Font Size: a A A

The Fault Diagnosis And Life Prediction Of Type 3NB-1300C Drilling Pump

Posted on:2008-11-26Degree:MasterType:Thesis
Country:ChinaCandidate:H M LuoFull Text:PDF
GTID:2121360218963524Subject:Safety Technology and Engineering
Abstract/Summary:PDF Full Text Request
Drilling pump is an important piece of equipment in drilling engineering. So its working conditions monitoring and remaining life predictions are of great importance to drilling engineering. In this dissertation tracking test was carried out on five type 3NB-1300C drilling pumps, which are at work in one of the drilling engineering companies of SINOPEC. Their vibration signals were acquired. Combined with the historical data acquired on fault pumps before, lots of parameters were calculated like time-domain parameters, frequency-domain parameters, wavelet packets energy and so on. Characteristic parameters, which can most reflect the working conditions of pumps, were determined through comparison and selection between all these parameters. Sample of relationships between remaining life and characteristic parameters was built, and the sample model was trained using artificial neural network. Weight and threshold, which can be applied to the remaining life predictions of these drilling pumps, were obtained. Finally, the remaining life predictions system of drilling pumps based on artificial neural network was developed, mainly aimed to predict the remaining life of seven bearings, which were the key parts of drilling pumps. In this dissertation the function of condition monitoring and fault diagnosis was improved. With the application of fuzzy mathematics, frequency-domain analysis and diagnosis were carried out on vibration signal, which was acquired after demodulated resonance on bearings in the power end. Valuable vibration law, general characteristics and fault characteristics of tested parts were obtained. In this dissertation Matlab of MathWork and LabWindows/CVI of NI were well combined to achieve the integrated programming. CVI is adapted as a developing tool, thus more satisfactory operation and friendlier user interface were achieved. The system developed can monitor the working conditions of important parts on drilling pumps and predict the remaining life of bearings on line, which can facilitate the management and maintenance of equipment.
Keywords/Search Tags:Drilling Pump, Life Prediction, Wavelet Packets Analysis, Characteristic Parameters, Neural Networks
PDF Full Text Request
Related items