Whole-Condition Dynamic Alarm Research of Hydropower Unit Operating Parameters

Abstract:

Set the alarm value in the condition monitoring system of hydropower unit to monitor unit state parameters, if the difference between the two values exceeds the safety threshold, the alarm works, it is an effective method. Now the state parameters alarm values are set up mainly for the stable operation area in the unit condition monitoring system, which were not suit for low load conditions and in the process of load. To realize the full condition dynamic alarm of hydroelectric generating units, they are two ways: (1), build unit state parameter prediction model, compared predict values with the measured values. (2), set the alarm line, when the measured value exceeds the alarm line, the alarm works,. These two methods are considered the working condition head and output .This paper introduces several different alarm line settings, such as line alarm, partition line alarm, dynamic alarm curve and dynamic alarm surface.

• Periodical: Applied Mechanics and Materials (Volume 607)
• Pages: 677-680
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.607.677
• Online since: July 2014
• Authors: Xin Feng Ge*, Yuan Feng, Yi Zhou Liu, Yu Quan Zhang, Yuan Zheng
• Keywords: Dynamic Alarm, Hydropower Unit, Parameter Prediction, State Monitoring

References

[1] WANG Tie-sheng; ZHANG Li-ping, Application of immune wavelet network model to fault diagnosis of hydro-turbine generating units, [J]. Journal of hydraulic engineering, 2009, (06) (In Chinese).

[2] GUI Zhong-hua, HAN Feng-qin. NEURAL NETWORK BASED ON WAVELET PACKET-CHARACTERISTIC ENTROPY FOR FAULT DIAGNOSIS OF DRAFT TUBE[J]. Proceedings of the CSEE, 2005, 25(4): 99-102. (In Chinese).

[3] OU Jian , SUN Caixin , BP Neural Network Based on Genetic Algorithm for Vibration Fault Diagnosis of Turbine-generator Set [J] . Voltage Technology, 2006, 32(7): 46-48, 68. (In Chinese).

[4] Artificial neural network tutorials, [M], university of California press, Beijing post and telecommunications 2006，P67.

[5] Neural network theory and MATLABR2007 [M], Electronics industry press, 2007. (In Chinese).

[6] MATLAB neural network 30 case analysis [M], Beijing aerospace university press, 2010(In Chinese).