Digital Transformation for Best Practice Maintenance

Since last year I call 24th of July (24/7) the World Reliability Day. It’s a day to celebrate the men and women that keep all the ‘things’ in complex plants working 24 hours a day, 7 days a week. I’m in awe of these people. How can we digitally transform plants to make the job for these men and women easier? Here are my personal thoughts:

Plants are expected to reduce downtime by avoiding unscheduled downtime due to equipment failures, extending the period between scheduled turnarounds, and shortening the duration of the turnaround. At the same time, the maintenance cost is expected to be reduced and all of the above accomplished with a smaller team. A second layer of automation focused on equipment condition can improve plant maintainability and help with new data-driven reliability programs. However, most plants were built with minimal input from operations & maintenance for reliability programs beyond the very largest machinery like turbines. Most other equipment relies on scheduled inspections and preventive maintenance. Part of the problem is that 4-20 mA and on-off cabling and system I/O is expensive. However, fieldbus and wireless sensors now dramatically reduce the cost of automation. Plants are now using this second layer of automation to improve plant availability, save time, and reduce maintenance cost based on more predictive information, a digital transformation of how the plant is run and maintained.

Availability Challenges

Failures that cause downtime or slowdowns are difficult to predict without timely information. Affected equipment include for instance:

• Pumps
• Blowers
• Centrifugal compressors
• Cooling towers
• Air cooled heat exchangers (ACHX)
• Heat exchangers
• Valves
• Steam traps

Service Contracts

Plants typically have maintenance contracts with third-parties that periodically carry out inspection and manual data collection with portable testers on rotating machinery and other equipment. However, due to the long time between checks and the difficulty of detecting early "markers" and health trends based on infrequent data, failures still occur between checks. Digital plants now use sensors and software to drive maintenance planning. Using wireless sensor networks an existing plant can be modernized to a digital plant.

Inspection Rounds

All plants have online protection systems on their large machines like steam turbines, gas turbines, and large compressors which are most critical, some even on their largest pumps.

Plants have a “second tier” of hundreds of smaller machines like motors, pumps, gearboxes, cooling tower fans, blowers, air cooled heat exchanger fans, and smaller centrifugal and reciprocating compressors etc. essential for production. However, these “balance of plant” equipment don’t have online protection systems due to the high cost. Some machines have vibration switches, but by the time the switch trips the machine it is already too late, the plant unit is no longer producing or have to slow down. The problem may also have gone pretty far, requiring more extensive, expensive, and longer duration repairs. The challenge is to ensure developing issues are not allowed to go so far as to cause slowdown or shutdown. Monthly manual data collection using portable testers for vibration and temperature etc. is not sufficient to capture developing issues sufficiently early because a small issue can emerge and rapidly escalate to a failure in a few weeks, or even in a couple of days.

The original intention may be for weekly rounds, but due to workload it may instead become monthly. Collected data may not always be analyzed in a timely manner. Additionally, manually collected data may not be stored electronically which means deteriorating trends are not spotted. There has even been cases were data is collected manually, the value is outside the limits, but no action gets taken anyway because the abnormal condition is not electronically flagged and logged as an alarm. Lastly, some plant areas can be harmful so visits by personnel should be minimized.

Automatic Data Collection

A wireless transmitter is a new way to automatically collect data for equipment monitoring. This includes wireless transmitters for vibration, temperature, acoustic noise, fluid levels, pressures, motor current, and many other measurements. A vibration transmitter is important, but not the only measurement required. Other input is also required to get a full picture of equipment condition. Additional transmitters for condition monitoring may not be required for every piece of equipment in the plant. A plant modernization audit can be carried out to identify which essential equipment is currently unmonitored and should be instrumented.

A wireless vibration transmitter is simpler than a machinery protection system and very much lower cost. A wireless transmitter is ideal for smaller machines with roller element anti-friction bearings such as pumps, fans, gearboxes, and motors. Because no power cables, no signal wires, no cabinets, and no floor space is required, wireless vibration transmitters are also very much easier and lower cost to deploy. Wireless vibration transmitters with onboard analytics are also very much easier to use.

By collecting vibration data automatically every hour and other parameters once a minute instead of monthly, weekly, or daily using a portable tester, condition monitoring becomes far more predictive than with manual data collection. It is possible to detect that something is going to happen because before something fails there are "markers" and degrading trends that can be seen for vibration, temperature, fluid levels, motor current, and pressures etc. Capturing the early warning signs when they first appear requires automatic data collection. This also makes root cause analysis easier. By digitally transforming from reactive repairs to condition-based maintenance for currently unmonitored equipment the maintenance costs can be reduced by addressing the problem prior to failure.

Onboard Analytics

Automatic data collection doesn’t mean an avalanche of data for vibration experts to analyze, quite the opposite actually. Wireless vibration transmitters monitor large numbers of equipment to automatically get an overview of which pieces of equipment are in good condition, which ones are starting to develop issues, and which ones are closer to failure. This way limited analyst resources can be prioritized to not spend precious time on equipment that is healthy but instead focus equipment most urgently in need of attention and on root-cause analysis.

Predicting failure by analyzing and interpreting infrequently collected vibration waveforms and spectrums is very difficult and requires a certified vibration analyst with expert knowhow. Root-cause analysis based on infrequent data also requires expertise, especially if process condition data is not available in the vibration analysis software. The site may not have sufficient category III and IV certified vibration analysts to review all the data in time to catch a developing problem. By alarming on vibration reading taken every hour, with an alarm limit set lower than the vibration switch trip point, the maintenance and reliability departments get an early warning, before the machine trips. This gives them advance notice to fix the problem.

Significant increase in total vibration is often seen only in the late stages close to failure. However, PeakVue technology embedded in some vibration transmitters, alarms on peak value acceleration. This provides an even earlier warning of onset of a problem, thus giving the maintenance department more time to respond. Moreover, it is automatic, reducing the workload for the vibration analysts. 

Machines vibrate more or less depending on the type of machine, the mounting base, and alignment etc. For instance, two identical pumps will vibrate differently depending on how the coupling between pump and motor is aligned. The alarm limit must be set appropriately to avoid nuisance alarms. Some modern multi-parameter condition monitoring software has automatic baseline to set alarm limits correctly for this purpose.

Depending on the kind of equipment, vibration is also a function of flow, speed, or fan pitch. The alarm level should therefore be correlated with operating conditions such as for variable speed pumps. Some multi-parameter condition monitoring software uses statistical methods to alarm only on meaningful increases in vibration. That is, in some applications sensors alone are not enough; analytics software is sometimes required. The vibration spectrum and waveforms are also available from the wireless vibration transmitter at a click of a button.

By capturing developing problems early, downtime and production loss can be avoided, and the required repairs will be less extensive, lower cost, and shorter duration. Reliability engineers can now also get information from the software event log and vibration trend when investigating the root cause of machinery trips and failures in order to see what has happened to the machine over the days and hours leading up to the failure. This is useful for instituting measures to prevent reoccurrence. Also important is that by automating, the solution saves time for everyone.

Wireless vibration transmitters will not take the place of online machinery protection systems and cannot completely eliminate portable testers. Online machinery protection systems typically have a 100 ms update period. Portable testers are used for checking once a month. That is a 7 orders of magnitude difference. Wireless vibration transmitters is a new complimentary tool ideal for hourly updates fitting machines with a reliability methodology where 100 ms is overkill and 1 month is too infrequent.

Multi-Parameter Condition Monitoring

Traditionally reliability engineering has mainly been focused on vibration. However, there are many other equipment failure modes not detected through vibration monitoring. Vibration monitoring alone is not enough. For instance, a pump mechanical seal failure will not be detected or diagnosed with a vibration transmitter. Getting a complete picture of equipment health requires multiple measurements. To improve maintainability, plants need highly instrumented assets. This means additional sensors for equipment and process data such as pressures, temperature, flow, fluid levels, motor current, and position etc., which together with analytics software enable fault detection and simplify root-cause analysis. Many of these sensors are non-intrusive and can be installed while the plant is running while some sensors are best installed during the next turnaround. There now is condition monitoring software which has specialized analytics algorithms for each equipment type such as pump, fan/blower, centrifugal compressor, cooling tower, heat exchanger, and air cooled heat exchanger etc. - an early warning system which has easily understood status flags which does not require a category III or IV vibration analyst to interpret.

Modernizing process equipment with wireless transmitters and built-for-purpose analytics software give plants the ability to track vibration and other parameters for each piece of equipment, and to visualize in trend charts how rapidly it is getting worse over time. If vibration or peak value exceeds the pre-limit, an alarm is triggered in the condition monitoring software prompting attention, before the vibration switch trips causing downtime. Some of the required measurements are usually available from the DCS, but missing measurements such as temperatures of bearings or motor windings may need to be added using WirelessHART or fieldbus transmitters. For plants have a fieldbus digital ecosystem the additional transmitters can be connected to existing junction boxes. For plants built on 4-20 mA and on-off signals, WirelessHART is easy to implement and a low risk to deploy.

The additional transmitters and software turns ordinary process equipment into smart connected equipment by providing condition indication. Note that this detail information does not need to go to the operators, but instead to the reliability and maintenance personnel responsible for keeping the equipment running. The operators see simple status indication, not the same level of detail.

Air Cooled Heat Exchangers (ACHX)

ACHX fan bearings, belts, gearboxes, and couplings will experience wear and tear, which can eventually lead to failure. Some of the fans may inadvertently be running at their resonant speed thereby accelerating mechanical failure. Or the louvers may have issues. This will reduce cooling capacity and may in turn result in lost production, flaring, and vapor release, which is cost over and above the repair costs. Precipitates like salts cause fouling on the inside of the tubes, accumulation of dust or other airborne debris cause fouling on the outside of the tubes, both which reduce the heat transfer capability of the exchanger. 

ACHX diagnostics include:

• High motor/fan vibration
• High motor/fan bearing temperature
• Resonance
• Bearing fault
• Exchanger fouling
• Reduced cooling
• Louver mechanical defects
• Fan pitch actuator defect

The ACHX diagnostics detect fouling enabling cleaning to be scheduled to restore efficiency and process throughput and to reduce power consumption. Developing problems are detected allowing overhaul to be scheduled before trip or failure occurs.

Pumps

A pump will experience wear and tear on bearings which may lead to vibration and failure of pump and seal. Restricted discharge flow or plugged suction strainer could result in cavitation which leads to pump failure and mechanical seal failure due to vibration. Over and above the repair costs, this may result in production loss, process leak and fire. 

Pump diagnostics include:

• High vibration
• Bearing fault
• Low head
• Low discharge pressure
• Seal pressure
• Strainer cleaning required
• Seal flush system faults
• Liquid hydrocarbon leak

The pump diagnostics detects developing problems allowing switchover to the standby pump or overhaul to be scheduled before trip or failure occurs.

Pump Seal Instrumentation

The fourth edition of API Standard 682 for mechanical seal piping plans has changed the seal system alarm instrumentation from switches to transmitters. Transmitters provide a more reliable method of alarming and also reveals gradual deterioration of the seals over time not revealed with switches. This enables more predictive approaches to monitoring pump health. Wireless transmitters make seal flush systems upgrade from switches to transmitters easy as there is no need to rewire, reconfigure, or expand the system I/O cards.

Heat Exchangers

Fouling builds up in heat exchanger bundles in heating or cooling service over time impeding the heat transfer and thus reducing efficiency. Fouling is not always long term. For instance, in a refinery certain crudes cause fouling faster. When heat exchangers are fouling or plugging, operations basically has three options: shutdown to clean or slowdown throughput meaning production loss, or increase make-up heat raising the energy cost.

Common practice in the past was to routinely pull out heat exchanger bundles one by one for cleaning without knowing which bundles are fouling causing the heat exchanger to underperform. However, dismantling and cleaning heat exchangers this way is very labor intensive and can consume considerable time. Heat exchanger bundles require a crane for removal.

A somewhat better approach is to manually collect temperature data between bundles from time to time using a portable tester for input into a spreadsheet to compute performance. Periodically testing dozens of heat exchanger bundles manually with a portable temperature probe is still very labor intensive and therefore does not get done frequently enough. As a result fouling continues for many months. Moreover, since the four temperature points are not measured simultaneously, and due to process dynamics, the result are often inaccurate and confusing. As a result the heat exchanger fouling may cause production loss.

Some plants already have heat exchanger monitoring from their DCS, but this is usually measured across an entire heat exchanger, obscuring the individual condition of each bundle. Generally the bundles are not individually instrumented. Temperature at the inlet of the first bundle and outlet of the last bundle is not sufficient. The measurements for individual bundle condition monitoring are missing.

By modernizing with wireless transmitters on heat exchangers in conjunction with built-for-purpose analytics software with heat exchanger condition monitoring algorithm, the condition monitoring is automated and plants get the ability to track the fouling in each heat exchanger bundle, with the ability to visualize in trend charts how rapidly the heat transfer deteriorates over time. If fouling is accelerated, an alarm is triggered in the condition monitoring software prompting attention. Some of the required measurements may be available from the DCS. For instance, the process flow is measured for process control, but not the cooling water flow or heating steam flow. The missing measurements such as temperatures between bundles, or DP across each bundle, and missing flow, must be added. In a plant built on fieldbus these additional transmitters can simply be added to the nearest junction box since networks are usually not fully loaded. In plants built on 4-20 mA and on-off signals, using WirelessHART is usually the easiest way to add transmitters. 

Temperature sensors can make use of existing thermowells, or non-intrusive clamp-on pipe surface temperature sensors can be used. These are very easy to install and a low risk to deploy. 

Heat exchanger diagnostics include:

• Cleaning required
• Fouling
• Low flow
• Decreasing duty

The fouling information can be used to decide to inject anti-foulant chemicals, or determine which of the heat exchanger bundles needs cleaning, and the optimum time to clean. This enables the maintenance manager to strike a balance between the need to clean the heat exchanger in order to maintain throughput on the one hand, and on the other hand to avoid the outage associated with cleaning. The temperature measurements can also be used to alarm operators if heat exchangers are operating near design limits.

Fans/Blowers

A blower, such as forced draft blowers on fired heaters, will experience wear and tear on bearings which may lead to vibration and failure of the fan. The fan may inadvertently be running at its resonant speed thereby accelerating mechanical failure. Or the louvers may have issues. The intake filter may be plugging. Over and above the repair costs, a forced-draft blower failure may result in fired heater trip causing some units to slowdown and others to shutdown. 

Blower diagnostics include:

• Bearing and gear wear
• Operating near known resonance
• Louver mechanical defect
• Plugged suction filter ?

The blower diagnostics detects developing problems allowing overhaul to be scheduled before trip or failure occurs.

Centrifugal Compressors

A compressor will experience wear and tear on bearings which may lead to vibration and failure. Restricted discharge flow or plugged suction strainer results in instability which leads to pump failure due to vibration. Or the vanes may have issues. Over and above the repair costs, this may result in production loss. Large process compressors typically have an online protection system, but the smaller compressors usually have nothing and therefore should be instrumented.

Centrifugal compressor diagnostics include:

• Bearing and gear wear
• Compressor instability
• Control vane defect
• Low flow
• Plugged suction filter?

The compressor diagnostics detects developing problems allowing overhaul to be scheduled before trip or failure occurs.

Cooling Towers

If a cooling tower is fouling, cooling water temperatures may rise, and capacity may be limited during hot days. This in turn may require production to be slowed down to not overload column condensers and product coolers. Worse still, failure of a cooling tower fan or pump could result in a slowdown or shutdown of a process unit for repairs. Issues that can affect pumps and fans are already explained above. Poor cooling water quality may accelerate fouling, corrosion, and microbiological growth in the cooling tower and the heat exchangers, again affecting throughput and requiring shutdown for cleaning. 

Cooling tower diagnostics include:

• High vibration
• Bearing fault
• Bearing temperature
• Low head
• Low discharge pressure
• Low suction pressure
• Strainer fault
• Water temperature
• Fouling
• Corrosion
• Scaling / water quality
• Low flow
• Windage/water loss

The cooling tower condition information can be used plan maintenance to strike a balance between minimizing unscheduled downtime due to failure and scheduled downtime for preventive maintenance. Plants can thus reduce downtime and maintenance cost.

Other Reliability Solutions

Once a plant has been modernized with a plant-wide WirelessHART network, additional transmitters can easily be added to detect other reliability and maintenance related issues. This may include for instance:

• Reciprocating compressor valve manifold temperature to detect valve issues
• Conveyor belt vibration or misalignment; coal, ash, ore, etc.
• Control valves, intelligent on-off valves, electric actuators / Motor Operated Valves (MOV)
• Steam traps
• Centrifuges
• Chillers
• Furnace hot-spots

Common Infrastructure

The reliability engineers and maintenance manager can get maintenance information and reliability data straight to their own workstations, including additional information currently not available from the DCS. They do not need to use the DCS operator consoles. The wireless instrument network is shared by multiple departments that need information from around the plant. The reliability and maintenance department, energy team, HS&E department, and operations share the investment in the plant-wide WirelessHART network. WirelessHART gateways can integrate with any control system. It is not necessary to upgrade or migrate the existing DCS, or make any changes to the controls, to take advantage of wireless transmitters.

IIoT-Based Remote Monitoring Services

Apart from plant personnel monitoring the condition of the equipment themselves, once the equipment has been instrumented it is also possible to enable secure access across the Internet, but this is optional. This concept is known as the Industrial Internet of Things (IIoT). IIoT in turn enables the plant to subscribe to third-party remote monitoring services. This step should only be taken when the organization is ready. For instance, an external service provider or original equipment manufacturer with a pool of experts can monitor the condition and performance of equipment in the plant, as a paid service providing reports directing the maintenance team in the plant to which equipment need service, and what needs fixing. This reduces not only the data collection burden on the plant, but also the analysis work. Lastly, even if sensors are not permanently connected up through the Internet, temporarily enabling access enables remote experts on valves, vibration, and analyzers etc. to assist the plant personnel in problem solving when they are stuck.

Remote locations

Condition monitoring on smaller essential equipment in addition to the large critical machines is valuable in any plant. However, service techs for remote sites like oil & gas fields, mines, and offshore installations spends as much as 80% of their time traveling and only 20% doing actual troubleshooting. Once equipment has been instrumented it can be monitored from anywhere. Experienced engineers in a central location can support a junior technician in the field without having to travel to site. This can potentially save days of downtime. It is not unusual for a technician to come to site only to find that the required spares are not available and another trip with the right parts is required. If an experienced engineer can instead diagnose the problem remotely from a central location to determine which spares to bring, the technician can settle the issue in one trip.

Data-Driven Smart Plant

Modernizing with additional instrumentation and built-for-purpose condition monitoring software enables the plant to detect issues with process equipment earlier, that would otherwise lead to failure and downtime had it not been detected.

Not designing a new plant or modernizing an existing with a second layer of automation for condition monitoring sets the stage for a reactive maintenance culture. Therefore, make sure these maintenance and reliability solutions are incorporated into the project and budgeted prior to kickoff also for greenfield plants. Consult with a reliability expert also experienced in digital sensor networks to audit your existing plant or new plant design to identify which reliability and maintenance productivity solutions are needed. EPC contractors or process licensors will not put in these maintainability and reliability solutions on their own accord. It is up to the plant owner to specify these requirements to ensure a top performing plant. See further explanation of these applications in this article:

https://www.ceasiamag.com/2015/06/multi-parameter-condition-monitoring/

Well, that’s my personal opinion. If you are interested in how the digital ecosystem is transforming process automation click “Follow” by my photo to not miss future updates. Click “Like” if you found this useful to you and share it with others if you think it would be useful to them.