Method and device for predicting the failure of a vehicle component using fault codes
The predictive maintenance method uses fault codes to project vehicle data into a space of healthy vehicles, enabling real-time failure risk estimation and proactive maintenance to prevent component breakdowns, addressing the inefficiencies of current reactive and preventive maintenance approaches.
Patent Information
- Application Number
- FR2023004622
- Authority / Receiving Office
- FR · FR
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-05-10
- Publication Date
- 2024-11-15
AI Technical Summary
Existing vehicle maintenance methods are reactive, leading to costly and inconvenient corrective maintenance due to unpredictable component failures, while preventive maintenance does not guarantee avoidance of defects.
A predictive maintenance method using fault codes, projecting vehicle data into a space representative of fault codes from healthy vehicles to determine component failure probability through a learned model, enabling real-time risk estimation based on indicators from the vehicle and its environment.
Enhances the reliability and timeliness of predicting component failures, allowing proactive maintenance to prevent breakdowns and reduce downtime.
Smart Images

Figure 00000000_0000_ABST
Abstract
Description
Title of the invention: Method and device for predicting the failure of a vehicle component using fault codes. Technical field
[0001] The invention relates to methods and devices for predicting the failure of one or more components of a vehicle, in particular a motor vehicle. The invention also relates to a method and device for predictive maintenance of a vehicle. Technological background
[0002] Contemporary vehicles are made up of a large number of components or parts, each of which is susceptible to failure or defect.
[0003] When a vehicle driver encounters a defect or malfunction in one or more of their vehicle's components, they must contact a specialized center to determine whether the malfunction or defect requires corrective maintenance. Such corrective maintenance most often involves intervention or repair at a garage, for example. This type of maintenance entails costs and inconvenience due to the vehicle being out of service for a period that can sometimes last several days or even several weeks.
[0004] To limit the risks of breakdown or defect and prevent associated problems, preventive vehicle maintenance operations are scheduled. However, these maintenance operations do not guarantee that a vehicle component will not subsequently encounter a defect or experience a breakdown. Summary of the invention
[0005] One object of the present invention is to solve at least one of the problems of the technological background described above.
[0006] Another object of the present invention is to improve the prediction of failures or defects of organ(s) or component(s) of a vehicle.
[0007] According to a first aspect, the invention relates to a method for predicting the failure of a component of a first vehicle, the method comprising the following steps: - receiving first data representative of a set of first fault codes detected by the first vehicle via a wireless link; - projection of the first data into a representative space of occurrences of a set of fault codes from a first part of a set of second vehicles, the first part comprising second vehicles, said to be without fault, of the set of second vehicles having a level of occurrences of the set of fault codes lower than a level of occurrences of the set of fault codes of a second part of the set of second vehicles, the second part comprising second vehicles, said to be with failure, of the set of second vehicles; - determination of a set of initial representative characteristics of the first vehicle from a result of the projection; - prediction of a probability of failure of the component by feeding a failure prediction model with the first characteristics, the failure prediction model having been learned in a so-called learning phase from second data representative of a set of second fault codes obtained for the set of second vehicles, the space having been determined during said learning phase.
[0008] Such a method makes it possible to estimate, more reliably and in real time, the risk of failure associated with a vehicle component based on indicators received from that vehicle and representative of its actual use and / or environment. This estimation is based on a model learned from the same indicators of a set of other vehicles, some of which are in good working order or without faults. Using a representative sample of fault code occurrences for second, so-called fault-free vehicles (i.e., those with a low occurrence of fault codes) makes it possible to detect vehicles that deviate from this group of so-called healthy or fault-free vehicles and to detect a potential problem with the component in advance.
[0009] According to one variant, the method further comprises a step for learning the failure prediction model, the learning step comprising the following steps: - identification of each second vehicle of the first part and of each second vehicle of the second part from representative data of requests for support for repair of components of at least part of the set of second vehicles and from representative configuration data of each second vehicle of the set of second vehicles; - reception of the second set of data; - for each second fault code in the set, determination of a first occurrence of the existence of the second fault code for the second vehicles of the first part and of a second occurrence of the existence of the second fault code for the second vehicles of the second part from the second data; - selection of at least some of the second fault codes in the set by comparing, for each second fault code in the set, the first occurrence of existence with the second occurrence of existence, a second fault code being selected when the result indicates a difference between the second occurrence of existence and the first occurrence of existence exceeding a determined threshold; - determination of the space from the first occurrences of existence for at least a part of the second selected fault codes; - projection of the second data in space; - determination of a set of secondary characteristics representative of the set of secondary vehicles from a result of the projection; and - generation of the failure prediction model by classification of second characteristics.
[0010] According to another variant, the determination of the space includes a transformation of the first occurrences of existence into principal components by method of principal component analysis applied to the first occurrences of existence.
[0011] According to a further variant, the space is formed of 10 principal components obtained from the transformation.
[0012] According to another variant, the space is formed of a number of principal components between 3 and 20 principal components.
[0013] According to yet another variant, the first and second characteristics include: - a value corresponding to a logarithm of an absolute value of the result of the projection to which 1 is added; - a result of a convolution product between the principal components; - a result of a convolution product of the first 3 principal components among the principal components; - a result of convolving the first 3 principal components with the logarithm of each of the first 3 principal components; and - a result of a convolution of the logarithms of the first 3 principal components.
[0014] According to an additional variant, each fault code in the set of first fault codes and in the set of second fault codes corresponds to a fault code detected during a determined life phase of the first vehicle and of each second vehicle, respectively.
[0015] According to another variant, the method further includes a step of transmitting data representative of an alert when the probability of failure reaches a determined threshold.
[0016] According to a second aspect, the invention relates to a device for predicting the failure of a component of a vehicle, the device comprising a memory associated with a processor configured for the implementation of the steps of the process according to the first aspect of the invention.
[0017] According to a third aspect, the invention relates to a vehicle, for example of the type with tomobile, comprising a device as described above according to the second aspect of the invention.
[0018] According to a fourth aspect, the invention relates to a computer program which includes instructions adapted for the execution of the steps of the process according to the first aspect of the invention, in particular when the computer program is executed by at least one processor.
[0019] Such a computer program may use any programming language, and be in the form of source code, object code, or an intermediate form between source code and object code, such as in a partially compiled form, or in any other desirable form.
[0020] According to a fifth aspect, the invention relates to a computer-readable recording medium on which is recorded a computer program comprising instructions for carrying out the steps of the process according to the first aspect of the invention.
[0021] On the one hand, the recording medium can be any entity or device capable of storing the program. For example, the medium can include a storage means, such as a ROM, a CD-ROM or a microelectronic circuit-type ROM, or a magnetic recording means or a hard disk drive.
[0022] On the other hand, this recording medium can also be a transmissible medium such as an electrical or optical signal, such a signal being able to be transmitted via an electrical or optical cable, by conventional or radio frequency, by self-directing laser beam, or by other means. The computer program according to the invention can, in particular, be downloaded from an Internet-type network.
[0023] Alternatively, the recording medium may be an integrated circuit in which the computer program is incorporated, the integrated circuit being adapted to execute or to be used in the execution of the process in question. Brief description of the figures
[0024] Other features and advantages of the invention will become apparent from the description of the following non-limiting embodiments of the invention, with reference to the attached Figures 1 to 4, in which:
[0025] [Fig-1] schematically illustrates a vehicle communication environment, according to a particular embodiment of the present invention;
[0026] [Fig. 2] schematically illustrates a process of learning a failure prediction model for a vehicle component and predicting a failure probability for a vehicle component of [Fig. 1] based on the learned model, according to a particular and non-limiting embodiment of the present invention
[0027] [Fig.3] schematically illustrates a device configured for predicting the probability of failure of a component of the vehicle of [Fig.1], according to a particular embodiment of the present invention;
[0028] [Fig.4] illustrates a flowchart of the different steps of a method for predicting the probability of failure of a component of the vehicle of [Fig.1], according to a particular embodiment of the present invention. Description of the implementation methods
[0029] A method and device for predicting the probability of failure of one or more components of a vehicle will now be described in what follows with joint reference to Figures 1 to 4. The same elements are identified with the same reference signs throughout the description that follows.
[0030] According to a particular and non-limiting example of an embodiment of the invention, the prediction of a failure of one (or more) component(s) of a first vehicle, for example the on-board charger of a first vehicle of the electric or hybrid type including a traction battery, includes the reception, for example from the first vehicle via a wireless link, of first data representative of a set of first fault codes detected by the first vehicle and for example transmitted to a server of the "cloud" (or "cloud" in French).This initial data is projected onto a space representing the occurrences of a set of fault codes from the first part of a set of second vehicles. These second vehicles are considered "healthy" or "without faults," meaning they have no faults in the component in question. The occurrence level of fault codes for the second vehicles in this first part of the set is lower than the occurrence level of fault codes for the second part of the set of second vehicles, which corresponds to vehicles with faults. A set of initial characteristics representative of the first vehicle is determined from the projection result. Finally, the component's failure probability is determined by feeding these initial characteristics into a failure prediction model.Such a failure prediction model was previously learned during a so-called machine learning method training phase, also called automatic learning, from second data representative of a set of second fault codes obtained for the set of second vehicles, the space having been determined during the training phase.
[0031] A breakdown corresponds, for example, to a failure of a component or part of the vehicle, that is to say, a malfunction resulting in a cessation of the operation of the component or part in question. A breakdown corresponds also to a degraded operation of the component or organ in question compared to normal operation, for example, operation or only part of the services that the component or organ provides in normal operation is provided.
[0032] The implementation of a machine learning method makes it possible to estimate more reliably and in real time the risk of failure associated with a component of a vehicle based on indicators received from that vehicle and representative of the actual use made of that vehicle and / or the environment of that vehicle, on the basis of a model learned from the same indicators of a set of other vehicles which have not experienced failures of that component, or at least have experienced a low number of fault codes, and of a set of other vehicles which have experienced this failure, or at least have experienced a higher number of fault codes.
[0033] Predicting the risk of failure makes it possible to anticipate the problem and potentially correct it before it occurs, improving failure prevention.
[0034] Fig. 1 schematically illustrates a communication environment 1 of a first vehicle 10, according to a particular and non-limiting embodiment of the present invention.
[0035] The first vehicle 10 corresponds, for example, to a vehicle with an internal combustion engine, with electric motor(s), or even a hybrid vehicle with an internal combustion engine and one or more electric motors. Vehicle 10 thus corresponds, for example, to a land vehicle, such as a car, a truck, a bus, or a motorcycle.
[0036] The first vehicle 10 also advantageously incorporates a communication system configured to communicate with one or more remote devices 101 via a wireless communication network infrastructure. The remote device 101 advantageously corresponds to a device configured to process data, for example, data stored in the memory of the remote device 101 and / or data received from the vehicle 10. The remote device 101 corresponds, for example, to a "cloud" server 100.
[0037] The communication system of the first vehicle 10 includes, for example, one or more communication antennas connected to a telematic control unit, known as a TCU (Telematic Control Unit), which is itself connected to one or more computers of the embedded system of the first vehicle 10. The antenna(s), the TCU, and the computer(s) form, for example, a multiplexed architecture for providing various services useful for the proper functioning of the first vehicle 10 and for assisting the driver and / or passengers of the first vehicle 10 in controlling the first vehicle 10 and / or for diagnosing the operation of one or more components of the first vehicle 10. The computer(s) and the TCU communicate and exchange data with each other via one or several computer buses, for example a CAN data bus communication bus (from the English "Controller Area Network" or in French "Réseau de contrôlers"), CAN FD (from the English "Controller Area Network Flexible Data-Rate" or in French "Réseau de contrôlers à débit de données flexible"), FlexRay (according to the ISO 17458 standard) or Ethernet (according to the ISO / IEC 802-3 standard).
[0038] The mobile communication infrastructure enabling wireless data communication between the first vehicle 10 and the remote device 101 includes, for example, one or more communication devices 110 such as a relay antenna (cellular network) or a roadside unit, known as a RSU. In a communication mode using such a network architecture, the data is, for example, transmitted by the first vehicle 10 to the remote device 101 in the "cloud" 100 via a relay antenna 110 (the antenna 110 being, for example, connected to the "cloud" 100 via a wired link).
[0039] The wireless communication system enabling data exchange between the first vehicle 10 and the remote device 101 corresponds, for example, to: - a vehicle-to-infrastructure (V2I) communication system, for example based on the 3GPP LTE-V or IEEE 802.1 IP standards of ITS G5; and / or - a cellular network communication system, for example an LTE (Long-Term Evolution) network, LTE-Advanced (also called LTE 3G, 4G or 5G); and / or - a Wifi type communication system according to IEEE 802.11, for example according to IEEE 802.lin or IEEE 802.1 lac.
[0040] A process for predicting the failure of one or more components of the first vehicle 10, an example of which is illustrated in [Fig.2], is advantageously implemented by a system comprising the vehicle 10 (i.e. by one or more devices embedded in the vehicle 10, for example one or more computers) and the remote device 101 connected in communication with the vehicle 10 via a connection or a wireless link.
[0041] The failure prediction process presented below is described with regard to a particular and non-limiting example relating to the prediction of a failure of a component of the first vehicle 10 corresponding to the onboard charger controlling the charge of the traction battery of the first vehicle 10 corresponding according to this example to an electric or hybrid vehicle.
[0042] Of course, such a failure prediction process applies to any component or part of a vehicle, for example a starter battery (or 12 V battery), an ADAS (Advanced Driver-Assistance System), an alternator, a starter, one or more calculators, etc.
[0043] The prediction process advantageously comprises two phases, each of these phases comprising one or more operations.
[0044] The first phase 20 corresponds to a so-called learning or training phase of one or more prediction models and the second phase 21 corresponds to a so-called production or prediction phase (also called the inference phase) based on the model(s) learned in the learning phase and data feeding the learned model(s).
[0045] The first phase and the second phase are for example implemented by the remote device 101.
[0046] According to one embodiment, the first phase is implemented by the remote device 101 (or by a server outside the "cloud", for example a server hosted in a center (for example a design office) adapted to do the learning) and the second phase by the first vehicle 10, for example by a computer of the embedded system of the first vehicle 10.
[0047] According to another embodiment, learning is implemented in the first phase and the parameters of the prediction model(s) are refined in real time from the data (including the first indicators) received during the second phase by one or more first vehicles.
[0048] Learning phase
[0049] The learning implemented in the first phase advantageously corresponds to supervised learning from a set of data associated with a set or group of second vehicles 11 comprising, for example, a few dozen, a few hundred, thousands or tens / hundreds of thousands of second vehicles 11. According to an alternative embodiment, the learning implemented in the first phase corresponds to unsupervised learning from the set of data associated with the set of second vehicles 11.
[0050] According to a particular example, the second vehicles 11 included in the set of second vehicles all have the same or a similar configuration, that is to say they all correspond to the same type of vehicle (for example the same series of a particular model of vehicle) with identical or similar components or parts (same version for example).
[0051] In a first operation 201 of the learning phase 20, the second vehicles, referred to as non-defective (forming a first part of the set of second vehicles 11), and the second vehicles 11, referred to as defective (forming a second part of the set of second vehicles 11), are identified from: - representative data of requests for vehicle component repair support, corresponding for example to requests or applications for handling of work carried out by a garage or dealer and transmitted to the manufacturer of each second vehicle concerned, such a request being also called a "credit request", this data being obtained for example from a server 12 of the manufacturer receiving these credit requests and / or from the servers 12 of the garage or dealer transmitting these credit requests; and - representative configuration data of each second vehicle, for example obtained from a server 13 of the car manufacturer.
[0052] Second vehicles identified as not having experienced any breakdowns are identified and grouped to form the first part of the second vehicle set 11 and second vehicles identified as having experienced breakdowns or a high breakdown rate are identified and grouped to form the second part of the second vehicle set 11.
[0053] The second vehicles of the first and second part are identified from data received from one or more servers 12, 13 of a car manufacturer and / or after-sales service, the second vehicles of the first and second part being selected for example in such a way that their configurations match.
[0054] The first part of the set of second vehicles and the second part of the set of second vehicles form, for example, coherent parts in that the second vehicles which compose them have an identical or similar configuration (for example same make and same model, or even same make, same model and same level of equipment).
[0055] The identification of these second vehicles is, for example, associated with each component of a plurality of components, the groups or parts being formed according to the component. For example, according to the example relating to the on-board charger, the first part of the second vehicles comprises the second vehicles that have not experienced a failure of the on-board charger, and the second part of the second vehicles comprises the second vehicles that have experienced at least one failure of the on-board charger.
[0056] In this first operation 201, each second vehicle 11 transmits to a server such as the remote device 101 second data representing the (second) fault codes encountered when such a fault code is detected or generated by a control unit of the second vehicle 11. The second vehicle 11 corresponds to a so-called connected vehicle, just like the first vehicle 10, configured to communicate data via a wireless link or connection. According to one embodiment, the fault codes generated by all or some of the second vehicles 11 are collected during a diagnostic of the second vehicle 11, these fault codes then being transmitted to the remote device 101 by the entity that performed the diagnosis (for example a garage or a car dealership).
[0057] A fault code is also called a DTC code (from the English "Diagnostic Trouble Code" or in French "Diagnostic fault code").
[0058] According to one embodiment, each second vehicle 11 further transmits data representative of a specific life phase of the second vehicle, such data including, for example, an identifier of the specific life phase. According to another embodiment, this data further includes data representative of a time interval associated with the life phase (for example, start and end times), depending on the life phase.
[0059] For example, each second vehicle 11 transmits data representative of a traction battery charging session when the component concerned by the prediction corresponds to the on-board charger.
[0060] The data received by the remote device 101 are for example stored in memory and associated with each second vehicle 11, each second vehicle 11 being for example identified by a unique identifier (for example the VIN (Vehicle Identification Number)) obtained from the configuration data.
[0061] The transmission of data relating to the determined life phase allows the remote device to select the fault codes obtained during this determined life phase only from the set of fault codes detected for the second vehicles 11.
[0062] The configuration data of each second vehicle 11 include, for example, the vehicle type, a vehicle identifier, the vehicle make, the vehicle model, the vehicle hardware and / or software configuration, etc.
[0063] In a second operation 202 of the learning phase 20, fault codes having a high (or high) conditional probability of occurrence (or existence) are identified.
[0064] The list of second fault codes generated and obtained for each second vehicle 11 are extracted, for example, from the memory of the device 101 in which these second fault codes are recorded and stored.
[0065] A first occurrence of the appearance or existence of fault codes is calculated or determined for the second vehicles 11 of the first part (without failure) and a second occurrence of the appearance or existence of fault codes is calculated or determined for the second vehicles 11 of the second part (without failure), the second vehicles 11 of the first part and of the second part having been identified in the previous operation 201.
[0066] Depending on the variant in which a particular life phase associated with the component for which the failure prediction model is calculated is taken into account, only the Second fault codes appearing during this particular life phase (e.g. traction battery charging sessions) are counted to determine the first and second occurrences.
[0067] A probability of occurrence of each fault code in the fault code list is calculated or determined for all second vehicles in Part 1 and for all second vehicles in Part 2. The probability is, for example, expressed as a percentage obtained based on the first and second occurrences relative to the number of second vehicles in Part 1 and Part 2 of the vehicle set.
[0068] For each second fault code in the fault code list, a difference between the probability of occurrence or existence of a fault code during an identified fault for the second vehicles of the second part (vehicles with fault) and the probability of occurrence or existence of a fault code for the second vehicles 11 of the first part (vehicles without fault) is calculated or determined.
[0069] A portion of the secondary fault codes is selected based on the calculated differences; for example, only the secondary fault codes for which the difference exceeds a threshold are selected. Such a threshold is, for example, equal to 5%. In other examples, the threshold is set at 7%, 10%, or 15%, for example.
[0070] In a third operation 203 of the learning phase 20, a space representing the first part of the second vehicles 11, that is to say the second healthy or fault-free vehicles, is calculated or determined.
[0071] The space is determined from the first occurrences of appearance or existence for the second selected fault code(s).
[0072] The determination of the space is obtained, for example, via the method known as principal component analysis, abbreviated PCA (from the English "Principal Component Analysis"), which consists of transforming related variables (the occurrence of fault codes) into new variables that are uncorrelated with each other. These new variables are called principal components, and such a method allows for the number of variables to be taken into account.
[0073] According to one embodiment, the space is determined according to the t-SNE algorithm (from the English "t-distributed stochastic neighbor embedding" or in French "insertion stochastique des voisins t-ditribués").
[0074] The space corresponds to a multidimensional space whose number of dimensions depends on the number of principal components selected. For example, the number of principal components selected to represent the space is equal to 10 (the first 10 principal components). According to other examples, the number of principal components selected is equal to 5, 8, or 15.
[0075] The selected principal components correspond to the eigenvectors more representative of fault code occurrences for second vehicles of the first part of second vehicles, i.e. healthy or fault-free vehicles.
[0076] Selecting the first 10 principal components allows, for example, to reconstruct about 98% of the variance of the dataset (the first occurrences of fault codes, which makes it possible to establish a proper reference space for the healthy or fault-free vehicles considered.
[0077] The second data representing the occurrence of each fault code of each second vehicle 11 are then projected onto the space thus defined, and 10 components are obtained for each second vehicle 11 (each second vehicle of the first part and each second vehicle of the second part).
[0078] In a fourth operation 204 of the learning phase 20, a set of second characteristics representative of the set of second vehicles 11 is determined from the components obtained for each second vehicle 11 in the previous step, that is to say from the result of the projection of the second data onto the proper space determined in operation 203.
[0079] The second set of characteristics includes, for example, all or part of the following characteristics or values: - a value corresponding to the logarithm of the absolute value of the result of the projection onto each of the components plus 1; - a result of a convolution product between the principal components, for example between the 10 principal components, for example for each pair of principal components; - a result of a convolution product of the first 3 principal components of the principal component set; - a result of convolving the first 3 principal components with the logarithm of each of the first 3 principal components; and - a result of a convolution of the logarithms of the first 3 principal components.
[0080] When the set of main components comprises 10 components and when the set of above characteristics are calculated, the number of second characteristics per second vehicle 11 is equal to 30.
[0081] In a fifth operation 205 of the learning phase 20, the failure prediction model of the component concerned, for example the on-board charger, is generated or trained by classification of the second characteristics.
[0082] For this purpose, the dataset used, i.e. the second characteristics, is processed to generate a training dataset, a validation dataset and a test dataset.
[0083] For example, the dataset formed from the second set of features is divided to form a test set (e.g. 30% of the total dataset) and a set for training and validation (e.g. 70% of the total dataset).
[0084] Then the set obtained for training and validation is again split to form a training dataset (for example 70% of the set for training and validation) and a model validation dataset (for example 30% of the set for training and validation).
[0085] The failure prediction model (also called the classification model) is then trained using the training dataset, for example according to a "random forest" type learning method.
[0086] Such a model makes it possible to differentiate between vehicles likely to break down, i.e., to have the component concerned present a defect, from those likely not to break down, with for example a probability associated with each class at the output of the model (the number of classes is for example equal to 2 or more, with a probability or a probability range that the component breaks down associated with each class).
[0087] The prediction model thus generated is then validated using the validation dataset.
[0088] The performance of the generated prediction model is then tested, for example, using the test dataset.
[0089] When the prediction phase 21 is implemented by the first vehicle 10, the parameters defining the prediction model obtained in the fifth operation are implemented through the remote device 101 (which will have previously received the data from the first vehicle 10) to predict a possible failure on the first vehicle 10.
[0090] Several prediction models are for example generated or learned as described above, a prediction model being generated for each component or group of components of a set of determined vehicle components.
[0091] Prediction phase
[0092] In a first operation 211 of the prediction phase 21, the device in charge of the prediction phase (i.e. the remote device 101 or one or more computers of the embedded system of the first vehicle 10) receives or collects first data representative of a set of first fault codes detected by the first vehicle 10. The first data corresponds for example to the second data having been used in the learning phase 20.
[0093] The first data are for example transmitted by the first vehicle 10 to the remote device 101 via the wireless network infrastructure over time when a first fault code is generated and detected by the first vehicle 10.
[0094] When the prediction phase is implemented by the first vehicle 10, the first data is stored in a memory associated with the computer of the embedded system of the first vehicle 10, or in a "cloud" server to which the first vehicle 10 is connected.
[0095] According to one embodiment, information representing the time at which each first fault code was detected, measured, or transmitted is associated with each fault code. In other words, the first data are time-stamped.
[0096] When the prediction phase is implemented by the remote device, the first vehicle 10 transmits, according to an alternative embodiment, its identification number, for example its VIN number.
[0097] The receipt or collection of the first data is similar to the receipt or collection of the second data.
[0098] According to one embodiment, the first vehicle 10 further transmits data representative of a specific life phase of the first vehicle, such data including, for example, an identifier of the specific life phase. According to another embodiment, this data further includes data representative of a time interval associated with the life phase (for example, start and end times), depending on the life phase.
[0099] For example, the first vehicle 10 transmits data representative of a traction battery charging session when the component concerned by the prediction corresponds to the on-board charger.
[0100] In a second operation 212 of the prediction phase 21, the first data are projected into the space representing the occurrences of a set of fault codes of vehicles without breakdown obtained during the learning phase 20.
[0101] According to an embodiment where one or more particular life phases of the first vehicle 10 are taken into consideration, the first projected data correspond to those of the first fault codes that occurred during this or these life phases, the set of first fault codes being, for example, filtered to select only those appearing during the life phase(s) considered.
[0102] In a third operation 213 of the prediction phase 21, a set of first representative characteristics of the first vehicle 10 are determined from the result of the projection of the first data into the space representative of healthy or fault-free vehicles.
[0103] The determination of the first characteristics is similar to the determination of the second characteristics described for learning phase 20 and is not repeated in detail here.
[0104] In a fourth operation 214 of the prediction phase 21, a probability of Failure of the component under consideration is predicted by feeding the failure prediction model obtained during the prediction phase with the first characteristics obtained in the third operation 213.
[0105] The initial data are thus classified, and a class is associated with the initial data provided as input to the prediction model. The failure probability associated with the resulting class provides a failure probability associated with the component under consideration.
[0106] According to a particular embodiment, an alert is for example generated and transmitted to the first vehicle 10 and / or to a mobile communication device 103 (for example a smartphone, a tablet or a connected object) connected wirelessly to the remote device 101 and / or to the first vehicle 10 and / or to a server of an after-sales service 12.
[0107] The alert is generated for example when the probability of failure exceeds a determined threshold (for example 0.6 or 0.7 or 0.8). - Data representing the alert for reaching a threshold are transmitted, for example: - to the first vehicle 10 and / or the mobile communication device 103 to alert the driver of the risk of breakdown and / or inform them of the procedure to follow to prevent the breakdown; and / or - to a server 101 associated with an after-sales service 12 in charge of contacting the first vehicle 10 (or its owner) to assist him in the measures to be taken to prevent the breakdown or defect. - When the prediction phase is implemented by the first vehicle 10, the data representing the alert of reaching a threshold are transmitted, for example: - intended for the driver or owner of the first vehicle 10 to alert them to the risk of breakdown and / or inform them of the procedure to follow to prevent the breakdown; this data is transmitted, for example, to a display device integrated into the first vehicle 10 and / or to a remote device such as a smartphone 103; and / or - to a server 101 associated with an after-sales service 12 in charge of contacting the first vehicle 10 (or its owner) to assist him in the measures to be taken to prevent the breakdown or defect.
[0108] The prediction phase is advantageously implemented in each first vehicle of a set or group of first vehicles comprising several first vehicles, for example, tens, hundreds, or thousands of first vehicles. According to this example, the data obtained from these first vehicles (first data, fault codes) are used to enrich in real time the learned prediction model(s) and refine the parameters of this or these prediction models. The data obtained during the prediction phase are thus also used in real time for continuous or real-time learning of the prediction model(s), according to this particular example.
[0109] Figure 3 schematically illustrates a device 3 configured to predict a failure of one or more components of a vehicle, for example, the first vehicle 10, according to a particular and non-limiting embodiment of the present invention. The device 3 corresponds, for example, to a device embedded in the vehicle 10, for example, a computer. According to another example, the device 3 corresponds to a computing or data processing device, for example, the remote device 101.
[0110] Device 3 is, for example, configured to carry out the operations described opposite Figures 1 and 2 and / or the steps of the process described opposite [Fig. 4]. Examples of such a device 2 include, but are not limited to, embedded electronic equipment such as a vehicle's on-board computer, an electronic control unit such as an ECU (Electronic Control Unit), a telematic control unit (TCU), a smartphone, a tablet, a laptop computer, a server, or a combination of several of the devices listed above. The elements of device 3, individually or in combination, may be integrated into a single integrated circuit, into several integrated circuits, and / or into discrete components.Device 3 can be implemented in the form of electronic circuits, software (or computer) modules, or a combination of electronic circuits and software modules.
[0111] Device 3 includes one (or more) processor(s) 30 configured to run Instructions for carrying out the steps of the process and / or for executing the instructions of the software embedded in the device 3. The processor 30 may include integrated memory, an input / output interface, and various circuits known to those skilled in the art. The device 3 further includes at least one memory 31, for example, volatile and / or non-volatile memory, and / or includes a memory storage device that may include volatile and / or non-volatile memory, such as EEPROM, ROM, PROM, RAM, DRAM, SRAM, flash, magnetic disk, or optical disk.
[0112] The computer code of the embedded software(s) including the instructions to be loaded and executed by the processor is for example stored on memory 31.
[0113] According to various particular embodiments, the device 3 is coupled in communication with other similar devices or systems and / or with communication devices, for example a TCU (Telematic Control Unit), for example via a communication bus or through dedicated input / output ports.
[0114] According to a particular and non-limiting embodiment, the device 3 comprises a block 32 of interface elements for communicating with external devices, for example a remote server or the cloud, a TCU unit. The interface elements of the block 32 comprise one or more of the following interfaces: - RF radio frequency interface, for example of type Bluetooth® or Wi-Fi®, LTE (from the English "Long-Term Evolution" or in French "Evolution à long terme"), LTE-Advanced (or in French LTE-avancé); - USB interface (from the English "Universal Serial Bus" or "Universal Serial Bus" in French); - HDMI interface (from the English "High Definition Multimedia Interface", or "High Definition Multimedia Interface" in French); - LIN interface (from the English "Local Interconnect Network", or in French "Réseau interconnecté local").
[0115] According to another particular embodiment, the device 3 includes a communication interface 33 which allows communication to be established with other devices (such as other computers of the embedded system or other servers) via a communication channel 330. The communication interface 33 corresponds, for example, to a transmitter configured to transmit and receive information and / or data via the communication channel 330. The communication interface 33 corresponds, for example, to a wired network of the CAN (Controller Area Network) type, CAN FD (Controller Area Network Flexible Data-Rate) type, FlexRay (standardized by ISO 17458) or Ethernet (standardized by ISO / IEC 802-3).
[0116] According to a further particular embodiment, the device 3 can provide output signals to one or more external devices, such as a display screen 340, one or more loudspeakers 350 and / or other peripherals 360 (projection system) via respective adapted output interfaces 34, 35, 36. In another variant, one or more of the external devices is integrated into the device 3. The display screen corresponds, for example, to a screen, whether touch-sensitive or not.
[0117] Figure 4 illustrates a flowchart of the various steps in a method for predicting the failure of one or more components of a vehicle, for example, the first vehicle 10, according to a particular and non-limiting embodiment of the present invention. The method is implemented, for example, by a device embedded in the vehicle 10, by the remote device 101, or by a system comprising the first vehicle 10 and the remote device 101.
[0118] In a first step 41, initial data representative of a set of initial fault codes detected by the first vehicle are received via a wireless link thread.
[0119] In a second step 42, the first data are projected into a space representing occurrences of a set of fault codes from a first part of a set of second vehicles, the first part comprising second vehicles, said to be without failure, from the set of second vehicles having a level of occurrences of the set of fault codes lower than a level of occurrences of the set of fault codes from a second part of the set of second vehicles, the second part comprising second vehicles, said to be with failure, from the set of second vehicles.
[0120] In a third step 43, a set of first representative characteristics of the first vehicle is determined from a result of the projection.
[0121] In a fourth step 44, a probability of failure of the component is predicted by feeding a failure prediction model with the first characteristics, the failure prediction model having been learned in a so-called learning phase from second data representative of a set of second fault codes obtained for the set of second vehicles, the space having been determined during said learning phase.
[0122] According to one embodiment, the variants and examples of the operations described in relation to Figures 1 and 2 apply to the steps of the process in [Fig.4].
[0123] Of course, the invention is not limited to the embodiments described above but extends to a method of warning of a failure before the failure occurs and the device configured for the implementation of such a method.
[0124] The invention also relates to a vehicle, for example an automobile or more generally an autonomous land-powered vehicle, comprising device 3 of [Fig.3],
[0125] The invention also relates to a system comprising device 3 of [Fig.3] (or a set of devices 3) mounted in a vehicle and a remote device, for example remote device 101, the device(s) 3 being connected in wireless communication with remote device 101.
Claims
Claims
1. A method of predicting failure of a component of a first vehicle (10), said method comprising the following steps: - reception (41) of first data representative of a set of first fault codes detected by said first vehicle (10) via a wireless link; - projection (42) of said first data into a space representative of occurrences of a set of fault codes of a first part of a set of second vehicles (11), said first part comprising second vehicles, said to be without fault, of said set of second vehicles having a level of occurrences of said set of fault codes lower than a level of occurrences of said set of fault codes of a second part of said set of second vehicles, said second part comprising second vehicles, said to be with fault, of said set of second vehicles; - determination (43) of a set of first characteristics representative of said first vehicle (10) from a result of said projection (42); - prediction (44) of a probability of failure of said component by feeding a failure prediction model with said first characteristics, said failure prediction model having been learned in a so-called learning phase (20) from second data representative of a set of second fault codes obtained for said set of second vehicles (11), said space having been determined during said learning phase (20).
2. The method of claim 1, further comprising a step of training said failure prediction model, said training step comprising the following steps: - identification (201) of each second vehicle of said first part and of each second vehicle of said second part from data representative of requests for support for repair of components of at least one part of said set of second vehicles and from data representative of configuration of each second vehicle of said set of second vehicles; - reception of said second data; - for each second fault code of said assembly, determination
3.
4.
5. of a first occurrence of existence of said second fault code for the second vehicles of said first part and of a second occurrence of existence of said second fault code for the second vehicles of said second part from said second data; - selection (202) of at least a part of the second fault codes of said set by comparison, for each second fault code of said set, of said first occurrence of existence with said second occurrence of existence, a second fault code being selected when said result indicates a difference between said second occurrence of existence and said first occurrence of existence greater than a determined threshold; - determining (203) said space from said first occurrences of existence for said at least one part of the second selected fault codes; - projection of said second data into said space; - determination (204) of a set of second characteristics representative of said set of second vehicles from a result of said projection; and - generation (205) of said failure prediction model by classification of said second characteristics. Method according to claim 2, for which the determination (203) of said space comprises a transformation of the first occurrences of existence into principal components by a principal component analysis method applied to said first occurrences of existence. Method according to claim 3, wherein said space is formed of 10 principal components obtained from said transformation. Method according to one of claims 3 to 4, for which said first and second characteristics comprise: - a value corresponding to a logarithm of an absolute value of said result of said projection to which 1 is added; - a result of a convolution product between said principal components; - a result of a convolution product of 3 first principal components among said principal components; - a result of a convolution product of said first 3 principal components and a logarithm of each of said first 3 principal components; and - a result of a convolution product of the logarithms of said first 3 principal components.
6. Method according to one of claims 2 to 5, for which each fault code of said set of first fault codes and of said set of second fault codes corresponds to a fault code detected during a determined life phase of the first vehicle (10) and of each second vehicle (11), respectively.
7. Method according to one of claims 1 to 6, further comprising a step of transmitting data representative of an alert when said probability of failure reaches a determined threshold.
8. Computer program comprising instructions for implementing the method according to any one of the preceding claims, when these instructions are executed by a processor.
9. Device (3) for predicting failure of at least one component of a first vehicle, said device (3) comprising a memory (31) associated with at least one processor (30) configured for implementing the steps of the method according to any one of claims 1 to 7.
10. Vehicle (10) comprising the device (2) according to claim 9.
Citation Information
Patent Citations
Method and device for predicting the failure of at least one component of a vehicle
FR3120724A1
