PROGNOSIS - Predictive Diagnostic and Prognostic Tools Integration for Decision Support Systems in Gas Turbine Fleets
In this project, the consortium based on MDU, Saab, and Siemens will explore the use of learning systems and artificial intelligence combined with physics-based and expert-driven predictive tools for intelligent and automated decision support systems.
Predictive and real-time analysis of turbomachinery health conditions allows to increase operation efficiency, optimize maintenance scheduling, and avoid unexpected failures. This is important for power generation plants and even more critical for aero-engines.
Operators and manufacturers have a strong need for simulating and monitoring the condition of their individual unit or fleets. Diagnostics and prognostics currently relies heavily on human expertise, which may lead to human error, delayed solutions, and sub-optimal decisions.
In this project, the consortium based on MDU, Saab, and Siemens will explore the use of learning systems and artificial intelligence combined with physics-based and expert-driven predictive tools for intelligent and automated decision support systems. In particular, fleet data will be used through machine learning methods to enhance predictability via cross-comparison of engines in the same fleet and analysis of historical data.
The final aim is a prompt and accurate estimation of key component remaining-life for improved operations and optimal maintenance planning. The project will demonstrate the capability of a diagnostics and decision support system to improve decision-making on real power plants and aero-engines and to increase the knowledge of the industrial partners on their fleets.
The project goal is to implement and test an intelligent decision support system for gas turbine
applications and fleet analysis based on the integration of advanced diagnostic and prognostic
tools. The decision support system will be able to automate tedious tasks with the supervision of
the human expert, identify the root cause of performance deviation, make reliable predictions within
the fleet, and provide suggestions to operators and service engineers regarding optimal operations
and maintenance schedule.
Activities included in the project
The project will address the following scientific challenges:
- Fuse diagnostic information and multiple sensors to identify the root cause of detected deviations in performance.
- Fuse diagnostic information and prognostic prediction within a fleet of different engines to enable continuous forecasting
of system performance and decision-making.
- Apply machine learning techniques to develop an intelligent decision support system that learns from human experience.
- Identify and quantify the benefit of using historical fleet data for anomalies identification and prognostics.