GEAR TEETH SIGNAL PROCESSING SYSTEM BY WINDOWING AND FOURIER TRANSFORM, METHOD AND PROGRAM BASED ON SUCH A SYSTEM
The system processes gear tooth signals using windowing and Fourier transform to analyze gear micro-geometry, addressing the limitations of existing methods by providing a detailed frequency analysis and identifying phantom orders, ensuring accurate gear noise detection.
Patent Information
- Authority / Receiving Office
- FR · FR
- Patent Type
- Patents
- Current Assignee / Owner
- STELLANTIS AUTO SAS
- Filing Date
- 2024-07-04
- Publication Date
- 2026-05-22
AI Technical Summary
Existing gear noise detection methods, such as 'gravel measurement' and sine interpolation, fail to provide a detailed analysis of gear noise parameters, leading to inadequate detection of noisy gears and insufficient normative criteria for defining and measuring gear noise behavior.
A system and method for processing gear tooth signals using windowing and Fourier transform techniques, involving data selection, rectification, angular basis change, weighting, and Fourier transform to identify phantom orders, allowing for a detailed analysis of gear micro-geometry frequency content.
Enables the detection of the entire frequency content of gear micro-geometry, providing a continuous and faithful representation of gear noise characteristics, effectively identifying and removing phantom orders.
Smart Images

Figure 00000008_0000 
Figure 00000008_0001 
Figure 00000008_0002
Abstract
Description
Title of the invention: GEAR TEETH SIGNAL PROCESSING SYSTEM BY WINDOWING AND FOURIER TRANSFORM, PROCESS AND PROGRAM BASED ON SUCH A SYSTEM
[0001] The invention relates to the field of gear noise detection systems by physical measurement of gear teeth.
[0002] In acceptance testing of gearboxes and reducers, noise and vibration, vibration or transmission errors, or acoustic measurements are carried out to guarantee the good quality of the final product. To perform a complete test, it is necessary to accelerate and decelerate all the gears of the gearbox to ensure the detection of any failure or noise in order to protect the user.
[0003] This speed variation involves numerous emitted frequencies that are difficult to link to a specific component. One solution is to calculate all the emitted frequencies in order. Order can be defined as the number of events per revolution of a rotating part. To compare the results from the test bench with what is measured on the gear, a similar order analysis approach must be applied based on an extended definition of the roll angle.
[0004] Traditionally, gears are measured on a dedicated measuring machine consisting of a rotating table on which the gear is clamped and a transverse axis on which a probe is placed. The combination of rotation and translation allows the probe to be in contact with the flank of the gear and consequently to measure defects along a profile, a helix, or division errors; or any other characteristic according to any gear measurement standard.
[0005] Typically, in the automotive industry, gears are measured over 4 to 6 equally spaced teeth. Some developments in the analysis of gear micro-geometry show that the normative criteria for defining and measuring gears are insufficient to describe the noise behavior of gears.
[0006] A solution known as "gravel measurement" exists. This solution is unsatisfactory because its analytical developments are not known. Furthermore, in some cases, it fails to detect noisy gears. The corresponding noise calculation method appears to be based on a sine interpolation method according to US patent 10,753,730 B2, but this type of solution does not appear to suitable because it does not allow for a detailed analysis of the parameters related to the measurements.
[0007] An objective of the present invention is to remedy the defects of the prior art, and in particular to propose a solution for analyzing gear tooth noise based on precise parameters enabling the detection of the entire frequency content of a micro-geometry of a gear.
[0008] To achieve this objective, the invention proposes a system for processing a topographic measurement signal of the teeth of a gear mechanism, the teeth each comprising a root at the level of a tooth base, a head at the level of a tooth apex following the root, the signal comprising a root zone corresponding to the roots, a head zone corresponding to the heads, and an intermediate zone between the root zone and the head zone, the intermediate zone comprising tooth corrections, the processing system comprising: - a means of receiving data from the measurement signal; - a selection method allowing the selection of data corresponding to the intermediate zone in order to obtain selected data; - preferably a rectification method allowing the removal of denture corrections from the selected data in order to obtain rectified data; - a means of changing the angular basis allowing the angular basis of the selected data or rectified data to be changed in order to obtain shifted data; - a weighting and windowing method allowing the shifted data to be weighted in order to obtain weighted data; - a means of calculating a Fourier transform of the weighted data so as to obtain a spectrum of transformed data; - an identification method enabling the identification of phantom orders in the spectrum of transformed data.
[0009] Advantageously, the invention proposes a new calculation method based on windowing and weighting of measurements and then the application of a Fourier transform.
[0010] This makes it possible to detect the entire frequency content of the micro-geometry of a gear.
[0011] Preferably, the rectification means allows a 2nd order polynomial to be applied to the selected data.
[0012] This allows the signals relating to the bulges of the tooth above the head of the tooth to be removed.
[0013] Preferably, the weighting and windowing means allows a Hann window to be implemented on the shifted data so as to obtain the weighted data.
[0014] This makes it possible to obtain continuity of the signal at the level of the overlaps, and to reconstruct a continuous signal.
[0015] More preferably, the calculation means allows a Fourier transform to be applied to weighted data corresponding to one revolution of a gear wheel in order to obtain the spectrum of transformed data.
[0016] This makes it possible to obtain a one-dimensional spectrum of the microgeometric transmission error.
[0017] Another object of the invention relates to a method for processing a topographic measurement signal of teeth of a gear mechanism, by means of a processing system according to the invention, characterized in that it comprises the following steps: - a data reception step of the measurement signal; - a selection step in which the data corresponding to the intermediate zone are selected in order to obtain selected data; - preferably a rectification step in which the denture corrections are removed from the selected data so as to obtain rectified data; - an angular basis change step in which the angular basis of the selected data or rectified data is changed so as to obtain shifted data; - a weighting and windowing step in which the shifted data is weighted in order to obtain weighted data; - a step of calculating a Fourier transform on the weighted data, corresponding in particular to a revolution of a gear wheel, so as to obtain a spectrum of transformed data; - an identification step in which phantom orders are identified in the spectrum of transformed data.
[0018] Preferably, in the rectification step, a polynomial of order 2 is applied to the selected data.
[0019] Preferably, in the weighting and windowing step, a Hann window is implemented on the shifted data so as to obtain the weighted data.
[0020] The invention further relates to a computer program comprising program code instructions for executing the steps of the processing method according to the invention, when said program is running on a computer.
[0021] The invention will be further detailed by describing non-limiting embodiments, and based on the accompanying figures in which: - [Fig.l] schematically illustrates the indices of the measured points derived from the raw measurements; - [Fig.2] schematically illustrates indices according to [Fig.1] without the root zones and the head zones; - [Fig.3] schematically illustrates indices according to [Fig.1] without the root zones, head zones and bulge zones on the tip of the teeth; - [Fig.4] schematically illustrates the positions and amplitudes of the rectified data in the same angular basis of all the teeth; - [Fig.5] schematically illustrates a close-up view of the curves of [Fig.4]; - [Fig.6] schematically illustrates an implementation of Hann's weighting and windowing; - [Fig.7] schematically illustrates a one-dimensional spectrum of micro-geometric transmission error, based on deviation as a function of angular position; - [Fig.8] schematically illustrates a spectrum like that of [Fig.7], based on amplitude as a function of angular position; and - [Fig.9] schematically illustrates a spectrum in order, based on amplitude as a function of order.
[0022] The invention proposes a system and a method for processing gear measurement data. It is based on measuring all the gear teeth along a tooth profile or a tooth helix. These measurements are then placed in a common angular basis and post-processed to apply a Fourier transform.
[0023] The invention proposes to modify the post-processing of measurement data; and to obtain an ordered spectrum using a Fourier transform.
[0024] With this solution, the user has control over the parameters of the spectrum calculation algorithm. Furthermore, since the measurement data are not interpolated, they are more faithful to the physical reality of the gear.
[0025] The signal includes a root component corresponding to a first working part at the base of the tooth. This can be referred to as root R. The signal further includes a tip component corresponding to a second working part at the apex of the tooth. The corresponding signal is also visible in the raw indices.
[0026] The algorithm is implemented based on the raw measurement data illustrated in [Fig. 1]. The abscissas of this figure represent an index I, and the ordinates represent a deviation D in pm. The left part of the figure corresponds to the roots R of the teeth, and the right part corresponds to the heads T of the teeth.
[0027] In the context of the invention, only the data useful during meshing (intermediate zone E in [Fig. 1]) is retained, namely the data relating to the measurements corresponding to the positions between the root and tip positions of the gear teeth. Indeed, these zones are not relevant for the identification of phantom orders. [Fig. 2] illustrates the raw data without the tips and roots R; and the [Fig.3] illustrates the raw data without the T heads, R roots and the bulge at the top of the tooth.
[0028] These data are then rectified with a polynomial, preferably a second-degree polynomial, with the aim of removing the corrections of the teeth (chamfer of the tooth).
[0029] The rectified data are illustrated in [Fig. 4]. The abscissas of this figure represent an angular position P in radians, and the ordinates represent an amplitude A in pm. [Fig. 5] is a zoom of [Fig. 4] to show the overlaps.
[0030] Alternatively, these corrections to the teeth (domed) can be retained
[0031] The parts corresponding to the roots of the teeth are removed.
[0032] The parts corresponding to the heads of the teeth are removed.
[0033] The data are then placed in the same angular basis.
[0034] The data are then post-processed by weighting and Hann windowing to obtain signal continuity. This allows for the processing of the overlap due to the tooth helicoid. Figure 6 illustrates the Hann windowing. H corresponds to the Hann curve, B to the raw data, and W to the profile after weighting.
[0035] Alternatively, other types of window may be used in certain cases.
[0036] Figure 7 illustrates a one-dimensional spectrum of the microgeometric transmission error (or "1D Micro Geometry Transmission error" - abbreviated 1D MGTE), based on the deviation D as a function of the angular position P. Figure 8 illustrates a one-dimensional spectrum of the microgeometric transmission error based on the amplitude A as a function of the angular position P.
[0037] Finally, a Fourier transform of the signal is calculated over one revolution of the gear wheel (2*pi). This makes it possible to obtain an ordered spectrum by keeping only the magnitudes and thus revealing the different harmonic contents.
[0038] Figure [Fig.9] illustrates a one-dimensional spectrum of the microgeometric transmission error based on the amplitude A as a function of the order O.
[0039] Reference numeral G relates to a phantom order identified within the scope of the invention. This phantom order is not periodically found in measurements such as [Fig. 9], and represents phantom noise or unwanted vibrations in the corresponding gear wheel during operation. Such a gear wheel should therefore not be used for the sake of user comfort.
[0040] The invention further relates to a pre-checking method and program implementing the elements and tools discussed above. The program can be loaded into a computer's memory.
Claims
Demands
1. A system for processing a topographic measurement signal of teeth of a gear mechanism, the teeth each comprising a root at the level of a tooth base, a head at the level of a tooth apex following the root, the signal comprising a root zone (R) corresponding to the roots (R), a head zone (T) corresponding to the heads, and an intermediate zone (E) between the root zone (R) and the head zone (T), the intermediate zone comprising tooth corrections, the processing system comprising: - a means for receiving data from the measurement signal; - a selection means for selecting the data corresponding to the intermediate zone (E) so as to obtain selected data; - preferably a rectification means for removing the tooth corrections from the selected data so as to obtain rectified data;- a means for changing the angular basis, allowing the angular basis of the selected or rectified data to be changed in order to obtain shifted data; - a means for weighting and windowing, allowing the shifted data to be weighted in order to obtain weighted data; - a means for calculating a Fourier transform of the weighted data in order to obtain a spectrum of transformed data; - an identification means for identifying phantom orders (G) in the spectrum of transformed data.
2. Processing system according to claim 1, characterized in that the rectification means allows a 2nd order polynomial to be applied to the selected data.
3. Processing system according to any one of claims 1 to 2, characterized in that the weighting and windowing means makes it possible to implement a Hann window on the shifted data so as to obtain the weighted data.
4. A processing system according to any one of claims 1 to 3, characterized in that the calculation means allows for the application of a Fourier transform on weighted data corresponding to one revolution of a gear wheel in order to obtain the spectrum of transformed data.
5. A method for processing a topographic measurement signal of teeth of a gear mechanism, by means of a processing system according to any one of claims 1 to 4, characterized in that it comprises the following steps: - a data reception step of the measurement signal; - a selection step in which the data corresponding to the intermediate zone (E) are selected so as to obtain selected data; - preferably a rectification step in which the tooth corrections are removed from the selected data so as to obtain rectified data; - an angular basis change step in which the angular basis of the selected data or the rectified data is changed so as to obtain shifted data; - a weighting and windowing step in which the shifted data is weighted so as to obtain weighted data;- a step of calculating a Fourier transform on the weighted data, corresponding in particular to one revolution of a gear wheel, so as to obtain a spectrum of transformed data; - an identification step in which phantom orders (G) are identified in the spectrum of transformed data.
6. Processing method according to claim 5, characterized in that in the rectification step, a polynomial of order 2 is applied to the selected data.
7. Processing method according to any one of claims 5 to 6, characterized in that in the weighting and windowing step, a Hann window is implemented on the shifted data so as to obtain the weighted data.
8. Computer program comprising program code instructions for carrying out the steps of the processing method according to any one of claims 5 to 7, when said program is running on a computer.