Measurement Error Influence on Gas Turbine Operability for Condition-Based Maintenance and Reliability/Availability Improvement

Abstract

Nowadays the major oil & gas companies are forced to move to a condition-based maintenance in order to extend mean time between maintenance (MTBM) and increase productivity, predicting gas turbine behavior and preventing unexpected failures. Gas turbine operability is influenced by many technological limitations, such as high temperature exposure limit of hot components, requested fuel composition flexibility, combustion chamber pulsations, pollutant emissions, rotor vibrations, compressor surge limit etc. In order to prevent exceeding these limitations certain significant measurable variables are monitored and controlled during machine operation; the instrumentation accuracy and stability have a strong impact on gas turbine reliability, availability and control. The aim of this paper is to demonstrate how the accuracy and stability of the instrumentation installed on a gas turbine has a direct influence on its operability. Monte Carlo simulations have been carried out to assess each instrument deviation effect and the level of impact. The second phase is oriented towards establishing an improvement process for a safe and reliable machine operation. Among the above-mentioned critical limits the main focus of this analysis is the firing temperature estimation and its control, which is performed by measuring gas turbine exhaust temperature and compressor discharge pressure. This study has highlighted how an error on compressor discharge pressure measurement during unit operation may provoke gas turbine over-firing or under-firing, respectively causing fast life reduction of hot components and the limitation of maximum delivered power. Furthermore, an error on compressor discharge pressure measurement would strongly decrease firing temperature estimation confidence, which is the main parameter for components residual life assessment to target a condition-based maintenance approach. In this paper the corrective actions to increase the accuracy of firing temperature estimation and control will be described. Based on the results of this study, reliability and availability improvements will also be presented. Using the present work as a guideline, similar analyses can be carried out for all other critical aspects that can be related to measurable variables.