<p>The blades of steam turbines, gas turbines, and compressors in thermal power plants are deteriorating in the long-time operation. The periodical maintenance, repair, and overhaul (MRO) are essential to maintain a stable power supply and its high efficiency. Thus, plants need to be shut down for a number of weeks during the MRO, replacing the power by alternative powers to maintain the electric supply. Generally, older and lower efficient power plants usually stopped are put into operation. It results in wasting additional fuels, increasing the cost. In Japan, thermal power plants are currently indispensable to suppress the electric invariance due to renewable energies such as solar and wind powers. Such a situation forces the thermal power plants to work under off-design conditions. The off-design operation obviously accelerates the blade deterioration, which may lead to sudden failure. Therefore, developing a method to avoid unexpected failures is crucial and quite valuable. In this study, a method to detect the failure signs of an intermediate-pressure steam turbine is presented.</p><p>The analysis of huge amounts of big time-series is required in order to detect the signs of machinery failures. However, it would take a huge amount of time to collect a variety of data for various blade conditions in the actual operating steam turbine. Moreover, getting abnormal data or failure data from the operation is quite difficult because of the necessity of long-time monitoring. This paper proposes a classification method of the big time-series data alternatively collected from the numerical simulation. First, the time-series data of various cases including usual cases and abnormal cases of the actual intermediate-pressure steam turbine operation are obtained from the numerical simulation and stored as a database. Then, the target range of the data is reduced and classified using K-means. The classification results are used to judge whether the state of the turbine operation is normal or abnormal. Finally, the proposed classification method can appropriately classify them into a usual case or abnormal case. Therefore, this method can decide appropriate MRO timing by judging the blade condition in intermediate-pressure steam turbines.</p>

Detection of Machinery Failure Signs From Big Time-Series Data Obtained by Flow Simulation of Intermediate-Pressure Steam Turbines

Paper Proceedings: https://asmedigitalcollection.asme.org/GT/GT2021/volume/84898

Turbo Expo 2021

Supporting the need among researchers worldwide for additional online and multimedia content, ASME now offers the ASME Conference Video Collection. This platform will collect and provide long-term access to conference presentation videos, encompassing all technology areas within ASME's technical conference and events program. About 30,000-40,000 engineers across all mechanical engineering and related disciplines participate in ASME conferences each year. This includes authors of approximately 6,000 to 8,000 technical papers published by ASME, the majority of whom also submit video presentations for our virtual conferences. The video collection will be available as an add-on to existing product packages or as a separate subscription.