Method for determining the deformation of a tire subjected to a load during rolling

Description

Field of invention

[0001] The present invention relates to the field of measurement signals delivered by measurement means on board the mounted assembly of a land vehicle during rolling. Technological background

[0002] Recent developments in connected mounted assemblies measuring physical quantities of the mounted assembly using sensors embedded in the mounted assembly lead to determining the state of the mounted assembly and therefore open the door to the development of services related to monitoring the state of the mounted assembly. If general measured quantities such as the inflation pressure of the mounted assembly or the temperature of this mounted assembly are not very sensitive to measurement noise generated by the rotation of the mounted assembly on a ground of random roughness since these general quantities vary little during the rotation of the mounted assembly, finer quantities are highly sensitive to physical phenomena related to the rotation of the mounted assembly. In addition, the mounted assembly is subjected to external forces. Some are related to the movement of the mounted assembly such as rolling resistance.Others are applied at any time, particularly in static conditions, such as load. These applied forces can influence the fine-grained quantities that we are trying to measure. Finally, new services require cleaning the directly measured physical quantities before collecting useful information from measurement signals, such as the deformation of the tire casing.

[0003] One of the objects of the invention which follows aims to resolve the problems of disturbances of the measurement signals recorded by a sensor in order to collect only a measurement cleaned of the disturbances of certain physical phenomena in order to obtain a magnitude, vector or scalar, of the deformation of the pneumatic envelope.

[0004] In order to better understand the invention, the circumferential direction S, axial direction A and radial direction R are understood here to mean directions defined relative to the rotating reference frame of the tire around its natural axis of rotation. The radial direction R is the direction moving perpendicularly away from the natural axis of rotation. The axial direction A is the direction parallel to the natural axis of rotation. Finally, the circumferential direction S forms a direct trihedron with the predefined radial and axial directions. Description of the invention

[0005] The invention relates to a method for obtaining the deformation of a pneumatic envelope. The pneumatic envelope is in a wheel-mounted state in order to constitute a mounted assembly in rolling condition at a rotational speed W subjected to a load. The pneumatic envelope has a top, intended to be in contact with the ground which is of revolution around a natural axis of rotation. The method comprises the following steps: Fixing at least one sensor on the tire casing at the top of the tire casing capable of generating at least one output signal sensitive to acceleration in the direction normal to the top experienced by said sensor in the tire casing; Acquiring at least one first time signal Sig comprising at least the amplitude of the at least one output signal during rolling; Delimiting the first signal over a number N TDR< of wheel revolutions, N TDR< being greater than or equal to 1, in order to construct a wheel revolution signal Sig TDR<; Determining at least one reference speed W reference< associated with at least one part of the wheel revolution signal Sig TDR<; Normalizing the at least one part of the wheel revolution signal Sig TDR< by a quantity which is a function F proportional to the square of the reference speed W reference<, over a number of wheel revolutions N' TDR<, N' TDR< being greater than or equal to 1;Angularly resample the at least one part of the wheel rotation signal Sig TDR<; Define at least one first energy density S from the at least one part of the angularly resampled normalized wheel rotation signal Sig TDR<, named S +< when the at least one part of the angularly resampled normalized wheel rotation signal Sig TDR< is greater than a threshold A, or named S -< when the at least one part of the angularly resampled normalized wheel rotation signal Sig TDR< is less than or equal to said threshold A; Identify the deformation of the tire casing Def % as a function G of the at least one first energy density S.;

[0006] The signal recovered from the sensor is the temporal amplitude of the acceleration of the sensor during rolling of the assembled assembly under the stated conditions. In the direction normal to the top As a result, the acquired signal displays the variations in amplitude over a part of the wheel revolution of the tire casing, including potentially those associated with the crossing of the contact patch by the part of the tire casing where the sensor is fixed, but also those associated with other specific areas of the wheel revolution such as that corresponding to the angular sector opposite the contact patch which is sensitive to counter-deflection, or those corresponding to the angular sectors located at 90 degrees from the contact patch relative to the axis of rotation. In all these areas, variations in the movement of the sensor, of the accelerometric type, are potentially observable on the output signal depending on the sensitivity of the sensor.

[0007] On the first acquired signal, a reference speed is associated which can be identified on this first signal or from another source such as another signal or the output of a quantity of a system external to the assembled assembly. This reference speed is necessarily associated with the same time frame as the part of the first signal. This reference speed is used to normalize the amplitude of the first signal using a function F whose variable is the reference speed. The function F is the power squared function. Depending on the dependence of the amplitude of the sensor signal on the reference speed if this dependence is perceived as a parasitic signal of the deformation of the tire casing, the normalization of the sensor signal is undertaken. Thus, the first normalized signal becomes independent of this reference speed.For example, this reference speed can be the rotation speed of the mounted assembly or the translation speed of the mounted assembly depending on the direction of movement of the mounted assembly. Therefore, the first signal can be used independently of the reference speed which is linked to the rotation of the mounted assembly.

[0008] The method also includes a step of delimiting the first signal Sig on a number of wheel revolutions in order to take advantage of the periodicity of the sensor signal with the natural rotation of the tire casing in rolling conditions. However, it is not essential that the number of wheel revolutions be whole at this step, we can have a delimitation of the signal on a real number of wheel revolutions as long as this number of wheel revolutions is at least greater than 1. Preferably, several wheel revolutions are required.

[0009] The method also includes an angular resampling of the first signal or the wheel revolution signal which can take place before or after the normalization step. This step makes it possible to transform the time signal into a spatial signal by phasing the time signal with respect to one or more angular references of the mounted assembly. This angular reference can first of all be taken from the first signal by a specific response of the sensor at a particular azimuth at the wheel revolution. But, this angular reference can also come from another signal of a sensor which shares a common clock with the first signal. This clock sharing or synchronization of the signals is natural if the two sensors come from the same device or if the signals pass through a common device. This angular resampling naturally makes it possible to generate a spatially periodic signal at the wheel revolution.To do this, it is sufficient to interpolate the signals on a fixed angular division to generate a perfectly angularly periodic signal. However, if the mounted assembly is animated by a variable speed movement, this resampling still allows the generation of an angularly periodic signal. It is not essential for the method that the angular resampling generates an output signal with a fixed angular step.

[0010] Then, the simple comparison of the amplitude level of the wheel revolution signal with a threshold A makes it possible to generate an energy density S. The amplitude of the wheel revolution signal with respect to a threshold A which can, for example, simply be the unit value potentially generates a doublet of positive and negative deformation energy densities (S +< , S -< ) from the wheel revolution signal. Thus, the method only defines a deformation energy density of the tire envelope and distributes it into two subsets according to its position with respect to the threshold A. These are simple operations to perform which consume few resources.

[0011] Finally, the method determines the deformation of the tire casing as a function of the calculated energy density. Thus, the deformation represents a normalized deformation energy over a physical wheel revolution of the tire casing. As a result, an energy invariant related to the deformation of the tire casing subjected to a load in rolling conditions is identified. Of course, only one wheel revolution is necessary for the method. However, preferably, the number of wheel revolutions will be at least 5, or even 10, in order to average the results, which will make it possible to overcome random phenomena on the signal such as obstacles on the road on which the tire casing rolls. Thus, the accuracy of the method is improved in industrial mode.

[0012] Advantageously, the step of determining the reference speed W reference< consists of calculating the ratio of the angular variation to the time duration separating two azimuthal positions of the sensor in the pneumatic envelope around the natural axis of rotation from the wheel revolution signal Sig TDR< or from a signal phased with the wheel revolution signal Sig TDR<, according to the following formula: W reference = Δ α / Δ t Where α is the angular position and t is the time abscissa associated with the angular position.

[0013] In the case where the reference speed corresponds to the angular rotation speed of the tire, this reference speed is calculated on an angular variation of the signal between two known positions. Preferably, this reference speed is evaluated over a signal duration of less than one wheel revolution, which makes it possible to define it quickly and to carry out the step of normalizing a part of the first signal at the level of the electronic device associated with the sensor. In addition, this then makes it possible to angularly resample the part of the first signal with better precision if the tire is driven by a variable angular speed. Indeed, at the level of one wheel revolution, the variation in angular speed is necessarily small for a tire whose development can extend to 2 meters for a tire for a private vehicle or 3 meters for a tire for heavy goods vehicles.The acceleration or deceleration applied to the tire over this length is, by nature, low with the drive and braking systems of current vehicles. Of course, it is entirely possible to integrate a variation in angular velocity during the wheel revolution with finer azimuth in order to take into account, for example, micro-variations in angular velocity that appear during the wheel revolution, for example, before and after crossing the contact patch or when encountering a discontinuity in ground movement, such as a transverse bar on the ground.This precision on the reference speed during the wheel revolution then allows a more precise normalization of the signal but also an increased angular precision on the angular position of the measurement points of the first signal during the angular resampling step, which improves the precision sought to capture minimal variations during the wheel revolution.

[0014] According to a particular embodiment, the azimuthal positions of the pneumatic casing are included in the group comprising an angular position detectable on the wheel rotation signal Sig TDR< corresponding to the entry into the contact area, the exit from the contact area or the central position of the contact area or any angular position defined from the signal phased with the wheel rotation signal Sig TDR<.

[0015] These are azimuthal positions that affect the acceleration sensor signal and correspond to specific angular positions. As a result, these positions are easily identifiable on the sensor signal. In addition, it is easy to assign them their azimuthal references. Indeed, the central position of the contact patch corresponds to an azimuthal position of 0 or 180 degrees relative to the ground normal. If we determine a length of the contact patch through the entry and exit points of the contact patch, we can know the angle formed by the contact patch as the ratio of the length of the contact patch to the development of the tire envelope reduced to one wheel revolution, i.e. 360 degrees. We distribute the sector formed by the contact patch equally on either side of the ground normal.Of course, access to a signal other than the first signal also allows for finer angular sectorization than the wheel revolution like an angular encoder.

[0016] Advantageously, the angular pitch is less than 18 degrees.

[0017] This ensures that one of the measurement points is located at the contact area. As a result, significant acceleration variations will at least be observed between this sampling point and its nearest neighbors, making it possible to determine an entry and exit point of the contact area on the first signal. Very advantageously, the angular pitch is less than 6 degrees, preferably less than 3 degrees.

[0018] Using a finer angular pitch ensures that several measurement points are captured at the contact area. This fineness of observation ensures greater accuracy of the method by avoiding the spatial discretization of the points, which is not necessarily regular here. The multitude of points also ensures that you are not disturbed by an inconsistent measurement from the sensor.

[0019] According to a preferred embodiment, the method comprises a step of aggregating the data of the at least one part of the angularly resampled normalized wheel rotation signal Sig TDR< on at least one sub-part of the at least one part of the angularly resampled normalized wheel rotation signal Sig TDR<, the sub-part of the at least one part of the angularly resampled normalized wheel rotation signal Sig TDR< becoming the at least one part of the angularly resampled normalized wheel rotation signal Sig TDR<.

[0020] Preferably, the sub-part of at least one part of the angularly resampled normalized wheel rotation signal Sig TDR< is a wheel rotation.

[0021] This step makes it possible to identify a wheel rotation signal taking into account various random variables in the wheel rotation, which makes it possible to reduce the size of the vectors to be handled for the last two steps of the method, in particular that used to identify the first energy density S. For this purpose, the sub-part of the angularly resampled normalized wheel rotation signal is limited to a single wheel rotation.

[0022] According to a preferred embodiment, the data aggregation step comprises one of the methods included in the group comprising the average over a decile interval, the median, the selection or the decile interval, the interpolation methods, the weighted or unweighted average, the optimization of the parametric model of the deformation of the tire. The aggregation aims to align the measurements carried out on a new angular distribution of the first signal in order to give meaning to all of the raw measurement data by not favoring one area over another due to an abundance of measurement points. The aggregation step aims to deliver a balanced signal in terms of measurement points with an angular pitch chosen by the operator according to the deformations of the tire envelope that one wishes to observe.For this purpose, the parametric model optimization method of the tire deformation is well suited since this parametric model can be theoretical not taking into account the raw measurements related to the entire applied measurement chain. The output signal of the aggregation step is a theoretical output of the parametric model having the minimum dispersion with all the recorded measurement points.

[0023] Advantageously, having phased the first signal Sig relative to an angular position of the pneumatic envelope, a correction Corr is made to the first signal Sig to take into account the effect of Earth's gravity before the normalization step.

[0024] The disadvantage of the accelerometric signal is that it is sensitive to Earth's gravity if it is oriented substantially parallel to Earth's gravity. In the case of the pneumatic envelope, the sensor is rotationally linked to the pneumatic envelope. Therefore, when the sensor is oriented radially, the amplitude of the sensor signal is influenced by Earth's gravity during a wheel revolution. Indeed, this is found in the signal in the form of a sinusoidal function of amplitude linked to Earth's gravity presenting its nodes at azimuths of the pneumatic envelope separated by 180 degrees when the orientation of the sensor is aligned with the gravitational vector, i.e. substantially perpendicular to the ground.Conversely, when the sensor is oriented parallel to the ground, which corresponds to two azimuthal positions 180 degrees apart from each other and generally + / - 90 degrees from the gravitational vector, the sensor signal is not influenced by the Earth's gravity. In order to eliminate this parasitic component of the accelerometric signal, the amplitude of the signal should be combined by a corresponding sinusoidal function by having phased the first sensor signal with the vertical to the ground corresponding to the direction of the gravitational vector.

[0025] According to a particular embodiment, the method comprises a step of filtering at least one part of the angularly resampled normalized wheel rotation signal Sig TDR<.

[0026] High-frequency disturbances may persist on the angularly resampled normalized signal, which will be processed in the next step of defining the first energy density S using the threshold A. However, filtering the signal will simplify the next step by minimizing possible errors.

[0027] Preferably, threshold A is between 0.5 and 0.9.

[0028] The threshold is intended to distribute the discretized points of the angularly resampled normalized Sig TDR< wheel rotation signal between the energy densities S +< and S -< . If the signal is highly noisy, as in the absence of a data aggregation step or a filtering step, this distribution of points can be influenced by these disturbances. The threshold A is intended to correct this imperfection linked to the measurement signal. The value of threshold A depends on the quality of the angularly resampled normalized Sig TDR< wheel rotation signal. If the method takes the optional steps and the road roughness is low, a value in the upper part of the range will be preferred.

[0029] Very preferably, the definition of the positive energy densities S +< and negative S -< is obtained by the following formulas: S + = ∑ Sig TDR > A Sig TDR − 1 ∗ N ′ TDR N U ; And S − = ∑ Sig TDR ≤ A Sig TDR − 1 ∗ N ′ TDR N U ;

[0030] Where u is the abscissa of the angularly resampled normalized wheel rotation signal Sig TDR<.

[0031] This is a simple way to obtain a scalar value of each energy density from the discretized signal of the angularly resampled normalized wheel revolution signal using elementary mathematical and logical operations. These operations can take place at the level of the electronic device coupled to the sensor.

[0032] Preferably, the acquisition of the first signal is carried out at a constant sampling frequency and the spatial discretization of sampling of the first signal is less than 6 degrees, preferably less than 3 degrees, very preferably less than 1 degree.

[0033] For example, if we want the evaluation of the deformation of the tire envelope to take place at the level of the mounted assembly, the sensor must be associated with an electronic component comprising a microcontroller, a memory space, a battery and a clock. Then, the spatial discretization envisaged with a constant sampling frequency makes it possible to carry out elementary operations at the microcontroller level while minimizing battery consumption. In addition, the minimal discretization of the order of 60 points per wheel revolution makes it possible to limit the number of operations and transfers to the memory space. However, the precision obtained on the deformation of the tire envelope is good while having saved the battery of the electronic component. This makes it possible to store or transfer only intermediate scalar values ​​of the method.

[0034] Advantageously, the function G is a linear function of the spectral density S according to the following formula: G X = X / N ′ TdR

[0035] Thus, it is an elementary formula for the deformation of the tire casing which applies either to S +< or to S -< . Necessarily, S -< corresponds to the energy density calculated from the material points of the tire development, which at a precise instant T, include those of the contact patch or in its immediate vicinity. Indeed, these have an absolute acceleration close to zero when crossing the contact patch, they are therefore necessarily lower than the threshold A. By default, the energy density S +< corresponds to the energy density of the other points of the tire development and in particular those outside the contact patch. This highlights that we are in the presence of an invariant linked to the deformation of the tire casing subjected to a load Z. In the case where we only use S +< , it is not necessary to have a high spatial discretization since the variations outside the contact patch are less strong.This has the advantage of reducing the necessary sampling frequency of the electronic device coupled to the sensor or of being able to obtain precise information on the deformation of the pneumatic envelope at high rotation speeds.

[0036] Very advantageously, the function G is a linear function of the spectral densities S +< and S -< according to the following formula: G X Y = X + Y / 2 * N ′ TdR

[0037] In this case, the tire deformation energy must be accumulated over the entire tire development. In order to minimize measurement uncertainties, all of the normal acceleration measurement points at the top are used to determine the tire deformation, which reduces energy consumption compared to an analysis with a signal with a higher sampling frequency. Brief description of the drawings

[0038] The invention will be better understood on reading the following description, given solely as a non-limiting example and made with reference to the appended figures in which the same reference numbers designate identical parts throughout and in which: There Figure 1 presents a synopsis of the method according to the invention. The Figure 2 shows an illustration of a first signal from a sensor. The Figure 3 presents the angular resampling of the wheel rotation signal. The Figure 4 shows an illustration of the resampled and normalized wheel revolution signal. The Figure 5 presents an illustration of the final signal after aggregation of the data on a sub-part of the wheel revolution signal. The Figure 6 is an illustration of evaluating the energy densities S from the angularly resampled normalized wheel rotation signal. Detailed description of embodiments

[0039] There Fig. 1represents a block diagram of the method according to the invention. Starting from a first signal Sig obtained by time acquisition 201 of the amplitude output of a motion sensor during rolling of the tire on which the sensor is mounted, a certain number of steps are carried out following various possible paths to finally obtain a scalar representative of the deformation of the tire.

[0040] The first path consists of starting from the time signal at the output of step 201, to determine a reference speed W reference< 202 of the tire in its assembled configuration, that is to say tire mounted on rim and inflated. Here, the first signal Sig 101 is already delimited over a certain number of wheel revolutions, 12 exactly. Consequently the first signal Sig 101 is confused with the wheel revolution signal Sig TDR< . This reference speed can be an angular speed linked to the natural rotation of the tire around its axis of rotation but it can also be the linear translation speed of the tire according to the direction of movement of the latter. This quantity can be determined from the wheel revolution signal Sig TDR< but can also be determined from another signal temporally phased with the first signal and therefore the wheel revolution signal Sig TDR< .

[0041] Then, the wheel revolution signal 203 should be normalized from the first signal from step 201 by a function F of the variable W reference< from step 2. This function is the power squared function. At the output of this step 203, a normalized signal of the movement of the tire envelope is obtained in a temporal description.

[0042] It is then appropriate to angularly resample the normalized signal in order to find a signal angularly periodic per wheel revolution through step 204. Thus, at the end of this step 204, we obtain a normalized signal and angularly resampled over several wheel revolutions.

[0043] The second path consists of starting from the first signal Sig, which is also the wheel revolution signal Sig TDR< , from the output of step 201, resampling the first signal Sig angularly by phasing this first signal using the shape of the first signal or by having another signal temporally phased with the first signal. The other signal emanating from another sensor, or another channel of the same sensor such as the circumferential acceleration of a three-dimensional accelerometer. This angular resampling of the first signal makes it possible to generate a periodic signal around the wheel at the output of step 204.

[0044] After phasing this angular signal using another time signal, a reference speed is determined from another time signal phased with the first signal. Preferably, it is the same other signal used to angularly resample the first signal in step 204. A reference speed W reference< is thus identified at the output of step 202.

[0045] Then, the reference speed is used to normalize the angularly resampled signal from step 204 using a function of the reference speed variable. This gives a wheel rotation signal Sig TDR< angularly resampled normalized at the output of step 203.

[0046] Optionally, regardless of the path taken, the data of the angularly resampled normalized wheel revolution signal Sig TDR< from step 204 on the first path or from step 203 on the second path are aggregated. This aggregation of the data is done on a sub-part of the input signal which is a single wheel revolution, since the angularly resampled and normalized signal is periodic to the wheel revolution by nature.

[0047] Alternatively, if the first signal 101 is polluted by known physical phenomena such as an accelerometer signal influenced by Earth's gravity, it is sometimes useful, although not essential, to perform a correction of the first signal of this physical phenomenon to limit the parasitic noise generated by this physical phenomenon. This correction can take place between any step between step 201 and 204 but necessarily before step 205 of aggregation of the data, which makes it possible to improve the quality of the signal of deformation of the tire envelope. If the correction takes place after the normalization step, it will also be necessary to normalize the correction so as not to introduce a correction error.

[0048] Then the method comprises a step of identifying a deformation energy density of the tire casing 205 from the angularly resampled normalized wheel revolution signal. Although this can be done on only a part of the wheel revolution through the positive energy density S +< or the negative energy density S -<, it is preferable to take at least one complete wheel revolution, which allows access to the two aforementioned quantities. It is also important not to forget to count the number of wheel revolutions N' TDR< on the angularly resampled normalized wheel revolution signal. If this signal is delimited on a wheel revolution, the identification of the energy densities is linked to the quality of this signal, which justifies going through the optional step of aggregating the data.However, if this signal is delimited over a large number of wheel revolutions, this signal may contain a fraction of an additional wheel revolution which will only slightly modify the value of the energy densities. In this case, it will be preferable to count the wheel revolutions from the azimuthal position located 180 degrees from the center of the contact area. Indeed, the additional fractions of wheel revolution will provide additional points on the positive energy density S +< whose variations between the points are lower than during the phases of entry and exit from the contact area which will have a greater impact on the negative energy density S -<.

[0049] Finally, the method includes a step 206 of identifying the deformation of the tire casing in rolling conditions under static load Def %. This is done by exploiting the energy density(s) S +< , S -< evaluated in step 205 through a function G.

[0050] THE Fig. 2 to 4 are an illustration of the method by the second path described in the synopsis of the Fig. 1 . The illustration is made on an accelerometer fixed to the top of a tire attached to the inner rubber of the tire (in English "inner liner"). Here, the tire is of the MICHELIN CrossClimate brand in size 265 / 65R17 under a static load of 800daN when mounted on a motor vehicle. The mounted assembly was inflated to 3 bars. The measurements were taken while the vehicle was driving on asphalt circuits of varying roughness under standard speed and load conditions applied according to the tire marking. The mounted assembly is located on the front axle of the vehicle. The measurements were taken here mainly while driving in a straight line.

[0051] On the Fig. 2, we visualize a time signal 101 acquired with a signal acquisition frequency of 3200Hz allowing a very fine discretization of the signal. This therefore records all the variations in acceleration at the top of the tire envelope during rolling. This was delimited over 12 wheel revolutions in order to constitute the wheel revolution signal Sig TDR<.

[0052] The recording of the Fig. 2was made during the vehicle acceleration phase, which results in an increase in the amplitude of the accelerometric signal. The sensor here is a single-axis accelerometer mounted radially relative to the top of the tire casing before constituting the assembly mounted by standard fixing techniques known from the state of the art. The data transmission was made by wireless communication between an electronic device galvanically connected to the accelerometer and a second radiofrequency device placed at the vehicle. In this particular case, the measurement post-processing was carried out at the vehicle. However, it is entirely possible to carry them out at the level of a first electronic device linked to the sensor, equipped with a microcontroller, or even a microprocessor, coupled with sufficient memory space to carry out the elementary mathematical operations required by the method.

[0053] Here, the first step consists of determining the reference speed by taking as reference speed the angular speed of rotation. To do this, it is necessary to phase the first time signal 101 with a reference azimuthal position of the wheel revolution. For this purpose, the first signal 101 regularly presents drops in the level of fairly strong amplitude 111, 112 which translates the crossing of the contact area by the angular sector where the accelerometer is located. By nature, these descents, respectively these rises, of these drops 111, 112 represent the entry, respectively the exit of the contact area. We will define the center of the contact area as the middle of the interval separating the entry and the exit of the contact area. We will assign to this center the azimuthal position 0 degrees which will be our azimuthal reference.By taking a second angular reference on the next drop of signal 112 for example, we determine on signal 101 a wheel rotation of 360 degrees and a time interval associated with this wheel rotation. We will define the reference speed W reference< as the ratio of the angular variation between the two centers of the contact area by the time interval separating these two azimuthal positions. We will assign the reference speed W reference< to the part of the signal located between these two centers of the contact area. Of course, we can also consider two non-contiguous drops 111, 115 of the time signal 101 to determine a second reference speed W reference< and assign this second speed to the part of signal 101 located between drops 111 and 115.

[0054] There Fig. 3presents the result of the angular resampling step of the time signal 101. Thus, by taking advantage of the determination of the centers of the contact area for each drop in the time signal carried out in the previous step, we easily phase the time signal to the wheel revolution over 360 degrees. It is then agreed to linearly distribute the discretized measurement points over the wheel revolution. Even if an angular positioning error is made at this step, a linear interpolation carried out for example during the data aggregation step will smooth the results and minimize the angular positioning error. In a more sophisticated way, having evaluated a reference speed at each wheel revolution. It is possible to assign angular speeds which evolve over the wheel revolution by taking into account the reference speeds of the contiguous revolutions.For example, having determined the reference speeds over three consecutive turns, we can assign to the central wheel turn, a first reference speed on the first quarter of a wheel turn as being the barycenter speed of the reference speed of the previous turn weighted 2 and the reference speed of the current turn weighted 1. The following quarter will have a reference speed as being the barycenter speed of the reference speed of the current turn weighted 2 by the reference speed of the previous turn weighted 1. The third quarter of a wheel turn will have a reference speed as being the barycenter speed of the reference speed of the current turn weighted 2 by the reference speed of the next turn weighted 1. Finally, the last quarter of a wheel turn will have a reference speed as being the barycenter speed of the reference speed of the current turn weighted 1 by the reference speed of the next turn weighted 2.All the discretized measurement points are distributed over each quarter turn of the wheel proportionally to the ratio of the reference speeds of each quarter turn to the reference speed of the current turn. Other methods of smoothing the points can also be carried out. Here, the spatial discretization of the points is not regular due to the variable rolling speed. It is entirely possible to make this discretization of the points of the signal 102 regular by applying a method of interpolation of the measurement points on a given angular distribution per wheel turn. This then makes it possible to obtain an angularly resampled signal 102 with a regular angular step. The . Fig. 3 shows the angularly resampled signal 102 which is periodic at the wheel revolution with any discretization of the measurement points.

[0055] There Fig. 4presents the result of the normalization step of the first angularly resampled signal 102 without interpolation of the points. Thus, by taking advantage of the periodicity per wheel revolution of the angularly resampled wheel revolution signal, the angular signal is easily cut per wheel revolution or on a multiple of the wheel revolution as illustrated in Fig. 4, here over 12 wheel revolutions. The normalization step consists of dividing the amplitude of the signal by the power squared function of the reference speed associated with each part of the wheel revolution. The reference speed having been determined during the first stage of signal processing 101 for example. The reference speed being an angular speed. The result observed on curves 103 and 103 bis is that between each wheel revolution, the amplitude of the normalized signal is similar. We no longer observe the strong amplitude variations between the various wheel revolutions carried out at different speeds and on different roads. In addition, the signal is centered on the unit value. Then, we superimpose the segments of wheel revolutions on the same angular interval of a length which is an integer multiple of 360 degrees which are materialized by the gray curves forming here a curve bundle 103.This allows us to realize the measurement dispersion between wheel revolutions, which is accentuated by the fact that the signals have not been corrected for Earth's gravity. However, if we apply a low-pass filter, we obtain the black curve 103 bis which is much less chaotic since it is cleaned of certain parasitic noises. This allows us to see that the 103 bis signal is periodic at the wheel revolution with weak variations between wheel revolutions. At the end of this normalization of the 102 signal, we obtain here an angularly resampled and normalized 103 signal. The . Fig. 4 shows the angularly normalized resampled signal 103 which is centered on the unit value as evidenced by the filter applied to the 103bis curve.

[0056] There Fig. 5is the result of the step of aggregating the data from signal 103 of the previous step which is an optional step. Here, we then superimpose the segments of each wheel revolution on the same angular interval of a length of 360 degrees which are materialized by the gray curves forming here a bundle of curve 104. This makes it possible to realize the dispersion of measurement between each wheel revolution, which is accentuated by the fact that the signals have not been corrected for Earth's gravity. However, if we apply a correction for Earth's gravity to each wheel revolution before the normalization step since the accelerometer is here sensitive to Earth's gravity, the aggregation of the data by an average method over a decile interval determines curve 104 bis which is much more stable at the wheel revolution.This makes it possible to obtain the deformation signal of the tire envelope subjected to external forces, in particular the static load in this case. This signal 104bis is representative of the measurement of the tire envelope in rolling conditions at variable speed on ground of any roughness. This curve is an invariant of the tire envelope in rolling conditions under static load in a state mounted on a rim.

[0057] There Fig. 6 is an illustration to explain the calculation of positive S +< and negative S -< energy densities on a normalized angularly resampled Sig TDR< 10 wheel revolution signal corresponding to a single wheel revolution. Of course, the method is identical if the angularly resampled normalized Sig TDR< wheel revolution signal is delimited over several wheel revolutions.

[0058] Threshold A is determined here as being the unit value. This threshold is shown by the continuous line 11. In fact, it is preferable for real signals to take a value equal to 0.7. If the signals are highly disturbed, a value equal to 0.5 or 0.6 can be chosen. On the other hand, for signals obtained on generally smooth roads, a value of the order of 0, 8 or 0.9 can be used. This value A must be fixed for all stages of the method.

[0059] The positive energy densities S +< or negative S -< are calculated as the summation of the absolute values ​​of the differences between the wheel revolution signal 10 and the unit value, represented by the continuous curve 11. Necessarily, the area delimited by the surfaces S +< is equal to the area delimited by the surface S -<.

[0060] From the estimation of these energy densities S, it is easy to determine the deformation of the tire envelope Def % subjected to a static load in rolling conditions.

Claims

1. Method for ascertaining the deformation of a tyre casing subjected to a load when mounted on a wheel so as to constitute a pneumatic mounted assembly in rolling state with rotation speed W, said tyre casing having a crown, intended to be in contact with the ground, and in revolution about a natural rotational axis, comprising the following steps: - Fastening at least one sensor to the tyre casing at the crown of the tyre casing so as to generate at least one output signal sensitive to the acceleration, in the direction normal to the crown, applied to said sensor in the tyre casing; - Acquiring (201) at least one first temporal signal Sig comprising at least the amplitude of the at least one output signal while rolling; - Delimiting the first signal Sig over a number NTdR of wheel turns so as to construct a wheel-turn signal SigTdR, NTdR being greater than 1; - Determining at least one reference speed Wreference (202) associated with at least one portion of the wheel-turn signal SigTdR; the method being characterized by the following steps: - Normalizing (203) the at least one portion of the wheel-turn signal SigTdR; by a variable which is a function F proportional to the square of the reference speed Wreference, over a number of wheel turns N'TdR, N'TdR being greater than or equal to 1; - Angularly resampling (204) the at least one portion of the wheel-turn signal SigTdR; - Defining at least one first energy density S (205) from the at least one portion of the angularly resampled normalized wheel-turn signal SigTdR, named S+ when the at least one portion of the angularly resampled normalized wheel-turn signal SigTdR; is greater than a threshold A, or named S- when the at least one portion of the angularly resampled normalized wheel-turn signal SigTdR; is less than or equal to said threshold A; - Identifying the deformation Def% (206) of the tyre casing as a function G of the at least one first energy density S.

2. Method for ascertaining the deformation of a tyre casing subjected to a load according to Claim 1, wherein the step of determining the reference speed Wreference (202) consists of establishing the ratio of the angular variation to the temporal duration separating two azimuthal positions of the sensor in the tyre casing around the natural rotational axis, from the first signal Sig or from a signal in phase with the first signal Sig, according to the following formula: W Reference = Δ α Δ t wherein α is the angular position and t is the temporal abscissa associated with the angular position.

3. Method for ascertaining the deformation of a tyre casing subjected to a load according to either of the preceding claims, wherein the angular pitch is less than 18 degrees, preferably less than 6 degrees, very preferably less than 3 degrees.

4. Method for ascertaining the deformation of a tyre casing subjected to a load according to one of the preceding claims, wherein the method comprises a step of aggregating the data from the at least one portion of the angularly resampled normalized wheel-turn signal SigTdR; over at least one sub-portion of the at least one portion of the angularly resampled normalized wheel-turn signal SigTdR, the sub-portion of the at least one portion of the angularly resampled normalized wheel-turn signal SigTdR; becoming the at least one portion of the angularly resampled normalized wheel-turn signal SigTdR.

5. Method for ascertaining the deformation of a tyre casing subjected to a load according to the preceding claim, wherein the sub-portion of the at least one portion of the angularly resampled normalized wheel-turn signal SigTdR; is one wheel turn.

6. Method for ascertaining the deformation of a tyre casing subjected to a load according to one of the preceding claims, wherein the data aggregation step (205) comprises one of the methods included in the group comprising: mean over a decile interval, the median, the selection or interval of deciles, the methods of interpolation, the weighted or non-weighted mean, optimization of the parametric model of tyre deformation.

7. Method for ascertaining the deformation of a tyre casing subjected to a load according to one of the preceding claims, wherein having phased the first signal Sig with respect to an angular position of the tyre casing, a correction Corr is made to the first signal Sig to take account of the effect of terrestrial gravity before the normalization step.

8. Method for ascertaining the deformation of a tyre casing subjected to a load according to one of the preceding claims, wherein the method comprises a step of filtering the at least one portion of the angularly resampled normalized wheel-turn signal SigTdR.

9. Method for ascertaining the deformation of a tyre casing subjected to a load according to one of Claims 1 to 8, wherein the threshold A is between 0.5 and 0.9.

10. Method for ascertaining the deformation of a tyre casing subjected to a load according to one of Claims 1 to 9, wherein the positive energy density S+ and negative energy density S- are obtained by the following formulae: S + = ∑ Sig TDR > A Sig TDR − 1 ∗ N ′ TDR N U ; and S − = ∑ Sig TDR ≤ A Sig TDR − 1 ∗ N ′ TDR N U ; where N is the total number of discretized points on the angularly resampled normalized wheel-turn signal SigTDR.

11. Method for ascertaining the deformation of tyre casing according to Claim 10, wherein the first signal is acquired at a constant sampling frequency and the spatial discretization for sampling the first signal is less than 6 degrees, preferably less than 3 degrees, very preferably less than 1 degree.

12. Method for ascertaining the deformation of a tyre casing subjected to a load according to one of Claims 1 to 11, wherein the function G is a linear function of the spectral density S according to the following formula: G X = X / N ′ TdR .

13. Method for ascertaining the deformation of a tyre casing according to one of Claims 1 to 11, wherein the function G is a linear function of the spectral densities S+ and S- according to the following formula: G X , Y = X + Y / 2 * N ′ TdR .