Sensing a better maintenance solution for pump systems

Machine learning (ML) and artificial intelligence (AI) can support better maintenance and improved OEE for pump systems.

By Trenton Roncato Juraszek February 23, 2024

Pump systems insights

  • The article explores the integration of machine learning (ML) and artificial intelligence (AI) in enhancing maintenance strategies for pump systems, emphasizing their role in improving overall equipment effectiveness (OEE) by providing better predictive maintenance insights.
  • It highlights the success of applying smart sensors and digital platforms for monitoring pump performance, resulting in significant reductions in unplanned downtime and costs, and mentions a case where a customer experienced reduced motor failures from once a month to once every six months after implementation.

According to a report from OneServe, 53% of all machinery downtime is caused by hidden internal faults that maintenance engineers didn’t spot. This is why it is important to have an effective plant maintenance strategy in place.

Today, industrial software is playing an increasingly useful role in helping manage and analyze data more effectively for maintenance purposes and it is possible for both new and legacy pump systems to be integrated into a software platform that links directly to plant equipment through smart sensors, installed at the device level.

The software platform, combined with sensors linked to motors and drives, for example, make it possible to evaluate the frequency and duration of asset problems, and source their origin, to support predictive maintenance.

Today, a sensor that might have been used to monitor a drive or motor can also be retrofitted to other equipment to create a monitoring platform that can enable legacy pump systems to benefit from industrial Internet of Things (IIoT) enabled preventive maintenance strategies.

Monitoring pumps

So what should a specialized pump monitoring platform look like? For industrial motors, software and sensors apply 3D vibrational analysis to estimate the energy consumption and load a motor is carrying. Historical and real-time data is gathered by the sensors and software and relayed to the plant’s supervisory control and data acquisition (SCADA) and manufacturing execution systems (MES). Through artificial intelligence (AI) and machine learning (ML), this data can be used to automatically adjust and optimize the performance of the motor.

In such a scenario, OEE and sustainability are closely connected and it becomes possible to analyze the forecasted deterioration of equipment and advise on predictive maintenance requirements.

One customer in the petrochemical segment applied these technologies after experiencing at least one motor failure per month. After applying the digital suite, the customer can now go for six months without a failure. A return on investment (ROI) was also achieved within six months.

The same technologies can be used to directly monitor other electric-driven equipment such as pumps, compressors, fans and gearboxes. Using dosing pump systems as an example. While traditional pump systems do have fault-logging systems, these are usually managed according to a planned maintenance strategy which means that assets are replaced and maintained according to a predetermined schedule, with the risk that faults can worsen in-between maintenance checks. In some instances, the system might not be regularly checked which can lead to a reactive maintenance approach where faults are not addressed until it’s too late.

A pump monitoring platform sees smart sensors attached directly to pumps to provide fast and accurate data collection at the device level. This data can be processed to generate valuable real-time insights.

As a result, operators can access the data to make immediate and informed decisions relating to system performance. Their decision might relate to changing pump operating modes such as manual, batch and timed or making instant efficiency improvements.

Using these technologies and processes, it is possible to significantly reduce unplanned downtime and costs. It also is a crucial element of modern maintenance strategies and OEE, and therefore manufacturing sustainability.

Original content can be found at CE Europe.

Author Bio: Trenton Roncato Juraszek is an application engineer at WEG.