirement point.

Why is manual measurement so ineffective? Most consecutive periodic measurements are performed in slightly different locations on the pipe, by different technicians, often armed with different inspection equipment. Such variances in measurements (See figure 1) add up to data noise, rendering the information largely useless. Variability of ±1 mm (.040 in.) is expected, but if a pipe wall is 5 mm (.200 in.) thick, the engineer evaluating the data will lack confidence when trying to determine when that pipe will reach its retirement thickness.

Predicting damage rates is challenging, especially in areas where the corrosivity or erosivity of the process fluid varies frequently. Nowhere in the production chain is this experienced more than refineries, which therefore have the highest variability in corrosion and erosion load. Traditional inspection methods just discussed simply don’t provide adequate quality or sufficient measurement frequency to drive predictive maintenance able to keep equipment running safely.

From periodic to continuous monitoring

Permanently installed, ultrasonic wall thickness monitoring sensors (See figure 2) on a wireless network are meant to work in harsh environments. The installation cost of ultrasonic sensors is low because they are non-intrusive and can therefore be mounted just about anywhere. Data retrieval via a new or existing WirelessHART network reduces installation and operating costs. Internal power modules last up to nine years, so no maintenance is required between turnarounds.

Figure 2: Ultrasonic corrosion sensors are permanently installed on this crude tower overhead line. Insulation is subsequently replaced around the sensor legs. Courtesy: Emerson Automation Solutions

Figure 2: Ultrasonic corrosion sensors are permanently installed on this crude tower overhead line. Insulation is subsequently replaced around the sensor legs. Courtesy: Emerson Automation Solutions

If a WirelessHART gateway needs to be added, it is usually installed near the process unit. It collects data from all nearby WirelessHART transmitters and sends it via a wired backhaul to the plant’s operations and information technology systems. Software processes the corrosion data, stores it for historical analysis and makes it available for viewing. Advanced processing software can compare previously recorded ultrasonic waveforms with new data to improve measurement resilience when the internal meta-surface morphology is very rough. This problem is one of the main causes of variability among manual ultrasonic wall thickness measurements.

Figure 3: Corrosion sensors helped identify the optimal dosage for corrosion inhibitors. Courtesy: Emerson Automation Solutions

Figure 3: Corrosion sensors helped identify the optimal dosage for corrosion inhibitors. Courtesy: Emerson Automation Solutions

Emerson’s Rosemount corrosion monitoring systems (See figure 3), formerly known as Permasense corrosion systems, which use the adaptive cross correlation processing technique, further enhance measurement analyses, so even small levels of corrosion or erosion are detected in a matter of days. The advanced signal processing method used by this software, along with its data visualization and analysis features, makes data interpretation significantly easier and quicker. Data from multiple sensors can be visualized on a single screen as a corrosion heatmap.

Three brief case studies

Here are three brief examples that illustrate best practices for corrosion monitoring:

Optimize corrosion inhibitor use: Real-time corrosion data from permanently mounted ultrasonic thickness sensors can help evaluate the effectiveness of a chemical treatment program. One European refiner used a network of wireless ultrasonic sensors installed across its crude overhead system to optimize the dosage of corrosion inhibitor additives. Prior to optimization, corrosion rates of up to 1.2 mm per year (.050 in.) were observed. Over a month-long period, the refiner increased the additive dosage in steps and tracked the effect. Once optimized, sensor data (see figure 4) showed the corrosion trend had stabilized, negating the need to replace the piping.

Figure 4: Data projected forward helps determine when piping will need to be replaced. Courtesy: Emerson Automation Solutions

Figure 4: Data projected forward helps determine when piping will need to be replaced. Courtesy: Emerson Automation Solutions

Don’t overestimate a problem: Preventive maintenance can identify what equipment does not need attention, which can be a major benefit. A deep-water production platform in the Gulf of Mexico deployed ultrasonic thickness monitoring sensors across 20 high-risk locations. Operators used data to verify a lack of corrosion activity, enabling continued safe production. Operators and engineers were confident that the current operating conditions were not adversely damaging the equipment on the platform, alleviating concerns regarding asset availability.

Look ahead to the next turnaround: Data from continuous corrosion sensors can help predict the timing of equipment replacement very accurately. One Middle Eastern oil producer examined data (See figure 5) during the preceding year from a single sensor, which showed a highly variable corrosion rate in an amine unit. It helped identify the root causes of the corrosion variability, while predicting when the vessel would need replacement. Applying lines-of-best-fit to the data (Figure 6) enabled projections regarding minimum allowable thickness.

Figure 5: Continuous corrosion monitoring data from a single corrosion sensor installed on an amine unit vessel. Courtesy: Emerson Automation Solutions

Figure 5: Continuous corrosion monitoring data from a single corrosion sensor installed on an amine unit vessel. Courtesy: Emerson Automation Solutions

The data showed that, if the plant continued to operate in the same way as the previous month, the unit would need to replace this piping at the end of 2019. If the plant operated on average as it had for the past 12 months, then the piping would need replacement six months later. The two projections gave confidence to the date range when the equipment would need replacing. This unit is due for a planned shutdown in Summer 2019 and the monitoring data shows that the operator can continue to operate safely under historic operating conditions until then. The operator can include replacement equipment in the scope of the shutdown, pre-ordering the replacement piping and vessel.

Final words

Successful predictive maintenance for fixed equipment in refineries is essential for plant safety, availability and profitability. One of the biggest threats to the health of oil & gas assets comes from the inside in the form of corrosion and erosion. If left unchecked, these hidden threats eat away metal until leaks form, releasing hydrocarbons. Resulting disasters can cause fatalities, environmental impacts and financial losses.

Fixed equipment repairs or, more frequently, replacements are major capital investments and invariably involve shutting down parts of the plant. Actions need to be prioritized, planned and executed to minimize production impact along with immediate and opportunity costs. Therefore, management of equipment health, including its replacement, requires the best available data.

Figure 6: Data projected forward helps determine when piping will need to be replaced. Courtesy: Emerson Automation Solutions

Figure 6: Data projected forward helps determine when piping will need to be replaced. Courtesy: Emerson Automation Solutions

Continuous corrosion and erosion monitoring systems, based on ultrasonic thickness monitoring sensors, deliver this required data and enable operators to better understand the following:

  • Current equipment health
  • How the plant is coping with corrosion and erosion demands
  • When equipment will need replacing.

Operators with access to data of this quality and frequency can operate safer plants with enhanced availability, reduced maintenance costs and increased profitability.


Jake Davies, Ph.D.
Author Bio: Jake Davies is global marketing director for Rosemount corrosion monitoring systems at Emerson Automation Solutions, formerly known as Permasense Ltd. Jake holds a PhD from Imperial College London, an MEng from Oxford University and an MBA from The Open University.