60. Assessing the Condition of Hydro Turbines by Analyzing Operational Data

Pumped-storage hydro power plant plays an important role in power system operation. Its main roles are to meet peak and emergency demand of electricity, to maintain system frequency and to realize economical operation of whole power system. However, equipment in pumped-storage plant tends to deteriorate with early age, because of its large capacity, high head and strenuous operational condition such as frequent start and stop. For this reason, development of condition based maintenance technology is highly required especially for a pumped storage plant. Conventionally, condition of a hydraulic turbine and generator is checked through routine inspection performed by operation staff with his sense and by analysis of operation data collected by CMMS (Computerized Monitoring and Management System). CMMS is quite effective to check the machine health because it can obtained real time data from sensors installed on each part of the machine such as temperature of bearing metal, oil level, pressure, vibration and so on. It usually raises an alarm if the data exceed a certain threshold value. The purpose of these activities is to detect abnormalities at an early stage. However, it is difficult by following reasons. (1) Operational data have complicated characteristics because equipment in pumped storage power plant repeats start & stop and its output fluctuates continuously. So, threshold value tends to be set far enough from normal operation data to avoid unnecessary alarm. (2) The data collected through routine inspection can not be compared simply with the past data because the data is fragmentary information at a certain operational situation such as operation time, output and so on. In order to solve these problems, we collected and analyzed many operation data of hydraulic turbines by using the advanced CMMS which was installed in Okutataragi pumped-storage power plant, located in Hyogo Prefecture, Japan. As a result, effective sets of data for comprehension of the machine condition are selected. This paper describes a practical method to monitor machine condition and some interesting examples of correlation among certain operation data.