Storage Tank – Thermal Image Chronograph

NOTE: Each image is adjusted to provide the best contrast for visual representation of issues being observed. Temperature scales will vary for this reason.

Synopsis

This case study provides a chronological review of the issues found within a customer’s storage tank. The first signs of possible issues within the storage tank were found in thermal images of the north side of the tank (Figure 1). The bean line can be seen just above the warm spot.

Thermal images of north side of tank. Both images show the same spot from different angles.

Time Context

Discovery: Wednesday, July 6, 2011
Correction: Saturday, July 30, 2011

Summary of Action

The images in Figure 1 prompted the need for additional imaging within the tank. Images taken from removed air vents on the north side were the first images captured within the storage tank and indicated temperatures in excess of 300°F in the roof support system (Figure 2).

Thermal images of storage tank interior.

Two spots indicating a thermal anomaly were identified on the beans, but later could not be located (Figure 3).

Thermal anomaly in beans.

It was believed that debris had fallen from the roof support and self-extinguished upon contact with the beans.
Images taken on July 7, 2011 from the man access hatch confirmed many of the prior areas seen on July 7, 2000. Several new areas were found, indicating that the issue had spread throughout the tank roof support system (Figure 4).

Areas of concern in storage tank.

Temperatures identified were in excess of 500°F.
In addition, a new thermal anomaly was found on the southwest portion of the roof (Figure 5).

Thermal anomaly in southwest portion of roof.

It was determined that this was above one of the roof supports identified from internal scans. Future scans failed to re-identify this anomaly, indicating that it had moved away from the roof.

Thermal image of beans in storage tank.

Figure 6 is the first image that showed an issue had developed within the beans. The area of increased temperature is clearly seen to be sitting on or near the surface of the beans. Later images showed the progression of this area down into the beans.
Taken one day after discovery, Figure 7 showed that the temperature had increased from 91°F to 173°F.

Changes in thermal anomaly one day after discovery.

Additionally, the area was sitting lower within the beans. As shown in Figure 8, the growth of the increased thermal activity indicated that a much larger area within the tank was involved. This only represented the area directly in contact with the wall of the tank. At this point, nothing was known as to the extent of any issues beyond this area.

Growth of increased thermal activity.

On July 15, the temperature readings on the exterior of the tank were maintaining themselves in the 130°F range (Figure 9).

Temperature readings of exterior of storage tank.

The total area had increased, but appeared to have ceased the initial rapid growth seen in the first few days.
Over the next several days, the tank was filled with nitrogen in an attempt to extinguish any issues within the tank. The large spot on the northeast side of the tank continued to present itself with varying temperatures and sizes, but nothing larger or higher than had been previously seen. The original small spot on the north side also continued to present itself, but there was no increase in temperature or size. For this reason, it was uncertain if this was related to issues within the tank or due to the tank itself.

Internal thermal video screen capture of beans and storage tank.

Figure 10 is a screen capture from an internal thermal video provided by the Gainesville Fire Department that shows a long trail of actively smoldering beans. The upper left portion of this still image is the area just inside the large hot area seen on the tank exterior. The estimated length was over 60 feet.

Storage tank with 500+ gallons of water added.

The screen capture in Figure 11 shows the area adjacent to the tank wall after showering over 500 gallons of water onto the area. While this image appears to indicate a positive result, images taken of the exterior indicated otherwise:
1. Figure 12 was taken at 7:30AM, prior to any water being induced into the tank. The maximum temperature at this time was 154.9°F.

Storage tank before adding water, 7:30AM.

2. Figure 13 was taken at 10:40AM, prior to any water being induced into the tank. The maximum temperature at this time was 176.1°F. This was a normal increase due to solar gain throughout the day.

Storage tank before adding water, 10:40AM.

3. Taken at 11:01AM, 20 minutes into applying water, Figure 14 shows that the exterior temperature had increased almost 30 degrees. Additionally, the area displaying intense heat had greatly increased. The dark blue area to the upper left of the spot shows the water contacting the inside of the tank.

Storage tank after adding water, 11:01AM.

4. By 11:13AM, over 500 gallons of water had been applied onto the beans. The external temperature had increased to 202.9°F (Figure 15). The dark blue area shows where water was applied and indicates that it was not penetrating the beans as expected.
It was decided to end applying water due to the lack of positive results.

Storage tank after adding water, 11:13AM.

One day after applying water, the temperature settled in the mid 120°F range (Figure 16). The larger light blue area was believed to be water surrounding the area and absorbing temperature from the hot beans.

Thermal image taken one day after adding water to the storage tank.

As originally seen, the small spot on the north side of the tank was still present (Figure 17). At this time, it was believed to be an indication of the tail end of the smoldering ribbon seen in the firefighters’ video.

Thermal anomaly on north side of tank after adding water.

Plan of Action

Upon reviewing the application of water to the beans and the tank, it was decided to access and remove the beans from the side of the tank. This task involved cutting a minimum of two holes into the side of Tank 100. The first hole was to be directly over the observed hot spot with the second hole at a point that would allow sufficient draining of the beans. The ability to use infrared imaging ensured accurate placement of both holes and the success of this process.
Figure 18 was taken prior to any holes being cut into the tank. This image shows the extent to which the water had saturated into the beans, as well as the level of the beans. Figure 19 shows issues directly beneath the first hole cut into the tank, while Figure 20 shows greater issues than previously indicated.

Thermal activity prior to cutting holes in tank wall.

Screen shot from the firefighters’ video, taken below first hole.

Based on the information from the first hole, the second hole was determined to be necessary. This hole was to be cut 10 feet to the right of and 10 feet down from the hot spot on the northeast side of the tank. The hole was to be 24 inches in diameter.
Figures 21-30 are a few selected still images and video screen captures taken throughout the day.

Working on storage tank.

22 through 27 were taken during the bean extraction process.

This screen capture shows one of the final areas inside the tank.

This screen capture shows the same area from Figure 28 as it was being dragged out of the tank. Future imaging failed to find this area.

This screen capture shows the area inside the tank where the hot spot existed. The arrow shows a faint remnant of the previous issue. This small area was removed from the tank.

Supporting Data

A total of 201 still images and 86 videos were recorded. A portion of the images and videos were provided by the Gainesville Fire Department.
All on-site images were captured with a FLIR P660 (s/n 404000174), calibration date December 22, 2010.

Resolution

Thermal imaging was conducted of the internal area of the tank two days after all issues had been removed from Storage Tank 100.
No issues were identified in either the roof supports or the beans.

Allied Reliability is pleased to announce that we have been named a recipient of the Thomas M. Drury Award for Achieving a perfect incident record in a high-exposure industry for the second year in a row.  Each year, the Thomas M. Drury award for Contractor Safety is presented at the South Central Joint Mine Safety and Health Conference (MSHA) to contractors that have shown exceptional safety leadership and performance while working within operations.

Allied Team Members from left to right: Cary Weaver- IR Analyst, Chris Ozanich- IR Analyst, Christopher Rohde- Program Manager, Steve Toomey- SME, Leonard Bonderenko- Vibration Analyst

Reliably Safe

As a company that sends Full Time Equivalent, part time and mentor consultants to client sites around the country it is imperative that our employees conduct their duties in the safest way possible not only to uphold the stringent safety expectations of Allied Reliability but so as to never put one of our clients in danger or compromise the safety policies of our clients.  This two- fold level of expectations could be a difficult promise to live up to. But as an company that holds our performance to the highest standards possible we expect our team to go beyond compliance to exceed safety regulations and become an example for all of their on the job colleagues as well as an on-site resource for best safety practices.

Keeping up
In order to make sure Allied Reliability Team members can act as safety benchmarks within their environment we make a point to consistently provide them with important safety tips and reminders.  As a company we share weekly ‘Safety Tool Box Talks’ with our team in the field for them to share with their team.  Each week our field technicians and analysts review a useful single point lesson that can be shared and discussed in a 5-10 minute time frame.  Each tip is centered around safety, whether it be a Tool Box Talk on ladder safety or lacerations, each week we share something new for our team to reflect on.  This type of continuous improvement and learning is what sets Allied Reliability apart from our competition and why we provide Inspired Reliability.

By Stacy Heston, CLS, PMP, CMRP, OMA, Lubrication Subject Matter Expert, Allied Reliability Inc.

Lubrication practices are often overlooked as a potential time and money saver for a plant.  However Cargill Charlotte has found that by optimizing their lubrication task intervals, utilizing specific grease volumes, and implementing an oil analysis program, the overall reliability of their manufacturing process, as well as, the healthy status of their lubricated components has increased.
In 2009, Allied Reliability performed a Reliability Centered Lubrication Current State Analysis at the site to determine the state of the lubrication program.  Through the assignment of a score to different lubrication activity and practices, the site’s lubrication program was benchmarked on several parameters including program documentation, management structure and practices, storage and handling, etc.  By assigning a score, the benchmarking activity becomes more objective and less subjective allowing for a definitive path forward.
When Allied Reliability arrived on site, “the site had run amuck with lubricants everywhere, it was all about the flavor of the month…” where lubrication was concerned, and “no structure existed for the scheduling of lubrication tasks,” says Andy Nolan.  As a result of the Current State Analysis, it was determined that a RCL walk down was needed to bring the site up to speed with industry best practices for lubrication.
Shortly after the Current State Analysis, Allied Reliability began the development of a Reliability Centered Lubrication Program at the site.  The program design began with a full walk down of all lubricated components on site.  During the walk down phase, all data pertinent to the development of lubrication specifications, interval calculations, and lubricant quantity calculations is gathered.  This phase sets the stage for the remaining program design as the accuracy and content of this data is the core to a strong program.
During the walk down, opportunities for optimizing intervals are identified as well as opportunities for optimizing the lubricant list.  Optimization does not always mean an increase in intervals or a decrease in the number of lubricants on site, rather it is the assignment of calculated intervals that are specific to the component and takes into account the specific operating conditions the component is exposed to.  Essentially, the same bearing may be exposed to wash down practices and heavy debris requiring an increased interval, such as 30 days to account for these environmental exposures.  However, the same type of bearing operating at the same speed with limited exposure to moisture and no exposure to debris may have an interval of 360 days.  This deviation in intervals means that the bearing with the more aggressive environment and operating conditions receives the more aggressive treatment.
When the Charlotte site began the implementation phase of their RCL program, the lubricated components were in poor health.  Approximately 60% of the oil lubricated assets were considered to be in the red, or in a state that did not meet industry best practices for contamination control, operating temperature, etc.  Within three months, the percentage decreased to approximately 40%.  By the end of 2010, 100% of all oil lubricated assets had achieved a green status.

Components were easily converted from a red state to a green state through the implementation of oil analysis, off line filtration, and monthly inspections.  Oil analysis aids in the identification of in use lubricants that required filtration to clean up solid contaminates, needed refreshing due to changes in additive levels and lubricant properties, or need complete replacement.
Previous to the implementation phase, the site utilized a time based oil change methods, had a multitude of lubricants with no specific requirements for each component, utilized screen breathers, and had no structure for the inclusion of filtration.  Lubricants were stored in drums and in areas that did not provide a clean, temperature controlled atmosphere.  The overall program lacked structure with regards to regrease intervals and amounts, as well as, oil analysis and oil change intervals.
The implementation phase of the RCL Program Design included recommendations for component modification with items such as desiccant breathers, quick connect couplings, level gauges, and sample valves.  These modifications provide multiple benefits for the lubrication program as well as the plant operating structure as a whole.  Level gauges and sample valves allow lube technicians to perform level checks and sample collection tasks with the machine running.  Without these modifications, the machine and possibly the product line would need to be down for the completion of these tasks.  It also saves time from a labor perspective.  A dipstick level indicator involves several steps to determine the level and exposes the lubricant to the possibility of contamination, where as a visual level gauge is nothing more than a quick peek then walk on to the next component.
The Charlotte site followed some very basic steps to manage the program implementation starting with the incorporation of desiccant breathers and other types.  After some experimentation with different types of breathers, there was an eventual shift to the use of trap and hybrid breathers based on the components size and location within the plant.  Of all the modifications, breathers are the least evasive to install, and limits the amount of debris able to enter the sump down to 3µ in size.
The second step included the conversion of mineral based oils to synthetic based oils in the smaller sump components.  Using the current practice of 6 or 12 month oil drains, each sump was converted during the next scheduled oil drain task.  At that time, new intervals of 12 or 18 month were implemented for each sump.  This step applies only to sumps that are too small for oil sampling.
The third step was the implementation of oil sampling practices on sumps within the 4 to 22 gallon range.  This implementation allowed for the extension of oil drains through the monitoring of the lubricant health.  Unlike the second step, these components were not immediately converted to a synthetic oil.  Rather the current lubricant was monitored and cared for until such time the oil analysis results showed a break down in the lubricant properties.  Once the lubricant properties began to degrade, an oil change with conversion to a synthetic occurred.  This maximized the life of the lubricant that was in service before change over without harm to the component.
*During lubricant drains, additional modifications were implemented while the sumps were empty such as the installation of level gauges, quick connects, and sample valves.
Once the oil lubricated components had been addressed, the fourth step was implemented.  The fourth step was the mass implementation of the grease lubrication routes that were developed based on the walk down phase.  These routes were provided in an electronic format which is downloadable to a pda, or other handheld device.  The routes were grouped together based on location within the plant, lubricant required, and interval.  The interval and re-grease volumes are calculated using gathered data to ensure the right amount of lubricant is delivered at the right time.
At the onset of the RCL design, this location had 1 overworked lubrication technician.  Now, they have 1 lubrication technician who is able to complete all required tasks within a timely manner because a solid program structure has been implemented to address all aspects of lubrication.  Task intervals have been optimized to ensure over and under lubrication is not an issue, and oil drain intervals have been extended.  Additionally, lubrication related failures have decreased allowing focus to shift from putting out fires to maintaining the current state.
“We now have a set program that anyone can follow with component state data being collected regularly to ensure appropriate actions are taken to avoid lubricant and component failure.  We have structure and focus.”

By Angie Meinsma, Oil Analysis & Lubrication Technician, Allied Reliability

It is often said that oil is the “lifeblood” of machines and equipment. Taking oil samples from rotating equipment for condition monitoring is like taking blood samples from a human. Analysts look for abnormalities within the fluid. The goal is to find faults, treat the condition, and extend the life expectancy.

When used properly, fluid analysis becomes a valuable diagnostic tool that can reduce maintenance costs, increase productivity, and boost company profits. When used in conjunction with other diagnostic technologies such as vibration analysis and thermography, well-executed fluid analysis strategies can detect a variety of equipment problems before they become failures and give users the valuable time necessary to make decisive, well-informed maintenance decisions. Time has allowed most people and organizations to see the value of integrating the condition monitoring resources for root cause management; they team up to provide control over the root causes of machine failure. Vibration analysis detects abnormal running conditions, such as unbalance, misalignment, and looseness, while oil analysis detects overall lubricant quality and contamination.

The three major reasons for lubricant failure are:
• Contamination – Component life is dependent on the cleanliness of the lubricating fluid. The cleanliness of any lubricant is dependent on oil handling practices, top up procedures, and the quality of both air breather and oil filtration.
• Oil Degradation, specifically oxidation – Oxidation occurs when atmospheric oxygen combines with hydrocarbon molecules and undergoes a chemical change. This chemical change results in the catastrophic and permanent change to a different chemical makeup for the oil molecule. The rate at which the oil molecules react with the oxygen depends on a number of factors, but the most prevalent are temperature and/or additive depletion.
• Additive Depletion – Additives are consumed or chemically depleted while performing their function. After being totally consumed, the additive can no longer provide the special property it performs for the base oil. The lubricant’s performance then suffers, and again the oil must be changed. Each one of these additives has a finite life, and when they reach the end of that life, you can forget about any advantage they helped provide. Some machines rely heavily on this advantage, and when it goes, so does the life expectancy of the machine.

These three factors are why we change oil. No matter what you do, eventually you will have to change it. However, the cleaner, cooler, and drier it is kept, the longer you will be able to go between those changes.

In order to establish appropriate maintenance strategies for your equipment, it is necessary to first understand the failure modes that are inherent in the parts and components that make up your asset base. Once the failure modes are identified, the proper condition monitoring and preventive maintenance tasks can be put in place to identify the warning signs of those failure modes or defects, keeping in mind that the earlier we identify an issue, the better. Unfortunately it is not uncommon for lubrication to be overlooked during the maintenance strategy design session. This is often due to the fact that lubrication is rarely thought of as an integral “part” that comprises your asset. The reality of the situation is that it is in fact a vital part and lubrication failure holds the same consequences as a bearing or coupling failure. Ultimately, the asset fails to perform according to it’s designed function. As with all failure modes, certain condition monitoring technologies are better suited to identify certain failures. This is why lubrication analysis is an integral piece of your maintenance and reliability initiatives.

When condition monitoring technologies are used in conjunction with lubrication analysis, the findings and recommendations are typically far superior. This approach ensures comprehensive coverage consistent with the the likely failure modes of your critical equipment and allows the earliest detection capabilities possible, which foster root cause issue identification. This approach affords ample time to plan, schedule, and perform recommended corrective actions and provides the highest possible return on your maintenance investments.

By: Reggie Fett, Motor Circuit Analyst

Synopsis

Summary of Action

Supporting Data

Conclusion

By: T.J. Garten, IR/MCA Specialist

Bean Conditioner – Horizontal Rotary Vessel

Synopsis

Fault

Summary of Action

Supporting Data

Figure 1: Pre-repair Demodulation Spectrum

Figure 2: Pre-repair Historical Data

Figure 3: Pinion Wear

Figure 4: Broken Welds on Pinion Gear Pillow Block Pedestal

Figure 5: Post-repair Demodulation Data

Conclusion

The corrective action taken during the repair of the failed trunnion bearing resulted in an altered elevation of the rotary vessel. The attempts to adjust for this were inadequate and resulted in a changed mating pattern between the pinion gear and bull gear assemblies. Over time, the increased impacting caused the pinion shaft to deflect, breaking the welds on the pedestals. This deflection also caused an angular misalignment between the pinion shaft and the gearbox output shaft. The level of misalignment may also have been affected by utilizing dial indicators for precision shaft alignment rather than the on-site laser alignment equipment. As the trunnion rollers wear, the clearance between the pinion gear and bull gear’s high spot decreases, resulting in increased shaft deflection. The increased shaft deflection results in transmitted vibration to the feed screw platform and torsional stress on the gearbox output coupling.Future corrections need to take into account all clearances and their effect on correlating components. Monitoring of adjusted components should be within days of adjustments and load changes. Installation of remote accelerometers on the pinion gear shaft pillow blocks would allow for quicker analysis of shaft misalignment, bearing condition, and pinion gear mesh frequencies.

By: Aubrey Green, Lead Analyst, Allied Reliability

Case Study Information:

While working as a Predictive Maintenance Technician in a paper mill, an analyst received a telephone call from the Pulping area. Their samples of white liquor were showing particles (turbidity). These particles were affecting the quality of pulp that was being produced. When these sand-like particles were processed in the paper machine, they produced a large mark. This mark lowered the quality of the paper. The department wanted to know if the analyst had any information on the process that could be causing the problem. They were particularly interested in pump vibration that would show the origins of the particles.

A quick look through the vibration data showed no issues that could be noted. The analyst asked for a simple piping diagram of the route of the white liquor from the clarifier to the digester. Armed with this information, the analyst decided to start at the beginning of the system and work to the end.

The system began with white liquor clarifiers started the system. Samples of liquor at the clarifier stage showed a few expected particles in the system. There was not enough turbidity to cause the problem the Pulping Department was experiencing. The analyst continued to study the diagram, looking for potential problem areas.

A holding tank was located before the digester. Between the clarifier and the tank there were only pipes and pumps, and the analyst decided to focus his attention on the tank. He knew he needed to find an accurate way to see what was in the tank. White liquor is extremely caustic, so dropping a weight into the tank would be dangerous. In addition, there were only one or two places to access the tank from the top.

Based on past experience, the analyst knew that the tank level could be determined with an Infrared (IR) camera. The temperature differential between the tank contents and the gas above the contents can be enough to show the tank level. The analyst had used this procedure many times in the past to confirm actual tank levels versus instrument readings. In this situation, he thought that if there was material in the bottom of the tank, it may have a temperature differential also.

Since the tank was not insulated, the image was not difficult to acquire. Using a thermal imager, the analyst took an image of the tank (Figure 1). The level of the tank was quickly noticeable. In addition, there was a wedge of material in the bottom of the tank, which was evident because the tank was known to have a flat bottom. This deposit was the result of years of particles from the clarifier settling in this tank. The inlet piping shown on the left of the image was actually acting as an agitator, mixing the particles with the white liquor before sending it to the digester.

Figure 1: Image of White Liquor Tank Showing Deposit

The analyst showed the image to the operator, who remarked, “Now we have an x-ray machine to see inside our tanks.” In this case, his comment was true.

An immediate resolution to the problem was needed, but the tank could not be taken out of service. The concentration of particles increased as the tank level lowered. Therefore, the tactical plan was to keep the tank level high enough to keep the concentration of particles at an acceptable level. The strategic plan was to clean the tank when it could be taken out of service.

Conclusion:

The deposits were removed from the White Liquor Tank during the next mill outage. Following that, the issues with turbidity were determined to be eliminated.

Dealing with hazardous chemicals in tanks and discovering problems within a containment may seem to be a daunting issue. Using the right technology in the right manner can give you insight to internal conditions and help you solve the problems in an efficient and effective manner.

Don’t Let the title fool you, this is a great and very short case study!  Allied Reliability SME Aubrey Green examines a dilemma at a salt evaporation plant in upper NY.  Read it for yourself here:

Radio Frequency Interference on Commtest VBx Series Data Collectors from Improperly Grounded Equipment

Allied Reliability Analyst Tim Greismer offers this case study on the motor of a Coal Pulverizer which was failing and creating operational problems for the coal plant.  Vibration analysis failed to isolate the problem, Tim and his team use Motor Circut Analysis to identify the root cause of the issue.  In this study he walks you through his process.  Read it for yourself here:

MCA Detects Where Vibration Analysis Could Not

As published in SMRP Solutions Magazine August 2011, GPAllied SME Darrin Wikoff addresses the Increasing Pace of Change in today’s dynamic business climate.  Read it for yourself here:

The Increasing Pace of Change