(146c) Digital approach for fouling predictive analysis on Steam turbines through operational data

Industrial Internet of Things (IIoT) applied to oil & gas equipment is nowadays a must-have-tool for engineering maintenance and operation teams to monitor and manage the installed fleet to ensure the highest availability and reliability. Operational data combined with subject matter expertise and smart analytics have consolidated capabilities for anomaly detection and early warning. Industrial market is more and more heading to production, maintenance and performance optimization horizons.

Deep domain expertise combined with physics-based analytics allows to meet this expectation.

Such analytics, based on design tool and OEM models (digital twin), let to highlight sub-optimal operation, related root cause and actionable recommendation to restore the expected design condition.

Fouling is one of the most common and relevant limitation for turbines and compressors operability producing performance deterioration, thrust bearing load increase and, depending on contaminants, possibility of corrosion cracking.

The paper highlights analytics techniques and capabilities to provide assessments related to operation, performances and maintenance that could be affected by machine fouling.

Steam turbine fouling is typically due to contaminants deposit and affected by steam purity, pressure and temperature in the different machine sections. Presence of deposits translates in a different geometry in the steam path producing an increase of pressures detectable through wheel chamber pressure transmitters. Main effects could be higher steam consumption, reduced power capability and reduced aerodynamic efficiency.

The analytic is based on a comparison between measured wheel chamber pressure and the expected one (in new and clean condition).

Furthermore, the wheel chamber pressure is proportional to the thrust bearing load: the analytic outputs also allow to estimate the stresses on the bearing with respect to margin to design allowable load.