Session: 06-03 Pressure gain combustion and propulsion cycles I

Submission Number: 179173

A Comparative Study of Time-Series Models for Gas Turbine Inlet Filter Condition Prediction 

In the energy field, availability is the main factor for power plant units, especially during high-demand periods. In Brazil, the price of electricity depends on its energy sources. Renewable sources, such as hydroelectric and wind, are always dispatched. However, when they cannot meet the country’ s demand, thermal power plants are dispatched, and the energy price increases. To maximize thermal power plant availability during high-demand periods, it is necessary to implement condition-based maintenance (CBM) concepts. A good strategy is to monitor equipment degradation, which allows for increasing the period between maintenance intervals and planning comprehensive interventions, thereby optimizing costs and preventing unavailability during critical periods. The present study uses time-series forecasting with different approaches, including Autoregressive Integrated Moving Average (ARIMA), Multi-Layer Perceptron (MLP), Convolutional Neural Networks (CNN), and Long Short-Term Memory (LSTM), to forecast how differential pressure in the inlet air filters influences gas turbine performance and to estimate the inlet air filter saturation. The studies were carried out using real engine data, and the simulations were performed using an in-house computer program specially developed in the Python programming language. The simulation results showed that the MLP approach presented the lowest error values and the least dispersion of error values across the experiments reported in this work. Using this approach, it is possible to predict the saturation of the inlet air filters, thereby enabling the operating and maintenance team to accurately forecast the inlet air filters’ remaining useful life.

Presenting Author: Ivan Da Costa Vieira Petrobras

Presenting Author Biography: Engineer of Control and Automation graduated from the Federal University of Minas Gerais (UFMG). In Belo Horizonte, he worked, among other roles, with software development and at an automation integration company, providing services to Vale.
He has been an Equipment Engineer (Electronics) at Petrobras since August 2010. He worked at the Cubatão Thermoelectric Power Plant and the Nova Piratininga Power Plant as an Automation Engineer in the areas of industrial instrumentation and control and automation of electric power Turbogenerators and Steam Generating Boilers. Furthermore, he served as a Maintenance Manager at the Nova Piratininga Power Plant and the Ibirité Power Plant.
He completed a Master's degree at the Aeronautics Institute of Technology (ITA), specializing in Gas Turbines.

Authors:

Ivan Da Costa Vieira Petrobras
Cleverson Bringhenti Aeronautics Institute of Technology - ITA
Jesuíno Takachi Tomita Texas A&M University
Alexandre Mendonça Krul Aeronautics Institute of Technology - ITA
Antonio Bruno De Vasconcelos Leitão Aeronautics Institute of Technology
Franco Jefferds Dos Santos Silva Aeronautics Institute of Technology - ITA