HP Tagline--Column--Digital Technologies

Do not just monitor bad actor machines: Start when they are healthy

R. Divita, ITT Goulds Pumps, Singapore

Anyone who has ever watched a medical drama knows the story. By the end of the hour, the exasperated doctor pores over charts to find the nugget of information they are looking for—the information in the patient’s medical history that will help them discover their ailment.

What if the doctor could not access the patient’s medical history? Without this important background, the chances of identifying the cause of the illness and ending the episode at the 60-min mark would be slim to none.

Unfortunately, in many cases, this scenario applies to monitoring the health and condition of machines in process industries. Often, facilities hesitate to invest in monitoring rotating equipment until the machine fails. By that time, there is no background information, such as vibration or temperature data, to help site engineers diagnose the machine’s failure.

However, by implementing a remote monitoring solution on healthy equipment, users can set a baseline of machine health to gain essential insights when the machine begins to fail.

The way it has always been done. The risks associated with pump leakages are significant in process facilities that handle materials like hydrocarbons or liquefied natural gas (LNG). To prevent leaks, facilities like these typically do maintenance on pumps at standard intervals (e.g., every 2 yr). In many cases, the pump is shut down as soon as a leak appears.

However, this leads to an unnecessary drain on resources. The lack of insight into pump repairs may lead to healthy pumps being shut down for maintenance, while inefficiently operating pumps are neglected until it is too late. When healthy pumps are repaired prematurely, the facility wastes time and money.

Conversely, when unhealthy pumps are not promptly addressed, they can fail, resulting in further production loss and safety concerns, which is the bane of any pumping facility. To unlock access to a plethora of important machine health data (e.g., temperature and vibration), many facilities have begun to turn to remote machine health monitoring.

Remote machine health monitoring. Remote monitoring enables process facilities to practice predictive maintenance and solve potential problems before they become major issues. Plants can prevent surprise equipment failure and extend machine life by integrating a remote monitoring system with suitable sensors and data logging functionality.

Additionally, many new sensor systems can provide an Industrial Internet of Things (IIoT) asset intelligence platform that can utilize automated diagnostics to securely monitor and assess—locally or remotely—the health of machines. Overall vibration, vibration spectrums, temperature and pressure data can be analyzed to detect and resolve issues that might otherwise lead to equipment breakdowns.

Many facilities that elect to use remote monitoring begin by identifying bad actors or machines that have chronic issues and consistently fail; sensors are attached to these machines to identify any anomalies that must be addressed.

However, this strategy does have one flaw. Remember the scenario where the physician cannot determine the patient’s health issues because they cannot review the medical history? Imagine the machine is the patient, and the medical history is all the data of the machine’s past performance, including vibration and temperature. If it is not understood what the machine used to look like, how and when it will fail cannot be accurately determined.

By monitoring bad actor equipment when it is already experiencing issues, IIoT sensors often simultaneously highlight a range of machine problems. The user then finds it difficult to decide when to act, as no data is available to contextualize the issue. The facility is essentially starting at square one. Therefore, it is crucial to begin monitoring when the equipment is healthy.

Setting the baseline for success. Installing remote machine health monitoring sensors on healthy equipment identifies and tracks changes over time. This includes general trends and vibration spectrum changes. A good monitoring system will incorporate automated diagnostics of the vibration spectrums to ensure that higher volumes of data are successfully managed and reviewed. The longer the facility waits to begin this process, the more data the operation is missing from the equipment’s health history.

For example, if sensors are installed on five new pumps in a LNG facility, a health baseline for these machines is set. These pumps’ vibration, temperature and pressure data can be instantly recorded, identifying how they appear when healthy (FIG. 1).

Digital-Fig-01

One year later, before the overall vibration has increased, this continuous collection of data compared with the initial baseline may show that one of the pumps is experiencing a slight increase in bearing fault frequencies. In addition, the sensor is set to routinely capture vibration spectrum data, which shows the degradation speed over time. The operations team can set the sensor alarm to a level just above the known trend to receive early notification of the next issue—in this case, a stage 4 bearing defect. Over time, alarms can be re-adjusted to higher thresholds.

If the vibration steadily increases in the following months, the user can determine when to shut down the machine for repairs, ensuring no further damage to other rotating components and allowing them to minimize lost production time and save on repairs.

Additionally, with the data collected from the healthy machine’s operation through its repair stage, the user has a picture of the machine’s estimated lifespan. This enables them to practice predictive maintenance, establish more accurate maintenance budgets and forecast workloads without unplanned events. For example, if the pump reaches a Stage 3 bearing defect, maintenance can be scheduled when it begins to approach the known threshold again.

Practicing predictive maintenance means less equipment failure, reduced machine downtime and a decreased drain on resources. However, to get started, facilities must implement a remote machine health monitoring system to monitor trends, gather data and identify when machines are approaching the danger zone. There is no better time than now. HP

First Author Rule Line

ROBERTO DIVITA is the Manager of monitoring and control for ITT PRO Services’ Asia Pacific region. Divita is based in Singapore and has been with ITT for 15 yr, having joined the company in 2007 as a mechanical engineer. He oversees the technical application of ITT Monitoring and Control solutions throughout the APAC region, including applying IIoT solutions to pumps and other rotating equipment, conducting site audits, leading energy studies and utilizing RCA support and practical solutions to lower the total cost of ownership for clients. Divita works with clients in oil and gas, power, mining and pharmaceuticals. He earned a BS degree in mechanical engineering from Curtin University in Perth, Australia.