Research On The Key Technologies Of On-line Detection Of Wear Debris In Lubrication

The gearbox in the mainstream doubly-fed wind turbine undertakes the function of power transmission and speed transform,and is also one of the main components with the most faults at present.Its reliability directly affects the normal operation of the wind turbine and the profit of the wind farm.At present,online vibration monitoring is often used to monitor the health of the gearbox and to locate the fault using spectrum analysis.It has also achieved good results.Gearbox failure caused by the gear wear,vibration method is not sensitive to early gear wear,wear only when the tooth surface more serious vibration through the perception.When the gear wear occurs,the abrasive particles into the circulating oil,through the detection of abrasive oil can be an earlier understanding of the gear box conditions,to achieve early warning.The oil abrasive particle sensor in the paper is based on the principle of electromagnetic induction and can be used to detect wear particles in the gearbox on-line.The sensor is mainly composed of sine excitation source,excitation and induction coil,rectifier and filter circuit,signal processing circuit and communication module.Among them,the design of excitation and induction coil structure,sine excitation source,weak signal processing methods directly affect the detection sensitivity and stability of the final sensor,the paper intends to study the following key technologies.First of all,the structural design of the coil has the greatest impact on the sensitivity of the abrasive inspection.This article explores some of the structures that can improve the inspection capability.In this paper,the electromagnetic field simulation software Maxwell + Simplorer co-simulation software,the existing double excitation single induction coil structure simulation analysis,and on this basis,to improve the arrangement of the coil structure and different electromagnetic factors under the structure of the simulation,and ultimately get a Double excitation double induction coil structure with magnetic permeability medium with higher detection sensitivity.Secondly,in view of the long time running of the traditional RC oscillator circuit,the oscillation resistance and the capacitance value change due to temperature drift,etc.,resulting in the change of its output frequency,resulting in the increase of the signal noise and the influence of the detection sensitivity.In this paper,a direct digital frequency synthesis(DDS)method is proposed to generate a pure and stable sinusoidal excitation source.By simulating and analyzing the symmetry of the sine wave and the Sunderland algorithm,the look-up table in DDS is compressed to effectively suppress the spurs Signal generation.The use of single chip and DDS integrated circuit design of the corresponding circuit,after testing,this method will not lead to temperature drift disorders.Again,in order to improve the sensitivity of detection,we can identify the abrasive signals submerged in the noise.In this paper,we use the mallat algorithm in wavelet transform to design the wavelet threshold filtering.By using MATLAB to analyze different base wavelet,wavelet decomposition level and threshold Select the comparative analysis to get the best parameters of the abrasive signal processing characteristics.Finally written in C language program to achieve DSP,the filtering effect is obviously better than five or three times smoothing filter.Finally,the host computer monitoring interface is programmed in Lab VIEW to realize the communication with the grit sensor,which can receive the grit classification data sent by the grit sensor and has the functions of trend display and alarm.After testing,the abrasive particle sensor detection sensitivity and operational stability than before has been significantly improved.