Development
Advanced Data Analytics
By utilizing the underlying physics of assets,
get the right answers to achieve useful analytical results.
Training, support, and case-studies using advanced analytics to detect underlying root causes of issues.
Automatic Pattern Recognition (APR) software usage has seen a heavy increase in usage in recent years. Typically, APR software is deployed in M&D centers to analyze gas turbine data and detect when performance has strayed from what is expected for a “normal” unit. Unfortunately, these APR models are severely limited in diagnostics ability as they are only as good as the data they are trained off of. Since many significant faults are rare in gas turbines, APR models are generally trained off of “good” data and frequently cannot give more information beyond the fact that something in the gas turbine has changed. With our Fault Signature Database, we train a gas turbine model based on your data, and use that model to predict performance changes as various hardware degrades. This synthetic data can then be fed into your preferred APR software to create more appropriately trained models for common issues that arise when operating a gas turbine.
Perspective of Power Industry
-
Better Informed APR Software
Most APR models are limited to data from operational history, which leaves many common faults out of the model’s training data. By fitting your data to a gas turbine model, we can generate synthetic data as if there faults were present, and expand the training data set for your APR software.
-
Detect Unexperienced Failures
Since APR models are typically trained from your operational data, they lack the ability to predict and diagnose faults that you have not experienced before. Our Fault Signature Database allows you to train these unexperienced faults into your APR models.
-
Improve Diagnostics
Since APR models are typically trained from your operational data, they lack the ability to predict and diagnose faults that you have not experienced before. Our Fault Signature Database allows you to train these unexperienced faults into your APR models.