Font Size: a A A

Research On Fault Diagnosis Of Drilling Pump Powertrain

Posted on:2013-09-11Degree:MasterType:Thesis
Country:ChinaCandidate:D Z ZhouFull Text:PDF
GTID:2271330467452853Subject:Vehicle Engineering
Abstract/Summary:PDF Full Text Request
The drilling pump plays a crucial role in the process of drilling, and the normal running of its power end has guaranteed the power resource of the pump’s normal work. So the diagnosis is significant no matter to the pump or to the whole drilling process.Firstly, before the specific analysis of the power end of the drilling pump failure was started, the literature obtained the vibtation signal of drilling pump power end was non-stationary and low noise ratio. So de-noise and improve the noise ratio were needed before dealing with the signal. Then studied the method of wavelet de-noising and established evaluation criteria of wavelet de-noising effect threshold forms.Analyzed the de-noising effect of using common nonlinear wavelet effect threshold in different noise ratio and type.The rigrsure showed much better. It did good to the pretreatment of drilling pump fault signal and so the exteaction of characteristics.Secondly, this paper studied the main fault of the drilling pump pwer end (bearings, herringbone gear, crosshead, main bearing vagay main bolt). Characteristic quantities obtained when failure occurs in the vibration signal in the frequency shown by numerical modeling and analysis of the failure characteristics of fault vibration signal response.Selected the fault diagnostic methods compatiable with the quantiyies and used the real vibration signal which measured by the well team in site to verify the method. The result showed that the Hilbert transform and Cepstrum Analysis which got in empirical mode decomposition had good effect on bearings failure analysis. Fault features could be obviously found after refining and analyzing the spectrum which was thinned and picked out when it passed the band-pass filter. The accuracy rate of Composite characteristic that composed of time-domain characteristic and energy got in wavelet decomposition could be more than90%after RBF neutral network fault identification. The two dimensional holospectrum had obvious advantage to the main bearing bushing and main bolt failure diagnosis. At the same time noise and temperature signals as a secondary analysis of fault diagnosis was used. Finally these methods were combined to the overall scheme of the set of a drilling pump power end fault diagnosis.In the end, this paper had developed a set of implemental overall program designs, which combined analysis method with field experiment of drilling pump power end failure diagnosis.The designs included acquisition program design, fault diagnosis principle and process design, diagnostic software program design.
Keywords/Search Tags:drilling pump, power end, fault diagnosis, waveletde-noise, vibration responsemodel, Action threshold
PDF Full Text Request
Related items