Ansaldo Energia is Italy’s largest supplier, installer and service provider for power generation plants and components, with the capabilities to build turnkey power plants on green field sites using its own technology and its own independent design, production, construction, commissioning and service resources. Ansaldo Energia has an installed capacity of 218,000 MW in 90 countries, it has around 60 combined cycles (roughly 300 machines) under Long Time Service Agreement which required a continuous monitoring, data elaboration and maintenance. Ansaldo Energia implemented central diagnostic infrastructure about 15 years ago to calculate and monitor on-line gas-turbine, steam-turbine efficiency KPIs, and to provide a condition based maintenance service to its customers. Nowadays the infrastructure has been developed and improved for obtaining a global diagnostic center based in Genoa and Miami, now the data core can integrate and post elaborate data coming from all the worldwide assets. In recent decades, the advances in technology of sensors, instrumentation and control, communication systems, have made available an increasing amount of data from systems, than need to be organized and specifically processed, to exploit the maximum available information. In this presentation, the latest developments of Ansaldo Energia remote monitoring and diagnostic network are described. Starting from the acquisition systems (ADA) with vibration and combustion integrated processing, to the database generation and sharing, until the introduction of predictive analysis systems (APEx). ADA(tm) (Advanced Diagnostic Analysis) is the Ansaldo Energia suite for condition-based maintenance covering all the critical diagnostic needs for energy industry machinery, including: Steam and gas turbine performance monitoring, Gas turbine combustion monitoring, Vibration analysis, Generator diagnostics, Electrical transient and Fast event recorder. Based on its modular design and leveraged by OSIsoft PI infrastructure, ADA allows for advanced monitoring of main equipment parameters like steam and gas turbine performances, gas turbine combustion, machinery vibrations, generator diagnostic, electrical transient and others. Computing modules, automatic report generation, alarms notification, large data storage capabilities are some of the key features of this state-of-the-art product in the field of remote monitoring and diagnostic. Moreover, AEN developed a WEB portal used for sharing diagnostic data with all AEN departments involved on data analysis. APEx (Ansaldo Predictive Expert) is Ansaldo Energia predictive tool used for predictive diagnostic and maintenance of combined cycle power plant component such as Gas Turbine, Steam Turbine, generator and BOP. APEx make use of machine learning algorithms to build diagnostic model for real-time monitoring of many combined cycle power plant operating parameters. The models are trained on historical normal behavior data using a large number of signals as input variables. After the training the model is released for the real-time monitoring to the diagnostic front-end. The system estimates an expected value of the monitored parameters comparing with the measured value. The deviations between the expected and measured values, called residuals, are used to detect incipient abnormal behavior or degradation of the machinery by the generation of Early Warnings. This procedure ensures high promptness level, minimizing false alarms. The system also provides a calculation of time to fault to evaluate the severity of the degradation, estimating the residual operating time before the fault or the trip. The developed APEx models are: combustion model, vibration model and performance model. Through those different models, Ansaldo Energia front-end is able to detect in advance abnormal behavior or underperformance of the machine under investigation such as: burners, compressor and air intake fouling, changes in combustion parameters (emissions and dynamic behavior), machine unbalancing, rotor rubbing, bearing issues, performance loss; in wider terms APEx helps to detect precursors of deviation from expected behavior in order to improve specialist analysis and plan the maintenance activities in advance.In this presentation, case studies of APEx models online monitoring, generation of Early Warnings and time to fault or trip evaluation are described.
