How to check the load on a rolling pneumatic tire

Abstract

A method for verifying a load on a pneumatic tire is disclosed, the method comprising: Fixing a sensor to the tire to generate an acceleration relative to a crown normal; Acquiring (201) a time signal SigTDR (101) containing the amplitude of acceleration during rolling; A velocity W associated with at least one portion of the signal SigTDR reference determining (202) W reference normalizing (203) that portion of the signal SigTDR by a variable that is a function F proportional to the square of Angularly resampling (204) the portion of the signal SigTDR; A step (205) of defining an energy density S using a threshold A or a spectral variable β by spectral analysis based on the angularly resampled normalized signal SigTDR; Specifying (206) a transformation Def% as a function G of S or β; A step (207) of determining the load Z by a function H of Def%; Includes.

Description

[Technical field]

[0001] The present invention relates to the field of measurement signals provided by measuring means mounted on a mounting assembly of a land vehicle during rolling in order to determine the static load applied to the mounting assembly. [Background technology]

[0002] Recent developments in coupled mounting assemblies, measuring physical variables of the mounting assembly with sensors mounted on the mounting assembly, provide a determination of the mounting assembly's condition and thus open the door to the development of services related to the monitoring of the mounting assembly's condition. While common variables measured, such as the air pressure of the mounting assembly or the temperature of this mounting assembly, only change slightly during the rotation of the mounting assembly and are therefore less sensitive to the measurement noise generated by the rotation of the mounting assembly on a surface of random roughness, more subtle variables are very sensitive to the physical phenomena associated with the rotation of the mounting assembly. Furthermore, the mounting assembly is subjected to external forces such as static load or air pressure. Other forces such as loads are applied at all times, especially while stationary. These applied forces can affect the subtle variables that are to be measured. Finally, new services require cleaning of the directly measured physical variables before useful information can be obtained from the measurement signals, such as the static load of the tire casing. Summary of the Invention [Problem to be solved by the invention]

[0003] One of the objects of the invention below is to solve the problem of disturbances in the measurement signal generated by the sensor in order to obtain only measurements free of disturbances of certain physical phenomena, aiming to obtain a scalar value for the static load on the tire casing.

[0004] For a better understanding of the invention, the circumferential direction S, the axial direction A and the radial direction R are directions defined relative to a rotational reference frame about the inherent axis of rotation of the tire casing. The radial direction R is a direction extending perpendicularly away from the inherent axis of rotation. The axial direction A is a direction parallel to the inherent axis of rotation. Finally, the circumferential direction S forms a regular triangle with the predefined radial and axial directions. [Means for solving the problem]

[0005] The present invention relates to a method for ascertaining the load on a tire casing, the tire casing being mounted on a wheel so as to form a mounting assembly in rolling motion at a rotational speed W. The tire casing has a crown in contact with the ground and rotating about a natural axis of rotation. The method comprises: - fixing at least one sensor to a crown of a tire casing to generate at least one output signal responsive to acceleration applied to said sensor in the tire casing in a direction perpendicular to the crown; - obtaining, during rolling, at least one first time signal Sig comprising at least an amplitude of at least one output signal; - Number of wheel revolutions N at least once TDR The first signal is divided into two parts, and the wheel rotation signal Sig TDR and -Wheel rotation signal Sig TDR At least one reference speed W associated with at least one portion of reference determining - Number of wheel revolutions N at least once TDR Over the entire time, the reference speed W reference The wheel rotation signal Sig is calculated by a function F proportional to the square of TDR normalizing at least a portion of -Wheel rotation signal Sig TDR Angularly resampling at least a portion of - at least one angularly resampled normalized wheel rotation signal Sig TDR, using a threshold A to obtain at least one first energy density S, or, if the angular pitch is fixed, an angularly resampled normalized wheel rotation signal Sig TDR defining at least one spectral variable derived from a spectral signal spect(Sig) of at least one portion of - determining a deformation Def% of the tire casing as a function G of at least one first energy density S or of at least one spectral variable; - defining the load Z applied to the mounting assembly using a bijective function H that includes as a variable at least the deformation Def% of the tire casing; Includes.

[0006] The signal received from the sensor is the time amplitude of the sensor acceleration in a direction normal to the crown during rolling of the mounting assembly under specific conditions. The acquired signal thus represents amplitude variations over a portion of the wheel revolution relative to the tire casing, which may include not only those related to the traversal of the contact patch by that portion of the tire casing on which the sensor is mounted, but also those related to other specific zones of the wheel revolution, for example those corresponding to angular sectors opposite the contact patch that are subject to adverse deflections, or those corresponding to angular sectors located at 90 degrees from the contact patch relative to the axis of rotation. In all these zones, variations in the movement of the accelerometric sensor may be observed on the output signal, depending on the sensitivity of the acceleration sensor.

[0007] This first acquired signal is related to a reference speed that may be determined on this first signal or obtained from another source, such as another signal, or may be obtained from the output of a variable by a system external to the mounting assembly. This reference speed necessarily relates to the same time frame as the portion of the first signal. This reference speed serves to normalize the amplitude of the first signal with a function F whose variable is the reference speed. This function F is a squared power function. The sensor signal is normalized as a function of the dependency of the sensor signal amplitude on the reference speed, if this dependency is recognized as a parasitic signal of tire casing deformation. The first normalized signal is thus independent of this reference speed. For example, this reference speed may be the rotational speed of the mounting assembly or the translational speed of the mounting assembly in the direction of movement of the mounting assembly. As a result, the first signal can be used independently of the reference speed related to the rotation of the mounting assembly.

[0008] The method also includes a step of partitioning the first signal Sig over a number of wheel revolutions in order to exploit the periodicity of the sensor signal relative to the natural rotation of the tire casing under rolling conditions. However, this step does not require that the number of wheel revolutions is an integer number, as long as the number of wheel revolutions is at least greater than 1, and the signal can be partitioned over the actual number of wheel revolutions. Preferably, multiple wheel revolutions are used.

[0009] The method also includes an angular resampling of the first signal or the wheel rotation signal, which can be performed before or after the normalization step. This step allows the conversion of the time signal into a spatial signal by synchronizing the time signal with one or more angular references of the mounting assembly. This angular reference can first be obtained from the first signal by a specific response of the sensor to the individual azimuth angles of the wheel rotation. However, this angular reference can also be obtained from another signal of a sensor that shares a common timer with the first signal. This synchronization of the shared timer or signal is natural when the two sensors are from the same device or when the signals are transmitted to a common device. This angular resampling naturally allows the generation of a spatial signal that is periodic with respect to the wheel rotation. Therefore, to generate a completely angularly periodic signal, it is sufficient to interpolate the signal over a set angular interval. However, this resampling can also generate an angularly periodic signal if the mounting assembly is to undergo a movement with a variable speed. It is not essential for the method that the angular resampling generates an output signal with a constant angular pitch.

[0010] In a first option, the energy density S can be defined by simply comparing the amplitude level of the wheel rotation signal with a threshold value A. The amplitude of the wheel rotation signal with respect to the threshold value A (which may for example simply be a unit value) allows generating from the wheel rotation signal a pair of deformation energy density components, one positive and one negative (S+, S-). Thus, the method only defines the deformation energy density of the tire casing and distributes it between the two subsets depending on its position with respect to the threshold value A. These are simple operations to perform and consume little resources.

[0011] In a second option, the method includes a step of performing a spectral analysis from said portion of the angularly resampled normalized wheel rotation signal. This is useful to ensure that said portion of the initial signal is defined at a constant angular pitch, which ensures a regular spatial discretization of the sensor signal. If necessary, a step of angular resampling ensures that the angular pitch is constant to allow a good quality spectral analysis, which may require an interpolation method of the measurement points to redefine the signal at a constant angular pitch. The method then includes a step of defining a spectral variable or spectral variables related to the spectral signal obtained from the previous step.

[0012] With either option, the method determines the deformation of the tire casing as a function of the calculated energy density or at least one spectral variable. Thus, in the first option, the deformation represents the normalized deformation energy over one physical wheel revolution of the tire casing. As a result, an energy variable is identified that is related to the deformation of the tire casing under rolling load. In the second option, the deformation is expressed in the form of a scalar or vector that is an invariant of the tire casing under rolling static load.

[0013] Of course, one wheel revolution is necessary to determine the tire casing deformation Def%, but it is preferable to have at least 5, or even 10, wheel revolutions so that the results can be averaged, which makes it possible to overcome any unpredictable phenomena in the signal, such as obstacles on the roadway on which the tire casing is rolling. Therefore, in industrial mode, this improves the accuracy of the method.

[0014] Finally, whatever the option selected, the method includes a step of determining the static load Z on the mounting assembly using a function H that depends on the tire casing deformation Def%, itself expressed differently depending on the option selected. As a result, the representation spaces of the tire casing deformation Def% are different, so that the function H is related to the selection of the representation space of the tire casing deformation Def%, which in turn is related to the option selected.

[0015] Advantageously, the reference speed W reference The step of determining the wheel rotation signal Sig according to the following formula: TDR From, or the first signal Sig TDR and establishing from the signal synchronized with the ratio of the angular change to the duration separating two azimuthal positions about a unique axis of rotation for the sensor in the tire casing; [Formula 1] W reference =Δ(α) / Δ(t) where α is the angular position and t is the time abscissa related to the angular position.

[0016] If the reference speed corresponds to the rotational angular speed of the tire casing, this reference speed is calculated over the angular change of the signal between two known positions. Preferably, this reference speed is evaluated over a signal duration of less than one wheel revolution, which allows a quick definition of the reference speed and a normalization step to be performed on a portion of the first signal in an electronic device associated with the sensor. Furthermore, this allows an angular resampling of this portion of the first signal with better accuracy when the tire casing moves with a variable angular speed. Indeed, at the level of wheel rotation, the variations in angular speed are necessarily small for tires with a development length that can reach 2 meters for car tires and 3 meters for truck tires. The acceleration or deceleration applied to the tire casing over this length is naturally small in the drive and braking systems of current vehicles. Of course, it is entirely possible to integrate the angular velocity variations during the wheel rotation at finer azimuth settings in order to take into account small variations in angular velocity that occur during the wheel rotation, for example before and after passing over a contact patch, or when encountering discontinuities in the movement on the ground, such as a cross bar on the ground, etc. This precision with respect to the reference speed during the wheel rotation then allows not only to normalize the signal more accurately, but also to improve the angular precision of the angular position with respect to the measurement point of the first signal during the angle resampling step, thus improving the desired precision for sensing the smallest variations during the wheel rotation.

[0017] According to a particular embodiment, the azimuthal position of the tire casing is determined based on the position of the wheel rotation signal Sig. TDR The wheel rotation signal Sig corresponds to any specified angular position from the signal synchronized with TDR The angular positions detectable from are included in the group.

[0018] These are the azimuth positions that affect the signal from the acceleration sensor and correspond to specific angular positions. These positions are therefore easy to identify on the signal from the sensor. Moreover, it is also easy to assign their azimuth reference. In fact, the center position of the contact patch corresponds to an azimuth position of 0 degrees or 180 degrees with respect to the normal to the ground. If the length of the contact patch is determined from the entry and exit points into the contact patch, the angle that the contact patch forms can be established as the ratio between the length of the contact patch and the development length of one revolution or 360 degrees of the tire casing. The sectors that the contact patch forms on both sides of the normal to the ground are divided equally. Naturally, by using signals other than the first signal, angular sectoring finer than one revolution of the wheel is also possible, as with an angle encoder.

[0019] Advantageously, the angular pitch is less than 18 degrees.

[0020] It is therefore possible to ensure that one of the measurement points is located on the contact surface, so that at least the acceleration change between this sampling point and the closest point is observed, allowing the entry and exit points of the first signal into and out of the contact surface to be determined.

[0021] Highly advantageously, the angular pitch is less than 6 degrees, preferably less than 3 degrees.

[0022] Using a finer angular pitch allows sensing multiple measurement points within the contact area. This fine observation allows improving the accuracy of the method by avoiding the spatial discretization of the points, which is not necessarily regular here. The large number of points also ensures the absence of disturbances due to inconsistent measurements of the sensors.

[0023] According to a preferred embodiment, the method comprises the steps of: TDR The data from at least one portion of the TDRaggregating over at least one sub-portion of at least one portion of the normalized wheel rotation signal Sig TDR At least one sub-portion of the portion of the angularly resampled normalized wheel rotation signal Sig TDR At least one part of.

[0024] Preferably, the angularly resampled normalized wheel rotation signal Sig TDR A sub-portion of at least one portion is an integer multiple of a wheel revolution, and highly preferably one wheel revolution.

[0025] This step allows the identification of the wheel rotation signal taking into account the variations in wheel rotation and reduces the size of the vectors to be processed in the last two steps of the method, namely the steps used to identify the first energy density S or the spectral variables after the spectral analysis of the signal. To this end, the subparts of the angularly resampled normalized wheel rotation signal are integer multiples of one wheel revolution in order to exploit the inherent periodicity of the signal with respect to the wheel rotation, which is favorable for a good quality spectral analysis.

[0026] According to a preferred embodiment, the data aggregation step includes one of the methods included in the group consisting of average over a decile interval, median, selection of a decile or interval, methods of interpolation, weighted or unweighted average, optimization of a parametric model of the tire deformation.

[0027] The purpose of the aggregation is to set up a means to execute on a new angular distribution of the first signal in order to resolve the set of raw measurement data. The aggregation step is intended to provide a balanced signal for the measurement points of the angular pitch selected by the operator according to the tire casing deformation observed. To this end, a method of optimizing a parametric model of the tire deformation is ideal, since this parametric model is theoretical and may not take into account the measurement noise associated with the entire measurement chain to which it is applied. The output signal from the aggregation step is the theoretical output of the parametric model that has the smallest variance with the set of recorded measurement points.

[0028] Advantageously, after synchronising the first signal Sig with respect to the angular position of the tyre casing, and before the normalisation step, a correction Corr is made to the first signal Sig in order to take into account the influence of the earth's gravity.

[0029] A drawback of the acceleration signal is that it is sensitive to the earth's gravity when it is oriented in a direction approximately parallel to the earth's gravity. In the case of a tire casing, the sensor is rotationally coupled to the tire casing. As a result, when the sensor is oriented radially, the amplitude of the sensor signal is affected by the earth's gravity during wheel rotation. This is reflected in the signal in the form of a sine function of amplitude coupled to the earth's gravity, with nodes at the azimuth of the tire casing separated by 180 degrees when the sensor is oriented in line with the gravity vector, i.e. substantially perpendicular to the ground. Conversely, the sensor signal is not affected by the earth's gravity when the sensor is oriented parallel to the ground, i.e. corresponding to two azimuth positions separated by 180 degrees from each other and approximately ±90 degrees from the gravity vector. To remove this parasitic component of the acceleration signal, it is necessary to couple the amplitude of the signal with the corresponding sine function by synchronizing the first sensor signal with the vertical position relative to the ground, which corresponds to the direction of the gravity vector.

[0030] According to a particular embodiment, the method comprises the steps of: TDRThe method includes filtering at least a portion of the signal.

[0031] High frequency interference may remain in the angularly resampled normalized signal to be processed in the next step. For example, in the case of option 1, i.e., defining the energy density S using a threshold A, filtering the signal simplifies the step by minimizing possible errors.

[0032] According to a second preferred embodiment, the angularly resampled normalized wheel rotation signal Sig TDR The step of obtaining at least one spectral variable from the spectral signal spect(Sig) of at least a portion of the spectrum includes a step of identifying at least one spectral variable over at least one spectral block of the spectral signal spect(Sig), preferably over a first positive spectral block of the spectral signal spect(Sig).

[0033] Preferably, the at least one identified spectral variable is included in the group consisting of maximum, median, mean, first block passband, first block area under the curve, median frequency, mean frequency, maximum frequency.

[0034] According to a first preferred embodiment, the angularly resampled normalized wheel rotation signal Sig is calculated using a threshold A. TDR The step of obtaining at least one energy density S from at least one portion of the angularly resampled normalized wheel rotation signal Sig TDR When at least one portion of the first energy density S is greater than a threshold A, + Steps defining the angularly resampled normalized wheel rotation signal Sig TDR is equal to or less than the threshold A, - The method includes the step of defining:

[0035] Preferably, this ratio is between 0.5 and 0.9.

[0036] The purpose of threshold A is to distribute the discretized points of the angularly resampled normalized wheel rotation signal SigTDR between the energy densities S+ and S-. If the signal carries a lot of noise, as is the case in the absence of a data aggregation or filtering step, this point distribution can be affected by this interference. The purpose of threshold A is to correct this imperfection associated with the measurement signal. The value of threshold A is a quality function of the angularly resampled normalized wheel rotation signal SigTDR. If the method employs any step and the road roughness is small, values ​​at the top of the range will be preferred.

[0037] Highly preferred is a positive energy density S + and negative energy density S - The definition of is given by the following formula: [Formula 2a] JPEG2024532702000002.jpg15170[Formula 2b] JPEG2024532702000003.jpg16170

[0038] where u is the angularly resampled normalized wheel rotation signal Sig TDR is the abscissa value of

[0039] This is a simple method of obtaining a scalar value for each energy density from a discretized signal derived from the angularly resampled normalized wheel rotation signal using elementary mathematical and logical operations that can be performed within the electronic devices associated with the sensor.

[0040] Advantageously, the function G is a linear function.

[0041] In the first alternative, the function G is a linear function of the spectral density S according to the following formula:

[0042] [Formula 3a]

[0043] G(X)=X / N' TdR

[0044] Therefore, this is S + Or S - This is a basic formula for tire casing deformation that applies to the tire casing deformation. - corresponds to the energy density calculated from the mass points of the tire's development length, which contains mass points at or very close to the contact patch at a precise instant of time T. Specifically, these points have near-zero absolute acceleration when passing through the contact patch, and therefore are necessarily below the threshold A. By default, the energy density S + corresponds to the energy density at other points on the tire's development length, in particular at the points on the outer contact patch. It demonstrates that there exists an invariant related to the deformation of the tire casing under the load Z. S + In the case of using only the contact patch, the variations outside the contact patch are not so significant, so a high degree of spatial discretization is not necessary. The advantage of this is that it reduces the required sampling frequency of the electronic device connected to the sensor, or allows obtaining accurate information about the tire casing deformation at high rotation speeds.

[0045] However, the function G has a spectral density S according to the following formula: + and S - is a linear function of: [Formula 3b] G(X,Y)=(X+Y) / (2*N' TdR )

[0046] In that case, the tire deformation energy needs to be summed over the entire tire development length. To ensure that the measurement uncertainty is minimized, a set of measurement points is used to measure the acceleration normal to the crown to determine the tire casing deformation, which reduces energy consumption compared to high frequency analysis.

[0047] In the case of the first option, the applicant was surprised to find that examining the first positive frequency block of the spectral signal spect(Sig) is sufficient to identify one or more variables associated with this first block that are relevant for determining the tire casing deformation with adequate quality. The variables that are most sensitive to the tire casing deformation are specified in the prepared list. These are standard variables of the spectral signal that require little computational resources, which is favorable for the method. Moreover, these variables are primarily sensitive to the tire casing deformation and are less sensitive to secondary variables. As a result, these variables are ideally suited to the general deformation of the tire casing as a whole, for example deformations caused by global forces on the entire tire casing, such as static loads.

[0048] In this case, the function G does not need to be sophisticated, and the applicant has found that a linear function G of one or more spectral variables can adequately determine the deformation of the tire casing in response to various usage conditions of the tire casing subjected to varying static loads.

[0049] According to a preferred embodiment, the function H is an affine or power function using the following formula: [Formula 10a] JPEG2024532702000004.jpg11170 or [Formula 10b] JPEG2024532702000005.jpg11170 where (A,B) or (X,Y) are parameters relating to the mounting assembly.

[0050] Depending on whether the objective is to evaluate the load Z applied to the mounting assembly under normal or special conditions of use, one or the other formula for the function H must be used. Indeed, in the conventional field of use of tire casings applying the ETRTO (European Tire and Rim Technical Organization) rules, a simple affine function correctly describes the variation of the load Z as a function of the tire casing deformation Def%. As a result, the knowledge of the mounting assembly, and in particular the tire casing, allows a reliable determination of the load Z applied to the mounting assembly. However, if it is intended to extend the range of modeling the load as a function of the tire casing deformation for a specific application, for example very low or high loads Z, a power-type expression is more suitable. However, in the general area of ​​use, both functions give very similar results and are sufficient for the desired accuracy of less than 10%, preferably less than 5%.

[0051] According to a preferred embodiment, when the mounting assembly is inflated to an air pressure P, the parameter A or X depends on at least the air pressure P, and preferably the parameter A or X is an affine function of the tire pressure P according to the following formula: [Formula 11a] JPEG2024532702000006.jpg11170[Formula 11b] JPEG2024532702000007.jpg11170 where (a1,a2) or (x1,x2) are coefficients related to the mounting assembly.

[0052] According to a highly preferred embodiment, when the mounting assembly is inflated to an air pressure P, the parameter B or Y depends on at least the air pressure P, and preferably the parameter B or Y is an affine function of the tire pressure P according to the following equation: [Formula 12a] JPEG2024532702000008.jpg16170 or [Formula 12b] JPEG2024532702000009.jpg16170 where (b1,b2) or (y1,y2) are coefficients relating to the mounting assembly.

[0053] Most tire casings are mounted on the wheel and then inflated to an air pressure P that depends on the type of tire casing. This air pressure P influences the mechanical behavior of the mounting assembly and especially of the tire casing. As a consequence, the deformations of the tire casing are influenced by this variable. This effect then needs to be taken into account for the coefficients A or X. This is the simplest expression of the dependence of the parameter A on the air pressure P and is practical, especially in the traditional field of use of tire casings subject to ETRTO regulations.

[0054] The dependence of these second parameters B or Y of the function H on the inflation pressure P is similar to the change in slope of the function H as a function of the tire casing deformation Def%. This change in slope is not as obvious as the change in the stiffness of the tire casing at the inflation pressure P described by the first parameters A or X. However, the change in these second parameters B or Y with the inflation pressure P improves the estimation of the load Z on the mounting assembly and thus on the tire casing.

[0055] Thus, if the function H is an affine function that is completely dependent on the air pressure P, then up to four parameters a1, a2, b1 and b2 need to be identified to estimate the load Z on the mounting assembly. Of course, if the wheel is changed, the parameter set needs to be readjusted to give an accurate estimate. This parameter set can also be identified through characterization by digital simulation, or by experimental testing, or by a combination of the two.

[0056] The invention will be better understood from reading the following description, purely by way of non-limiting example, in connection with the accompanying drawings in which the same reference numbers represent the same parts in all cases, in which: [Brief description of the drawings]

[0057] [Figure 1] 1 shows an outline of the method according to the present invention. [Diagram 2] 1 shows an example of a first signal from a sensor. [Diagram 3] Angular resampling of the wheel rotation signal is shown. [Figure 4] 1 shows an example of a resampled normalized wheel rotation signal. [Diagram 5] 13 shows an example of the final signal after aggregating data over subportions of the wheel rotation signal. [Figure 6a] 13 is an illustration of the estimation of energy density S from an angularly resampled normalized wheel rotation signal. [Figure 6b] 1 is an illustration of the spectral signal spect(Sig) of wheel rotation. [Figure 7] 4 shows an estimate of the load Z on the rolling mounting assembly. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0058] 1 shows an overview of the method according to the invention, which performs several steps along different possible paths in order to finally obtain a scalar representative of the load applied to the mounting assembly from a first signal Sig obtained by temporal acquisition 201 of the amplitude output of an acceleration sensor during the rolling of a tire casing equipped with the sensor.

[0059] The first path derives from the time signal at the output of step 201 a reference speed W of the tire casing in its mounting assembly configuration, i.e., the tire casing mounted on the rim. reference Here, the first signal Sig101 is already segmented over a certain number of wheel revolutions (12 to be precise). As a result, the first signal Sig101 is determined based on the wheel revolution signal Sig TDR This reference speed can be the angular speed associated with the inherent rotation of the tire casing about its axis of rotation, but it can also be the translational speed per unit length in the direction of travel of the tire casing. This value corresponds to the wheel rotation signal Sig TDR However, the first signal and therefore the wheel rotation signal Sig TDRIt may also be determined from another signal that is time-synchronized with the

[0060] Next, the wheel rotation signal Sig TDR is the variable W obtained in step 202 reference The first signal resulting from step 201 is normalized (203) by a function F of: This function is a square power function. After this step 203, a signal normalized for the movement of the tire casing in the time description is obtained.

[0061] The normalized signal then needs to be angularly resampled to find a signal that is angularly periodic with respect to the wheel revolutions, via step 204. Then, after this step 204, the result is a normalized, angularly resampled signal over several wheel revolutions.

[0062] The second path is the wheel rotation signal Sig resulting from step 201. TDR The method includes a step of angularly resampling the first signal Sig from a first signal Sig that is also periodic with respect to the wheel rotation, by synchronizing the first signal with the morphology of the first signal or by synchronizing in time with the first signal another signal, which may come from another sensor or from another track of the same sensor, such as the circumferential acceleration of a three-dimensional accelerometer. This angular resampling of the first signal results in a signal that is periodic with respect to the wheel rotation at the end of step 204.

[0063] After synchronizing this angular signal with another time signal, a reference velocity is determined from the other time signal synchronized with the first signal, preferably the same other signal that was used to angularly resample the first signal in step 204. As a result, at the end of step 202, a reference velocity W reference is identified.

[0064] In that case, the reference speed makes it possible to normalize the angularly resampled signal from step 204 using a function of the reference speed variable, so that at the end of step 203, the angularly resampled normalized wheel rotation signal Sig TDR is given.

[0065] Optionally, whichever path is taken, the angularly resampled normalized wheel rotation signal Sig resulting from step 204 on the first path or step 203 on the second path is TDR This data aggregation in step 208 is performed on sub-portions of the input signal that are multiples of the wheel revolutions since the angularly resampled normalized signal is periodic in nature with respect to the wheel revolutions.

[0066] Alternatively, if the first signal 101 is contaminated by a known physical phenomenon, such as an accelerometer signal affected by the Earth's gravity, it may be useful (although not essential) to perform a correction of the first signal to this physical phenomenon in order to suppress the parasitic noise caused by the physical phenomenon. This correction can be performed at any step between steps 201 and 204, but necessarily before the data aggregation step 205, thereby improving the quality of the signal regarding the deformation of the tire casing. If the correction is performed after the normalization step, it should also be normalized so as not to introduce any correction errors.

[0067] In a first alternative, the method then includes a step 205 of determining an energy density related to the tire casing deformation from the angularly resampled normalized wheel rotation signal. This is done by determining the positive energy density S + or negative energy density S - , may be performed for only a portion of the wheel revolution, but preferably at least one complete wheel revolution is performed, which provides both variables. TDR On top of the wheel rotation speed NTDR It should also be remembered to record the energy density S. If the signal is separated by one wheel revolution, the determination of the energy density is related to the quality of the signal, and the optional step of data aggregation is justified. However, if the signal is separated over many wheel revolutions, the signal may contain additional fractions of wheel revolutions that will slightly change the value of the energy density. In this case, it is preferable to count the wheel revolutions from an azimuth position located 180 degrees from the center of the contact patch. The additional fractions of wheel revolutions will result in a positive energy density S + The energy density variation between these points is the negative energy density S - This is smaller than the approach and exit phases, which have a major impact on the contact area.

[0068] According to a second option, a spectral analysis 205 is performed on the wheel rotation signal normalized and resampled in step 204 or 203 depending on the path, which signal is periodic with respect to the wheel rotation. If the angular pitch is not regular, the measurement points should be interpolated onto theoretical points regularly spaced across the signal. Possibly, the spectral analysis step 205 is performed after a data aggregation step 207 providing a signal with a fixed angular pitch. The spectral signal resulting from step 205 is analyzed to extract one or more spectral variables, preferably extracted from the first positive spectral block.

[0069] The method then comprises a step 206 of determining the tire casing deformation in rolling condition under a static load Def%, which is carried out by using the function G to determine the energy density(s) S evaluated in step 205. + , S - or the angularly resampled normalized wheel rotation signal Sig TDRThis is achieved by using one or more spectral variables evaluated from the spectral signal spect(Sig) of the tire casing, which may be one or more of the spectral variables that provide a function G that in turn provides a vector, preferably a scalar, as an invariant of the tire casing deformation in rolling conditions under external forces.

[0070] Finally, in step 207, the method determines the load Z applied to the mounting assembly using a function H relating the load Z to the tire casing deformation Def%. Due to the fact that the tire casing deformation is evaluated based on the response of the measurement signal, which may be much larger than the mere passage of the contact patch, as in the case where the deformation Def% is evaluated using the energy density S+, the wheel rotation signal Sig TdR It is possible to obtain this deformation accuracy with a spatial discretization of . This is less intensive in terms of energy and memory capacity and allows the determination of the load Z to be performed by a measuring device mounted on the tire casing, such as a TMS (Tire Monitoring Sensor). Since it is not the aim to determine the size of the contact patch, the spatial discretization of the method does not need to be as fine as in the prior art.

[0071] Figures 2 to 4 illustrate the method using the second path outlined in figure 1. An accelerometer is described which is fixed to the crown of a tire casing, mounted on the inner liner of the tire casing. The tire casing is a Michelin CrossClimate, size 265 / 65R17, with a static load of 800 daN when mounted on a vehicle. The mounting assembly was inflated to 3 bar. Measurements were performed while the vehicle was running on a circuit of asphalt with different roughnesses, under standard speed and load conditions according to the tire markings. The mounting assembly was installed on the front axle of the vehicle. Measurements were mainly performed in straight-line driving conditions.

[0072] FIG. 2 shows a time signal 101 acquired with a signal acquisition frequency of 3200 Hz, which allows a very fine discretization of the signal. As a result, it records all the acceleration variations at the crown of the tire casing during rolling. The wheel rotation signal Sig TDR This signal was sectioned over 12 wheel revolutions to construct

[0073] The recording in FIG. 2 was made during the acceleration phase of the vehicle, which is reflected by the increase in the amplitude of the acceleration signal. The sensor here is a 1-axis acceleration sensor mounted radially relative to the crown of the tire casing. The data was transmitted by wireless communication between an electronic device electrically connected to the accelerometer and a second radio frequency device located in the vehicle. In this particular case, post-processing of the measurements was performed in the vehicle. However, it is quite possible to perform these in a first electronic device connected to the sensor and equipped with a microcontroller or microprocessor and coupled to sufficient memory space to perform the elementary mathematical operations required by the method.

[0074] Here, the first step consists in determining the reference speed, taking the rotational angular velocity as the reference speed. For this purpose, the first time signal 101 must be synchronized with a reference azimuth position of the wheel rotation. For this purpose, the first signal 101 exhibits regular, very strong amplitude dips 111, 112, reflecting the passage of the contact patch through the angular sector in which the accelerometer is mounted. Naturally, these downward and upward slopes for the dips 111, 112 represent the entry and exit of the contact patch, respectively. The center of the contact patch is the middle of the section separating the entry and exit of the contact patch. This center is assigned to the 0° azimuth position, which becomes the azimuth reference. By adopting a second angular reference, for example at the next signal dip 112, the signal 101 is determined for a wheel rotation of 360° and for the time interval related to this wheel rotation. The reference speed W reference is defined as the ratio of the angular change between the two centers of the contact surface to the time interval separating these two azimuth positions. This reference velocity W referenceis assigned to the portion of the signal located between these two centers of the contact area. Naturally, taking into account the two non-adjacent drops 111, 115 of the time signal 101, the second reference speed W reference and assigning a second rate to the portion of the signal 101 located between the two drops 111, 115.

[0075] FIG. 3 shows the result of the step of angularly resampling the time signal 101. It is thus easy to synchronize the time signal with a wheel revolution over 360 degrees, using the determination of the center of contact for each drop of the time signal performed in the previous step. The discretized measurement points are then linearly distributed with respect to the wheel revolution. Even if angular positioning errors occur in this step, the result will be smoothed and the angular positioning errors will be minimized, for example by linear interpolation performed during the data aggregation step. In a more sophisticated way, a reference speed is evaluated for each wheel revolution. It is possible to assign an evolving angular speed to the wheel revolution by considering the reference speeds of successive revolutions. For example, if the reference speeds are determined over three successive revolutions, it is possible to assign to the central wheel revolution a first reference speed for the first quarter wheel revolution, which is the center of gravity speed weighted by 2 to the reference speed of the preceding revolution and by 1 to the reference speed of the current revolution. The next quarter will have a reference speed weighted by 2 to the reference speed of the current revolution and by 1 to the reference speed of the preceding revolution. The third quarter of the wheel revolution has a reference speed of the center of gravity weighted by 2 to the reference speed of the current revolution and weighted by 1 to the reference speed of the next revolution. Finally, the last quarter of the wheel revolution has a reference speed of the center of gravity weighted by 1 to the reference speed of the current revolution and weighted by 2 to the reference speed of the next revolution. All the discretized measurement points are distributed in each quarter wheel revolution in proportion to the ratio of the reference speed of each quarter revolution to the reference speed of the current revolution. Other methods of smoothing these points can also be applied. Here, the spatial discretization of the points is not regular due to the variable rolling speed. It is quite possible to make this discretization regular for the points of the signal 102 by applying a method of interpolating the measurement points over a given angular distribution with respect to the wheel revolution. An angularly resampled signal 102 with a regular angular pitch is then obtained. Figure 3 shows an angularly resampled signal 102 that is periodic with respect to the wheel revolution, with an arbitrary discretization of the measurement points.

[0076] FIG. 4 shows the result of a step of normalizing the first angularly resampled signal 102 without interpolation of points. Thus, using the periodicity of the angularly resampled wheel rotation signal with respect to the wheel revolution, it is easy to resolve the angular signal over a wheel revolution or over a multiple of a wheel revolution (here 12 wheel revolutions) as shown in FIG. 4. The normalization step involves dividing the amplitude of the signal by a squared power function of a reference speed associated with each part of the wheel revolution. The reference speed was determined for example during the first signal processing step 101. The reference speed is here the angular speed. The result observed for curves 103 and 103bis is that the amplitude of the normalized signal is similar for each wheel revolution. There is no longer a strong variation in amplitude between the various wheel revolutions made at different speeds and on different roads. Also, the signal is centered on a unit value. Then, over the same angular interval length, which is an integer multiple of 360 degrees, the wheel rotation segments are superimposed to form a curve bundle 103, here shown by the grey curve. This takes into account the spread of measurements between wheel revolutions, emphasized by the fact that here the signal is not corrected for the Earth's gravity. However, applying a low-pass filter results in a smoother black curve 103bis, since certain parasitic noises are removed. This makes it possible to see that the signal 103bis is periodic with respect to the wheel revolutions, with slight fluctuations between wheel revolutions. At the end of this normalization of the signal 102, an angularly resampled normalized signal 103 is obtained. Figure 4 shows the angularly resampled normalized signal 103, centered on a unit value, as confirmed by the filter applied to the curve 103bis.

[0077] FIG. 5 shows the result of a step of aggregating the data of the signals 103 from the previous step, which is an optional step. Here, the segments of each wheel revolution are superimposed over the same angular interval length of 360 degrees, as shown by the grey curves forming the curve bundle 104. This is emphasized by the fact that the signals are not corrected for the Earth's gravity, taking into account the spread of the measurements between each wheel revolution. However, since the accelerometer is now sensitive to the Earth's gravity, if a correction for the Earth's gravity is applied to each wheel revolution before the normalization step, the data aggregation by the method of averaging over a decile interval determines a very stable curve 104bis for the wheel revolution. This results in a signal for the deformation of the tire casing under external forces, in particular in this case a static load. This signal 104bis is representative of the measurements of the tire casing in rolling condition at variable speeds on ground of any roughness. This curve is an invariant of the tire casing in rolling condition mounted on a rim and under a static load.

[0078] FIG. 6A shows the angularly resampled normalized wheel rotation signal Sig corresponding to one wheel revolution. TDR For positive energy density S + and negative energy density S - FIG. 2 is a diagram for explaining the calculation of the angularly resampled normalized wheel rotation signal Sig TDR The method is exactly the same if x is spaced over multiple wheel revolutions.

[0079] A threshold A is determined here as a unit value. This threshold is shown by the solid line 11. In practice, it is preferable to adopt a value equal to 0.7 for real signals. If the signal is highly interfered with, values ​​equal to 0.5 or 0.6 can be selected. However, for signals obtained on a generally smooth road surface, values ​​of the order of 0.8 or 0.9 can be used. This value of threshold A must be set for all steps of the method.

[0080] Positive energy density S+ or negative energy density S - is calculated as the sum of the absolute values ​​of the differences between the wheel rotation signal 10 and the unit value represented by the continuous curve 11. + The area defined by the region S - is equal to the area defined by

[0081] From these estimates of energy density S, it is easy to determine the deformation Def% of the tire casing under static load in rolling condition.

[0082] Figure 6b shows the spectrum of the angularly resampled normalized wheel revolution signal with a fixed angular pitch of 0.1 degrees and spaced over 12 wheel revolutions. To suppress high frequency phenomena, the signal obtained from step 203 of the first pass or step 204 of the second pass was first filtered with a low pass filter of 1 / 30th of a wheel revolution.

[0083] Then, after spectral analysis of the filtered signal, i.e. here the signal from the aggregation step of step 208, by means of a Fourier transform, a curve 105 is obtained which represents the amplitude of the Fourier transform over a limited frequency band. This curve shows various spectral blocks, the first of which has a large amplitude, but each subsequent block is not negligible in itself.

[0084] It is possible to obtain several spectral variables from this spectral response 105. In this case, we focus on the first block, but the analysis can also be carried out on subsequent blocks.

[0085] To consider the sensitivity of the method, Fig. 6b shows a second dotted curve 106 corresponding to the spectral response of the same sensor fixed to the same mounting assembly for different static loads and different air pressures, where the mounting assembly is swapped between the front and rear axles of the vehicle. Thus, the mechanical response of the tire casing to the two variables, air pressure and static load, is necessarily different. However, the spectral response shows a similarity in shape with a continuous block-shaped response, where the width and height of the blocks are a function of the external forces applied to the tire casing.

[0086] This shows that while the analysis of the first block may not be sufficient for weak fluctuations in the external forces acting on the tire casing, it has sufficient discriminatory power to determine the deformation of the tire casing associated with such fluctuations in external forces.

[0087] Spectral variables such as maximum, median, mean, passband, area under the curve associated with the first block, etc., can all be criteria for identifying tire casing deformation. However, the median frequency, mean frequency, and maximum frequency are also secondary criteria in tire casing deformation and, although still discriminatory, indicate much weaker driving forces.

[0088] A tire casing deformation value Def% can then be assigned using a function of one or more spectral variables in vector or scalar form. Preferably, the maximum values ​​105bis and 106bis of the first block have been found to be very good indicators of the tire casing deformation, which allows the tire casing deformation to be determined through an affine function of the maximum values ​​of the first block. However, the determination of the tire casing deformation can be more sophisticated if other spectral variables associated with secondary spectral blocks are also taken into account.

[0089] FIG. 7 shows the estimated load Z on the mounting assembly in rolling motion at a rotational speed W. Two different casings were used. The first tire casing E1 is a 385 / 55R22.5 heavy load vehicle tire casing of the Michelin X Multiway T series with wear level D1, mounted on a 22.5 inch plate wheel. The second tire casing E2 is a 315 / 80R22.5 of the Michelin X Multiway 3D XDE series with wear level D2. Each casing is equipped with an on-board electronic device with a uniaxial accelerometer positioned in the crown on the inner liner at the height of the protruding sculpture elements, i.e. at a height different from the longitudinal grooves. The acquisition frequency of the accelerometer is 1200 Hz.

[0090] Each tyre casing undergoes a series of rolling scenarios with travel speeds varying around 20, 40 and 60 km / h, with air pressure P varying from 7 bar to 9 bar in 1 bar increments. The pressure is measured during rolling, in this case by means of a pressure sensor integrated into the TPMS mounted on the wheel valve. Finally, the load Z applied to the mounting assembly varies between 2000 kg and 5000 kg in 1 ton increments.

[0091] The four coefficients (a1, a2, b1, b2) of the affine function of function H for each tire casing were previously determined by means of digital simulations. In fact, since this is precisely the area of ​​use recommended by the ETRTO regulations, the affine representation of function H should be preferred.

[0092] Half of the rolling scenarios were performed at a constant rotational speed, and the other half were performed at a variable rotational speed around the target speed of ±15%.

[0093] Figure 7 shows a continuous line corresponding to the response given by the coefficients of the function H, which in this case depend on the air pressure P and on the mounting assembly including the tire casing, and also shows symbols of different shapes depending on the target rolling speed: a diamond for a speed of 20 km / h, a circle for a speed of 40 km / h and a cross for a speed of 60 km / h.

[0094] Curve 1001 corresponds to a mounting assembly with a tire casing E1 with an air pressure of 7 bar. Curve 1002 corresponds to a mounting assembly with a tire casing E2 with an air pressure of 8 bar. Finally, curve 1003 corresponds to a mounting assembly with a tire casing E1 with an air pressure of 9 bar.

[0095] A relatively good correlation is found between the estimated load Z and the true applied load, regardless of travel speed and inflation pressure. Furthermore, depending on the properties of the tire casing, the affine representation of the load is realistic for testing of the mounting assembly in this range of service conditions.

[0096] The same good results are obtained regardless of the nature of the tire casing, the load applied, the air pressure used, and the wear of the tire.

Claims

1. A method for checking the load applied to a tire casing when attached to a wheel so as to form an air-filled mounting assembly rolling at a rotational speed W, wherein the tire casing has a crown that contacts the ground and rotates about a natural axis of rotation, fixing at least one sensor to the crown of the tire casing to generate at least one output signal sensitive to an acceleration in a direction perpendicular to the crown and applied to the sensor within the tire casing; during rolling, obtaining at least one first signal Sig including at least the amplitude of the at least one output signal; The wheel rotation number N exceeding once TDR dividing the first signal over time to form a wheel rotation signal Sig TDR and steps of configuring; said wheel rotation signal Sig TDR at least one reference speed W related to at least one part of reference determining step and The number of wheel rotations N for one or more times TDR Over the period of, the reference speed W reference By a variable which is a function F proportional to the square of, the at least one part of the wheel rotation signal Sig TDR Normalizing the at least one part of; and the step of angularly resampling at least one part of the wheel rotation signal Sig TDR and At least one first energy density S is determined from the at least one angularly resampled normalized wheel rotation signal SigTDR using a threshold value A, or, when the angular pitch is fixed, from the at least one spectral variable derived from the spectral signal spect(Sig) of the at least one part of the angularly resampled normalized wheel rotation signal Sig TDR specifying at least one spectral variable derived from the spectral signal spect(Sig) of the at least one part of; specifying the deformation Def% of the tire casing as a function G of the at least one first energy density S or the at least one spectral variable; defining the load Z applied to the mounting assembly using a bijective function H including at least the deformation Def% of the tire casing as a variable; A method comprising.

2. the reference speed W reference The step of determining includes establishing a ratio of an angular change to a duration that separates two azimuth positions about the eigen-rotation axis for the sensor in the tire casing from the first signal Sig or from a signal synchronized with the first signal Sig, according to the following mathematical formula [Equation 1] where α is the angular position and t is the temporal abscissa related to the angular position, a method for checking the load applied to the tire casing according to claim 1.

3. The angular pitch is less than 18 degrees, preferably less than 6 degrees, very preferably less than 3 degrees, a method for checking the load applied to the tire casing according to claim 1 or 2.

4. The data from the at least one portion of the angularly resampled normalized wheel rotation signal Sig TDR is aggregated over at least one sub-portion of the at least one portion of the angularly resampled normalized wheel rotation signal Sig TDR including a step of aggregating, and the sub-portion of the at least one portion of the angularly resampled and normalized wheel rotation signal Sig TDR is the at least one portion of the angularly resampled and normalized wheel rotation signal Sig TDR A method for verifying a load applied to a tire casing according to claim 1 or 2, wherein the at least one portion of the angularly resampled and normalized wheel rotation signal Sig

5. The wheel rotation signal Sig TDR The method for verifying the load applied to the tire casing according to claim 4, wherein the sub - part of the at least one part of TDR is an integer multiple of the wheel rotation.

6. After synchronizing the first signal Sig with respect to the angular position of the tire casing, before the normalization step, a correction Corr is performed on the first signal Sig to account for the influence of the earth's gravity, a method for checking the load applied to the tire casing according to claim 1 or 2.

7. The step of filtering at least one portion of the angularly resampled normalized wheel rotation signal Sig TDR A method for verifying a load applied to a tire casing according to claim 1 or 2, comprising the step of filtering at least one portion of TDR .

8. The normalized wheel rotation signal Sig resampled angularly as described above TDR The step of obtaining the at least one spectral variable from the spectral signal spect(Sig) of the at least one portion of is as follows: across at least one spectral block of the spectral signal spect(Sig), preferably across the first positive spectral block of the spectral signal spect(Sig), the method for verifying the load applied to the tire casing according to claim 1 or 2, including the step of identifying the at least one spectral variable.

9. The at least one identified spectral variable is included in a group including a maximum value, a median value, an average value, a passband of the first positive spectral block, an area under the curve of the first positive spectral block, a frequency of the median value, a frequency of the average value, and a frequency of the maximum value, a method for checking the load applied to the tire casing according to claim 8.

10. Using the threshold value A, from at least one part of the angularly resampled normalized wheel rotation signal Sig TDR The step of obtaining the at least one energy density S from the at least one part is that when at least one part of the angularly resampled normalized wheel rotation signal Sig TDR is greater than the threshold value A, defining a first energy density S + or, when at least one part of the angularly resampled normalized wheel rotation signal Sig TDR is less than or equal to the threshold value A, defining a second energy density S - A method for verifying the load applied to a tire casing according to claim 1 or 2, comprising the step of

11. The threshold value A is between 0.5 and 0.9, a method for checking the load applied to the tire casing according to claim 10.

12. The method for checking the load applied to the tire casing according to claim 1 or 2, wherein the function G is a linear function.

13. The bijective function H is an affine function or a power function represented by the following mathematical formula: [Mathematical formula 10a] or [Mathematical formula 10b] Here, (A, B) or (X, Y) is a parameter related to the mounting assembly. The method for checking the load applied to the tire casing according to claim 1 or 2.

14. When the mounting assembly is inflated to the air pressure P, the parameter A or X depends at least on the air pressure P. Preferably, the parameter A or X is an affine function of the air pressure P represented by the following mathematical formula: [Mathematical formula 11a] [Mathematical formula 11b] Here, (a 1 , a 2 ) or (x 1 , x 2 ) is a coefficient related to the mounting assembly, and a method for checking the load applied to the tire casing according to claim 13.

15. When the mounting assembly is inflated to the air pressure P, the parameter B or Y depends at least on the air pressure P. Preferably, the parameter B or Y is an affine function of the air pressure P represented by the following mathematical formula: [Mathematical formula 12a] or [Mathematical formula 12b] Here, (b 1 , b 2 ) or (y 1 , y 2 ) is a method for checking the load applied to the tire casing according to claim 13, which is a coefficient related to the mounting assembly.

Images