An automatic detection method and system for pulley runout in water pumps

Abstract

This application discloses an automatic detection method and system for pulley runout in water pumps. The method includes: sequentially acquiring the angular position and displacement values ​​during the first forward rotation, reverse rotation, and second forward rotation of the pulley under the same clamping conditions; reversing the reverse displacement data and extracting low-order periodic components from both the reverse and first forward displacement data, and obtaining the backlash angle through cross-correlation to correct the angular coordinates of the reverse data; interpolating the three displacement data segments to the same angular grid, fitting the positional variation term based on the difference between the two forward data segments, and correcting the first forward displacement value accordingly; then using the corrected three displacement values ​​as input, solving for the common geometric components through robust least squares, and using the difference between the maximum and minimum values ​​within one revolution as the pulley runout value. This application can reduce the impact of commutation backlash, clamping positional variations, and abnormal sampling on the detection results, improving the repeatability and reliability of the detection.

Description

Technical Field

[0001] This application relates to the field of automatic detection technology for pulley runout, specifically to an automatic detection method and system for pulley runout of a water pump. Background Technology

[0002] In a water pump assembly, the pulley is usually press-fitted onto the pump shaft and forms a rotating support structure together with bearings, seals and other components. Radial runout or end face runout of the pulley can affect the smoothness of transmission and the reliability of the assembly operation. Therefore, it is usually necessary to set up a testing station to detect its runout during mass production.

[0003] Current detection methods typically involve driving the pulley to rotate one revolution during a single clamping operation and determining the runout value based on the difference between the maximum and minimum values ​​collected by the displacement sensor. However, in the case of a water pump assembly, the sensor signal not only includes the geometric runout of the pulley itself but also incorporates additional displacements caused by factors such as bearing clearance, seal friction, positioning changes after clamping, and hysteresis of the rotating mechanism.

[0004] The aforementioned additional displacement will cause differences in the displacement waveform of the same product under different rotation directions and rotation sequences, resulting in insufficient repeatability of the runout value calculated directly from single-turn data, which can easily lead to misjudgment. Especially in automated inspection scenarios, without the identification and correction of forward and reverse angle offsets, initial rotation positioning changes, and occasional sampling anomalies, the inspection results are difficult to stably reflect the true geometric state of the pulley. Summary of the Invention

[0005] The purpose of this application is to provide an automatic detection method and system for the runout of a water pump pulley, so as to solve the problems mentioned in the background art.

[0006] According to one aspect of this application, an automatic detection method for pulley runout in a water pump is provided, comprising the following steps:

[0007] In the same clamping, the rotating mechanism drives the pulley on the water pump assembly to perform the first forward rotation, the reverse rotation and the second forward rotation in sequence, and collects the angular position and displacement value of each segment to obtain the first forward displacement data, the reverse displacement data and the second forward displacement data;

[0008] The reverse displacement data is arranged in reverse order, and the low-order periodic components of the reversed reverse displacement data and the first forward displacement data are extracted respectively. The hysteresis angle is obtained by cross-correlation, and the angular coordinates of the reversed reverse displacement data are corrected by shifting the hysteresis angle.

[0009] The three displacement data segments are interpolated onto the same angle grid to obtain the first positive displacement value, the reverse displacement value, and the second positive displacement value at each grid point;

[0010] Calculate the difference between the first and second positive displacement values ​​at each grid point, obtain the in-situ change term by fitting the difference with a low-order period, and use the in-situ change term to correct the first positive displacement value.

[0011] The common geometric components are obtained by taking the corrected first positive displacement value, reverse displacement value and second positive displacement value as input, and using robust least squares to solve for them. The difference between the maximum and minimum values ​​of the common geometric components within one revolution is taken as the pulley runout value.

[0012] Preferably, the low-order periodic components are the first-order and second-order components of the displacement data with the basic period of one rotation of the pulley; the extraction method includes performing a discrete Fourier transform on the displacement data within one rotation range, retaining the amplitude and phase of the first-order and second-order frequency components, setting the remaining frequency components to zero, and obtaining the low-order periodic components through inverse transform.

[0013] Preferably, the backlash angle is obtained by cross-correlation as follows: a series of tentative angle offsets are applied sequentially to the low-order periodic components of the reverse displacement data after reversal; for each tentative angle offset, the mean square error between the offset reverse low-order periodic components and the low-order periodic components of the first forward displacement data within the overlap angle range is calculated; and the tentative angle offset that minimizes the mean square error is taken as the backlash angle.

[0014] Preferably, the search for the tentative angle offset includes performing a coarse search with steps of integer multiples of the angle grid interval to determine the range of the backlash angle, and then performing a search within that range with steps of angle grid intervals. A fine search is performed with the step size to obtain the backlash angle, where It is an integer greater than 1.

[0015] Preferably, the step of correcting the first positive displacement value with the in-situ change term includes: at each grid point, subtracting half of the in-situ change term value from the first positive displacement value of the corresponding grid point to obtain the corrected first positive displacement value.

[0016] Preferably, when solving for the common geometric components using robust least squares, a decomposition model is established for the displacement values ​​at each grid point, decomposing each displacement value into a superposition of common geometric components, intra-segment reference displacements, direction-dependent error terms, and residual components. The common geometric components are shared by three data segments and have a one-cycle periodicity. The intra-segment reference displacements are constants independent of angles within each data segment. The direction-dependent error terms are constrained to zero in the two forward data segments but take non-zero values ​​in the reverse data segments. The residuals are generated by the solution process.

[0017] Preferably, the robust least squares loss function is adopted. The scaling parameter of the Hubel loss function is determined during the equipment debugging stage. Specifically, the standard part is repeatedly tested several times according to the three-segment sampling process. The initial robust solution is obtained by ordinary least squares or by using the preset initial scaling parameter. The initial residuals of each test are obtained, the standard deviation of the residuals of each test is calculated, and the preset multiple of the standard deviation of the residuals is used as the scaling parameter.

[0018] Preferably, weights are assigned to the three segments of displacement data for robust least squares solution. The weights are determined during the equipment commissioning phase. Specifically, the standard part is repeatedly tested several times according to the three-segment sampling process. The repeatability standard deviations of the first positive displacement data, the reverse displacement data, and the second positive displacement data are calculated respectively. The reciprocal of the repeatability standard deviation of each segment is normalized as the weight of the corresponding segment, and the sum of the weights of the three segments is 1.

[0019] Preferably, a state consistency check is performed before outputting the pulley runout value. The check includes three types: positive repeatability check, which calculates the root mean square value of the difference between the corrected first positive displacement value and the second positive displacement value at each grid point and compares it with a preset positive repeatability threshold; steering consistency check, which calculates the difference between the reverse displacement value and the mean of the corrected first and second positive displacement values ​​at each grid point, fits the difference to a low-order period and takes the amplitude of the fitted curve, and compares it with a preset direction correlation error threshold; and angle alignment quality check, which compares the backlash angle with a preset normal range of backlash angle. The pulley runout value is output when all three checks pass, and an unstable detection state indicator is output when any one check fails.

[0020] In another aspect, this application also provides an automatic detection system for the runout of a water pump pulley, comprising:

[0021] Rotation mechanism, displacement acquisition unit, angle acquisition unit, and processing unit;

[0022] The rotating mechanism is used to drive the pulley on the water pump assembly to perform a first forward rotation, a reverse rotation, and a second forward rotation in the same clamping operation.

[0023] The displacement acquisition unit is used to acquire the displacement value of the corresponding measurement position of the pulley during each segment of rotation. The angle acquisition unit is used to acquire the angular position during each segment of rotation. The processing unit is used to generate the first positive displacement data, the reverse displacement data, and the second positive displacement data based on the angular position and displacement value of each segment.

[0024] The processing unit is also used to arrange the reverse displacement data in reverse order, extract the low-order periodic components of the reversed reverse displacement data and the first forward displacement data respectively, obtain the backlash angle through cross correlation, and use the backlash angle to shift and correct the angular coordinates of the reversed reverse displacement data.

[0025] The processing unit is also used to interpolate the three displacement data segments to the same angle grid, and obtain the first positive displacement value, the reverse displacement value and the second positive displacement value at each grid point;

[0026] The processing unit is also used to calculate the difference between the first positive displacement value and the second positive displacement value at each grid point, obtain the in-situ change term by fitting the difference with a low-order period, and correct the first positive displacement value with the in-situ change term.

[0027] The processing unit is also used to take the corrected first positive displacement value, reverse displacement value and second positive displacement value as input, solve the common geometric components through robust least squares, and take the difference between the maximum and minimum values ​​of the common geometric components within one revolution as the pulley runout value.

[0028] This application also provides an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the automatic detection method for pulley runout of a water pump as described above.

[0029] In another aspect of this application, a storage medium is provided that stores computer program instructions thereon, which can be executed by a processor to implement the automatic detection method for pulley runout of a water pump as described above.

[0030] Another aspect of this application provides a computer program product, including a computer program that, when executed by a processor, implements the automatic detection method for pulley runout of a water pump as described above.

[0031] This application acquires three segments of displacement data—first forward, reverse, and second forward—from the pump assembly in its initial clamping state. The reverse data is then reversed and corrected for backlash angles, ensuring that the forward and reverse sampling results can be compared at the same physical angle. This avoids angular misalignment caused by changes in the rotation mechanism's clearance and bearing contact state. Simultaneously, the application extracts the positioning variation term based on the difference between the two forward data segments, correcting for systematic deviations caused by fixture clamping, seal friction, and bearing load adjustment during the initial rotation. Building upon this, robust least squares is used to solve for the common geometric component from the three data segments. The range of this common geometric component is used as the pulley runout value, ensuring that the final result primarily reflects the pulley's inherent stable geometric runout, reducing the impact of directional correlation errors, random vibrations, and localized abnormal sampling on the detection results. This application improves the repeatability and reliability of pulley runout detection in the pump assembly state, reduces misjudgments in mass production testing, and can be combined with state consistency verification to assess the stability of the testing process, facilitating subsequent retesting, workstation maintenance, and quality traceability. Attached Figure Description

[0032] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0033] Other features, objects, and advantages of this application will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0034] Figure 1 This is a schematic diagram of an automatic detection method for pulley runout of a water pump provided in an embodiment of this application.

[0035] Figure 2 This is a schematic diagram of a three-segment rotational sampling process provided in an embodiment of this disclosure.

[0036] Figure 3 A flowchart illustrating the reverse displacement data reversal and backlash angle correction provided in this embodiment of the disclosure.

[0037] Figure 4 This is a schematic diagram illustrating the calculation process for pulley runout value provided in an embodiment of this disclosure.

[0038] Figure 5 This is a schematic diagram of an automatic detection system for pulley runout of a water pump, provided in an embodiment of this application.

[0039] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0040] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0041] This embodiment provides an automatic detection method for pulley runout in water pumps. It is applicable to situations where the pulley is press-fitted onto the water pump shaft and, together with the bearings and seals, forms the water pump assembly. The method detects radial or lateral runout of the pulley at an automated mass production testing station. In this testing state, the signal collected by the displacement sensor includes not only the geometric runout of the pulley itself but also additional displacements caused by factors such as water pump bearing clearance, seal friction, positioning adjustments after clamping, and hysteresis of the rotating mechanism. These additional displacements cause differences in the displacement waveform of the same product under different rotation directions or sequences, affecting the repeatability of the runout detection results. Therefore, this method performs three rotational sampling segments sequentially within the same clamping operation. During data processing, the additional displacements are separated from the actual geometric runout of the pulley. Finally, the range of the geometric components that stably and repeatedly appear at the same physical angle in the three data segments is used as the runout value.

[0042] The testing station uses conventional fixtures to position and clamp the water pump housing or a specified reference surface, keeping the water pump assembly fixed throughout the testing process. The displacement sensor can be installed at the testing station, and the measuring head is aligned with the specified measuring position of the pulley, which can be the outer circular surface or the specified measuring surface inside the pulley groove. The rotating mechanism is connected to the pulley drive, which can drive the pulley to rotate at low speed around the axis and switch directions. The angle signal comes from the servo motor encoder pulse or the angle encoding signal built into the rotating mechanism, as well as the zero position signal used to determine the starting point of one revolution, without the need for an additional angle detection device.

[0043] The following detailed description, in conjunction with specific embodiments, illustrates the implementation process of the automatic detection method for pulley runout in water pumps described in this application. It should be noted that these embodiments are merely illustrative of this application and not intended to limit its scope of protection. Any conventional adjustments or substitutions made by those skilled in the art to the steps without departing from the concept of this application should be included within the scope of protection of this application.

[0044] like Figure 1 As shown in the figure, this application discloses a schematic diagram of an automatic detection method for pulley runout of a water pump, including the following method steps:

[0045] S1, under the same clamping, the rotating mechanism drives the pulley on the water pump assembly to perform the first forward rotation, the reverse rotation and the second forward rotation in sequence, and collects the angle position and displacement value of each segment to obtain the first forward displacement data, the reverse displacement data and the second forward displacement data.

[0046] S2, arrange the reverse displacement data in reverse order, extract the low-order periodic components of the reversed reverse displacement data and the first forward displacement data respectively, obtain the backlash angle through cross correlation, and use the backlash angle to translate and correct the angular coordinates of the reversed reverse displacement data.

[0047] S3, interpolate the three displacement data segments to the same angle grid, and obtain the first positive displacement value, the reverse displacement value and the second positive displacement value at each grid point;

[0048] S4, calculate the difference between the first positive displacement value and the second positive displacement value at each grid point, obtain the in-situ change term by fitting the difference with a low-order period, and correct the first positive displacement value with the in-situ change term.

[0049] S5 takes the corrected first positive displacement value, reverse displacement value, and second positive displacement value as input, and solves the common geometric components through robust least squares; the difference between the maximum and minimum values ​​of the common geometric components within one revolution is taken as the pulley runout value.

[0050] In some embodiments, for step S1, under the condition that the positions of the fixture, sensor and workpiece remain unchanged, the rotating mechanism drives the pulley to complete three rotation samplings in sequence, and obtains the first positive displacement data, the reverse displacement data and the second positive displacement data respectively.

[0051] like Figure 2 As shown, Figure 2 This is a schematic diagram of a three-stage rotation sampling process provided in an embodiment of this disclosure, wherein in S201: the first forward rotation sampling. The rotation mechanism drives the pulley to rotate slowly in the forward direction, and the rotation angle covers at least one full circle. The actual sampling range can collect several degrees of overlapping angles on top of 360 degrees. During the rotation, the displacement sensor collects data synchronously at a fixed sampling frequency. Each sampling point records three data items: the current angle position, the corresponding displacement value, and the sampling time. All sampling points constitute the first forward displacement data.

[0052] In S202: Reverse rotation sampling. The rotation mechanism switches direction, driving the pulley to rotate slowly in the opposite direction. The coverage angle and sampling method are the same as the first forward rotation, and all the obtained sampling points constitute the reverse displacement data.

[0053] In S203: Second forward rotation sampling. The rotation mechanism switches back to the forward direction, covering at least one full rotation. The sampling method is the same as before, and all the obtained sampling points constitute the second forward displacement data.

[0054] During the three sampling periods, the fixture remains clamped, the sensor position remains unchanged, and the workpiece is not disassembled or re-clamped.

[0055] In one embodiment, during the start and stop phases of each rotation, the rotating mechanism is in an acceleration or deceleration state, resulting in uneven intervals for angle signals and potential influence of displacement signals due to inertial vibrations. This portion of the sampled data is considered transitional data, which is clipped in this embodiment and not included in subsequent fluctuation calculations. The range of transitional data is determined by monitoring the time interval between adjacent angle sampling points: after rotation begins, the time interval between adjacent sampling points is calculated point by point. When this interval is continuously within a preset proportion of the target sampling interval, the rotation speed is considered to have stabilized, and the sampled data before this moment is marked as the initial transitional data. After rotation deceleration, the sampled data from the moment the time interval deviates from the preset proportion to the moment the rotation stops is marked as the stop transitional data.

[0056] The preset ratio range can be configured to ±5% to 15% of the target sampling interval. The specific value is determined by the acceleration and deceleration characteristics of the rotating mechanism. This determination method adaptively matches the actual motion characteristics of the rotating mechanism, eliminating the need for different devices to set fixed angle exclusion values. In some embodiments, if the servo system directly outputs a rotation speed signal, the speed stabilization moment can also be determined based on the rotation speed signal, with the same effect.

[0057] After cropping the transitional data, each of the three sampling segments retains valid data covering at least one full circle within its stable velocity range.

[0058] In some embodiments, for step S2, mechanical backlash exists after the rotation mechanism reverses direction, and the contact position between the rolling elements and raceways inside the water pump bearing also changes slightly due to the change in direction. The combination of these two factors results in an offset between the angle reading when a physical position on the pulley passes the sensor measurement point during reverse rotation and the angle reading when that physical position passes the same measurement point during forward rotation. If this offset is not corrected, directly comparing the forward and reverse data using the nominal angle will result in the same nominal angle actually corresponding to different physical positions on the pulley, inevitably leading to angular misalignment errors in subsequent data fusion. Therefore, this embodiment accurately estimates this offset and corrects the angular coordinates of the reverse data accordingly.

[0059] like Figure 3 As shown, Figure 3The flowchart for reverse displacement data reversal and hysteresis angle correction provided in this embodiment describes the process in step S301: reversing the angle sequence of the reverse displacement data so that the angle increment direction is consistent with the forward data, resulting in reversed reverse displacement data. The reversal process only changes the data arrangement order and does not alter the correspondence between the displacement value and angle of each sampling point. After reversal, due to the aforementioned mechanical clearance and bearing contact state changes, the displacement waveform of the reversed reverse displacement data still exhibits an overall phase shift in the angular direction relative to the first forward displacement data; this shift is the hysteresis angle.

[0060] In S302: Low-order periodic components are extracted from the first forward displacement data and the reverse displacement data after reversal, respectively, as the benchmark for subsequent cross-correlation calculations. Among them, the low-order periodic components refer to the first-order and second-order components in the displacement data with one revolution of the pulley as the basic period: the first-order component corresponds to the eccentricity or runout feature that occurs once per revolution of the pulley, and the second-order component corresponds to the ellipticity feature that occurs twice per revolution.

[0061] Specifically, the extraction method involves first interpolating the corresponding displacement data within a circle to a temporary iso-angle sequence, which is used only for low-order periodic component extraction and hysteresis angle estimation; then performing a discrete Fourier transform (DFT) on the displacement data within a circle, retaining the amplitude and phase of the first-order and second-order frequency components, setting the remaining frequency components to zero, and finally obtaining the low-order periodic component curve through an inverse transform.

[0062] Therefore, local spikes, high-frequency vibrations, and sampling noise in the displacement signal are filtered out, retaining only the slowest changing parts that best represent the overall geometry. Lower-order components are chosen as the angle alignment reference because the main characteristics of pulley geometric runout are concentrated in the first and second-order components. The amplitudes of these two components should be consistent in both forward and reverse data, and they are least affected by local disturbances. Using them as cross-correlation inputs can yield stable and reliable hysteresis angle estimates.

[0063] In S303: Using the low-order periodic components of the first forward displacement data as a reference, the hysteresis angle is calculated through cross-correlation. Specifically, cross-correlation involves applying a series of tentative angular offsets sequentially to the low-order periodic components of the reversed displacement data. For each tentative offset, the root mean square error (RMSE) between the offset reversed low-order periodic components and the low-order periodic components of the first forward displacement data is calculated within the overlap angle range. After traversing all tentative offsets, the offset that minimizes the MSE is taken as the hysteresis angle. The smaller the MSE, the higher the degree of agreement between the two low-order periodic component curves at that offset; the corresponding offset is the most likely actual angular phase difference between the forward and reverse data.

[0064] In some embodiments, to balance search efficiency and accuracy, the search for the tentative angle offset is performed in two steps: first, a coarse search is conducted with a step size of integer multiples of the angle grid interval to determine the approximate range of the backlash angle; then, within this range, a search is performed with a step size of several multiples of the angle grid interval. Perform a fine search for the step size, where The value is an integer greater than 1 and is predetermined according to the backlash angle resolution requirements to obtain the precise value of the backlash angle; the angle grid interval is the distance between adjacent angles in a discrete angle sequence with equal intervals from 0 degrees to 360 degrees.

[0065] In S304: The obtained backlash angle is used to perform an overall translation correction on the angular coordinates of the reversed displacement data. Specifically, the tentative angular offset that minimizes the mean square error is taken as the signed backlash angle, and the angular value of each sampling point in the reversed displacement data is translated according to this signed backlash angle to align it with the angular coordinates of the first forward displacement data. After the correction, the physical position of the pulley corresponding to each angular value in the reverse data is consistent with the physical position corresponding to the same angular value in the forward data. When comparing the forward and reverse data at the same nominal angle, the comparison is of the displacement value at the same physical position on the pulley.

[0066] In some embodiments, for step S3, the actual sampling points of the three sampling segments may not fall on the same angle value, and the three data segments need to be unified into the same angle grid; wherein the angle grid is an equally spaced discrete angle sequence from 0 degrees to 360 degrees, and the grid interval is determined according to the angle signal resolution of the rotating mechanism and the production cycle, for example, it can be configured as 0.5 degrees, 1 degree or 2 degrees.

[0067] For the first positive displacement data, the reverse displacement data after backlash correction, and the second positive displacement data, linear interpolation is performed between adjacent sampling points to calculate the displacement value at each grid angle point. The interpolation is only performed between adjacent sampling points and does not cross data gaps where the angle interval exceeds a preset limit. The preset limit is determined based on the angle grid interval and is greater than the single angle grid interval, so that normal adjacent sampling points can cover adjacent grid points, while abnormal missing sampling intervals are excluded. If the distance between adjacent sampling points on both sides of a certain grid angle point exceeds the limit, it is marked as missing and excluded. After the interpolation is completed, three displacement values ​​are formed at each valid angle grid point: the first positive displacement value, the reverse displacement value, and the second positive displacement value.

[0068] In some embodiments, for step S4, when the pump assembly is first clamped and begins to rotate, a stable frictional state has not yet been formed between the seal and the shaft surface, the contact position of the rolling elements inside the bearing may also undergo slight adjustments due to the initial load, and there may be an initial positioning process between the fixture and the workpiece contact surface. These factors result in a systematic difference between the first and second positive displacement values. This difference is different from random noise and has continuously changing low-frequency characteristics. Therefore, this embodiment defines this systematic difference as the positioning variation term, that is, the portion of the displacement difference between the first and second positive displacement values ​​with low-order periodic characteristics caused by the gradual stabilization of the support state after the workpiece is clamped at the same direction and angle position.

[0069] The difference between the first positive displacement value and the second positive displacement value is calculated point by point on the angle grid to form a difference sequence. This difference sequence includes both the slow changes caused by clamping and positioning, and may also contain sampling noise and occasional vibrations.

[0070] The difference sequence is treated as a function of angle and fitted with a low-order periodicity. During fitting, the first-order, second-order, and DC components are retained, while the higher-order components are removed. The resulting low-order curve is the in-situ variation term. The reasons for retaining the low-order components are as follows: the displacement change caused by clamping and positioning is global, manifesting as a slowly changing offset within a one-circle range; higher-order fluctuations in the difference are more likely to come from sampling noise or random factors and should not be included in the in-situ variation term; the DC component represents the overall mean deviation between two positive displacements, which is also a manifestation of the in-situ process, and therefore is retained as well.

[0071] At each grid point, half of the in-situ change term value is subtracted from the first positive displacement value of the corresponding grid point to obtain the corrected first positive displacement value. Half is used as the correction because the in-situ change occurs gradually between the first and second rotations. The sampling time of the first rotation is at the beginning of the change process, and the in-situ change term represents the total change. Using half of it to estimate the actual in-situ deviation at the time of the first rotation is a reasonable approximation. The corrected first positive displacement value replaces the original first positive displacement value and is used together with the reverse displacement value and the second positive displacement value to solve for subsequent common geometric components. The in-situ change term itself is no longer involved in subsequent calculations.

[0072] As a result, the systematic deviation between the corrected first and second positive displacement values ​​caused by the initial rotation and positioning process is reduced, and the two positive data segments have higher consistency when participating in the solution of common geometric components.

[0073] In some embodiments, for step S5, according to the present disclosure, components that stably repeat at the same physical angle, i.e. common geometric components, are extracted from the corrected first positive displacement value, reverse displacement value and second positive displacement value.

[0074] Specifically, such as Figure 4 As shown, Figure 4 This is a schematic diagram of the pulley runout calculation process provided in this embodiment of the disclosure. In S401: a decomposition model is established for the displacement values ​​of each point on the angle grid. Specifically, for any angle grid point, the displacement value of each segment of data at that point is decomposed into the superposition of four components: common geometric component, intra-segment reference displacement, direction-related error term, and residual.

[0075] The common geometric component is a displacement component shared by three data segments and having a periodicity of one revolution. It represents the stable geometric position of the pulley at that physical angle and is the sole basis for calculating the runout value. The reference displacement within each segment is a constant independent of the angle, and different displacements are allowed between segments to absorb the slight changes in the sensor zero position and the overall displacement deviation caused by clamping. The direction-related error term is constrained to zero in the two positive data segments, and is allowed to take non-zero values ​​and have continuous variation characteristics in the reverse data. It represents the additional displacement caused by changes in bearing clearance state and seal friction due to changes in steering direction. The residual is generated naturally in the solution process and absorbs random noise, occasional vibrations, or local burrs.

[0076] The above decomposition relationship can be expressed as: for any angle on the angle grid , No. Displacement values ​​of segment data satisfy:

[0077]

[0078] in As common geometric components, For the first The intra-segment reference displacement of the segment. For the first The direction-related error term of the segment, For the first Segment residuals; subscript Take 1, r, and 2 respectively, which correspond to the first positive, negative, and second positive data segments after correction.

[0079] In S402: A robust least squares method is used to solve for the common geometric components. Specifically, the goal is to find a set of... Each section and the direction-dependent error term of the reverse data This minimizes the overall residuals between the three data segments and the decomposition model under the robust loss function; the directional correlation error terms of the two positive data segments... and Since the constraint is zero in the solution process, it is not considered as a variable to be solved.

[0080] In one embodiment, to ensure the uniqueness of the decomposition result, common geometric components are considered during the solution process. The mean constraint of the first circle is zero, and the global constant offset of the common geometric components is determined by the reference displacement within each segment. Absorption; Directional correlation error term of inverse data The grid is represented by an angle grid, and its mean value is constrained to zero. The overall bias is absorbed by the reference displacement within the reverse segment. The above constraint only determines the reference of each component and does not change the difference between the maximum and minimum values ​​of the common geometric component within a circle.

[0081] In one embodiment, the Hubel loss function is used as the loss function. Its characteristics are: when the absolute value of the residuals is less than the scaling parameter, the square of the residuals is used as the loss value, consistent with ordinary least squares; when the absolute value of the residuals exceeds the scaling parameter, the loss value becomes a linear function of the absolute value of the residuals. Thus, the impact of a small number of outlier sampling points with large deviations is effectively suppressed, eliminating the need for pre-identification or manual removal of outliers.

[0082] The scaling parameter is the boundary point that distinguishes the squared region from the linear region in the Hubel loss function. Its value is determined during the equipment commissioning phase: using a standard part with a known runout value, the sample is repeatedly tested several times at the testing station according to the three-segment sampling process. The initial residuals of each test are obtained by first solving the initial solution using ordinary least squares or by using the preset initial scaling parameter. The standard deviation of the residuals of each test is then calculated. The preset multiple of the standard deviation of the residuals is used as the scaling parameter. The preset multiple can be configured to a value between 1.0 and 2.0.

[0083] In one embodiment, during the solution process, the following constraints are applied to the three data segments: if a peak appears only in one data segment at a certain angle grid point, and there is no corresponding change at the same angle in the other two segments, then the peak is absorbed by the residual term; the part of the reverse data that is consistent with the two forward data segments at the same angle enters the common geometric component, and the continuous difference related to the direction enters the direction-related error term.

[0084] Optionally, weights are assigned to the three displacement data segments for robust least squares solution. The weights are defined as the relative contribution of each data segment to the objective function, with the sum of the three weights being one, and their values ​​ranging from 0 to 1. The weights are determined statistically during equipment commissioning. For example, a standard part is repeatedly tested using a three-segment sampling process, and the repeatability standard deviations of the first forward displacement data, the reverse displacement data, and the second forward displacement data are statistically analyzed. The repeatability standard deviation refers to the average of the standard deviations of the displacement values ​​at each angle point for the same data segment across multiple tests.

[0085] The reciprocal of the repeatability standard deviation of each segment is normalized and used as the weight of the corresponding segment. If the repeatability standard deviation of the reverse segment of a certain device is greater than that of the forward segment, the weight of the reverse segment is correspondingly lower than that of the forward segment, but the reverse data still participates in the solution of the common geometric components. If the repeatability of the three segments is close, the weights tend to be equal. In this way, different testing stations can adjust the contribution ratio of each segment of data according to their respective repeatability levels without changing the hardware.

[0086] The following results are output after the solution is completed: common geometric components ; reference displacement of each segment Record for future reference; Direction-related error term for reverse data Used for steering consistency verification or detection status tracing, and residuals at various angle points. Records are kept for future reference and for tracing abnormal sampling points.

[0087] In S403: The difference between the maximum and minimum values ​​of the common geometric component within one revolution is taken as the pulley runout value. (Common geometric component) The difference between the maximum and minimum values ​​within the range of 0 degrees to 360 degrees is the pulley runout value. This runout value comes from the geometric information that recurs stably at the same physical angle in the three data segments, reducing the impact of commutation state differences, initial positioning changes, and occasional sampling anomalies.

[0088] In some embodiments, the detection system simultaneously outputs the difference between the maximum and minimum values ​​of the original displacement data of the second forward rotation within one revolution, which is recorded in the detection database as the traditional single-revolution runout value. The runout value of the common geometric component is given priority in the product qualification judgment.

[0089] In some embodiments, before outputting the pulley runout value, the stability of the current detection process is verified; the verification result determines whether the current runout value can be accepted, and is divided into the following three categories.

[0090] Specifically, the first type is positive repeatability verification; the difference between the corrected first positive displacement value and the second positive displacement value is calculated point by point at each angle grid point, and the root mean square value of the difference is calculated. This root mean square value is compared with a preset positive repeatability threshold. The positive repeatability threshold is determined during the equipment debugging phase. Specifically, it is the upper limit of the distribution of the root mean square value of the positive segment difference after continuous testing using standard parts. The upper limit is a statistical upper limit value pre-selected during the debugging phase, determined by the maximum value of the sample or a preset quantile value. If the root mean square value does not exceed the threshold, it passes; if it exceeds the threshold, it fails.

[0091] The second type is steering consistency verification; firstly, the average of the first and second positive displacement values ​​after correction is calculated at each angle grid point, and then the difference between the reverse displacement value and the average value is calculated point by point; after low-order period fitting of the difference value, the amplitude of the fitted curve is taken. The amplitude of the fitted curve is the difference between the maximum and minimum values ​​of the low-order period fitted curve within one circle. The amplitude is compared with the preset direction correlation error threshold; the threshold is also calibrated during the equipment debugging stage. If the amplitude does not exceed the threshold, it passes; if it exceeds the threshold, it fails.

[0092] The third category is angle alignment quality verification; the backlash angle obtained in the previous steps is compared with the preset normal range of backlash angle. The normal range of backlash angle is calibrated during the equipment debugging stage and reflects the backlash angle distribution of the rotating mechanism and bearing under normal conditions; the backlash angle passes when it is within the normal range and fails when it exceeds the range.

[0093] When all three types of checks pass, the system outputs the common geometric component runout value as the final test result; when any type of check fails, the system outputs an unstable test status indicator and marks the category of failure, prompting for retesting or workstation inspection, and does not output a product runout judgment conclusion for this data.

[0094] According to the embodiments of this disclosure, in the same clamping, three segments of sampling data—forward, reverse, and forward again—are used, and the alignment of the forward and reverse data in terms of physical angle is achieved by combining backlash angle correction. The systematic deviation caused by the unstable support state during the first rotation is reduced by the correction of the positional change term. Common geometric components are extracted from the three-segment aligned data using a robust least squares method. As a result, the repeatability of pulley runout detection in the state of the water pump assembly is improved, and misjudgments caused by factors such as bearing clearance, seal friction, and clamping positional changes are reduced. At the same time, the state consistency verification provides a quantifiable reliability assessment for each test, and the direction-related error term and positional change term separated in the test results are also convenient for traceability analysis of the test state.

[0095] It should be noted that although the operations of the method of this application are described in a specific order in the accompanying drawings, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. On the contrary, the steps depicted in the flowchart can be performed in a different order. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.

[0096] Please see Figure 5 , Figure 5 This application provides an automatic detection system for the runout of a water pump pulley. The system embodiment is similar to... Figure 1 Corresponding to the illustrated method embodiments, this system can be specifically applied to various electronic devices. The system specifically includes:

[0097] Rotation mechanism 501, displacement acquisition unit 502, angle acquisition unit 503, and processing unit 504;

[0098] The rotating mechanism 501 is used to drive the pulley on the water pump assembly to perform a first forward rotation, a reverse rotation, and a second forward rotation in the same clamping operation.

[0099] The displacement acquisition unit 502 is used to acquire the displacement value of the corresponding measurement position of the pulley during each segment of rotation. The angle acquisition unit 503 is used to acquire the angular position during each segment of rotation. The processing unit 504 is used to generate the first positive displacement data, the reverse displacement data, and the second positive displacement data based on the angular position and displacement value of each segment.

[0100] The processing unit 504 is also used to arrange the reverse displacement data in reverse order, extract the low-order periodic components of the reversed reverse displacement data and the first forward displacement data respectively, obtain the backlash angle through cross correlation, and use the backlash angle to shift and correct the angular coordinates of the reversed reverse displacement data.

[0101] The processing unit 504 is also used to interpolate the three displacement data segments to the same angle grid, and obtain the first positive displacement value, the reverse displacement value and the second positive displacement value at each grid point;

[0102] The processing unit 504 is also used to calculate the difference between the first positive displacement value and the second positive displacement value at each grid point, obtain the in-situ change term by fitting the difference with a low-order period, and correct the first positive displacement value with the in-situ change term.

[0103] The processing unit 504 is further configured to take the corrected first positive displacement value, reverse displacement value, and second positive displacement value as input, solve for the common geometric components using robust least squares, and take the difference between the maximum and minimum values ​​of the common geometric components within one revolution as the pulley runout value.

[0104] Each processing unit and / or module in the embodiments of this application can be implemented by an analog circuit that implements the functions described in the embodiments of this application, or by software that executes the functions described in the embodiments of this application.

[0105] Based on the same inventive concept, this application also provides an electronic device, such as... Figure 6 As shown, Figure 6 This is a schematic diagram of an electronic device structure according to an embodiment of the present disclosure. The method corresponding to the electronic device can be the method in the foregoing embodiments, and its problem-solving principle is similar to that method. The electronic device provided in this application includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the methods and / or technical solutions of the foregoing embodiments of the present application.

[0106] Specifically, the methods and / or embodiments in this application can be implemented as computer software programs. For example, the embodiments disclosed in this application include a computer program product comprising a computer program carried on a storage medium, the computer program containing program code for performing the methods shown in the flowchart. When the computer program is executed by a central processing unit, it performs the functions defined in the methods of this application.

[0107] Another embodiment of this application provides a computer-readable storage medium having computer program instructions stored thereon, which can be executed by a processor to implement the methods and / or technical solutions of any one or more embodiments of this application described above.

[0108] The flowcharts or block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of electronic devices, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-specific system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0109] The above description represents the preferred embodiments of the present invention. It should be noted that, for those skilled in the art, various improvements and modifications can be made without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for automatically detecting runout of a pulley for a water pump, characterized by, Includes the following steps: In the same clamping, the rotating mechanism drives the pulley on the water pump assembly to perform the first forward rotation, the reverse rotation and the second forward rotation in sequence, and collects the angular position and displacement value of each segment to obtain the first forward displacement data, the reverse displacement data and the second forward displacement data; The reverse displacement data is arranged in reverse order, and the low-order periodic components of the reversed reverse displacement data and the first forward displacement data are extracted respectively. The hysteresis angle is obtained by cross-correlation, and the angular coordinates of the reversed reverse displacement data are corrected by shifting the hysteresis angle. The three displacement data segments are interpolated onto the same angle grid to obtain the first positive displacement value, the reverse displacement value, and the second positive displacement value at each grid point; Calculate the difference between the first and second positive displacement values ​​at each grid point, obtain the in-situ change term by fitting the difference with a low-order period, and use the in-situ change term to correct the first positive displacement value. The common geometric components are obtained by taking the corrected first positive displacement value, reverse displacement value and second positive displacement value as input, and using robust least squares to solve for them. The difference between the maximum and minimum values ​​of the common geometric components within one revolution is taken as the pulley runout value.

2. A method of automatically detecting runout of a pulley for a water pump according to claim 1, characterized by, The low-order periodic components are the first-order and second-order components of the displacement data with the basic period of one rotation of the pulley. The extraction method includes performing a discrete Fourier transform on the displacement data within one rotation range, retaining the amplitude and phase of the first-order and second-order frequency components, setting the remaining frequency components to zero, and obtaining the low-order periodic components through inverse transform.

3. The automatic detection method for pulley runout of a water pump according to claim 1, characterized in that, The method of obtaining the backlash angle through cross-correlation includes: applying a series of tentative angle offsets sequentially to the low-order periodic components of the reverse displacement data after reversal; for each tentative angle offset, calculating the mean square error between the offset reverse low-order periodic components and the low-order periodic components of the first forward displacement data within the overlap angle range; and taking the tentative angle offset that minimizes the mean square error as the backlash angle.

4. The automatic detection method for pulley runout of a water pump according to claim 3, characterized in that, The search for the tentative angle offset includes a coarse search with steps of integer multiples of the angle grid interval to determine the range of the backlash angle, and then performing a search within that range with steps of angle grid intervals. A fine search is performed with the step size to obtain the backlash angle, where It is an integer greater than 1.

5. The automatic detection method for pulley runout of a water pump according to claim 1, characterized in that, The step of correcting the first positive displacement value with the in-situ change term includes: at each grid point, subtracting half of the in-situ change term value from the first positive displacement value of the corresponding grid point to obtain the corrected first positive displacement value.

6. The automatic detection method for pulley runout of a water pump according to claim 1, characterized in that, When solving for the common geometric components using robust least squares, a decomposition model is established for the displacement values ​​at each grid point. Each displacement value is decomposed into a superposition of common geometric components, intra-segment reference displacements, direction-related error terms, and residual components. The common geometric components are shared by three data segments and have a periodicity. The intra-segment reference displacements are constants independent of angles within each data segment. The direction-related error terms are constrained to zero in the two forward data segments but take non-zero values ​​in the reverse data segments. The residuals are generated during the solution process.

7. The automatic detection method for pulley runout of a water pump according to claim 6, characterized in that, The robust least squares method uses the Hubel loss function. The scaling parameter of the Hubel loss function is determined during the equipment commissioning phase. Specifically, it involves repeatedly testing a standard part using a three-segment sampling process several times, obtaining the initial residuals of each test through ordinary least squares initial solution or by using a preset initial scaling parameter, and statistically analyzing the standard deviation of the residuals for each test. A preset multiple of the standard deviation of the residuals is then used as the scaling parameter.

8. The automatic detection method for pulley runout of a water pump according to claim 1, characterized in that, Weights are assigned to the three segments of displacement data for robust least squares solution. The weights are determined during the equipment commissioning phase. Specifically, the standard part is repeatedly tested several times according to the three-segment sampling process. The repeatability standard deviations of the first positive displacement data, the reverse displacement data, and the second positive displacement data are calculated. The reciprocal of the repeatability standard deviation of each segment is normalized and used as the weight of the corresponding segment. The sum of the weights of the three segments is 1.

9. The automatic detection method for pulley runout of a water pump according to claim 1, characterized in that, Before outputting the pulley runout value, a state consistency check is performed. The check includes three types: positive repeatability check, which calculates the difference between the corrected first positive displacement value and the second positive displacement value at each grid point, calculates the root mean square value of the difference, and compares it with the preset positive repeatability threshold. Steering consistency verification involves subtracting the reverse displacement value from the mean of the corrected first and second positive displacement values ​​at each grid point. The difference is then fitted with a low-order periodic curve, and the amplitude of the fitted curve is compared with a preset direction correlation error threshold. Angle alignment quality verification involves comparing the backlash angle with a preset normal range. If all three verifications pass, the pulley runout value is output; if any verification fails, an unstable detection status indicator is output.

10. An automatic detection system for pulley runout in a water pump, characterized in that, It includes a rotation mechanism, a displacement acquisition unit, an angle acquisition unit, and a processing unit; The rotating mechanism is used to drive the pulley on the water pump assembly to perform a first forward rotation, a reverse rotation, and a second forward rotation in the same clamping operation. The displacement acquisition unit is used to acquire the displacement value of the corresponding measurement position of the pulley during each segment of rotation. The angle acquisition unit is used to acquire the angular position during each segment of rotation. The processing unit is used to generate the first positive displacement data, the reverse displacement data, and the second positive displacement data based on the angular position and displacement value of each segment. The processing unit is also used to arrange the reverse displacement data in reverse order, extract the low-order periodic components of the reversed reverse displacement data and the first forward displacement data respectively, obtain the backlash angle through cross correlation, and use the backlash angle to shift and correct the angular coordinates of the reversed reverse displacement data. The processing unit is also used to interpolate the three displacement data segments to the same angle grid, and obtain the first positive displacement value, the reverse displacement value and the second positive displacement value at each grid point; The processing unit is also used to calculate the difference between the first positive displacement value and the second positive displacement value at each grid point, obtain the in-situ change term by fitting the difference with a low-order period, and correct the first positive displacement value with the in-situ change term. The processing unit is also used to take the corrected first positive displacement value, reverse displacement value and second positive displacement value as input, solve the common geometric components through robust least squares, and take the difference between the maximum and minimum values ​​of the common geometric components within one revolution as the pulley runout value.

Images