JP4617168B2 - Bearing damage evaluation apparatus and bearing damage evaluation method - Google Patents

Description

  The present invention relates to a bearing damage evaluation apparatus, a bearing damage evaluation method, a bearing damage evaluation program, and a storage medium on which this program is recorded.

  Bearings (especially rolling bearings) are well known as mechanical elements that support rotating shafts. In general, machines are often accompanied by vibration and impact during operation, and bearings cause damage to inner rings and the like as they are used. Therefore, the occurrence of damage is an important factor in predicting the life of the bearing, and the life of the bearing can be evaluated by evaluating the damage.

On the other hand, if the bearing damage occurs, the value of the load acting on the bearing (rolling element) also fluctuates (changes). Therefore, if the load acting on the rolling element can be known, the life of the bearing (remaining life) can be increased. A bearing damage evaluation apparatus or a bearing damage evaluation method has been proposed (for example, see Patent Document 1).
JP 2002-257797 A

  However, in such an evaluation method, it is important to determine whether or not the measurement waveform is abnormal (hereinafter referred to as “abnormality”). On the other hand, the measurement value at the rising or falling edge of the waveform constantly changes. Since it is necessary to make a judgment, it is very difficult to judge the abnormality. Since the stable region is very small even near the peak of the waveform, the same conditions can be said.

  Specifically, quantitative judgments such as securing standard echo patterns (waveforms) that serve as criteria for judging abnormalities, calculation processing for judging abnormalities, and methods for setting thresholds for judgment are objective. A method or system to do it was essential.

  Therefore, the object of the present invention has been made in view of such circumstances, and its task is to secure a criterion for judging abnormalities, establish a detection method in an evaluation apparatus, specify an abnormal range, and determine the degree of abnormality. It is to provide a bearing damage evaluation technology that can grasp and accurately evaluate a bearing damage. In particular, it is to provide a bearing damage evaluation technique capable of quantitatively evaluating bearing damage not only qualitatively.

  As a result of earnest research, the inventor has found that the above object can be achieved by a bearing damage evaluation apparatus, a bearing damage evaluation method, a bearing damage evaluation program, and a storage medium in which the program is recorded. It came to complete.

The present invention generates ultrasonic waves from an ultrasonic probe attached to a bearing housing on which a bearing is supported toward the bearing, and from the boundary between the bearing housing and the outer ring of the bearing or the outer ring of the bearing and a rolling element. Is a bearing damage evaluation apparatus that evaluates the damage generated in the bearing by the echo height ratio H calculated by the following equation:
H = (1-h / ho) × 100
Here, h is the echo height when the bearing load is acting, and ho is the echo height obtained when there is no load when the bearing load is not acting.
An echo height ratio calculating means for obtaining an echo height ratio from the reflected wave received by the ultrasonic probe;
From the obtained echo height ratio, a waveform signal having an echo height ratio using time as an index is obtained, and only the echo height ratio equal to or greater than a predetermined threshold value Ho is used as an object of calculation, which is the primary of the slope of the waveform signal. A difference value calculating means for obtaining a secondary difference value that is a difference between the difference value and the primary difference value, and further prominently by squaring the difference value to the second power or the fourth power;
Average value calculating means for obtaining a moving average value of the primary difference value and the secondary difference value of the echo height ratio;
The value of the moving average value of the primary difference value, which has become noticeable by the fourth power after further quadratic difference, or the moving average value of the primary difference value exceeds the range in which the threshold value is added or subtracted or multiplied. The waveform analysis means for identifying the damaged part using as an index, if the sign of the primary difference value at each measurement point and the sign of the moving average value of the primary difference value change,
The provided, in the waveform analysis means, for the echo height ratio of the neighborhood and the identified site of injury, significantly the site of injury by the multiplied 1-order difference value in the difference value calculation unit, or more than a predetermined threshold value ΔHo A binarization process is performed on the difference value of, a moving average value of the binarized difference value is obtained, a range of the moving average value equal to or greater than a predetermined threshold value Hs is determined, and a predetermined range is set in the determined range The range width multiplied by the coefficient is set as the abnormal range, the echo height ratio data in the abnormal range is deleted, and there is no abnormality based on the least square method using the echo height ratio data before and after the abnormal range. A waveform is estimated.

Further, the present invention generates an ultrasonic wave toward the bearing from an ultrasonic probe attached to a bearing housing on which the bearing is supported, and the bearing housing and the outer ring of the bearing or the outer ring of the bearing and the rolling element A bearing damage evaluation method for measuring a reflected wave from a boundary and evaluating damage generated in the bearing by an echo height ratio H calculated by the following equation,
H = (1-h / ho) × 100
Here, h is the echo height when the bearing load is acting, and ho is the echo height obtained when there is no load when the bearing load is not acting.
An echo height ratio calculating step for obtaining an echo height ratio from the reflected wave received by the ultrasonic probe;
From the obtained echo height ratio, a waveform signal having an echo height ratio using time as an index is obtained, and only the echo height ratio equal to or greater than a predetermined threshold value Ho is used as an object of calculation, which is the primary of the slope of the waveform signal. A difference value calculating step of obtaining a secondary difference value which is a difference between the difference value and the primary difference value, and further making the difference value squared or raised to the fourth power;
An average value calculating step for obtaining a moving average value of the primary difference value and the secondary difference value of the echo height ratio;
The value of the moving average value of the primary difference value, which has become noticeable by the fourth power after further quadratic difference, or the moving average value of the primary difference value exceeds the range in which the threshold value is added or subtracted or multiplied. If the sign of the primary difference value at each measurement point and the sign of the moving average value of the primary difference value change further, the step of identifying the damaged site using either as an index,
With respect to the identified damaged part and the echo height ratio in the vicinity thereof, the damaged part is made noticeable by the multiplied primary difference value, and binarization processing is performed on the difference value equal to or larger than a predetermined threshold value ΔHo. A moving average value of the difference value subjected to the valuation process is obtained, a range of the moving average value equal to or greater than a predetermined threshold Hs is determined, a range width obtained by multiplying the determined range by a predetermined coefficient is set as an abnormal range, Deleting the echo height ratio data in the abnormal range, and estimating a waveform having no abnormality based on the least square method using the echo height ratio data before and after the abnormal range. .

In order to improve the accuracy of bearing damage evaluation, the echo height ratio waveform that processed the output of the ultrasonic probe is combined with the improvement in the precision of the measurement waveform accompanying the increased sensitivity of the ultrasonic probe It is important how the degree of deformation can be objectively recognized from the signal. In the present invention, based on the echo height ratio, the primary difference value that is the gradient or the secondary difference that is the difference between the primary difference values. By calculating the difference value or the moving average value thereof, it is possible to identify the abnormal position and identify the damage with high accuracy. In other words, focusing on the fact that it is not possible to make a sufficient abnormality judgment only by primary processing such as comparison of the differential value of a waveform signal, which has been common in the past, the type of damage that occurs in the bearing (specifically, wear powder (convex ) Or damage (concave), etc., but in the following, usually referred to as “damage”, referred to as “convex damage” when referring only to the former, and “concave damage” when referring only to the latter) By using the characteristics of the waveform that appears, the secondary processing signal of the waveform signal, such as the change in the slope or the deviation from the moving average value, is obtained, so that an abnormal measurement site can be identified based on the objective calculation criteria. It became possible to identify and clearly indicate the damage state

  Specifically, the ultrasonic probe is attached to the bearing housing, generates ultrasonic waves toward the bearing, and generates reflected waves from the boundary between the bearing housing and the outer ring of the bearing or from the boundary between the outer ring of the bearing and the rolling element. Receive. And if the degree of adhesion between the bearing housing and the outer ring of the bearing or the outer ring of the bearing and the rolling element is large (the solid contact area is large), the emitted ultrasonic wave is transmitted from the boundary, and this transmittance is proportional to the degree of adhesion. . Here, if the bearing is damaged, the degree of adhesion changes, and when the degree of concave damage is large, the degree of adhesion is small, so the transmittance of ultrasonic waves is small, and the magnitude of the reflected wave is large. At the same time, the transmittance gradient between the damaged portion and the periphery thereof (expressed as a primary difference value in the present invention) or the gradient difference value (secondary difference value) also increases. On the other hand, when the degree of the concave damage is small, the magnitude of the reflected wave becomes small, and at the same time, the gradient of transmittance between the damaged portion of the bearing and the periphery thereof, or the difference value of the gradient becomes small. Therefore, the degree of concave damage can be estimated by measuring this reflected wave. Further, as a state opposite to the concave damage, there is a convex damage due to generation and adhesion of wear powder to the bearing. The degree of adhesion changes if there is convex damage on the bearing, and when the degree of convex damage is large, the degree of adhesion increases, so the transmission of ultrasonic waves increases and the magnitude of the reflected wave decreases. Become. At the same time, the primary difference value or the secondary difference value between the damaged portion of the bearing and its periphery also becomes small. Conversely, when the degree of convex damage is small, the magnitude of the reflected wave becomes large. Therefore, the degree of convex damage can be estimated by measuring this reflected wave.

As described above, the present invention is intended to evaluate the occurrence of bearing damage by using the result measured by the ultrasonic probe. In other words, the moving average value of the reflected wave information including the damaged part (the output waveform of the ultrasound probe) represents the outline of the change in the reflected wave, so it is assumed that the waveform is close to the reflected wave from a normal bearing. And a damaged portion or an abnormal portion can be estimated by comparison with the difference value.

Conventionally, the range of “abnormality” and the abnormal value may change depending on the setting of the threshold value and the number of data points used when obtaining the rate of change. However, the present invention is not dependent on these as much as possible. Aiming at the determination, the problem is solved by limiting to a minimum threshold, specifically, a weighting for a moving average value , which will be described later, and setting a predetermined width for the moving average value .

Therefore, by providing a bearing damage evaluation technique that can accurately estimate the presence / absence of damage, damage location, and degree of damage by measuring this reflected wave and calculating its primary difference value or moving average value. can do.
In particular, in the present invention, in the analysis of the waveform signal of such an echo height ratio, the damaged portion is made prominent by the primary difference value multiplied for the identified damaged portion and the echo height ratio in the vicinity thereof, and a predetermined value is obtained. A binarization process is performed on the difference value equal to or greater than the threshold value ΔHo, a moving average value of the binarized difference value is obtained, a range of the moving average value equal to or greater than a predetermined threshold value Hs is determined, and the determination is performed. A range width obtained by multiplying the range by a predetermined coefficient is set as the abnormal range, and the echo height ratio data in the abnormal range is deleted, and the echo height ratio data before and after the abnormal range is used based on the least square method. Estimate the waveform without abnormality, calculate the difference in echo height ratio between the abnormal range and the waveform without abnormality, estimate the size of the local concave or convex in the waveform signal of the echo height ratio, Bearing inner ring or rolling It is possible to estimate the magnitude of the damage caused to the body.

  As used herein, the “difference value” means, for a specific echo height, its gradient, etc., the values of several points before and after the data including the value are averaged, and the difference between the specific value and the average value, or The value obtained by dividing the specific value by the averaged value is referred to as “5-point difference value” (value divided by the previous and next 2 points and the 5-point average value used), “13-point difference value” (front and back 6 values). And the value divided by the average value of 13 points used).

The present invention is a program related to the bearing damage evaluation apparatus or the bearing damage evaluation method, wherein the measurement part related to the second order differential value exceeding a target bearing specific threshold is transformed into a waveform signal of an echo height ratio. It is characterized by judging that the degree is high.

  In the analysis of the waveform signal of the echo height ratio, it was found that the occurrence of damage and the secondary difference value of the waveform signal are very correlated, and the waveform signal of the echo height ratio at the end of the damaged portion It is possible to identify the damaged part by capturing the point where the change greatly occurs. Specifically, it is estimated that a measurement site where the secondary difference value of the waveform signal of the echo height ratio exceeds a predetermined range corresponds to a site where the degree of deformation of the waveform signal is high (that is, the end of the damaged portion). By performing these calculation steps and waveform analysis steps, a highly accurate bearing damage evaluation program can be formed.

The present invention is a program related to the bearing damage evaluation apparatus or the bearing damage evaluation method, which is a measurement site related to the primary difference value exceeding a threshold specific to the target bearing, and the primary difference value. It is determined that a measurement part exceeding the range in which the threshold value is added to or subtracted from the moving average value is determined to have a high degree of deformation of the waveform signal of the echo height ratio.

In the analysis of the waveform signal of the echo height ratio, a measurement site whose primary difference value exceeds a predetermined range and exceeds a range having a predetermined width above and below the moving average value of the primary difference value is a waveform signal. It has been found that there is a great correlation with the deformation (that is, the occurrence of damage), and it is estimated that such a measurement part corresponds to a part having a high degree of deformation of the waveform signal. By performing such a calculation step and a waveform analysis step, a highly accurate bearing damage evaluation program can be formed, and the evaluation can be further improved by evaluating together with the analysis result by the “secondary difference value of the waveform signal”. Accuracy can be obtained.

The present invention is the above-described bearing damage evaluation program, which is a measurement site related to the primary difference value exceeding the threshold value unique to the target bearing, and multiplies the moving average value of the primary difference value by the threshold value. It is characterized in that a measurement part exceeding the range is judged as having a high degree of deformation of the waveform signal of the echo height ratio.

As described above, it was found that the waveform signal of the echo height ratio can be analyzed with high accuracy by combining the primary difference value and the moving average value thereof. The threshold value is added to the moving average value of the primary difference value. Instead of setting the range for adding or subtracting, by setting the range in which the moving average value of the primary difference value is multiplied by the threshold value , it is possible to estimate a portion having a high degree of deformation of the waveform signal. Therefore, a bearing damage evaluation program with high accuracy can be formed by performing such calculation step and waveform analysis step. At the same time, it is possible to obtain higher accuracy by evaluating together with the analysis result by the “secondary difference value of the waveform signal” and / or “adjustment range of the primary difference value and its moving average value ”. It becomes.

  The present invention is the above-described bearing damage evaluation program, characterized in that the waveform analysis means calculates the magnitude of damage for a measurement site determined to have a high degree of deformation.

  The present invention also relates to a computer-readable storage medium storing the bearing damage evaluation program. That is, the bearing damage evaluation program can be recorded on a recording medium (CD-ROM or the like). By installing the recording medium in a computer, the computer can function as a bearing damage evaluation apparatus.

  As described above, according to the present invention, it is possible to secure a standard for judging the abnormality of a bearing due to damage, which has been difficult in the past, and establish a detection method in an evaluation device to identify an abnormal range and grasp the degree of abnormality. Thus, it is possible to provide a bearing damage evaluation technique capable of accurately evaluating bearing damage. Further, it is possible to provide a bearing damage evaluation technique capable of quantitatively evaluating bearing damage not only qualitatively.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a conceptual diagram illustrating the configuration of a bearing damage evaluation apparatus.

<Configuration of bearing damage evaluation device>
A bearing 2 is supported at the center of the bearing housing 1. A peripheral portion of the bearing housing 1 is partially cut, and an ultrasonic probe 3 is attached. The bearing 2 is a rolling bearing and includes a large number of rolling elements 22 sandwiched between an outer ring 20 and an inner ring 21). The rotary shaft 4 is fixed to the inner diameter portion of the inner ring 21 by an appropriate method such as press fitting. The outer diameter portion of the bearing outer ring 20 is also closely fitted in a hole formed in the bearing housing 1.

  The ultrasonic probe 3 generates ultrasonic waves in a direction perpendicular to the mounting surface. The generated ultrasonic wave is configured to reflect the boundary between the bearing outer ring 20 and the bearing housing 1 or between the outer ring 20 and the rolling element 22 and receive the reflected wave (reflection type).

  The ultrasonic probe 3 is connected to the ultrasonic flaw detector 5. The ultrasonic flaw detector 5 incorporates a drive circuit for driving the ultrasonic probe 3, a receiving circuit for receiving reflected waves, and the like. The ultrasonic flaw detector 5 is connected to a personal computer 6, and a signal received by the ultrasonic probe 3 is AD converted and transmitted to the personal computer 6. The personal computer 6 incorporates a program for evaluating bearing damage from the received reflected wave signal, and the personal computer 6 is configured to function as a bearing damage evaluation apparatus.

  As an element constituting the ultrasonic probe 3, an electromagnetic vibrator using Lorentz force and a vibrator using the piezoelectric effect of piezoelectric ceramic can be used. Specifically, using a vibrator having a diameter of about several millimeters to several tens of millimeters, ultrasonic waves of several MHz to several tens of MHz are incident as pulses several to tens of thousands of times per second. In FIG. 1, the transmission unit and the reception unit are illustrated as an integrated member from the viewpoint of convenience of mounting the sensor, but the present invention is not limited to this, and it is also possible to use a combination of separate bodies. . In addition, when the reflected wave from the contact surface between the rolling element 22 and the outer ring 20 is the target, the measurement site becomes a portion where the two surfaces are in contact (elastohydrodynamic lubrication portion). It is preferable to adopt a method (focal type) in which ultrasonic waves are emitted from the transmitter so as to tie.

The ultrasonic probe 3 is attached at a position opposite to the rolling element 22 and from a boundary between the bearing outer ring 20 and the bearing housing 1 or a contact portion between the rolling element 22 and the outer ring 20, so-called near field limit. It is preferable to set the distance D or more. This is to avoid the influence of the complicated sound pressure distribution in the ultrasonic irradiation region on the reflection characteristics. The near field limit distance D is
D = d ² / 4λ (Equation 1)
Represented as: Here, d represents the diameter of the vibrator, and λ represents the wavelength of the ultrasonic wave.

  By using a plurality of ultrasonic probes 3 and monitoring the boundary between the bearing outer ring 20 and the bearing housing 1 at a plurality of locations or the contact state between the rolling elements 22 and the outer ring 20, it is possible from a viewpoint that cannot be achieved by a single monitoring. Measurement is possible. For example, from the distribution state of the outputs from the plurality of ultrasonic probes 3, whether or not a load is evenly applied to the shaft center, or what behavior of the bearing 1 as a whole when an instantaneous load is applied ( For example, it is highly advantageous in that measurement in an unsteady state, such as whether or not the rotational position is shifted, is possible.

<Method of observing the operation of a bearing with an ultrasonic probe>
Next, the principle of a method for observing the operation of the bearing using the ultrasonic probe 3 will be described. In FIG. 1, the ultrasonic wave emitted from the ultrasonic probe 3 is directed to the boundary between the bearing housing 1 and the bearing outer ring 20 or between the rolling element 22 and the outer ring 20, and a part of the ultrasonic wave is transmitted through the boundary. Is reflected at the boundary. This reflected wave is received by the ultrasonic probe 3. For example, if the degree of adhesion between the bearing housing 1 and the bearing outer ring 20 is large (the solid contact area is large), the emitted ultrasonic wave is easily transmitted from the boundary, and this transmittance is substantially proportional to the degree of adhesion.

In the present invention, a physical quantity called an echo height ratio is used to quantitatively represent the magnitude of the reflected wave. What is echo height ratio (H)?
H = (1-h / h ₀ ) × 100 (Equation 2)
Defined by h is the echo height when an external bearing load (indicated by W in FIG. 1) is acting, and h ₀ is the echo when no external bearing load is acting (no load). It is height. Note that the magnification of 100 is to display%, and the present invention is not limited to this. As the bearing load increases, the degree of adhesion between the bearing 2 and the bearing housing 1 increases and h decreases (the magnitude of the reflected wave decreases), so the echo height ratio H increases.

FIG. 2 is a diagram illustrating an observation example when the rotating shaft 4 is rotationally driven. The vertical axis represents the echo height ratio H (%), and the horizontal axis represents time (μs). Although the echo height ratio curve is represented by the periodic repetition waveform, the rolling element 22 is the echo height ratio H when it came directly under the ultrasonic probe 3 shows the maximum value H _M, the rolling elements 22 And the rolling element 22 are directly under the ultrasonic probe 3, the echo height ratio H shows the minimum value Hm.

<Configuration of main parts of bearing damage evaluation device>
Next, the structure of the principal part of the personal computer 6 which functions as a bearing damage evaluation apparatus is shown in FIG. The personal computer 6 includes a display device 60, a CPU 61, and a RAM 62. In addition, a program file 63 storing a bearing damage evaluation program and a data file 64 are provided. These are connected via a data bus. The bearing damage evaluation program stores programs for realizing functions of the echo height ratio calculating means 63a, the difference value calculating means 63b, the average value calculating means 63c, the waveform analyzing means 63d, etc. in the personal computer 6. This program is executed while being read into the RAM 62. The program can be installed in the personal computer main body using a recording medium such as a CD-ROM or a floppy (registered trademark) disk.

  The function of the echo height ratio calculating means 63a is as described above.

The difference value calculation means 63b is a means for calculating the above-described difference value. Specifically, n unit times t _{−n to} t _{+ n} before and after the specific measurement time tx are set, and the difference value calculation means 63b at the measurement time tx is set. the difference between the average value Hav echo height ratio Hx and measuring the echo height ratio of the time _t -n _{~t + n} or echo height ratio Hx in the measurement time tx,, the measured time _t -n _{~t + n} echo height A value divided by the average value Hav of the thickness ratio is calculated (referred to as 2n + 1 point difference value). The rate of change can be calculated without being influenced by the echo height ratio as an absolute value, and the echo height ratio at each time can be standardized. In actual evaluation, it is preferable to use a result obtained by further raising the difference result to the fourth power (or the second power) for conspicuousness.

  In the average value calculation means 63c, there is a method of setting a predetermined time width and calculating the average value every time, but the predetermined time width is set and a new echo height ratio is taken in every unit time. It is preferable to calculate the moving average value to be calculated (discard the oldest data). This is because smoothed instantaneous data can be obtained. In the present application, it may be expressed as “n-point average value” depending on the parameter n to be averaged. Here, n can be arbitrarily set depending on the magnitude and accuracy of damage to be evaluated.

In the data file 64, the echo height (h ₀ ) data obtained when the bearing load is set to zero or when the center position of the two rolling elements is measured is written as the echo height data file 64a. It is. Further, a relational expression data file 64b is written as a relational expression (or a table representing the relation) representing the relation between the echo height ratio H and the bearing load W.

<Evaluation method of damaged part of bearing>
Next, a method for evaluating a damaged part of a bearing will be schematically described.

  (1) The waveform signal of the echo height ratio obtained during actual operation of the bearing is a schematic sine wave as shown in FIG. That is, during actual operation, the rolling elements of the bearing also rotate. The degree of adhesion at the boundary between the bearing housing and the bearing outer ring is different when the rolling element is positioned directly below the ultrasonic probe and when the rolling element is not. When the rolling element is located directly under the ultrasonic probe, the echo height ratio is maximum.When the space between the rolling element and the rolling element is directly under the ultrasonic probe, the echo height ratio is Minimal. Therefore, the waveform signal of the echo height ratio according to the movement of the rolling element is a sine wave or a cyclic repetition waveform close to a sine wave. When there is no abnormality in the bearing, the echo height ratio waveform is a smooth curve.

  Here, when the bearing is damaged, the waveform signal of the echo height ratio is deformed from a normal state. For example, as shown in FIG. 4, local (abrupt) convex portions (indicated by A) and local concave portions (indicated by B) are seen in the echo height ratio waveform in a state where an abnormality has occurred in the bearing. . It is considered that the local convex portion is an increase in the support load of the rolling element because the rolling element of the bearing bites the wear powder. Further, the local concave portion is generated because the inner ring of the bearing is damaged and the support load of the rolling element is lowered. Furthermore, the width of the convex part and the concave part, and the height of the convex part and the depth of the concave part also have a deep relationship with the magnitude of the damage. It is considered that the depth of the sheath and the recess is also increased. Therefore, it is possible to estimate the magnitude of damage from the width of the convex portion or the concave portion, the height of the convex portion, or the depth of the concave portion.

  Thus, bearing damage can be evaluated by analyzing the protrusions and recesses, and the size of the damage can be estimated by obtaining the width and height (depth) of the protrusions and recesses. In addition to qualitative as well as quantitative damage assessment.

In the evaluation of damage, when the echo height ratio H is a value close to zero, the difference value may fluctuate greatly in the following calculation, and accurate processing cannot be performed. A method in which a predetermined threshold value H ₀ is set for the echo height ratio H, and an echo height ratio H lower than that is excluded from calculation is preferable. The setting of the threshold value can be determined empirically depending on the target bearing, the size of the load, or how to apply the load.

  (2) When a sine wave waveform signal is deformed and a convex portion or a concave portion is generated in a part of the waveform, the gradient (first-order difference value) of the waveform signal corresponding to the end portion of the convex portion or the concave portion is rapidly changed. . Specifically, this corresponds to the case where the positive gradient suddenly becomes a negative gradient before and after, or the negative gradient suddenly changes to a positive gradient. Therefore, the magnitude of the primary difference value can be used as one index for damage evaluation. A sudden change in the primary difference value can be regarded as the magnitude of the secondary difference value of the waveform signal. Therefore, the magnitude of the primary difference value can be used as one index for damage evaluation (index A).

  Here, there is a method for calculating the gradient by approximating the curve to a multidimensional function and calculating the derivative thereof, but as an inventor's knowledge, the concave portion caused by the damage becomes various curves depending on the state of the damage. From the above, it was found that the evaluation based on the above difference value indicates the gradient (primary difference value) or the change rate of the gradient (secondary difference value) more accurately in the shape.

Specifically, taking as an example the case where the K portion is damaged as shown in FIG. 5 (A), in FIG. 5 (B), the echo height is about 10 as the threshold value H ₀ in FIG. 5 (A). % Indicates the result of performing the primary difference (specifically, the result of performing the 13-point difference is the fourth power). In this case, the primary difference suggests the presence of an abnormality at many measurement points (measurement time) including the K part. Suggests the possibility.

  In order to estimate the actual damaged part, the result of the primary difference is averaged by five points, and further the secondary difference (specifically, the fourth difference is raised to the fourth power). As a result, as shown in FIG. 5C, it is suggested that there is an abnormality in the vicinity of the measurement point 900, and the range of damage can be further limited.

  (3) As a next examination result, when a sine wave waveform signal is deformed, not only the instantaneous value of the waveform signal corresponding to the end of the convex portion or the concave portion but also the average value or the combination of the average values of the difference values Therefore, the deformation of the waveform signal can be estimated as one index for damage evaluation (index B).

  (3-1) A case where damage occurs near the measurement point 460 as shown in FIG. FIG. 6B shows the result of obtaining the gradient of the measured waveform signal and performing the first order difference (specifically, the result of performing the two-point difference). In addition, FIG. 6C shows the result of averaging the results of the primary difference by 10 points.

In this case, in the primary difference, a large change is observed at many measurement points including the vicinity of the measurement point 460, and it is difficult to specify damage. In particular, since the first order difference value of the data near the base may change significantly, it is preferable not to use such data for estimation of the damaged site. In the following evaluation procedure, as illustrated in FIG. 7, it removes the data near the slope zero in FIG. 6 (A), the example, the evaluation of the threshold value 20% of peak height H _M of the waveform output.

  (3-2) Next, as illustrated in FIG. 8, when the primary difference at the measurement point is performed and the range where the threshold value is added to or subtracted from the average value of the primary difference value is exceeded, A method for determining that the degree of deformation of the waveform signal is high can be given (index B1). At the apex of the waveform of the echo height ratio, the primary difference value is close to zero, and the change in the vicinity thereof is very sensitively influenced by the primary difference value, and abnormality determination is greatly different from other parts. By setting a threshold value as a width to the average value of the primary difference values as in the present invention, erroneous determination can be prevented. Specifically, the first-order difference value (two-point difference) is averaged by 10 points, a threshold value of ± 0.35 width is set for the result, and a value exceeding this is taken out and evaluated.

  In the example of FIG. 8, the primary difference suggests that there is a large change at many measurement points, but from the change exceeding the threshold range set from the average value of the primary difference values, This suggests that there is an abnormality in the vicinity of 460 and 470. Here, the threshold value is selected as ± 0.35, but it is preferable to set in advance if it can be set unique to the bearing, and if it cannot be set, it is assumed that the threshold value is changed sequentially. A method is also possible in which a point that coincides with a point where a large change in the primary difference value occurs is determined to be abnormal.

  (3-3) As a different determination method, as illustrated in FIG. 9, the primary difference value at a specific measurement point exceeds a threshold specific to the target bearing, and the A method of determining that the degree of deformation of the waveform signal of the echo height ratio is high when the range of multiplying the average value of the primary difference values by the threshold value is included (index B2). Specifically, in order to detect an abrupt change in the slope of the waveform signal, a primary difference (two-point difference) is obtained, and an average of the ten points is obtained. Next, if the primary difference value is not less than the value obtained by multiplying the 10-point average value at the measurement point by 1.8 times the upper threshold value, or if it is not more than 0.1 times the lower threshold value, the value is taken out. evaluate.

  The primary difference value suggests the presence of large changes at many measurement points, but in the range where the threshold set from the 10-point average value of the primary difference value is 0.1 to 1.8 This suggests that there is an abnormality in the vicinity of the points 365, 370, 460, 470, and 540.

  (3-4) Further, as can be seen from FIG. 9, the movement in the vicinity where the primary difference value or its average value crosses zero is a very dynamic change in which the sign of the gradient changes. It is considered to play an important role in judging abnormalities. Also in FIG. 7, it is suggested that there is an abnormality near the measurement points 460 and 470 (index B3). Therefore, a method of adding a zero-cross element as a criterion to a criterion based on the primary difference value and its average value is also effective in determining an abnormality.

  (3-5) As can be seen from FIG. 9, when the primary difference value at the measurement point where the average value of the primary difference value is in the negative region suddenly changes to the positive region, or conversely, the primary The case where the primary difference value at the measurement point where the average value of the difference values is in the positive region suddenly changes to the negative region is considered to play an important role in determining the abnormality. 9 also indicates that there is an abnormality in the vicinity of the measurement point 470 (index B4).

  (4) The index A and the indices B1 to B4 serving as the above criteria can be estimated to be sufficiently abnormal even when used alone, and these sites can be arbitrarily combined, and the damaged site can be determined on condition that a plurality of indices are satisfied. It is also possible to estimate, and more accurate estimation is possible.

  (5) In the above observation, if the deterioration of the bearing progresses, the backlash of the bearing itself increases, so the accuracy decreases. In other words, when the bearing is normal, the waveform is clean, so it is easy to separate from the noise and the noise is small, but as the deterioration progresses, the backlash of the bearing increases, so the increase in noise itself is judged to be abnormal. there is a possibility. At this time, it is possible to prevent a decrease in accuracy by changing the threshold value.

  As described above, in the present invention, the damaged portion generated in the bearing inner ring or the rolling element can be estimated from the local convex portion or concave portion in the waveform signal of the echo height ratio by the waveform analyzing means 63d. . At the same time, the width of damage can be estimated when both ends of the damage can be specified.

<Evaluation method of bearing damage size>
In evaluating damage, evaluation of the size of the concave portion is given as an evaluation item. Hereinafter, in the observation example shown in FIG. 10 (A), the method for evaluating the damage magnitude of the bearing, that is, the method for quantifying the damage is schematically described in the case where it is found that the K portion has a recess by the above method. To do.

  (1) The size of the recess can be represented by the width Lf of the recess and the depth -ΔHf of the recess as shown in FIG. Conversely, the size of the convex portion can be represented by the width Lf of the convex portion and the height ΔHf of the convex portion. In the following description, the concave portion will be described, but the same processing can be performed for the convex portion. That is, in the present invention, the occurrence of wear powder (convex shape) and damage (concave shape) can be evaluated at the same level.

Here, if the start point of the concave portion in the waveform signal is p ₁ and the end point of the concave portion is p ₂ , Lf indicates the width in the horizontal direction (time axis direction) between the points p ₁ and p ₂ . Further, the p ₁ and the point p ₃ an intermediate position of the point p ₂ points assuming no recess, the vertical distance of the deepest position of the point p ₃ and the recess and -DerutaHf. Lf and -ΔHf are calculated by the waveform analysis means 63d. With this function, it is possible to quantitatively evaluate bearing damage.

(2) The following calculation is performed for the range determined to be abnormal in the above.
(2-1) for the vicinity of the portion K of FIG. 10 (A), the same as the process in FIG. 5, to set the predetermined threshold value _{H 0,} except that the following data.
(2-2) calculating a primary differential value for the threshold value H ₀ or more echo height ratio. Here, for example, a result shown in FIG. 11A can be obtained by performing a 9-point difference and then performing an absolute value process on a value obtained by raising the fourth power. FIG. 11B shows an enlarged display of the data limited to the range determined as abnormal in the above. It can be seen that it has a large peak in the vicinity of 883.
(2-3) As shown in FIG. 12A, a predetermined threshold value ΔH ₀ is set for the primary difference value in the range shown in FIG. Here, the threshold value ΔH ₀ = 0.02.
(2-4) Next, as shown in FIG. 12B, “01” processing is performed on the difference value equal to or greater than the threshold value ΔH ₀ . That is, a binarization process is performed with “1” when the difference value is positive and “0” when the difference value is zero.
(2-5) The average value of the difference values is calculated for the above results (see FIG. 12C). For example, a moving average of 9 points is obtained.
(2-6) A predetermined threshold value Hs is set to this average value, and a range width Lf ′ of a value larger than that is obtained. As shown in FIG. 12D, a range that can be determined to be abnormal can be set.
(2-7) Based on the obtained width Lf ′, a range width Lf ″ giving a tolerance (multiplying the width Lf ′ by a predetermined coefficient m) is obtained as shown in FIG. 13A. Request _{p 4} and the point _{p 5} points on the waveforms that make up the ". It can be limited to a waveform having no abnormality. The coefficient m is arbitrarily set according to the damage state and measurement accuracy.
(2-8) Next, the data between the points p ₄ to p ₅ on the waveform of the echo height ratio is deleted, and the solid line partial data Ka up to the point p ₄ and the solid line partial data Kb from the point p ₅ Is used to estimate the broken line portion data between the points p ₄ to p ₅ based on the least square method (see FIG. 13B). Waveforms with no abnormalities can be determined. According to the knowledge of the present inventor, it has been found that highly accurate estimation can be performed by performing the least square method using a fifth-order or sixth-order approximation.
(2-9) As shown in FIG. 13B, the difference is calculated based on the broken line portion data of the estimated echo height ratio and the actual measurement data, and the abnormal value is obtained. The magnitude of the damage range can be quantitatively grasped from the obtained abnormal value.

  As described above, in the present invention, the size of the damage generated in the inner ring or rolling element of the bearing is estimated by estimating the size of the local recess in the waveform signal of the echo height ratio by the waveform analysis means 63d. Can be estimated.

<Evaluation method of bearing deterioration>
The above description relates to the evaluation of the damaged portion. However, when evaluating the waveform of the echo height ratio, it is possible to determine the deterioration of the bearing itself from the disturbance of the entire waveform. That is, as illustrated in FIG. 14 , when the deterioration progresses as compared to the normal waveform, a minute fluctuation component rides on the entire waveform. That is, it is possible to determine the deterioration of the bearing by evaluating the entire waveform, not the partial evaluation of the waveform.

Specifically, since the echo height ratio is theoretically proportional to the surface pressure, the primary difference in the echo height ratio represents a variation in the surface pressure, that is, a variation in the support load. Further, the secondary difference in the echo height ratio is related to how much the support load has changed, that is, in a sense, a phenomenon corresponding to impact force (hereinafter referred to as “impact”). Therefore, as shown in FIG. 15 , if the overall primary difference and secondary difference of the echo height ratio waveform are pursued over time in the evaluation of the same bearing, the impact due to the deterioration of the bearing itself, that is, the occurrence of the backlash of the bearing. Can be detected.

Further, at this time, when the damage evaluation data is also described, as shown in FIG. 15 , the damage evaluation data increased as a whole while repeatedly increasing and decreasing. In other words, this evaluation method is very effective in evaluating the deterioration of the bearing.

  As described above, the bearing damage evaluation method can be applied to the bearing deterioration evaluation method, and by comprehensively evaluating these, a bearing evaluation apparatus or evaluation with an unprecedented accuracy can be achieved. The system and evaluation program could be configured.

<Another embodiment>
(1) The bearing to which the present invention is applied is not limited to a bearing having a specific structure. For example, it can be applied not only to normal ball bearings but also to angular ball bearings. For example, the ball can be applied not only to a single row but also to a double row.

  (2) In the present embodiment, only the bearing damage evaluation program is described, but this program may be integrated with a program for other purposes. For example, if a relational expression between the bearing load and the echo height ratio is obtained in advance using a known bearing load, the amount of change in the rolling element support load caused by an abnormality from the echo height ratio measured during machine operation and the bearing The load can be estimated. Therefore, it may be integrated with a program for estimating the bearing load. In addition, since the offset load can be obtained from the data obtained from the left and right ultrasonic probes 3, it may be integrated with a program having such a function. Further, it may be integrated with a program having another function. Of course, the same applies to the recording medium on which this program is recorded.

The conceptual diagram which illustrates the structure of the bearing damage evaluation apparatus which concerns on this invention. Explanatory drawing which shows the observation result of an echo height ratio waveform. Explanatory drawing which shows the structure of the principal part of a bearing damage evaluation apparatus. Explanatory drawing which shows the observation result of the echo height ratio waveform at the time of damage generation | occurrence | production. Explanatory drawing which shows one of the evaluation methods of the damage site | part which concerns on this invention. Explanatory drawing which shows one of the evaluation methods of the damage site | part which concerns on this invention. Explanatory drawing which shows the other evaluation method of the damage site | part based on this invention. Explanatory drawing which shows the other evaluation method of the damage site | part based on this invention. Explanatory drawing which shows the other evaluation method of the damage site | part based on this invention. Explanatory drawing which shows the first step of the evaluation method of the magnitude | size of the other damage based on this invention. Explanatory drawing which shows the next step in the evaluation method of the other damage magnitude | size which concerns on this invention. Explanatory drawing which shows the next step in the evaluation method of the other damage magnitude | size which concerns on this invention. Explanatory drawing which shows the final step of the evaluation method of the magnitude | size of the other damage which concerns on this invention. Explanatory drawing which shows the state of deterioration of the other bearing which concerns on this invention. Explanatory drawing which shows the evaluation method of deterioration of the other bearing which concerns on this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Bearing housing 2 Bearing 3 Ultrasonic probe 4 Rotating shaft 5 Ultrasonic flaw detector 20 Outer ring 21 Inner ring 22 Rolling element H Echo height ratio 63a Echo height ratio calculation means 63b Difference value calculation means 63c Average value calculation means 63d Waveform Analytical means

Claims (2)

Ultrasonic waves are generated toward the bearing from an ultrasonic probe attached to a bearing housing on which the bearing is supported, and reflected waves from the boundary between the bearing housing and the outer ring of the bearing or the outer ring of the bearing and the rolling element are generated. A bearing damage evaluation apparatus that measures and evaluates damage caused to a bearing by an echo height ratio H calculated by the following equation:
H = (1-h / ho) × 100
Here, h is the echo height when the bearing load is acting, and ho is the echo height obtained when there is no load when the bearing load is not acting.
An echo height ratio calculating means for obtaining an echo height ratio from the reflected wave received by the ultrasonic probe;
From the obtained echo height ratio, a waveform signal having an echo height ratio using time as an index is obtained, and only the echo height ratio equal to or greater than a predetermined threshold value Ho is used as an object of calculation, which is the primary of the slope of the waveform signal. A difference value calculating means for obtaining a secondary difference value that is a difference between the difference value and the primary difference value, and further prominently by squaring the difference value to the second power or the fourth power;
Average value calculating means for obtaining a moving average value of the primary difference value and the secondary difference value of the echo height ratio;
The value of the moving average value of the primary difference value, which has become noticeable by the fourth power after further quadratic difference, or the moving average value of the primary difference value exceeds the range in which the threshold value is added or subtracted or multiplied. The waveform analysis means for identifying the damaged part using as an index, if the sign of the primary difference value at each measurement point and the sign of the moving average value of the primary difference value change,
In the waveform analysis means, the damaged portion is made remarkable by the primary difference value multiplied in the difference value calculation means with respect to the identified damaged portion and the echo height ratio in the vicinity thereof, and is greater than or equal to a predetermined threshold value ΔHo. A binarization process is performed on the difference value of, a moving average value of the binarized difference value is obtained, a range of the moving average value equal to or greater than a predetermined threshold value Hs is determined, and a predetermined range is set in the determined range The range width multiplied by the coefficient is set as the abnormal range, the echo height ratio data in the abnormal range is deleted, and there is no abnormality based on the least square method using the echo height ratio data before and after the abnormal range. A bearing damage evaluation apparatus characterized by estimating a waveform . Ultrasonic waves are generated toward the bearing from an ultrasonic probe attached to a bearing housing on which the bearing is supported, and reflected waves from the boundary between the bearing housing and the outer ring of the bearing or the outer ring of the bearing and the rolling element are generated. A bearing damage evaluation method for measuring and evaluating the damage generated in the bearing by the echo height ratio H calculated by the following equation,
H = (1-h / ho) × 100
Here, h is the echo height when the bearing load is acting, and ho is the echo height obtained when there is no load when the bearing load is not acting.
An echo height ratio calculating step for obtaining an echo height ratio from the reflected wave received by the ultrasonic probe;
From the obtained echo height ratio, a waveform signal having an echo height ratio using time as an index is obtained, and only the echo height ratio equal to or greater than a predetermined threshold value Ho is used as an object of calculation, which is the primary of the slope of the waveform signal. A difference value calculating step of obtaining a secondary difference value which is a difference between the difference value and the primary difference value, and further making the difference value squared or raised to the fourth power;
An average value calculating step for obtaining a moving average value of the primary difference value and the secondary difference value of the echo height ratio;
The moving average value of the first-order difference value is further subtracted to the fourth power and then the magnitude of the value (index A) that becomes noticeable, or the moving-average value of the first-order difference value increases or decreases the threshold value. A step of specifying a damaged site by using as an index either the sign of the primary difference value at each measurement point or the sign of the moving average value of the primary difference value at each measurement point changes. ,
With respect to the identified damaged part and the echo height ratio in the vicinity thereof, the damaged part is made noticeable by the multiplied primary difference value, and binarization processing is performed on the difference value equal to or larger than a predetermined threshold value ΔHo. A moving average value of the difference value subjected to the valuation process is obtained, a range of the moving average value equal to or greater than a predetermined threshold Hs is determined, a range width obtained by multiplying the determined range by a predetermined coefficient is set as an abnormal range, Deleting the echo height ratio data in the abnormal range, and estimating a waveform having no abnormality based on the least square method using the echo height ratio data before and after the abnormal range. Bearing damage evaluation method.

Images