Early warning of increased vibration on a boiler feed pump at a power generation

Author Sticky

Johannes Mahanyele

Customer Reliability Engineer

GE Vernova’s Software Business

As a Mechanical Engineer specialising in strategy and engineering within the Power Generation, Oil, and Gas sectors, Johannes holds a B-Tech in Engineering, an MBA, and has completed Strategy Execution Certification at Harvard Business School, among other institutions. With over 13 years of engineering experience, Johannes adeptly harnesses cutting-edge technology, data science, and industry best practices to revolutionize industrial processes. In his role as a Customer Reliability Engineer, he is at the forefront of utilizing APM and SmartSignal predictive analytics to avert equipment downtime by detecting, diagnosing, forecasting, and preventing critical asset failures.

Jacqueline Vinyard

Director, Product Marketing

GE Vernova’s Software Business

A professionally trained journalist, Jackie has a degree in journalism and has spent 15+ years’ experience as a researcher and launching innovative technology. She lives in Boulder, CO with her husband, three children and two dogs. Her latest passion is launching software at GE Vernova to accelerate the energy transition and to decarbonize the world.

Sep 04, 2025 Last Updated
3 Minute read

Predict: What did GE Vernova predictive analytics software find?

A boiler feed pump (BFP) is a vital component in thermal power generation.  Its purpose is to ensure the boiler has a continuous supply of high-pressure water to produce steam for power generation. The boiler feed pump’s failure can halt power generation. To improve reliability for this critical asset, a GE Vernova customer uses APM SmartSignal predictive maintenance software with Industrial Managed Services (IMS).  At the customer’s coal-fired plant site, SmartSignal’s digital twin detected a deviation and sent an alert for BFP-2.
 
The AI/ML powered digital twin showed that that the pump inlet bearing vibration increased from 32 µm to 37 µm, which exceeded the expected operational vibration range by several micrometers. Even though this change may seem insignificant numerically, this is a clear sign of emerging mechanical stress or imbalance. In high-speed rotating machinery, even slight deviations can signal serious underlying issues. Potential impacts of elevated vibration could result in accelerated bearing wear, leading to early replacement, seal damage due to misalignment, or shaft movement from the vibration. This can lead to leaks of high-pressure water or lubricants. It can also lead to reducing the pump efficiency due to uneven loading on impellers and reduced hydraulic efficiency.  
 
Additionally, the mechanical seal water leak-off increased in temperature by 6 degrees Celsius. This is a potential warning sign of thermal stress or inefficiencies in the seal system. In precision-engineered systems like boiler feed pumps, even small temperature deviations can have cascading effects. Mechanical seals rely on a thin film of fluid between seal faces to lubricate and cool the seal faces. This prevents dry running, which can cause rapid wear or failure. When the leak-off temperature rises, the fluid’s viscosity decreases, reducing its lubricating ability. The cooling effect weaking and evaporation risk increases. If not caught in time, this can shorten the seal lifespan, may increase leakage, or lead to seal failure with sustained overheating.
 
The assigned IMS engineer for this customer raised a case and included this item in the weekly report for discussion with the customer.

Diagnose & Prioritize: What was the underlying cause of the pump failure and risk urgency?

1. Did the customer plan an outage and then find the issue, or was the issue discovered first?

The elevated vibration readings and other abnormal conditions were first detected through monitoring by APM/SmartSignal, with a recommendation from GE Vernova’s IMS team for the site to check detailed vibration data from start for an issue such as misalignment due to thermal offset starting.
 
After the early alert from IMS, the customer’s maintenance team planned targeted inspections and ultimately an outage to conduct repairs. Therefore, the issue was discovered first through condition monitoring, and the outage was planned subsequently to address it.
 
2. Was the outage planned at a day or time that was less impactful for their business?
 
The team was able to closely monitor the situation and make decisions based on the evolving condition of the equipment. The outage and subsequent actions included input from the diagnostic data, allowing the team to schedule shutdowns and restarts with consideration for operational needs.
 
3. Did SmartSignal help beyond just the BFP (Boiler Feed Pump), for example, with identifying the exact problem and coordinating the response?
 
Yes, SmartSignal provided early detection of abnormal vibration and temperature trends, enabling the team to localize the issue to the mechanical seal and bearing areas. The system helped differentiate between problems at the suction and discharge ends, and highlighted changes in the seal water cooler's performance. This guided maintenance teams to focus their inspection, repair, and replacement efforts specifically on the mechanical seal and related components, rather than conducting broad or unfocused maintenance. The data and assigned engineer working with the customers help with resource planning by having the correct personnel engaged for the repair and supporting coordination among maintenance teams for targeted interventions.
 
When the customer replaced the mechanical seal on BFP-2, the mechanical seal temperature and pump inlet bearing vibration returned to the model-predicted values.

Customer Value

Due to the early notification from GE Vernova’s IMS team, the customer was able to replace the mechanical seal on BFP-2. Prolonged operation with high pump vibration and elevated mechanical seal temperature could have caused bearing damage and potential failure. Additionally, GE Vernova’s IMS team was able to verify that the maintenance actions were successful by observing the actual values returning to the expected values. This intervention is estimated to have avoided approximately $122,000.00 in costs.*
 
What are the Key Benefits of using SmartSignal with IMS for early detection of failures?
  • A hard-to-detect issue was predicted by SmartSignal prior to failure
  • The maintenance team had time to plan an outage with less business impact
  • Repairs were prioritized by pinpointing the problem location (mechanical seal, suction-side vibration)
  • More significant repairs or equipment failure were avoided through early intervention
*Avoided costs are based on North American average production loss

Author Section

Authors

Johannes Mahanyele

Customer Reliability Engineer
GE Vernova’s Software Business

As a Mechanical Engineer specialising in strategy and engineering within the Power Generation, Oil, and Gas sectors, Johannes holds a B-Tech in Engineering, an MBA, and has completed Strategy Execution Certification at Harvard Business School, among other institutions. With over 13 years of engineering experience, Johannes adeptly harnesses cutting-edge technology, data science, and industry best practices to revolutionize industrial processes. In his role as a Customer Reliability Engineer, he is at the forefront of utilizing APM and SmartSignal predictive analytics to avert equipment downtime by detecting, diagnosing, forecasting, and preventing critical asset failures.

Jacqueline Vinyard

Director, Product Marketing
GE Vernova’s Software Business

A professionally trained journalist, Jackie has a degree in journalism and has spent 15+ years’ experience as a researcher and launching innovative technology. She lives in Boulder, CO with her husband, three children and two dogs. Her latest passion is launching software at GE Vernova to accelerate the energy transition and to decarbonize the world.