FR3133940A1 - THERMAL MONITORING SYSTEM FOR AT LEAST ONE COMPONENT OF A MOTOR VEHICLE, METHOD AND PROGRAM BASED ON SUCH A SYSTEM - Google Patents

Abstract

The invention relates to a thermal monitoring system for at least one component (C1, C2,…, Cn) of a motor vehicle (V), comprising - data from an automatic learning model (M) for predicting aging thermals of said component (C1, C2,…, Cn), on the basis of experimental rolling parameters (P1, P2,…, Pn, E1, E2,…, En) including digital simulation data, and physical tests, - a monitoring means for analyzing current running parameters (P1, P2, ..., Pn, E1, E2, ..., En) and for determining aging data for said component (C1, C2, ..., Cn) on the basis of the data from the machine learning model (M), - an alert means (A) for alerting of premature aging of said component (C1, C2,…, Cn) if the aging data indicates aging premature. The invention also relates to a method and a program based on such a system. Fig. 1

Description

THERMAL MONITORING SYSTEM FOR AT LEAST ONE COMPONENT OF A MOTOR VEHICLE, METHOD AND PROGRAM BASED ON SUCH A SYSTEM

The invention is part of the implementation of a system and a method for preventing the degradation of vehicle components under the effect of thermal stresses which vary depending on the driving mode adopted.

The idea consists of using data from thermal simulations and physical tests (generally referred to as BIG DATA), recorded for different operating points (running modes and conditions) in order to detect parts which may be damaged following thermal constraints.

Thermal constraints mainly depend on the operating point adopted. Thermal results can be determined on the basis of a digital fluid dynamics simulation (abbreviated MFN or CFD designating the English term “computational fluid dynamics”), on an engine computer.

Based on numerical simulations and physical tests, we can deduce that the rolling mode has a considerable impact on the thermal resistance of a component, and consequently on its lifespan. However, current maintenance catalogs provide for a maintenance intervention based on the number of kilometers traveled without taking into account the driving mode.

Today, preventive maintenance intervention is carried out in 2 cases:
- or based on the preventive maintenance logbook which contains the interventions to be carried out depending on the number of kilometers traveled in normal use. For example, the maintenance booklet may mention that the alternator has a minimum lifespan corresponding to 150,000 km, and the same information for a starter.

Deterioration of parts can be detected beforehand if the driver goes to a specialist mechanic for an inspection or diagnosis on board the vehicle. This diagnosis may identify parts presenting degraded performance and which may break down soon even before the mileage mentioned in the preventive maintenance log. This is due to the fact that the state of an automobile component is closely linked to the driving mode adopted by the driver.

The defects and problems presented by the two solutions mentioned above are:
- preventive intervention is carried out once the indicated mileage has been covered. However, the majority of organs find themselves out of service even before the recommended mileage;
- the check or diagnosis carried out by a mechanic involves significant analysis time;
- the impact of the mode and frequent driving conditions are not considered in the preparation of maintenance logs.

To overcome these defects, the invention proposes a thermal monitoring system for at least one motor vehicle component, comprising
- data from an automatic learning model for predicting thermal aging of said component, on the basis of experimental rolling parameters including digital simulation data, and physical test data,
- monitoring means for analyzing current driving parameters and for determining aging data for said component on the basis of data from the automatic learning model,
- an alert means for alerting of premature aging of said component if the aging data indicates premature aging.

Advantageously, the invention relates to a process based on artificial intelligence for evaluating the operating state of vehicle components taking into account the driving mode frequently adopted by the driver.

In particular, we designate an autonomous neural network capable of identifying components which may be damaged by thermal effects, and of alerting the driver in advance.

The invention makes it possible to correlate data from digital simulations, physical tests and the driver's driving mode. This correlation makes it possible to predict the components which may be damaged due to overshooting or thermal aging and to alert the driver in advance.

The system makes it possible to increase the reliability of the vehicle; to avoid corrective maintenance interventions; to limit maintenance time.

According to one variant, the learning model is based on a deep neural network. This makes it possible to have a precise estimate of the aging of said component.

According to one variant, the digital simulation data are digital fluid mechanics simulation data, and include thermal maps for said component. This allows for a very precise simulation.

According to a variant, the thermal monitoring system can be connected to a maintenance plan, and the alert means is also configured to update the maintenance plan on the basis of the aging data of said component. One or more maintenance plans can be considered for one or more specific components. This makes it possible to update the preventive maintenance plan according to the driver's driving mode.

According to a variant, the alert means is further configured to indicate current driving parameters to be modified to limit the aging of said component. This encourages the driver to change the driving mode and offers driving that is more respectful of the components.

According to one variant, the current and/or experimental driving parameters include one or more parameters among:
- as input, a vehicle speed, an ambient temperature, an ambient humidity, a status and speed of the motor-fan unit, an engine torque, a slope of travel, a gearbox ratio, a condenser flow, flow rates and radiator inlet water temperatures, intake air flow rates and temperatures, exhaust gas flow rates and temperatures, engine and gearbox oil temperatures, material characteristics, such as emissivity, and specific heat; And
- at the output, maximum temperatures of the different components, temperatures of oil, exhaust gas, heat transfer fluid, fuel.

This allows you to have a faithful image of the driving conditions.

The invention further relates to a method for thermal monitoring of at least one motor vehicle component, comprising the following steps:
- produce an automatic learning model for predicting thermal aging of said component, on the basis of experimental running parameters including digital simulation data and physical test data,
- monitor said component by analyzing current driving parameters and determining aging data of said component on the basis of data from the automatic learning model,
- alert of premature aging of said component if the aging data indicates premature aging.

Preferably, the thermal monitoring method further comprises an updating step for updating a maintenance plan on the basis of the aging data of said component.

Another object of the invention relates to a computer program comprising program code instructions for executing the steps of the thermal monitoring method according to the invention, when said program operates on a computer.

The invention further relates to a motor vehicle comprising a thermal monitoring system according to the invention.

The invention will be further detailed by the description of non-limiting embodiments, and on the basis of the appended which illustrates a monitoring system according to a preferred embodiment of the invention.

The invention proposes to integrate artificial intelligence tools and massive data (or “Big Data” in English) to evaluate the effect of thermal on the state of vehicle components, denoted C1, C2, …, Cn in . This relationship is ensured by a model M based on a neural network which is autonomous after its training.

Training, test and validation data are recovered from the results of digital simulations of computational fluid mechanics models (CFD type) as well as physical tests. These results are established for different operating points which correspond to the different driving modes and conditions which may be adopted by a driver.

More generally, it is possible to determine external driving parameters (denoted E1, E2, ..., En) or internal (denoted P1, P2, ..., Pn) with reference to the vehicle or the component C1 considered.

The applicant aims to provide vehicles capable of capturing data from driving modes and analyzing them to offer personalized and refined services to drivers.

The proposed solution, for evaluating the operating state of vehicle components C1, C2, ..., Cn under the effect of thermal, will be deployed in real time.

The proposed invention involves data from the various sensors provided for this purpose, which are sent to an engine computer C.

Given the state of modern vehicles, we will not have constraints to embed the invention in the vehicle.

Today, modern cars are equipped with sensors and computers allowing real-time acquisition of data linked to each mode and driving conditions (parameters P1, P2, ..., Pn, E1, E2, ..., En).

The thermal mapping of the vehicle mainly depends on the driving mode adopted.

The system, object of the present invention, makes it possible to establish a correlation between the driving mode adopted by the driver and the thermal stresses of the different organs and components C1, C2, ..., Cn. Thanks to this correlation, the system will be able to predict and recalculate the longevity of the different parts (expressed for example in kilometers remaining, or another variable such as a percentage, an index, a gauge or a combination).

Based on this prediction, the system alerts the driver of the new mileage for which parts must be changed or suggests changing their driving mode to increase the life expectancy of the parts.

You can define an alert means A such as an indicator light or a notification.

The operation of our system can be described as follows:

Concerning firstly the analysis of the input data (training, testing and validation), for each vehicle, we can make a thermal synthesis of the complete vehicle V. This synthesis aims to validate the thermal resistance of the different components C1, C2, etc., Cn with respect to the different operating points. To do this, we mainly use CFD-type numerical simulations and physical tests in the wind tunnel and during driving (parameters P1, P2, …, Pn, E1, E2, …, En).

These two validation methods first make it possible to set up a database which contains the temperatures reached by each part and fluids C1, C2, etc., Cn (exhaust gas, water, oil, etc.) of the vehicles for the different points of operation, and then secondly to establish thermal aging laws for the different components C1, C2, ..., Cn and to predict the lifespan of each component, via the M model.

This database can be used as a training, testing and validation basis.

During CFD numerical simulations and physical tests, several operating points are proposed to represent the reality on the ground. Among the simulation and test parameters (parameters P1, P2, …, Pn, E1, E2, …, En), we find as input:
- vehicle speed,
- Room temperature,
- humidity,
- the status and speed of the motor-fan unit (abbreviated GMV),
- the engine torque,
- slope,
- the gearbox ratio,
- the condenser flow,
- the flow rates and water temperature at the radiator inlet,
- the flow rate and temperature of the inlet air,
- the flow rate and temperature of the exhaust gas,
- the temperature of the engine and gearbox oils,
- the characteristics of the materials (emissivity, specific heat, etc.).

Furthermore, at the output, we observe:
- the maximum temperatures of the different parts/components C1, C2, …, Cn,
- the temperature of the oils,
- the temperature of the exhaust gas,
- the heat transfer fluid temperature,
- the fuel temperature,
- the law of thermal aging,
- the impact of thermal aging on the longevity of components C1, C2, ..., Cn,
- etc.

Secondly, now regarding the use of the system, the M model goes through stages of training, testing and validation of the deep neural network.

Data from simulations and physical tests (parameters P1, P2, …, Pn, E1, E2, …, En) are available in a well-modeled database. Three subsets (training, test, validation) can be trained. The neural network is preferably trained until convergence.

Concerning thirdly the deployment of the valid neural network, during driving, via different sensors and computers on board the vehicle V, we can recover the neural parameters already validated, and reproduce the temperature maps for the different components C1, C2, …, Cn and vehicle fluids, recorded during CFD numerical simulations and physical tests.

For each driving mode, the system has a different thermal map, more or less harmful in relation to the different components of the vehicle C1, C2, ..., Cn. These thermal maps are linked to a thermal aging law, for each component C1, pre-established and recognized by the system. The system, based on this law, will be able to recalculate the longevity of components C1, C2, etc., Cn and update the maintenance plan by bringing the date of the maintenance intervention closer or further away depending on the frequently used driving mode. by the driver.

The system may also suggest that the driver change his driving mode to another less harmful one to increase the life expectancy of components C1, C2, etc., Cn.

The invention further relates to a method and a computer program based on such a system. The computer program can be loaded into a unit C, for example of the on-board computer type of the motor vehicle V.

Claims (10)

Thermal monitoring system for at least one component (C1, C2,…, Cn) of a motor vehicle (V), comprising
- data from an automatic learning model (M) for predicting thermal aging of said component (C1, C2,…, Cn), on the basis of rolling parameters (P1, P2,…, Pn, E1, E2 , …, In) experimental including numerical simulation data, and physical test data,
- a monitoring means for analyzing current rolling parameters (P1, P2, …, Pn, E1, E2, …, En) and for determining aging data of said component (C1, C2, …, Cn) on the basis data from the machine learning model (M),
- an alert means (A) for alerting of premature aging of said component (C1, C2,…, Cn) if the aging data indicates premature aging. Thermal monitoring system according to claim 1, characterized in that the learning model (M) is based on a deep neural network. Thermal monitoring system according to any one of claims 1 to 2, characterized in that the digital simulation data are digital fluid mechanics simulation data, and include thermal maps for said component (C1, C2,…, Cn). Thermal monitoring system according to any one of claims 1 to 3, connectable to a maintenance plan, characterized in that the alert means (A) is further configured to update the maintenance plan on the basis of the data from aging of said component (C1, C2,…, Cn). Thermal monitoring system according to any one of claims 1 to 4, characterized in that the alert means (A) is further configured to indicate driving parameters (P1, P2, ..., Pn, E1, E2, …, En) current to be modified to limit the aging of said component (C1, C2,…, Cn). Thermal monitoring system according to any one of claims 1 to 5, characterized in that the current and/or experimental driving parameters (P1, P2, …, Pn, E1, E2, …, En) comprise one or more parameters among :
- as input, a vehicle speed, an ambient temperature, an ambient humidity, a status and speed of the motor-fan unit, an engine torque, a slope of travel, a gearbox ratio, a condenser flow, flow rates and radiator inlet water temperatures, intake air flow rates and temperatures, exhaust gas flow rates and temperatures, engine and gearbox oil temperatures, material characteristics, such as emissivity, and specific heat; And
- at the output, maximum temperatures of the different components (C1, C2, ..., Cn), temperatures of oil, exhaust gas, heat transfer fluid, fuel. Method for thermal monitoring of at least one component (C1, C2, …, Cn) of a motor vehicle, comprising the following steps:
- produce an automatic learning model (M) for predicting thermal aging of said component (C1, C2, …, Cn), on the basis of rolling parameters (P1, P2, …, Pn, E1, E2, …, In) experimental including numerical simulation data, and physical test data,
- monitor said component (C1, C2, …, Cn) by analyzing current rolling parameters (P1, P2, …, Pn, E1, E2, …, En) and by determining aging data of said component (C1, C2 , …, Cn) based on the data from the machine learning model (M),
- alert of premature aging of said component (C1, C2, ..., Cn) if the aging data indicates premature aging. Thermal monitoring method according to claim 7, characterized in that it further comprises an updating step for updating a maintenance plan on the basis of the aging data of said component (C1, C2, …, Cn). A computer program comprising program code instructions for executing the steps of the thermal monitoring method according to any of claims 7 to 8, when said program is running on a computer. Motor vehicle comprising a thermal monitoring system according to any one of claims 1 to 6.