CN104819841A - Built-in-coding-information-based single sensing flexible angle-domain averaging method - Google Patents

Abstract

Based on the single-sensing flexible angle domain averaging method with built-in coding information, firstly, the encoder is used to read the angular position signal of the test shaft in the mechanical equipment, and the polynomial fitting method is used to obtain the transient motion information, and the third-order difference Jitter signal and the first-order difference are obtained For the speed signal, calculate the angle domain equal interval sampling signal of the Jitter signal and the fluctuation of the speed signal, judge whether the signal is a stable signal by setting the speed fluctuation threshold, and resample the non-stationary signal exceeding the threshold range, otherwise As a smooth signal processing, the original j (t) signal is used as an equiangular interval sampling signal, and the flexible angle domain average method based on Chirp-Z transformation is adopted to determine whether there is a fault in the mechanical equipment. The present invention controls the experimental data to a large extent. The number of equipment simplifies the data collection procedure, reduces the test cost, benefits from the automation of fault feature extraction and diagnosis and monitoring, saves time and has higher efficiency.

Description

A Single-Sensor Flexible Angle Domain Averaging Method Based on Built-in Coding Information

technical field

The invention relates to the technical field of mechanical equipment fault diagnosis, in particular to a single-sensing flexible angle domain averaging method based on built-in coded information.

Background technique

Vibration analysis is one of the most effective ways to diagnose mechanical equipment faults at this stage. In some occasions, vibration sensors are difficult to install due to environmental and working conditions. The CNC machine tool splashes a lot of metal chips and coolant during the working process, which will cause damage to the vibration sensor. For manipulators, robots and other equipment, their complex motion postures bring difficulties to the installation and wiring of vibration sensors. In addition, large-scale machining centers require a fully enclosed working environment during service, which makes it impossible to install vibration sensors. Therefore, in the case of unavailable vibration information, building a new diagnostic information source has become an urgent problem to be solved in mechanical fault diagnosis at this stage. With the development trend of automation and intelligence of mechanical equipment, encoders have been widely configured on mechanical equipment as built-in sensing units, and are used for position feedback of motion control. They have high measurement accuracy, fast response speed, wide measurement range, and The advantages of good reliability, non-contact measurement and easy control. The built-in coded information is a digital signal converted from physical quantities such as angular displacement, rotational angular position, and angular velocity on the mechanical rotation test shaft. As a new diagnostic information source, it has broad application in the fault diagnosis and health monitoring of mechanical equipment. Application prospect.

However, the original output signal of the encoder is a complex multi-component coupling signal, which not only includes the transient speed shock caused by early faults, but also includes the normal speed fluctuation caused by the load change and the time-varying stiffness of the gear. And the amplitude of the latter is often stronger, which brings difficulties to the extraction of early fault features. Time-domain averaging technology is a traditional mechanical fault diagnosis technology, which can extract the periodic components of interest in the signal and improve the signal-to-noise ratio, but it is only suitable for constant speed conditions. In fact, many important Its operating speed is often non-stable under work requirements or conditions. Compared with the steady state, the vibration signal of mechanical equipment under variable speed becomes more complicated, which greatly increases the difficulty of extracting fault features. In 2013, French scholar Leclère et al. proposed the angle domain averaging technology for fault diagnosis of internal combustion engines and gears at variable speeds. However, the angle-domain averaging method still relies on the cooperation of the key-phase signal, which requires the installation of additional sensors and increases the cost and difficulty of testing, and its averaging algorithm cannot avoid the adverse effects of truncation errors on the signal-to-noise ratio, so only a single sensor is used for testing At the same time, the angle-domain averaging method that can avoid the truncation error interference is of great significance.

Contents of the invention

In order to overcome the above-mentioned shortcomings of the prior art, the object of the present invention is to provide a single-sensing flexible angle-domain averaging method based on built-in coding information, so as to realize fault feature extraction and diagnostic monitoring of mechanical equipment under stable and variable speed conditions.

In order to achieve the above object, the technical scheme that the present invention's solution takes is:

A single-sensing flexible angle domain averaging method based on built-in coding information, including the following steps:

Step 1: Use the encoder data acquisition card to read the built-in encoder signal in the mechanical equipment, perform high-frequency sampling and preprocessing on the signal, and obtain the angular position signal of the test shaft, which is recorded as x(t);

Step 2: Use the polynomial fitting method to obtain the transient motion information of the mechanical equipment, denote the Jitter signal obtained by polynomial fitting and third-order difference of x(t) as j(t) and the first-order difference to obtain the speed signal v(t) ;

Step 3: Calculate the angular domain equidistant sampling sequence y(n) of the Jitter signal j(t), first calculate the rotational speed fluctuation of the rotational speed signal v(t), the rotational speed fluctuation is defined as the standard deviation of the rotational speed signal divided by the average rotational speed, By setting the threshold to 0.5%, it is judged whether it is a stable signal. The signal exceeding the threshold range is a non-stationary signal, and the Jitter signal j(t) needs to be resampled in the angle domain, otherwise it will be processed as a stable signal, and the original j(t) signal That is, as an equiangularly spaced sampling signal;

Step 4: Using the flexible angle domain averaging method based on Chirp-Z (Chirp-Z Transform, CZT) transformation, CZT sets the comb filter in the frequency spectrum to retain the order of interest to avoid truncation errors and obtain the desired output The frequency sampling value of the order is obtained by the following formula:

$\mathrm{<span\; class="notranslate">\; CTZ</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&Center\; Dot;</span>{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; AW</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; )</span>}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}}$

In the formula, y(n) is an equidistant sampling sequence in the angle domain, N is the length of the sequence, A is the polar coordinate value of the initial sampling point, W is the frequency interval between sampling points, k is the corresponding order,

Then arrange the sampling values in the frequency domain into a vector array in the form of discrete Fourier, and then accurately obtain the average result in the angle domain of the desired order through the inverse discrete Fourier transform, so as to determine whether there is a fault in the equipment.

Compared with the prior art, the present invention has the following beneficial effects:

a) The present invention can obtain the torsional vibration signal reflecting the health status of the machine through the angular position signal measured by the encoder with high precision, that is, the Jitter signal. Jitter is an effective signal that can be used for early fault detection of the machine, and can truly reflect the The kinematic and dynamic characteristics of the system.

b) The encoder itself is a high-precision key phase device, which can resample the Jitter signal in the angle domain under non-stationary conditions to eliminate the influence of speed changes, and is suitable for large speed fluctuations.

c) On the basis of the traditional time-domain averaging method, the present invention creatively proposes a flexible angle-domain averaging method based on Chirp-Z transform. CZT can efficiently and accurately retain the order of interest by setting a comb filter in the frequency spectrum To avoid truncation error and improve the signal-to-noise ratio.

d) The present invention only utilizes the built-in single encoder information of the mechanical equipment, can accurately extract the fault information under the working condition of the equipment, controls the number of experimental equipment to a large extent, simplifies the data collection procedure, reduces the test cost, and is beneficial to realize The automation of fault feature extraction and diagnostic monitoring saves time and improves efficiency.

Description of drawings

Fig. 1 is a schematic diagram of the structure of the test bench of the embodiment.

Fig. 2 is the root crack fault of the planetary gear in the embodiment.

Fig. 3 is a flow chart of the method of the present invention.

Figure 4 is the encoder information of the start-stop order of the planetary gearbox intercepted in the example.

Fig. 5 is the Jitter signal obtained by performing third-order difference on the encoder information in the example.

Fig. 6 is the speed information obtained by performing first-order difference on the encoder information in the embodiment.

Fig. 7 is the result of using the method of the present invention for the planetary gear during the steady operation of the planetary gearbox of the embodiment.

Fig. 8 is the result of using the method of the present invention for the planetary gear during the starting and stopping process of the planetary gearbox of the embodiment.

Fig. 9 is the result of traditional time-domain averaging of the planetary gear during the steady operation of the planetary gearbox of the embodiment.

Fig. 10 is the result of traditional time-domain averaging of the planetary gear during the start-stop process of the planetary gearbox of the embodiment.

Detailed ways

The present invention will be described in detail below in conjunction with the accompanying drawings and embodiments.

The following will take a planetary gearbox planetary gear fault detection test bench as an example to illustrate. The test bench is composed of a drive motor, a coupling, an encoder, a magnetic powder brake, a bearing, and a planetary gearbox, as shown in Figure 1. The planetary gearbox is composed of an inner ring gear 1, a sun gear 2, and three evenly distributed planetary gears 3. The planet carrier is connected to the output shaft. Two encoders are installed at the input shaft and output shaft of the planetary gearbox. The entire device Driven by a motor, the torque is transmitted from the input shaft along the planetary gearbox to the magnetic powder brake at the output end, and the magnetic powder brake completes the loading process.

The specific parameters are as follows: 1) Rated power of drive motor: 1.2kW, rated speed: 40Hz; 2) Transmission ratio of planetary gearbox: 5.1:1, number of teeth of inner ring gear 1: 82, module: 1, number of teeth of planetary gear 3: 31 , Modulus: 1, number of teeth of sun gear 2: 20, modulus: 1; 3) The fault type of the planetary gear is root crack, as shown in Figure 2; 4) The torque of the magnetic powder brake under the rated power is 0.06N*m.

The single encoder signal is used to diagnose the fault of the planetary gear in the planetary gearbox, and the invention is applied to analyze the experimental data and compare with the traditional time-domain average method.

As shown in Figure 3, the single-sensing flexible angle domain averaging method based on built-in coding information includes the following steps:

Step 1: Use the encoder data acquisition card to read the built-in encoder signal in the mechanical equipment, perform high-frequency sampling and preprocessing on the signal, and obtain the angular position signal of the test shaft. In order to obtain complete start-stop data, when using the data It is necessary to remove the initial noise part, and intercept the data of 6-10s in the entire signal for a total of 4s, as shown in Figure 4, denoted as x1;

Step 2: Use the polynomial fitting method to obtain the transient motion information of the mechanical equipment, denote the Jitter signal obtained by polynomial fitting and third-order difference of x1 as j(t) and the first-order difference to obtain the rotational speed signal v1, respectively, as shown in Figure 5 and As shown in Figure 6;

Step 3: Calculate the angular domain equidistant sampling sequence y(n) of the Jitter signal j(t), first calculate the rotational speed fluctuation of the rotational speed signal v1, and the rotational speed fluctuation is defined as the standard deviation of the rotational speed signal divided by the average rotational speed, by setting The threshold is 0.5% to judge whether it is a stable signal, and the signal exceeding the threshold range is a non-stationary signal. In this example, it is a start-stop signal of a planetary gearbox. According to the speed signal shown in Figure 6, the standard deviation of the speed signal is 108.8883, and the average speed It is 253.1542, and then the calculated rotational speed fluctuation is 43.01%, which is greater than 0.5%, so it is a non-stationary signal. It is necessary to resample the Jitter signal j(t) in the angle domain. In addition, the comparison signal is set to a stable signal, and the original j( t) The signal is taken as an equiangularly spaced sampling signal;

Step 4: In order to obtain the weak fault information in the encoder signal and improve the signal-to-noise ratio, a flexible angle domain averaging method based on Chirp-Z transform is adopted. CZT efficiently and accurately divides the order of interest by setting a comb filter in the frequency spectrum. Retain to avoid truncation error. In this planetary gearbox fault diagnosis example, the meshing frequency (since the number of teeth of the planetary gear is 31 teeth, so the meshing frequency is 31st order) is reserved for the left and right 5th order of the side frequency to obtain the expected output The frequency sampling value of the order is obtained by the following formula,

$\mathrm{<span\; class="notranslate">\; CTZ</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&Center\; Dot;</span>{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; AW</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; )</span>}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; no</span>}}$

In the formula, y(n) is the sampling sequence at equal intervals in the angle domain, N is the intercepted data length of 4 seconds, A is the initial sampling order corresponding to the meshing order minus 5 (that is, 26th order), and W is the sampling interval is the order of unit 1, k is the value of the corresponding order is 26, 27, 28...36,

Then, each frequency domain sampling value is arranged into a vector array in the form of discrete Fourier, and the angle domain average result of the desired order can be accurately obtained by further inverse discrete Fourier transform, as shown in Fig. 7 and Fig. 8. 7 is the average result of the planetary gear flexibility angle domain when the fluctuation threshold is set to 0.5% during the stable operation of the planetary gearbox. The average result of planetary gear flexible synchronization, while Figure 9 and Figure 10 are the results obtained by using the traditional time-domain averaging method. Figure 9 is the diagnostic result of time-domain averaging of single encoder signals during the smooth operation of the planetary gearbox; Figure 10 It is the diagnosis result of the time-domain average of the single encoder signal during the start-stop process of the planetary gearbox.

Comparing the two methods under the two conditions, it can be clearly seen that the method proposed in the present invention can not only show better results than the traditional time-domain averaging method under steady conditions, but also can diagnose the planetary gearbox The root crack fault of the planetary gear that cannot be diagnosed by traditional time-domain averaging during the start-stop phase. It can also be seen from the examples that the method of the present invention can also artificially adjust the fluctuation threshold of the specified analysis variable, and change its size within a certain range in combination with the actual situation, which has a certain degree of redundancy. In order to improve the operation efficiency when identifying different degrees of faults.

Claims (1)

1. the patrilineal line of descent with only one son in each generation based on built-in coded message feels flexible angle domain averaging method, it is characterized in that, comprises the following steps:

Step one: utilize scrambler number to adopt card and read built-in encoder signal in plant equipment, high frequency sampling and pre-service are carried out to signal, obtains the angle position signal testing axle, be designated as x (t);

Step 2: adopt polynomial fitting method to obtain the transient motion information of plant equipment, carries out Jitter signal that fitting of a polynomial and third order difference obtain and is designated as j (t) and first order difference obtains tach signal v (t) by x (t);

Step 3: angle domain equal interval sampling sequences y (n) calculating Jitter signal j (t), first calculate the fluctuation of speed amount of tach signal v (t), fluctuation of speed amount is defined as the standard deviation of tach signal divided by mean speed, be 0.5% determine whether stationary signal by setting threshold value, the signal exceeding threshold range is non-stationary signal, need to carry out angle domain resampling to Jitter signal j (t), otherwise as stationary signal process, former j (t) signal i.e. conduct angularly interval sampling signal;

Step 4: adopt the flexible angle domain averaging method based on Chirp-Z conversion, order interested, by arranging comb filter in frequency spectrum, carries out retaining and avoids truncation error by CZT, and the frequency sampling value obtaining desired output order is obtained by following formula:

$\mathrm{CTZ}\left(y\left(n\right)\right)=\underset{n=0}{\overset{N-1}{\mathrm{\&Sigma;}}}y\left(n\right)\&CenterDot;{\left({\mathrm{AW}}^{-k}\right)}^{-n}$

In formula, y (n) for angle domain equal interval sampling sequence, N be the length of sequence, A is the polar value of starting sample point, and W is the frequency interval between sampled point, and k is corresponding order,

Then each frequency domain sample value is arranged in vector array according to the form of discrete fourier, changes the angle domain average result that accurately can obtain and expect order further by inverse discrete Fourier transform, thus determine the fault whether plant equipment exists.