Predictive maintenance methods and systems

Abstract

This invention relates to a predictive maintenance method and apparatus for a component (11) of a highway vehicle (10). The method is implemented by at least one control unit (12) connected to the component and includes the following steps: a. vehicle data collection, including: first data and second data related to the use of first and second predetermined usage parameters of the component, and third data capable of determining changes in at least one control parameter representing component wear based on vehicle mileage; b. based on the first and second data of the vehicle, selecting a category from a plurality of vehicles having similar first and second data, where the first and second data of the vehicle are similar to those of the selected category; c. comparing the third data of the vehicle with reference data for the selected category, which is obtained from the aforementioned control parameter change data; d. inferring the future behavior of the component based on the comparison result.

Description

Technical Field

[0001] The information disclosed this time pertains to the field of predictive maintenance of highway vehicle components. Background Technology

[0002] Preventive maintenance for highway vehicles refers to regular maintenance to replace components before they fail. Maintenance operations are generally scheduled based on the component's operating time, usually expressed in operating hours, with the calculated operating hours sufficient to allow for component replacement before failure.

[0003] This method is not very effective because it bases its work on component ratings without taking into account the actual condition of each component. As a result, various components will be replaced when they reach their expiration date, even if some of them are still usable.

[0004] Furthermore, in cases of sudden component failure, users discover the component malfunction without receiving prior warning of the component damage.

[0005] US document US20180204393A1 proposes predicting the maintenance required for a vehicle component (such as the air filter) based on vehicle usage time or mileage. This prediction is implemented when the average and / or standard deviation of representative filter clogging data collected exceeds a threshold. The prediction is based on the variation in the average and / or standard deviation of representative filter clogging data. The threshold can be preset or estimated based on parameters such as vehicle mileage, vehicle driving records, or calibration values ​​obtained during filter installation.

[0006] This method has a drawback. The data obtained regarding filter clogging may not be relevant to the condition of the component or the vehicle, which makes the accuracy of detection and prediction low. Summary of the Invention

[0007] This disclosure improves the situation.

[0008] A primary purpose of this disclosure is to more reliably predict the future behavior of components, such as future failures.

[0009] It proposes a method for predictive maintenance of components of highway vehicles, implemented by at least one control unit connected to the component, comprising the following steps:

[0010] a. Collect vehicle data, including:

[0011] i. Use the first type of data in combination with the first type of predetermined usage parameters of the component.

[0012] ii. Use the second type of data in combination with the second type of predetermined usage parameters of the vehicle.

[0013] iii. A third type of data that can determine the change of at least one control parameter representative of component wear based on vehicle mileage.

[0014] b. Based on the first and second types of vehicle data, select a vehicle category from a pool of pre-determined vehicle categories whose first and second types of data are similar to the vehicle data of the selected category.

[0015] c. Compare the third type of data for the vehicle with reference data for the selected category, which is obtained from the variation data of the aforementioned control parameters of representative component wear for each vehicle in the selected category.

[0016] d. Based on the comparison results, infer the future behavior of the component.

[0017] On the other hand, an apparatus for predictive maintenance of at least one component of a highway vehicle is proposed, the apparatus comprising:

[0018] - There is at least one control unit connected to the component, which includes a processor and a memory for storing:

[0019] *For multiple predetermined vehicle categories, the first data related to a first predetermined combination of component usage parameters and the second data related to a second predetermined combination of vehicle usage parameters are similar to the data for vehicles of the same category, and

[0020] *Reference data associated with each category, derived from the aforementioned control parameter variation data for wear of representative components for each vehicle within that category.

[0021] At least one control unit with the following functions:

[0022] a. Collect a series of data from the vehicle, including:

[0023] i. Use the first type of data in combination with the first type of predetermined usage parameters of the component.

[0024] ii. Use the second type of data in combination with the second type of predetermined usage parameters of the vehicle.

[0025] iii. A third type of data that can determine the change in at least one control parameter representing the wear of a representative component based on vehicle mileage.

[0026] b. Based on the first and second types of vehicle data, select one category from a plurality of pre-established categories stored in memory, wherein the first and second types of vehicle data are similar to the data of the vehicle in the selected category.

[0027] c. Compare the third type of vehicle data with reference data of the selected category stored in memory.

[0028] d. Based on the comparison results, infer the future behavior of the component.

[0029] According to another aspect, a computer program is proposed, containing all or part of the instructions for implementing the methods defined herein when the program is executed by a processor. According to yet another aspect, a non-transitory, computer-readable recording medium on which such a program is recorded is proposed.

[0030] The features mentioned in the following paragraphs may be used selectively. They may be used independently or in combination.

[0031] - In addition, the method also includes, at the end of the comparison step, a detection step of detecting abnormal behavior of the component relative to the selected category of vehicle components, and a step of inferring the future behavior of the component, i.e., inferring the future failure of the component.

[0032] - The reference data for the selected category includes dispersed data on the variation of the average control parameters obtained for all vehicles in the selected category, and the comparison step includes comparing the control parameter values ​​of the vehicle at a certain mileage with the corresponding dispersed data.

[0033] - In addition, the method also includes issuing an alarm message when the comparison step ends and abnormal behavior of the vehicle component is detected.

[0034] - The first type of data includes multiple usage frequencies for each predetermined combination of usage parameters of the component; the second type of data includes multiple usage frequencies for each predetermined combination of usage parameters of the vehicle.

[0035] - Multiple preset categories were obtained by using an unsupervised classification algorithm on first and second types of data collected from multiple vehicles;

[0036] - In addition, the steps for collecting a range of vehicle data include collecting a fourth type of data related to the vehicle's mileage, while the steps for selecting a vehicle category include:

[0037] * Select multiple preset categories from a set of preset categories based on the vehicle's mileage, and * Select one category from multiple categories based on the vehicle's first and second data;

[0038] – The control unit is configured to detect abnormal behavior of a component that differs from that of a vehicle component of the selected category at the end of the comparison, and to infer the future failure of that component. Attached Figure Description

[0039] The following detailed description and analysis of the accompanying drawings will demonstrate other features, details, and advantages.

[0040] Figure 1 This is a schematic diagram of a predictive maintenance system according to one embodiment.

[0041] Figure 2 This is a schematic diagram of a predictive maintenance method according to one embodiment.

[0042] Figure 3A The first set of data collected by a vehicle according to one embodiment is shown.

[0043] Figure 3B This demonstrates a second set of data collected by a vehicle according to one embodiment.

[0044] Figure 3C This demonstrates a third set of data collected by a vehicle according to one embodiment.

[0045] Figure 4 It shows a combination of vehicle usage parameters.

[0046] Figure 5 express Figure 2 The set of categories used by the methods in the document.

[0047] Figure 6 This indicates a change in a control parameter for a selected category and associated vehicle. Detailed Implementation

[0048] The drawings and descriptions below primarily contain definitive information. Therefore, they not only aid in further understanding the content disclosed herein, but also, where necessary, in better defining it.

[0049] Figure 1 An embodiment is shown, suitable for implementation. Figure 2 A predictive maintenance system based on predictive maintenance methods. The predictive maintenance system includes a predictive maintenance device installed on a road vehicle 10 (here, a car), and a remote server 20 adapted to communicate with the predictive maintenance device installed on the vehicle. A road vehicle is any vehicle with an engine (typically an internal combustion engine or an electric motor) that travels on roads and is capable of carrying people or goods.

[0050] Vehicle 10 includes at least one component 11, at least one control unit 12, at least one electronic controller 13 of the component, and multiple sensors 14. Control unit 12 is connected to component 11, electronic controller 13, and multiple sensors 14, for example, via a Controller Area Network (CAN) or a FlexRay type data communication bus. When control unit 12 has a suitable communication interface, control unit 12 can be configured to communicate directly with remote server 20, or indirectly through another control unit with a suitable communication interface. In this case, data is transmitted between the two different control units via the aforementioned data communication bus. Note that the electronic controller 13 of the component can be independent (as shown here) or integrated into control unit 12.

[0051] In the example illustrated here, control unit 12 is an electronic control unit (ECU) including at least one processor, a memory, and communication interfaces with various actuators and sensors of the vehicle (particularly the engine), as well as a communication interface with a remote server 20. The electronic controller 13 of component 12 is configured to execute predetermined control laws based on various parameters measured by various sensors specific to, but not limited to, that component. Electronic controller 13 also includes at least one processor, a memory, and communication interfaces with various sensors, if necessary.

[0052] Furthermore, during the control of component 11, the electronic controller 13 measures a parameter representing component wear that participates in the control command; this parameter is also called the component's control parameter. This value can be collected periodically by the control unit 12 to understand the relationship between the component's wear parameter and mileage.

[0053] Some of the sensors 14 are capable of acquiring physical characteristics that describe the dynamic behavior of the vehicle 10. In one embodiment, the sensors 14 can acquire usage parameters such as vehicle speed, engine torque, engine temperature, accelerator pedal position, vehicle acceleration or deceleration, steering wheel angle, or steering angle.

[0054] Another portion of the sensor 14 can be used to obtain the operating parameters of the component. Advantageously, these operating parameters are those that affect the wear of the component. When the electronic controller 13 differs from the control unit 12 (here, the engine control unit), the operating parameters of the component can also be obtained through the electronic controller 13. For example, parameters such as the injection pressure, oil temperature, injection quantity, and injection pump speed of a motor vehicle injector can be obtained. Of course, other types of components can also be considered, thus allowing for other operating parameters of the component to be taken into account.

[0055] Sensor 14 may also include an odometer to determine the number of kilometers the vehicle has traveled.

[0056] The control unit 12 is configured to collect data on vehicle use, component use, and component wear that varies with mileage.

[0057] In particular, the relevant data on vehicle use and component use can correspond to usage data of predetermined combinations of component use and vehicle use parameters, respectively.

[0058] The collected component usage data may include the frequency of use for predetermined combinations of component usage parameters. For example... Figure 3A As shown, for each predetermined combination of component usage parameters CP1(1), ..., CP1(N), the control unit 12 calculates the usage frequency of the specific combination of component usage parameters for the wear parameters of the injector, similar to the method described in paragraphs 1.10 on page 7 to 1.17 on page 8 of the patent application filed in the name of the applicant, number FR1900865.

[0059] In favorable circumstances, such as Figure 4 As shown, each usage parameter, such as parameters A, B, and C of a component, is divided into multiple numerical ranges {a1, ..., a20}, {b1, ..., b3}, {c1, ..., c3}. Each predetermined combination of component usage parameters CP1(1), ..., CP1(N) corresponds to a specific combination of one or more numerical ranges of each usage parameter A, B, and C of the component, for example, CP1(1) = {a1, a2; b2; c3}, ..., CP1(N) = {a20; b1; c1}. The data collected thereby can create a usage characteristic file for the vehicle component.

[0060] The collected vehicle usage data may include the frequency of use for predetermined combinations of vehicle usage parameters. For example... Figure 3B As shown, for each predetermined combination of vehicle usage parameters CP2(1), ..., CP2(N), the control unit 12 counts the number of times a specific combination of vehicle usage parameters is used. It will be found that the combinations of vehicle usage parameters are similar to... Figure 4 The description is based on a reference method, which is a combination of numerical ranges for various vehicle usage parameters. The data collected from this can be used to create a usage characteristic file for the vehicle.

[0061] Furthermore, the data related to the wear changes of the component corresponds to data related to the change of at least one wear parameter P3 of the component with mileage, such as... Figure 3C This is an illustrative reference. For example, a representative parameter for injector wear could be the injector's shut-off time.

[0062] from Figure 3C As can be seen, component wear parameters, that is, representative control parameters of component wear, change with mileage. The changes in wear parameters are related to the usage patterns of the vehicle and components, and thus to their usage characteristics.

[0063] Furthermore, as previously mentioned, the predictive maintenance device containing control unit 12 is capable of communicating with remote server 20. Remote server 20 is configured to collect data from multiple vehicles regarding changes in vehicle usage, component usage, and component wear with mileage, such as data previously referenced. Figure 3A , 3B And the data described in 3C. In particular, a set of data collected by remote server 20 corresponds to data from a reference similar to... Figure 1 The data described is collected and transmitted by multiple vehicles in the vehicle 10 (which includes a control unit 12 connected to components 11, an electronic controller 13, and multiple sensors 14).

[0064] The remote server 20 is also configured to determine multiple categories based on component usage and vehicle usage data collected for multiple vehicles as described above, using an unsupervised classification algorithm. The unsupervised classification algorithm can identify vehicles with similar component usage and vehicle usage characteristics. Vehicle usage and component usage characteristics are related to the frequency distribution of different combinations of component usage and vehicle usage parameters CP1(1), ..., CP1(N) and CP2(1), ..., CP2(M), respectively. Therefore, unsupervised classification can be performed based on the frequency of use of combinations of component wear and vehicle wear parameters collected for each vehicle.

[0065] Remote server 20 is also configured to determine reference data associated with each category. The reference data relates to changes in component wear parameters as a function of mileage and is obtained from component wear parameter variation data for each vehicle in a category.

[0066] Finally, the remote server 20 is configured to transmit data related to pre-established categories and relevant reference data to the vehicle's predictive maintenance device for storage in the memory of the control unit 12.

[0067] Therefore, the control unit 12 can infer the future behavior of the component based on data collected from the vehicle, such as data related to changes in vehicle use, component use, and component wear, as well as data related to multiple pre-established categories stored in its memory (including reference data associated with each pre-established category). Figure 2 Detailed explanation. Regarding the future behavior of a component, if a component exhibits abnormal wear behavior compared to similar vehicles, it may indicate future failure of the component or predict the wear parameter values ​​of the component.

[0068] In one embodiment, the vehicle 10 may also include a display (not shown) connected to the control unit 12 for displaying alarm information to the driver or a specialized repair service provider, for example, displaying an alarm when the component exhibits abnormal wear behavior that differs from wear data collected from vehicles with similar component and vehicle usage parameters, to indicate that the component needs repair.

[0069] Figure 2 The steps of the predictive maintenance method implemented by the control unit 12 of the predictive maintenance device are described in more detail.

[0070] Referring to the foregoing, the predictive maintenance method includes a step S100, in which the data storage receives and stores relevant data for categories pre-established by the remote server 20, as well as reference data associated with each category.

[0071] Predefined categories of data can identify vehicles with similar component and vehicle usage characteristics, while reference data is obtained from component wear parameter variation data for each vehicle in the relevant category. This reference data can be obtained through statistical analysis of the variation data collected for each vehicle in the relevant category. For example, reference data may include positional characteristics (such as pattern, median, arithmetic mean, quantiles) and dispersion characteristics (such as range, mean deviation, inter-quantile deviation, variance, standard deviation, and coefficient of variation). Figure 6 The example shows how the mean μ and standard deviation σ change with the mileage of a certain type of vehicle.

[0072] The predictive maintenance method also includes a step S200, which involves collecting data related to vehicle use, data related to component use, and data related to changes in component wear over mileage, as previously mentioned. Figure 3A , 3B As shown in 3C. It is important to note here that by using relevant data on vehicle and component usage, we can establish component usage characteristics and vehicle usage characteristics, that is, the distribution of usage frequency for various combinations of vehicle and component usage parameters.

[0073] Specifically, the number of times each predetermined combination of vehicle usage parameters CP2(1), ..., CP2(M) and each predetermined combination of component usage parameters CP1(1), ..., CP1(N) are used is collected. This can be achieved by adding a counter for each parameter combination, where each component or vehicle usage parameter falls within the range of values ​​for the relevant combination each time. Simultaneously, changes in wear parameters, i.e., changes in representative wear control parameters, are collected based on changes in mileage. In one embodiment, to reduce the amount of data collected, the average value of wear parameters obtained within a predetermined mileage can be collected, for example, an average value is collected every 100 kilometers.

[0074] The predictive maintenance method also includes a selection step S300, which selects a category from a plurality of preset categories stored in memory, wherein the component usage data collected during step S200 and the vehicle usage data are similar to the data of the selected category of vehicles.

[0075] Specifically, this involves using usage frequency distribution data with different combinations of component usage parameters CP1 and vehicle usage parameters CP2, along with preset category data stored in memory, to select categories with similar vehicle and component usage characteristics. For example, the preset category data might correspond to the position of the center point of each preset category, and the category with the center point closest to it is selected.

[0076] Then, in step S400, the change data of vehicle wear parameters collected in step S200 are compared with reference data related to the category selected in step S300.

[0077] In one embodiment, each value of the wear parameter collected in step S200 is compared with one or more thresholds defined according to relevant category reference data. For example, thresholds S1 and S2 can be determined based on the aforementioned location and / or dispersion characteristics.

[0078] exist Figure 6 In the example shown, the values ​​of wear parameters collected for the vehicle are checked, indicated by crosses in the graph, to determine if:

[0079] - Greater than the first threshold S1, i.e., μ+σ, or

[0080] - Less than the second threshold S2, i.e., μ-σ

[0081] For all vehicles in this category, the mean value of component wear parameter P3 at a given mileage is μ and the standard deviation is σ.

[0082] To reiterate, the wear parameter of the fuel injector, which represents the control parameter of component wear, can be the fuel injector's closing time.

[0083] As an alternative or supplement, step S400 may include a sub-step of calculating a floating average value from data related to the wear variation of the component collected in step S200, which is the average value of the wear parameters collected every 100 kilometers, and a sub-step of comparing the floating average value with the average value.

[0084] Then, in step S450, it is determined whether the behavior of vehicle component 10 is abnormal compared to vehicle components of the selected category. For example, abnormal behavior of the component is detected when the representative values ​​of the collected wear parameters exceed the reference range a certain number of times. Figure 6 In the example shown, the reference range is defined by the two thresholds S1 and S2 mentioned above. We will note that the reference range, i.e., the thresholds S1 and S2, may change with the mileage traveled, such as... Figure 6 The point-like curve is shown in the figure.

[0085] Then, in step S500, the future behavior of the component is inferred.

[0086] In one embodiment, when abnormal behavior is detected in step S450, a future failure of the component is inferred from it. In step S600, an alarm message is issued to the driver or maintenance agency, indicating that maintenance of the component is required.

[0087] Step S200 can be performed continuously, while steps S300, S400, S450 and S500 can be performed periodically according to the mileage traveled by the vehicle.

[0088] Therefore, the vehicle data collection step S200 may also include collecting data related to the number of kilometers the vehicle has traveled. When the vehicle has traveled a predetermined number of kilometers, step S300 and subsequent steps may be implemented, for example, when the number of kilometers the vehicle has traveled since the last implementation of these steps is less than 10,000 kilometers.

[0089] Furthermore, in one embodiment, step S300 of selecting a category from a plurality of categories pre-established and stored in memory on the remote server 20 may include:

[0090] - In the first selection sub-step S310, based on the mileage traveled by vehicle 10 determined in step S200, multiple preset categories are selected from a set of preset categories, and

[0091] - The second selection sub-step S320 selects one category from the multiple categories selected in sub-step S310.

[0092] In fact, as referenced Figure 5As shown, the category groups pre-established and stored in the memory of the electronic maintenance device by the remote server 20 may include multiple subcategories, each of which is established for a pre-determined range of driving kilometers.

[0093] exist Figure 5 In the example shown, C1,1,C1,2,C1,3,C1,4,C1,5,C1,6 are subcategories created for vehicles that have traveled 0 to 10,000 kilometers.

[0094] -C2,1,C2,2,C2,3,C2,4,C2,5,C2,6 are subcategories established for vehicles that have traveled 0 to 20,000 kilometers;

[0095] -C3,1,C3,2,C3,3,C3,4,C3,5 are subcategories established for vehicles that have traveled 0 to 30,000 kilometers;

[0096] -C4,1,C4,2,C4,3,C4,4,C4,5,C4,6,C4,7,C4,8,C4,9 are subcategories established for vehicles that have traveled 0 to 40,000 kilometers;

[0097] -C5,1,C5,2,C5,3,C5,4,C5,5 are subcategories established for vehicles that have traveled 0 to 50,000 kilometers.

[0098] Each subcategory includes vehicles and components with similar usage characteristics within a predetermined mileage range. Similarly, the reference data associated with these subcategories is calculated for the corresponding mileage range.

[0099] exist Figure 5 In the example shown, when step S300 is executed, a vehicle that has traveled 35,000 kilometers will be classified into one of the subcategories C4,1, C4,2, C4,3, C4,4, C4,5, C4,6, C4,7, C4,8, or C4,9.

[0100] Therefore, dividing a pre-defined set of categories into different subcategories based on vehicle mileage allows for the implementation of unsupervised classification algorithms using currently available data. This is particularly advantageous when the vehicles for which data is collected have not all accumulated a very long mileage. This makes it possible to obtain reference data at different mileage ranges, which are more accurate within the initial few mileage ranges. In fact, many more vehicles have traveled within the initial few mileage ranges, such as... Figure 5 The width of the box in the middle, which symbolizes the number of vehicles contained in each subcategory, is shown.

[0101] In advantageous cases, step S100, which involves receiving and storing information about preset categories and reference data associated with each category, can be updated periodically, particularly to refine the model as the mileage driven by various vehicles increases, i.e., to refine the categories and related reference data.

[0102] Furthermore, we will note that the data collected by control unit 12 can also be transmitted to a remote server to establish categories and calculate reference data associated with these categories using an unsupervised classification algorithm. The category establishment then includes a selection sub-step, where all data collected from different vehicles is selected based on the number of kilometers traveled. Therefore, after each vehicle has traveled a predetermined number of kilometers, the data collected and used by remote server 20 will be transmitted, where appropriate. In the example described here, this is done every 10,000 kilometers.

[0103] In another different embodiment, in step S450, if no abnormal behavior of the component is detected, it is inferred that the component behaves the same as all components in the selected category. In step S500, the wear parameter value of the component can be predicted after the vehicle has traveled a specified additional distance, thereby inferring the future behavior of the component. In particular, reference data for subcategories established for a range of mileage including a predetermined mileage can be used to predict at least one wear parameter value of the component within a predetermined mileage.

[0104] Advantageously, the selected relevant subcategory should include vehicles whose wear and component wear characteristics are similar to those of vehicles currently traveling at the time of step S100. Step S100 can then include receiving subcategory data and storing it in memory, particularly relevant reference data, for predicting wear parameter values ​​at a predetermined mileage based on the subcategory selected in step S320. For example, according to the foregoing example, if subcategory C4,3 is selected in step S320, and wear parameter values ​​for 100,000 kilometers are to be predicted, then subcategory C4,3 will be associated with subcategory C10,k, established for vehicles traveling between 0 and 100,000 kilometers, where the majority of vehicles in subcategory C4,3 have similar vehicle and component use characteristics to those in subcategory C10,k.

[0105] Then, by using at least one wear parameter, it becomes possible to deduce the end of the component's lifespan.

Claims

1. A method for predictive maintenance of a component (11) of a road vehicle (10), the method being implemented by at least one control unit (12) connected to the component (11), the method comprising the steps of: Vehicle data collection includes: First data relating to a first predetermined combination of usage parameters of the component. Second data relating to the second predetermined combination of usage parameters of the vehicle. It is possible to determine a third type of data based on the vehicle's mileage, which contains changes in at least one control parameter representing component wear. Based on the first and second types of vehicle data, from a pre-established pool of vehicle categories with similar first and second types of data, select a category whose first and second types of data are similar to those of the vehicles in the selected category. The third type of data for the vehicle is compared with reference data for the selected category, which is obtained from the variation data of the aforementioned control parameters for wear of representative components for each vehicle in the selected category. Based on the comparison results, the future behavior of the component can be inferred.

2. A method for predictive maintenance of components (11) of a road vehicle according to the preceding claims, wherein: The method also includes, at the end of the comparison step, a step for detecting anomalous behavior of the component relative to components of the selected category of vehicles, and The steps to infer the future behavior of the component are the steps to infer the future failure of the component.

3. A method for predictive maintenance of a component (11) of a road vehicle according to any of the preceding claims, wherein: - The reference data for the selected category includes the average dispersion data of control parameters obtained from all vehicles in the selected category, and - The comparison step includes comparing the control parameter values ​​of the vehicle at a certain mileage with the corresponding distributed data.

4. The method for predictive maintenance of components (11) of a highway vehicle according to claim 1, further comprising issuing an alarm message when the comparison step ends and abnormal behavior of the vehicle component is detected.

5. The method for predictive maintenance of components (11) of highway vehicles according to claim 1, wherein: - The first type of data includes multiple usage frequencies for each predetermined combination of usage parameters of the component; - The second type of data includes multiple usage frequencies for each predetermined combination of usage parameters of the vehicle.

6. The method for predictive maintenance of components (11) of a highway vehicle according to claim 1, wherein: - Multiple pre-established categories were obtained by implementing an unsupervised classification algorithm on first and second types of data collected from multiple vehicles.

7. The method for predictive maintenance of components (11) of a highway vehicle according to claim 1, wherein: - The vehicle data collection process also includes collecting a fourth type of data related to the vehicle's mileage. - The steps to select a vehicle category include: * Select multiple pre-defined categories from a set of pre-defined categories based on the vehicle's mileage, and * Select one of several categories based on the first and second types of data for the vehicle.

8. A device for predictive maintenance of at least one component (11) of a road vehicle (10), the device comprising: - There is at least one control unit (12) connected to component (11), which includes a processor and a memory for storing: *Multiple pre-established vehicle categories, each having first data related to the use of a first predetermined combination of component usage parameters and second data related to the use of a second predetermined combination of vehicle usage parameters, wherein the first and second data are similar to usage data for vehicles of the same category, and *Reference data associated with each category, derived from variations in control parameters representing component wear conditions for each vehicle within that category. The configuration of the at least one control unit is as follows: A series of data are collected from the vehicle, including: The first type of data is used in combination with the first type of predetermined usage parameters of the component. The second type of data is used in combination with the second type of predetermined usage parameters of the vehicle. A third type of data that can determine the variation of at least one control parameter representing component wear based on vehicle mileage. Based on the first and second types of vehicle data, a category is selected from a plurality of pre-established categories stored in memory, wherein the first and second types of vehicle data are similar to the data of the vehicle in the selected category. The third type of vehicle data is compared with reference data of the selected category stored in memory. Based on the comparison results, the future behavior of the component can be inferred.

9. The apparatus for predictive maintenance of at least one component (11) of a highway vehicle (10) according to claim 8, wherein the control unit is configured to detect abnormal behavior of the component relative to selected category of vehicle components at the end of a comparison and to infer future failures of the component.

10. A computer program product comprising instructions for a method implemented according to any one of claims 1 to 7 when the program is executed by a processor.

11. A non-transitory computer-readable recording medium having a program recorded thereon for performing the method of any one of claims 1 to 7 when executed by a processor.

Images