Using Analytics to Assess Health Status of DLE Combustion Gas Turbines

Abstract

The trend towards digitalization is becoming stronger and disruptive. Thanks to the experience gained over the past two decades, rotating equipment OEMs are now able to connect assets from any location, transfer data with cyber secure protocols, run analytics on the fly and manage remotely alerts. The above helps to provide valuable insights to customers, partner with them and drive high availability and reliability, optimize operations and support maintenance decisions (Allegorico, 2014). This paper addresses the problem of assessing the health status of a Dry Low Emissions (DLE) combustion system, one of the most critical components of a Gas Turbine. It describes how the combined use of remotely-acquired operational data and different types of analytics, which represents a digital replica of the system, is used in conjunction with expert's domain knowledge to drive planned and unplanned maintenance decisions. We applied this strategy to an Oil&Gas plant and the results of the integrated service delivered have been observed for several months, providing feedback on the methodology as well as points of reflection for further enhancements. The methodology presented here can be summarized in three phases: "Anamnesis"a set of data-driven models and analysis of historical events to characterize the DLE combustion system of remotely-connected units"Diagnosis"a cognitive investigation of the DLE mapping to detect instability symptoms, instrumentation failures, performance issues and emissions level and to identify the root cause of the combustion issues (hardware or software)"Therapy"Combustion issues resolution and relevant recommendations to support maintenance activities on-site Furthermore, some real cases of combustion problems on the connected fleet and their respective solutions will also be presented.