Method for assessing preload deterioration of a ball screw

Abstract

A method for assessing preload deterioration of a ball screw (4), the method being to be implemented by a computer device (1), the ball screw (4) comprising a nut (41), a screw shaft (42), a plurality of balls (44), and a recirculation mechanism (43) for recirculating the balls (44), the computer device (1) being in signal communication with a first sensor (2) mounted on the nut (41) adjacent to the recirculation mechanism (43) and periodically sending a vibration signal to the computer device (1) related to vibrations of the balls (44) in the recirculation mechanism (43), the computer device (1) further being in signal communication with a second sensor (3) mounted on the nut (41) and periodically sending an inertial force signal to the computer device (1) related to an inertial force along a direction,in which the nut (41) is moved relative to the spindle shaft (42), the method comprising the following steps: , A) obtaining an entry of time-domain vibration data based on the vibration signal received from the first sensor (2); B) obtaining at least one entry of frequency domain vibration derived data based on the entry of time domain vibration data; C) for each of the at least one entry of frequency domain vibration-derived data, obtaining a vibration eigenvector based on the entry of frequency domain vibration-derived data; D) performing a preload assessment based on a plurality of reference vibration vectors, a preload assessment range, and the vibration eigenvector(s) obtained for the at least one entry of data derived from a frequency domain vibration to obtain a preload assessment result; and E) Determine, based on the result of the preload assessment, whether preload deterioration has occurred on the ball screw (4), wherein the method is further characterized by the following steps following step E): H) obtaining at least one entry of time-domain inertial force data based on the inertial force signal received from the second sensor (3) when it is determined that preload degradation has occurred on the ball screw (4); I) for each of the at least one entry of time-domain inertial force data, obtaining an inertial force eigenvector based on the entry of time-domain inertial force data; J) performing a game assessment based on a plurality of reference inertial force vectors, a game assessment range, and the inertial force eigenvector(s) obtained for the at least one item of time-domain inertial force data to obtain a game assessment result; and K) Determine, based on the result of the clearance assessment, whether there is any clearance on the ball screw (4).

Description

[0001] The disclosure relates to a method for evaluating preload deterioration and, more particularly, to a method for evaluating preload deterioration of a ball screw.

[0002] Due to the advantages of high-precision movement, ball screws are widely used as transmission components in machine tools that require relatively high positioning accuracy. Generally, a ball screw comprises a plurality of balls, a nut, and a screw shaft. The nut engages the screw shaft through the balls, thereby performing a linear movement relative to the screw shaft.

[0003] Most ball screws are preloaded to eliminate backlash between input (i.e., rotation) and output (i.e., linear motion). Insufficient preload can lead to unwanted vibration when the ball screw is in motion. Furthermore, the preload of ball screws gradually decreases over a period of use, eventually leading to backlash. This unwanted vibration and backlash can consequently shorten the service life of ball screws and reduce their precision.

[0004] TW I653410 B discloses a method for determining whether backlash is developing in a ball screw based on the result of determining whether preload is present on the ball screw. Backlash is determined to be present when it is determined that no preload is present on the ball screw. However, such a method is not suitable for determining whether preload deterioration has occurred when no backlash has yet developed.

[0005] From US 7 680 565 B2 a method for determining fluctuations in operating characteristics is known in which operating pulses generated by a mechanical system are monitored and analyzed.

[0006] Therefore, an object of the disclosure is to provide a method for assessing preload deterioration of a ball screw which can at least mitigate the disadvantage of the prior art.

[0007] According to the disclosure, the method is adapted to be implemented by a computer device. The ball screw comprises a nut, a screw shaft, a plurality of balls, and a recirculation mechanism for recirculating the balls. The computer device is in signal communication with a first sensor mounted on the nut, adjacent to the recirculation mechanism, which periodically sends a vibration signal to the computer device related to vibrations of the balls in the recirculation mechanism. The computer device is further in signal communication with a second sensor mounted on the nut, which periodically sends an inertial force signal to the computer device related to an inertial force exerted on the nut along a direction in which the nut is moved relative to the screw shaft. The method comprises the following steps: A) obtaining an entry of time-domain vibration data based on the vibration signal received from the first sensor (2); B) obtaining at least one entry of frequency domain vibration derived data based on the entry of time domain vibration data; C) for each of the at least one entry of frequency domain vibration-derived data, obtaining a vibration eigenvector based on the entry of frequency domain vibration-derived data; D) performing a preload assessment based on a plurality of reference vibration vectors, a preload assessment range, and the vibration eigenvector(s) obtained for the at least one entry of data derived from a frequency domain vibration to obtain a preload assessment result; and E) Determine, based on the result of the preload assessment, whether preload deterioration has occurred on the ball screw; wherein the method is further characterized by the following steps following step E): H) obtaining at least one entry of time-domain inertial force data based on the inertial force signal received from the second sensor when it is determined that preload degradation has occurred on the ball screw; I) for each of the at least one entry of time-domain inertial force data, obtaining an inertial force eigenvector based on the entry of time-domain inertial force data; J) performing a game assessment based on a plurality of reference inertial force vectors, a game assessment range, and the inertial force eigenvector(s) obtained for the at least one item of time-domain inertial force data to obtain a game assessment result; and K) Determine, based on the result of the clearance assessment, whether there is any clearance in the ball screw.

[0008] Other features and advantages of the disclosure will become apparent from the following detailed description of the embodiment with reference to the accompanying drawings, in which: Fig. 1 is a block diagram illustrating one embodiment of a system used to implement a method for assessing preload degradation of a ball screw according to the disclosure; Fig. 2 is a perspective view illustrating an embodiment in which a first sensor and a second sensor of the system are mounted on the ball screw; Fig. 3 is a flowchart illustrating one embodiment of a training procedure for detecting preload degradation in the method according to the disclosure; Fig. 4 is a flowchart illustrating one embodiment of a training procedure for detecting a game in the method according to the disclosure; Fig. 5 and Fig. 6 are flowcharts cooperatively illustrating one embodiment of a preload degradation and backlash assessment procedure in the method according to the disclosure; Fig. 7 is a flowchart illustrating one embodiment of substeps for obtaining data derived from a frequency domain oscillation in the method according to the disclosure; and Fig. 8 is a flowchart illustrating one embodiment of substeps for obtaining an inertial force eigenvector in the method according to the disclosure.

[0009] With reference to Fig. 1 and Fig. 2 is an embodiment of a method for assessing preload deterioration of a ball screw 4 according to the disclosure adapted to be determined by a Fig. 1. The system 100 includes a computing device 1, a first sensor 2, and a second sensor 3. The ball screw 4 includes a nut 41, a screw shaft 42, a plurality of balls 44, and a recirculation mechanism 43 for recirculating the balls 44. The computing device 1 is in signal communication with the first sensor 2 and the second sensor 3.

[0010] In this embodiment, both the first sensor 2 and the second sensor 3 can be implemented by an accelerometer, however, an implementation thereof is not limited to that disclosed herein, and each sensor can be implemented by a displacement counter or a velocity measuring device in other embodiments. Specifically, an effective bandwidth of the first sensor 2 covers the frequency between 0.1 Hz and 5 Hz; an effective bandwidth of the second sensor 3 covers a frequency range (e.g., 0.1 Hz to 250 Hz) that is within ten times a bandwidth of a spin frequency of the spindle shaft 42, and the second sensor 3 has a digital resolution of 20 bits.

[0011] In this embodiment, the ball screw 4 is an externally rotating ball screw as shown in Fig. 2, and the return mechanism 43 includes a return pipe.

[0012] The first sensor 2 is mounted on the nut 41 and is adjacent to the return mechanism 43. The first sensor 2 periodically sends a vibration signal to the computing device 1 related to vibrations of the balls 44 in the return mechanism 43. The second sensor 3 is mounted on the nut 41 and periodically sends an inertial force signal to the computing device 1 related to an inertial force exerted on the nut 41 along a direction in which the nut 41 is moved relative to the spindle shaft 42.

[0013] It is worth noting that for a ball screw having a different structure than the externally circulating ball screw 4 described herein, the first sensor 2 and the second sensor 3 may be mounted at appropriate positions (may be the same position or different positions) on the ball screw, respectively, to obtain the vibration signal and the inertial force signal, respectively.

[0014] In this embodiment, the computing device 1 may be implemented to be a personal computer, a laptop computer, a notebook computer, a tablet computer, a computing server, or a cloud server, however, an implementation thereof is not limited to that disclosed herein and may vary in other embodiments.

[0015] The computer device 1 comprises a memory module 12, a display module 13, a communication module 11 which is signal-connected to the first sensor 2 and the second sensor 3, and a processing module 14 which is electrically connected to the communication module 11, the memory module 12 and the display module 13.

[0016] The communication module 11 is implemented to be a network interface controller or a wireless transceiver that supports, but is not limited to, wired communication standards and / or wireless communication standards (e.g., Bluetooth technology standards or cellular network technology standards).

[0017] The memory module 12 may be implemented by, but is not limited to, flash memory, a hard disk drive (HDD), a solid-state disk (SSD), an electrically erasable programmable read-only memory (EEPROM), or any other non-volatile memory devices.

[0018] The display module 13 may be a liquid-crystal display (LCD), a light-emitting diode (LED) display, a plasma display panel, a projection display, or the like. However, an implementation of the display module 13 is not limited to the disclosure herein and may vary in other embodiments.

[0019] The processing module 14 may be a processor, a central processing unit (CPU), a microprocessor, a micro control unit (MCU), a system on a chip (SoC), or any circuit that is configurable / programmable in a software manner and / or a hardware manner to implement functionalities discussed in this disclosure.

[0020] The memory module 12 of the computer device 1 stores a plurality of first training vibration eigenvectors and a plurality of second training vibration eigenvectors, a plurality of first training inertial force eigenvectors and a plurality of second training inertial force eigenvectors.

[0021] Each of the first training vibration eigenvectors (or each of the second training vibration eigenvectors) includes one of the following: a training kurtosis eigenvector indicating a kurtosis of a piece of frequency-domain data corresponding to the first training vibration eigenvector (or the second training vibration eigenvector), a training maximum peak eigenvector indicating a maximum peak value of the piece of frequency-domain data, a training total energy eigenvector indicating a total energy of the piece of frequency-domain data, and any combination thereof.

[0022] Each of the first training inertial force eigenvectors (or each of the second training inertial force eigenvectors) includes one of the following: a training peak-to-peak eigenvector indicating a peak-to-peak value of an entry of time-domain data to which the first training inertial force eigenvector (or the second training inertial force eigenvector) corresponds, a training maximum peak eigenvector indicating a maximum peak value of the entry of time-domain data, a training average peak eigenvector indicating an average peak value that is an average value of an absolute value of a positive maximum peak value and an absolute value of a minimum peak value of the entry of time-domain data, or any combination thereof.

[0023] It should be noted that implementations of the first and second training vibration eigenvectors and the first and second training inertial force eigenvectors are not limited to the disclosure herein and may vary in other embodiments.

[0024] The method for assessing preload deterioration of the ball screw 4 according to the disclosure includes a training procedure for detecting preload deterioration (see Fig. 3), a training procedure for capturing a game (see Fig. 4) and an assessment procedure for preload deterioration and play (see Fig. 5 and Fig. 6).

[0025] With reference to Fig. 1 and Fig. 3, the training procedure for detecting preload deterioration includes steps 50 to 53 discussed below.

[0026] At step 50, the processing module 14 of the computing device 1 uses the first training vibration eigenvectors as inputs to machine learning performed using an unsupervised learning algorithm to obtain reference vibration vectors located in a data space spanned by the first training vibration eigenvectors.

[0027] The unsupervised learning algorithm may include a clustering algorithm (e.g., k-means clustering) and / or a self-organizing map (SOM) algorithm.

[0028] In a scenario where the unsupervised learning algorithm is the clustering algorithm, the reference oscillation vectors thus obtained include central vectors, each representing a plurality of oscillation clusters obtained by performing the unsupervised learning algorithm.

[0029] In a scenario where the unsupervised learning algorithm is a SOM algorithm, the reference oscillation vectors obtained in this way include vectors each corresponding to neurons obtained by performing the SOM algorithm and which have been updated more than a preset number of times.

[0030] In step 51, for each of the second training vibration eigenvectors, the processing module 14 of the computing device 1 calculates a plurality of first candidate vibration distances, each between the second training vibration eigenvector and a respective one of the reference vibration vectors obtained in step 50. In particular, each of the first candidate vibration distances calculated in this way is the Euclidean distance, but is not limited thereto.

[0031] At step 52, the processing module 14 of the computing device 1 determines, for each of the second training vibration eigenvectors for which the corresponding first candidate vibration distances were calculated, a shortest of the first candidate vibration distances to serve as the first target vibration distance.

[0032] At step 53, the processing module 14 of the computing device 1 obtains a preload assessment range based on the first target vibration distances determined for the second training vibration eigenvectors, which is used in determining whether preload deterioration has occurred on the ball screw 4. Specifically, a distribution of the first target vibration distances is considered a normal distribution, and a confidence interval (CI) of 95% of the distribution of the first target vibration distances serves as the preload assessment range. However, an implementation of the preload assessment range is not limited to the disclosure herein and may vary in other embodiments.

[0033] With reference to Fig. 1 and Fig. 4, the training procedure for capturing games includes steps 60 to 63 explained below.

[0034] At step 60, the processing module 14 of the computing device 1 uses the first training inertial force eigenvectors as inputs to machine learning performed using an unsupervised learning algorithm to obtain reference inertial force vectors located in a data space spanned by the first training inertial force eigenvectors.

[0035] In a scenario where the unsupervised learning algorithm is the clustering algorithm, the reference inertial force vectors thus obtained include central vectors each representing a plurality of inertial force clusters obtained by performing the unsupervised learning algorithm.

[0036] In a scenario where the unsupervised learning algorithm is the SOM algorithm, the reference inertial force vectors obtained in this way include vectors each corresponding to neurons obtained by performing the SOM algorithm and updated more than another preset number of times.

[0037] At step 61, the processing module 14 of the computing device 1 calculates, for each of the second training inertial force eigenvectors, a plurality of first candidate inertial force distances, each between the second training inertial force eigenvector and a respective one of the reference inertial force vectors obtained at step 60. In particular, each of the candidate inertial force distances calculated in this way is the Euclidean distance, but is not limited thereto.

[0038] At step 62, the processing module 14 of the computing device 1 determines, for each of the second training inertial force eigenvectors for which the corresponding first candidate inertial force distances were calculated, a shortest of the candidate inertial force distances to serve as the first target inertial force distance.

[0039] At step 63, the processing module 14 of the computing device 1 obtains a play assessment range based on the first target inertial force distances determined for the second training inertial force eigenvectors. Specifically, a distribution of the first target inertial force distances is considered a normal distribution, and a confidence interval (CI) of 95% of the distribution of the first target inertial force distances serves as the play assessment range. However, an implementation of the play assessment range is not limited to the disclosure herein and may vary in other embodiments.

[0040] With reference to Fig. 1 and Fig. 5 and Fig. 6, the assessment procedure for preload deterioration and backlash includes steps 70 to 85 below.

[0041] At step 70, the processing module 14 of the computing device 1 obtains an entry of time-domain vibration data based on the vibration signal received from the first sensor 2.

[0042] At step 71, the processing module 14 of the computing device 1 obtains at least one entry of data derived from a frequency domain oscillation based on the entry of time domain oscillation data.

[0043] In detail, step 71 comprises sub-steps 710 to 714, as described in Fig. 7 and are outlined below.

[0044] In sub-step 710, the processing module 14 of the computing device 1 performs envelope processing on the entry of time-domain oscillation data to result in an entry of processed time-domain oscillation data. Since the implementation of the envelope processing is well known to those skilled in the art, a detailed explanation thereof is omitted here for the sake of brevity.

[0045] In sub-step 711, the processing module 14 of the computing device 1 retrieves from the processed time-domain vibration data a target time-domain vibration data entry corresponding to a period during which the nut 41 of the ball screw 4 moves at a constant speed. Note that the period during which the nut 41 of the ball screw 4 moves at the constant speed can be determined based on a preset speed of a motor that drives the movement of the ball screw 4.

[0046] In sub-step 712, the processing module 14 of the computing device 1 receives at least one entry of data derived from a time-domain oscillation based on the entry of target time-domain oscillation data.

[0047] It is worth noting that when the at least one entry of time-domain oscillation-derived data is in the plurality, the plurality of entries of time-domain oscillation-derived data should each correspond to time periods having an identical preset time length.

[0048] When a length of a period to which the entry of target time-domain vibration data corresponds is longer than the preset time length, the processing module 14 divides the entry of target time-domain vibration data into the plurality of entries of time-domain vibration-derived data such that each of the plurality of entries of time-domain vibration-derived data corresponds to a period having the preset time length.

[0049] On the other hand, when the length of the time period to which the target time-domain vibration data entry corresponds is shorter than the preset time period, the processing module 14 combines multiple target time-domain vibration data entries to form one of the multiple time-domain vibration-derived data entries, such that the multiple time-domain vibration-derived data entries each correspond to time periods having the identical preset time period. Note that when a total length of the time periods to which the multiple target time-domain vibration data entries to be combined correspond is longer than the preset time period, one of the thus-formed multiple time-domain vibration-derived data entries must be truncated to have the preset time period.

[0050] At sub-step 713, the processing module 14 of the computing device 1 performs a bandpass filtering operation on the time-domain oscillation-derived data entry for each of the at least one time-domain oscillation-derived data entry. Note that performing the bandpass filtering operation aims to retain a portion of the time-domain oscillation-derived data entry that falls within a frequency range within ten times the bandwidth of the spin frequency of the spindle shaft 42, while attenuating the remaining portion of the time-domain oscillation-derived data entry.This means that if a frequency range that can be detected by the first sensor 2 is exactly ten times the bandwidth of the spin frequency of the spindle shaft 42, the bandpass filtering process in step 713 can be omitted, and the process flow of the method proceeds directly to sub-step 714 after sub-step 712.

[0051] In sub-step 714, for each of the at least one entry of data derived from a time domain oscillation that has been subjected to the bandpass filtering process, the processing module 14 of the computing device 1 performs a Fourier transform on the entry of data derived from a time domain oscillation to result in the entry of data derived from a frequency domain oscillation.

[0052] At step 72, the processing module 14 of the computing device 1 obtains, for each of the at least one entry of data derived from a frequency domain vibration, a vibration eigenvector based on the entry of data derived from a frequency domain vibration.

[0053] Each of the vibration eigenvector(s) includes a kurtosis eigenvector indicating a kurtosis (kurtosis) of the entry of data derived from a frequency-domain vibration, a maximum peak eigenvector indicating a maximum peak value of the entry of data derived from a frequency-domain vibration, a total energy eigenvector indicating the total energy of the entry of data derived from a frequency-domain vibration, and any combination thereof. It should be noted that component(s) of the vibration eigenvector(s) is / are not limited to the disclosure herein and may vary in other embodiments.

[0054] Next, the processing module 14 of the computing device 1 performs a preload judgment described in steps 73 to 75 based on the vibration eigenvector(s) obtained in step 72, the reference vibration vectors obtained in step 50, and the preload judgment range obtained in step 53 to obtain a preload judgment result.

[0055] In step 73, the processing module 14 of the computing device 1 calculates, for each of the vibration eigenvector(s), a plurality of second candidate vibration distances, each between the vibration eigenvector and a respective one of the reference vibration vectors.

[0056] At step 74, the processing module 14 of the computing device 1 determines, for each of the vibration eigenvector(s) for which the corresponding second candidate vibration distances were calculated, a shortest of the second candidate vibration distances to serve as the second target vibration distance.

[0057] At step 75, the processing module 14 of the computing device 1 obtains the result of the preload assessment based on the preload assessment range and the second target vibration distance(s) determined for the vibration eigenvector(s).

[0058] In one embodiment, the processing module 14 of the computing device 1 calculates an average of the second target vibration distance(s) to obtain a vibration average value, and then determines whether the vibration average value falls within the preload assessment range to obtain the preload assessment result (e.g., the vibration average value is within or outside the preload assessment range).

[0059] In one embodiment, the processing module 14 of the computing device 1 determines a mode of the second target vibration distance(s) to obtain a vibration mode, and then determines whether the vibration mode falls within the preload assessment range to obtain the preload assessment result (e.g., the vibration mode is within or outside the preload assessment range).

[0060] It should be noted that the implementation of the preload assessment is not limited to the disclosure herein and may vary in other embodiments.

[0061] At step 76, the processing module 14 of the computing device 1 determines whether preload deterioration has occurred on the ball screw 4 based on the preload judgment result. If it is determined that preload deterioration has not occurred on the ball screw 4 (e.g., the vibration mean value is within the preload judgment range), the process flow of the method proceeds to step 77. However, if it is determined that preload deterioration has occurred on the ball screw 4 (e.g., the vibration mean value is outside the preload judgment range), the process flow of the method proceeds to step 78.

[0062] At step 77, the processing module 14 of the computing device 1 generates a first message indicating that no preload degradation has occurred on the ball screw 4 and controls the display module 13 to display the first message.

[0063] At step 78, when it is determined that preload degradation has occurred on the ball screw 4, the processing module 14 of the computing device 1 obtains at least one entry of time-domain inertial force data based on the inertial force signal received from the second sensor 3.

[0064] It should be noted that each of the at least one piece of time-domain inertial force data corresponds to a revolution period within which the nut 41 of the ball screw 4 moves on the screw shaft 42 from a start position to a final position and then from the final position back to the start position. In one embodiment, each of the at least one piece of time-domain inertial force data corresponds only to an acceleration (or deceleration) sub-period of the revolution period when the nut 41 is accelerated (or decelerated) during the movement.

[0065] At step 79, the processing module 14 of the computing device 1 obtains, for each of the at least one entry of time-domain inertial force data, an inertial force eigenvector based on the entry of time-domain inertial force data.

[0066] Each of the inertial force eigenvector(s) comprises one of the following: a peak-to-peak eigenvector indicating a peak-to-peak value of the time-domain inertial force data entry, a maximum-peak eigenvector indicating a maximum peak value of the time-domain inertial force data entry, an average-peak eigenvector indicating an average peak value that is an average value of an absolute value of a positive maximum peak value and an absolute value of a negative minimum peak value of the time-domain inertial force data entry, or any combination thereof.

[0067] In detail, step 79 comprises, with reference to Fig. 8 Substeps 790 to 792 outlined below.

[0068] In sub-step 790, for each of the at least one entry of time-domain inertial force data, the processing module 14 of the computing device 1 performs the envelope processing on the entry of time-domain inertial force data to result in an entry of processed time-domain inertial force data.

[0069] In sub-step 791, for each of the at least one entry of processed time-domain inertial force data obtained from the at least one entry of time-domain inertial force data, the processing module 14 of the computing device 1 performs a low-pass filtering operation on the entry of processed time-domain inertial force data.

[0070] It should be noted that performing the low-pass filtering operation aims to retain a portion of the processed time-domain inertial force data input that falls within a frequency range between 0.1 Hz and 5 Hz, while attenuating the remaining portion of the processed time-domain inertial force data input. That is, if a frequency range detectable by the second sensor 3 is precisely the frequency range between 0.1 Hz and 5 Hz, performing the low-pass filtering operation in step 791 can be omitted, and the process flow of the method proceeds directly to sub-step 792 after sub-step 790.

[0071] In sub-step 792, the processing module 14 of the computing device 1 obtains, for each of the at least one entry of processed time-domain inertial force data that has been subjected to the low-pass filtering operation, the inertial force eigenvector based on the entry of the time-domain inertial force data.

[0072] Next, the processing module 14 of the computing device 1 performs a backlash judgment described in steps 80 to 82 based on the inertia force eigenvector(s) obtained in step 79, a plurality of reference inertia force vectors obtained in step 60, and the backlash judgment range obtained in step 63 to obtain a backlash judgment result.

[0073] At step 80, the processing module 14 of the computing device 1 calculates, for each of the inertial force eigenvector(s), a plurality of second candidate inertial force distances, each between the inertial force eigenvector and a respective one of the reference inertial force vectors.

[0074] At step 81, the processing module 14 of the computing device 1 determines, for each of the inertial force eigenvector(s) for which the corresponding second candidate inertial force distances were calculated, a shortest of the second candidate inertial force distances to serve as the second target inertial force distance.

[0075] At step 82, the processing module 14 of the computing device 1 obtains the result of the play judgment based on the second target inertial force distance(s) and the play judgment range.

[0076] In one embodiment, the processing module 14 of the computing device 1 calculates an average of the second target inertial force distance(s) to obtain an inertial force average, and then determines whether the inertial force average is within the play evaluation range to obtain the play evaluation result (e.g., the inertial force average is within or outside the play evaluation range).

[0077] In one embodiment, the processing module 14 of the computing device 1 determines a mode of the second target inertial force distance(s) to obtain an inertial force mode, and then determines whether the inertial force mode is within the game judgment range to obtain the game judgment result (e.g., the inertial force mode is within or outside the game judgment range).

[0078] It should be noted that the implementation of performing the game evaluation is not limited to the disclosure herein and may vary in other embodiments.

[0079] At step 83, the processing module 14 of the computing device 1 determines whether backlash has occurred on the ball screw 4 based on the backlash judgment result. If it is determined that no backlash has occurred on the ball screw 4 (e.g., the inertial force average value is within the backlash judgment range), the process flow of the method proceeds to step 84. Otherwise, if it is determined that backlash has occurred on the ball screw 4 (e.g., the inertial force average value is outside the backlash judgment range), the process flow proceeds to step 85.

[0080] At step 84, the processing module 14 of the computing device 1 generates a second message indicating that preload deterioration has occurred on the ball screw 4, but that no backlash has occurred on the ball screw 4, and controls the display module 13 to display the second message.

[0081] At step 85, the processing module 14 of the computing device 1 generates a third message indicating that backlash has occurred on the ball screw 4 and controls the display module 13 to display the third message.

[0082] In summary, the method for judging preload deterioration of a ball screw according to the disclosure uses the computer device 1 to perform the preload judgment based at least on the vibration eigenvector(s); further, the computer device 1 performs the backlash judgment based at least on the inertial force eigenvector(s). Each of the vibration eigenvector(s) is obtained from the vibration signal generated by the first sensor 2 mounted adjacent to the return mechanism 43 of the ball screw 4 and related to vibrations of the balls 44 of the ball screw 4, and each of the inertial force eigenvector(s) is obtained from the vibration signal generated by the first sensor 2 mounted adjacent to the return mechanism 43 of the ball screw 4.The inertial force eigenvectors are obtained from the inertial force signal generated by the second sensor 3 mounted on the nut 41 of the ball screw 4 and related to the inertial force applied to the nut 41. Thereafter, the computing device 1 determines whether preload deterioration has occurred on the ball screw 4 based on the preload judgment result, and determines whether backlash has occurred on the ball screw 4 based on the backlash judgment result.

[0083] In the above description, for purposes of explanation, numerous specific details were set forth in order to provide a thorough understanding of the embodiment. However, it will be apparent to those skilled in the art that one or more other embodiments may be practiced without some of these specific details. Furthermore, it will be understood that reference throughout this specification to "one embodiment," an embodiment with an indication of an ordinal number, and so on, means that a particular feature, structure, or characteristic may be incorporated in the practice of the disclosure.It should also be understood that in the description, various features may sometimes be grouped together in a single embodiment, figure, or description thereof in order to streamline the disclosure and to aid in understanding various aspects of the invention, and that in practicing the disclosure, one or more features or specific details from one embodiment may be practiced together with one or more features or specific details of another embodiment where appropriate.

Claims

[1] A method for assessing preload deterioration of a ball screw (4), the method being to be implemented by a computer device (1), the ball screw (4) comprising a nut (41), a screw shaft (42), a plurality of balls (44), and a recirculation mechanism (43) for recirculating the balls (44), the computer device (1) being in signal communication with a first sensor (2) mounted on the nut (41) adjacent to the recirculation mechanism (43) and periodically sending a vibration signal to the computer device (1) related to vibrations of the balls (44) in the recirculation mechanism (43), the computer device (1) further being in signal communication with a second sensor (3) mounted on the nut (41) and periodically sending an inertial force signal to the computer device (1) related to an inertial force acting along one direction,in which the nut (41) is moved relative to the spindle shaft (42), the method comprising the following steps:, A) obtaining an entry of time-domain vibration data based on the vibration signal received from the first sensor (2); B) obtaining at least one entry of frequency domain vibration derived data based on the entry of time domain vibration data; C) for each of the at least one entry of frequency domain vibration-derived data, obtaining a vibration eigenvector based on the entry of frequency domain vibration-derived data; D) performing a preload assessment based on a plurality of reference vibration vectors, a preload assessment range, and the vibration eigenvector(s) obtained for the at least one entry of data derived from a frequency domain vibration to obtain a preload assessment result; and E) Determine, based on the result of the preload assessment, whether preload deterioration has occurred on the ball screw (4), wherein the method is further characterized by the following steps following step E): H) obtaining at least one entry of time-domain inertial force data based on the inertial force signal received from the second sensor (3) when it is determined that preload degradation has occurred on the ball screw (4); I) for each of the at least one entry of time-domain inertial force data, obtaining an inertial force eigenvector based on the entry of time-domain inertial force data; J) performing a game assessment based on a plurality of reference inertial force vectors, a game assessment range, and the inertial force eigenvector(s) obtained for the at least one item of time-domain inertial force data to obtain a game assessment result; and K) Determine, based on the result of the clearance assessment, whether there is any clearance on the ball screw (4). [2] Method according to claim 1, characterized by that step B) comprises the following sub-steps: B-1) performing envelope processing on the entry of time-domain vibration data to result in an entry of processed time-domain vibration data; B-2) retrieving, from the set of processed time-domain vibration data, a set of target time-domain vibration data corresponding to a period in which the nut (41) of the ball screw (4) moves at a constant speed; and B-3) Obtaining the at least one entry of frequency domain vibration derived data based on the entry of target time domain vibration data. [3] Method according to claim 2, characterized by that step B-3) comprises the following sub-steps: B-3-1) Obtaining at least one entry of time-domain oscillation-derived data based on the entry of target time-domain oscillation data; and B-3-2) performing, for each of the at least one entry of time-domain oscillation-derived data, a Fourier transform on the entry of time-domain oscillation-derived data to result in the entry of frequency-domain oscillation-derived data. [4] Method according to claim 1, characterized by in that in step C), each of the vibration eigenvector(s) comprises at least one of the following: a kurtosis eigenvector indicating a kurtosis of the entry of data derived from a frequency domain vibration, a maximum peak eigenvector indicating a maximum peak value of the entry of data derived from a frequency domain vibration, and a total energy eigenvector indicating the total energy of the entry of data derived from a frequency domain vibration. [5] The method according to claim 1, wherein the computer device (1) stores a plurality of first training oscillation eigenvectors and a plurality of second training oscillation eigenvectors, the method being further characterized by the following steps before step D): F) using the first training vibration eigenvectors as inputs to machine learning performed using an unsupervised learning algorithm to obtain reference vibration vectors located in a data space spanned by the first training vibration eigenvectors; and G) Obtaining the preload assessment range based on the reference vibration vectors and the second training vibration eigenvectors. [6] Method according to claim 5, characterized by that in step F): the unsupervised learning algorithm includes a clustering algorithm; and the reference vibration vectors comprise central vectors each representing a plurality of vibration clusters obtained by performing the unsupervised learning algorithm. [7] Method according to claim 5, characterized by that in step F): the unsupervised learning algorithm comprises a self-organizing map algorithm; and the reference oscillation vectors comprise vectors each corresponding to neurons obtained by performing the self-organizing map algorithm and updated more than a preset number of times. [8] Method according to claim 5, characterized by that step G) comprises the following sub-steps: G-1) for each of the second training vibration eigenvectors, calculating a plurality of first candidate vibration distances each between the second training vibration eigenvector and a respective one of the reference vibration vectors obtained in step F); G-2) for each of the second training oscillation eigenvectors for which the corresponding first candidate oscillation distances were calculated, determining a shortest of the first candidate oscillation distances to serve as the first target oscillation distance; and G-3) Obtaining the preload assessment range based on the first target vibration distances determined for the second training vibration eigenvectors. [9] Method according to claim 1, characterized by that step D) comprises the following sub-steps: D-1) for each of the vibration eigenvector(s), calculating a plurality of second candidate vibration distances each between the vibration eigenvector and a respective one of the reference vibration vectors; D-2) for each of the vibration eigenvector(s) for which the corresponding second candidate vibration distances were calculated, determining a shortest of the second candidate vibration distances to serve as the second target vibration distance; and D-3) Obtaining the preload assessment result based on the preload assessment range and the second target vibration distance(s) determined for the vibration eigenvector(s).

Images