CN105510010A - Characteristic parameter model for rotating machinery in misaligned structural anomaly state - Google Patents

Abstract

The invention discloses a characteristic parameter model for the rotating machinery in the misaligned structural anomaly state and belongs to the diagnostic field of mechanical faults, which mainly solves the fault detecting and fault type distinguishing problem during the fault occurrence early stage of the rotating machinery. Through analyzing the low-frequency spectrum characteristics of the rotating machinery in the misaligned structural anomaly state, associated characteristic parameters are constructed. In this way, the vibration characteristics of the rotating machinery in the above state can be reflected. Based on a resolution index (DI), the suitability of the characteristic parameter model and the resolution sensitivity of the characteristic parameter model are evaluated. Meanwhile, the problem that the fault type cannot be determined only based on conventional characteristic parameters can be solved.

Description

A kind of for the abnormal characteristic parameter model for misaligning of rotary machine configuration

Technical field

The present invention relates to a kind of for the abnormal characteristic parameter model for misaligning of rotary machine configuration, belonging to mechanical fault diagnosis field.

Background technology

Structural anormaly is also referred to as structural failure, and namely rotary machine breaks down, and at the low frequency region place indicating characteristic frequency spectrum of excited frequency.Structural anormaly causes the axle portion of rotating machinery overworked, is also the main cause causing bearing, gear etc. to break down.That is, structural anormaly may cause mechanical system to break down, and may cause huge production loss.Therefore, detection and specification configuration fault are very important to guarantee production efficiency and factory safety.In some sense, feature extraction can be described as the bottleneck problem in the research of Current mechanical fault diagnosis, and it is directly connected to the accuracy of fault diagnosis and the reliability of fault early prediction.In actual applications, often there is not simple relation one to one between fault and sign, if the corresponding multiple sign of a kind of fault possibility, otherwise a kind of sign also may caused by various faults, and these extract to fault signature and bring difficulty.

For a certain concrete fault type, its feature is showed by some physical parameter usually, and there is certain contacting with the strong and weak situation of each physical parameter, as long as the state of mechanical system changes, just inherently have influence on each dynamic physical parameter be associated with it, concern is comparatively wide, and these sensitive to fault, reliable and stable physical parameters are then used to build corresponding characteristic parameter model.Although the generation of certain fault type can cause the change of multiple physical parameter, the parameter that can be used as fault signature is limited, and Failure Characteristic Parameter model must possess high susceptibility, reliability and practicality.

Misaligning is the structural failure that the axle portion of rotating machinery often occurs.What is called misaligns, and refers to that two shaft centre lines with shaft coupling links up exist deviation, and as produced axis being parallel skew, axis produces cheap or both combination in angle.Though dimensionless group can the stage carries out fault diagnosis in early days for traditional characteristic parameter model such as margin index, flexure index, kurtosis index etc., be difficult to determine fault type.

Summary of the invention

Based on the technical matters of characteristic parameter model existence in the past, the present invention proposes a kind of novel characteristic parameter model for rotary machine configuration extremely for misaligning.Being the spectrum signature at low frequency place under condition of misalignment by analyzing rotary machine configuration abnormal, building 2 higher harmonics rates, direction of principal axis vibration index, vertical and horizontal amplitude ratio three parametric models.

Wherein, 2 higher harmonics rates are the spectrum value of 2 times of gyro frequencys that horizontal or vertical direction detects and gyro frequency

Wherein, the ratio of the spectrum value sum on the odds ratio turning the spectrum value sum under the spectrum value and gyro frequency 2 ~ 20 times of order harmonic frequencies that frequently component ratio is the gyro frequency that horizontal or vertical direction detects under the spectrum value of normal condition gyro frequency and gyro frequency 2 ~ 20 times of order harmonic frequencies, namely

Higher harmonics ratio be the spectrum value of 3 times of gyro frequencys and the spectrum value of gyro frequency that horizontal or vertical direction detects odds ratio on the ratio of the spectrum value of normal condition 3 times of gyro frequencys and the spectrum value of gyro frequency, namely $N_2=\frac{T_d\left(3f_r\right)/T_d\left(f_r\right)}{T_n\left(3f_r\right)/T_n\left(f_r\right)}.$

The present invention has the following advantages:

This characteristic parameter model resolution is high, practical, reliability is high.

Accompanying drawing explanation

Fig. 1 is the spectrogram of horizontal direction under centrifugal fan normal condition.

Fig. 2 is the spectrogram of vertical direction under centrifugal fan normal condition.

Fig. 3 is the spectrogram in centrifugal fan normal condition lower axis direction.

Fig. 4 is the spectrogram of horizontal direction under centrifugal fan non-equilibrium state.

Fig. 5 is the spectrogram of vertical direction under centrifugal fan non-equilibrium state.

Fig. 6 is the spectrogram in centrifugal fan non-equilibrium state lower axis direction.

Specific embodiments

Below in conjunction with accompanying drawing and concrete process of the test, the invention will be further described.

The diagnostic test of centrifugal fan: three accelerometers (PCBMA352A60) export from 5HZ to 60KHZ with the bandwidth measurement level of 10mV/g, vertical and axial vibration signal respectively under normal and non-equilibrium state.Vibration signal measured by accelerometer is transformed into signal recorder (scope scrambler DL750) after being amplified by transducer signal conditioner (PCBICP model 480C02).Former vibration signal in a frequency domain respectively as shown in figs. 1 to 6.These signals are all measured with constant speed (600 turns).The sample frequency of signal measurement is 50KHZ, and the sampling time is 20s.

Each parameter value is calculated as following table according to measurement result:

Utilize resolving index (DI) to carry out the resolution sensitivity of evaluating characteristic parameter as quality index, concrete proof procedure is as follows:

Suppose x ₁and x ₂be the value of the characteristic parameter that in measuring state 1 and state 2, signal calculates respectively, and meet normal distribution N (μ respectively ₁, σ ₁) and N (μ ₂, σ ₂).Here, μ and σ is mean value and the standard deviation of characteristic parameter.| x ₂-x ₁| value larger, distinguish the sensitivity of two states by characteristic parameter higher.Due to z=x ₂-x ₁also normal distribution N (μ is met ₂-μ ₁, σ ₁+ σ ₂), there is the following density function about Z:

$f\left(z\right)=\frac{1}{\sqrt{2\mathrm{\π}({\mathrm{\σ}}_1^2+{\mathrm{\σ}}_2^2)}}\exp \left\{\frac{{\{z-({\mathrm{\μ}}_2-{\mathrm{\μ}}_1)\}}^2}{2({\mathrm{\σ}}_1^2+{\mathrm{\σ}}_2^2)}\right\}---\left(1\right)$

Here, μ ₂>=μ ₁(in like manner, work as μ ₁>=μ ₂), probability can be calculated as follows:

$P_0={\∫}_{-\mathrm{\∞}}^{0}f\left(z\right)d_z---\left(2\right)$

Here, 1-P ₀be called as " resolution (DR) ".By following formula:

$\mathrm{\μ}=\frac{2-({\mathrm{\μ}}_2-{\mathrm{\μ}}_1)}{\sqrt{{\mathrm{\σ}}_1^2+{\mathrm{\σ}}_2^2}}---\left(3\right)$

In substitution formula (1) and (2), P ₀obtained by following formula:

$P_0=\frac{1}{\sqrt{2\mathrm{\π}}}{\∫}_{-\mathrm{\∞}}^{-DI}\exp (-\frac{{\mathrm{\μ}}^2}{2})d_{\mathrm{\μ}}---\left(4\right)$

Here, being calculated as follows of DI (resolving index):

$DI=\frac{{\mathrm{\μ}}_2-{\mathrm{\μ}}_1}{\sqrt{{\mathrm{\σ}}_1^2+{\mathrm{\σ}}_2^2}}$ Or $DI=\frac{\stackrel{\‾}{x_2}-\stackrel{\‾}{x_1}}{\sqrt{{\mathrm{\σ}}_1^2+{\mathrm{\σ}}_2^2}}---\left(5\right)$

Clearly, DI value is larger, " resolution (DR=1-P ₀) " value also will be larger, therefore, characteristic parameter also will be better.

According to above-mentioned proof procedure, the DI value of the character pair parameter calculated is as following table:

	M 1	M 2	M 3
				DI N-M	2.17765	10.2071	7.48199

Show the DI value of three characteristic parameters is all greater than 2.17 from above, according to area and the ordinate table of standardized normal distribution curve, as DI=2.17, resolution reaches 98.499%, therefore, three characteristic parameter models all reach more than 98.499% to rotary machine configuration is abnormal for the diagnosis resolution misaligning fault.

Claims (4)

1. the characteristic parameter model for rotary machine configuration extremely for misaligning, it is characterized in that: by analyzing the abnormal spectrum signature for low frequency place under condition of misalignment of rotary machine configuration, specifically comprising the characteristic parameter model of 2 higher harmonics rates, direction of principal axis vibration index, vertical and horizontal amplitude ratio.

2. the characteristic parameter model for rotary machine configuration extremely for misaligning according to right 1, it is characterized in that 2 times described higher harmonics rates be the spectrum value of 2 times of gyro frequencys and the spectrum value of gyro frequency that horizontal or vertical direction detects odds ratio on the ratio of the lower spectrum value of 2 times of gyro frequencys of normal condition and the spectrum value of gyro frequency, namely

3. the characteristic parameter model for rotary machine configuration extremely for misaligning according to right 1, it is characterized in that described direction of principal axis vibration index be the root-mean-square value of vibration signal on the axis direction and horizontal direction that detect odds ratio on the ratio of root-mean-square value of vibration signal in normal condition lower axis direction and horizontal direction, namely

4. the characteristic parameter model for rotary machine configuration extremely for misaligning according to right 1, it is characterized in that described vertical and horizontal amplitude ratio be the root-mean-square value of vibration signal in the vertical direction and horizontal direction that detect odds ratio on the ratio of root-mean-square value of vibration signal under normal condition in vertical direction and horizontal direction, namely