Innovative Solutions for Predicting and Preventing Equipment Failures

1. Predicting the remaining useful life of engineering assets through capturing both temporal and spatial dependencies of multiple sensor signals, for multiple degradation trends across multiple equipment health stages.
2. Detecting anomalies in operations of systems and engineering assets through continuous monitoring of multi-variate signal patterns and eliciting the indicators of anomaly for ease of failure diagnosis.
3. Prognosis and diagnosis of failures in engineering system and assets with limited labelled data that is also seamlessly transferable to a new domain.  

The technologies used in these tools have evolved with times to build more robust PHM models for multiple components in multiple applications.

In the aircraft industry, safety is paramount, and unplanned maintenance of aircraft components can lead to significant operational disruptions and financial losses. Such operational disruptions and losses can be substantially reduced through prognostics and health management solutions that A*STAR’s I2R has been developing since 2018, in collaboration with Singapore Airlines through the A*STAR-SIA Joint lab.

The technologies used in these solutions include Supervised Machine Learning with comprehensive feature engineering that represent the operational characteristics of the engineering system, Supervised Deep Learning for end-to-end prognostics and diagnostics and Unsupervised Deep Learning for anomaly detection.

These PHM solutions have successfully identified potential failures in advance, reducing significant downtime of SIA’s aircrafts. 

Sudden failures of critical components on a marine vessel could cause costly offshore downtime, as planned operations are inadvertently interrupted. 

Read more about A*STAR I²R ’s solutions for Predictive Maintenance for Marine Assets here.

In manufacturing, equipment downtime can severely impact production schedules and profitability. Our predictive maintenance solutions excel in managing equipment and assets equipped with time-series sensors such as vibration sensors, temperature monitors, and other diagnostic tools. By harnessing the continuous data streams from these sensors, our solutions enable proactive maintenance strategies and early detection of potential failures. This proactive approach ensures optimal performance and extends the longevity of critical assets.