Deterioration determination device

Abstract

The present invention relates to a deterioration determination device for determining deterioration of a rubber member with higher accuracy. A deterioration determination device determines the deterioration of a rubber member used in a travel system of a vehicle using a learning completion model obtained by machine learning, first data, second data, position information, and humidity or a humidity reflection parameter reflecting the humidity to determine the presence or absence of deterioration of the rubber member. The first data is time-series data of a temperature of the rubber member or a temperature-reflecting parameter reflecting the temperature of the rubber member, and the second data is time-series data of a steering angle of the vehicle.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to a deterioration determination device. BACKGROUND

[0002] In the past, as such a deterioration determination device, a device that determines deterioration of a non-pneumatic tire has been proposed (for example, refer to Patent Literature 1). In this device, it is determined whether or not the non-pneumatic tire is deteriorated based on a vibration value of the tire in a case where the non-pneumatic tire is mounted to a vehicle for use, an acceleration applied to the tire, a rotational speed of the tire, a temperature of a tread portion of the tire, a wear amount of the tread portion of the tire, a travel distance of the vehicle, a travel time of the vehicle, a travel speed of the vehicle, an acceleration of the vehicle, a travel path of the vehicle, a steering angle of the vehicle, a number of times of braking of the vehicle, and an operation amount of a brake of the vehicle.

[0003] [Patent Literature 1] Japanese Patent Application Publication No. 2023-53568

[0004] In the above-described deterioration determination device, it is desirable to determine deterioration of a rubber member with higher accuracy. SUMMARY

[0005] The main object of the deterioration determination device of the present disclosure is to determine deterioration of a rubber member with higher accuracy.

[0006] In order to achieve the above-described main object, the deterioration determination device of the present disclosure employs the following means.

[0007] The deterioration determination device of the present disclosure determines deterioration of a rubber member used for a travel system of a vehicle, in which it is determined whether or not the rubber member is deteriorated using a learning completed model obtained by machine learning in which first data, second data, position information of the vehicle, and humidity or a humidity reflection parameter that reflects the humidity are used as input variables and whether or not the rubber member is deteriorated is used as an output variable, the first data being time series data of a temperature of the rubber member or a temperature reflection parameter that reflects the temperature of the rubber member, the second data being time series data of a steering angle of the vehicle, the position information, and the humidity or the humidity reflection parameter.

[0008] In the degradation device disclosed herein, a learned model obtained through machine learning, using first data, second data, vehicle location information, and humidity or a humidity response parameter reflecting the humidity as input variables and the presence or absence of degradation of the rubber component as the output variable, is used to determine whether the rubber component has deteriorated. The first data is time-series data of the rubber component's temperature or a temperature response parameter reflecting the rubber component's temperature, and the second data is time-series data of the vehicle's steering angle. The deterioration of the rubber component is determined using a learned model that, in addition to the first data, second data, and location information, also uses humidity as an input variable. As a result, the deterioration of the rubber component can be determined with higher accuracy. Attached Figure Description

[0009] Figure 1 This is a configuration diagram showing a simplified configuration of a vehicle 20 equipped with a degradation determination device as an embodiment.

[0010] Figure 2 This is a flowchart illustrating an example of a decision routine executed by ECU60.

[0011] Figure 3 This is an explanatory diagram used to illustrate the construction method of the learned model. Detailed Implementation

[0012] Figure 1 This is a configuration diagram showing a simplified structure of a vehicle 20 equipped with a degradation determination device as an embodiment. As shown, the vehicle 20 includes a power source 32, tires 30a to 30d as rubber components, a body 38, suspension systems 40a to 40d, a navigation system 50, and an electronic control unit (hereinafter referred to as "ECU") 60.

[0013] The power source 32, for example, is configured as a motor or engine, which outputs the power for driving to the drive shaft 36, which is connected to the tires 30a and 30b, which serve as drive wheels, via a differential gear 35. The power source 32 is controlled by the ECU 60.

[0014] Tires 30a to 30d are constructed as rubber components with a tread, which is formed by adding sulfur to the surface of raw rubber and heating it.

[0015] The suspension devices 40a-40d are known suspension devices installed between the tires 30a-30d and the vehicle body 38 to mitigate impacts transmitted from the road surface to the vehicle body 38. The suspension devices 40a-40d include arms, springs, shock absorbers, and rubber bushings 42a-42d. The rubber bushings 42a-42d are inserted into the portion connecting the springs and shock absorbers to the vehicle body, absorbing small vibrations.

[0016] The navigation system 50 displays the route to the destination and the vehicle position Pv (which is the location information of the vehicle itself) on a display device (not shown) based on map information.

[0017] Although not shown in the diagram, the ECU60 is configured as a CPU-centric microprocessor. In addition to the CPU, it also has a ROM for storing processing programs, RAM for temporarily storing data, flash memory, input / output ports, and communication ports.

[0018] Signals from various sensors are input to the ECU 60. Examples of signals input to the ECU 60 include detection signals from sensors used to detect the state of the power source 32, wheel temperature Tw from temperature sensor 31a which detects the temperature of the tire 30a, acceleration αv from acceleration sensor 70 which detects acceleration in the forward and backward directions, steering angle θ from steering angle sensor 72 which detects the steering angle of a steering wheel (not shown), and humidity Hu from humidity sensor 74 which detects relative humidity.

[0019] Various control signals are output from the ECU 60 via the output port. Examples of signals output from the ECU 60 include control signals for controlling the power source 32. The ECU 60 stores vehicle specifications such as vehicle height H, vehicle weight, and the model of the power source 32 in its ROM.

[0020] In this embodiment of the vehicle 20, the ECU 60 controls the power source 32 in a manner that drives based on the required load rate and required power. Furthermore, during driving, the ECU 60 stores time-series data of the wheel temperature Tw from the temperature sensor 31a (i.e., first data Dt) and time-series data of the steering angle θ from the steering angle sensor 72 (i.e., second data Dθ) in a flash memory. The first data Dt and the second data Dθ are not deleted regardless of whether the ignition switch is turned on or off, but are stored as time-series data in the flash memory of the vehicle 20.

[0021] Next, the operation of the vehicle 20 configured in this way, especially the operation when determining the deterioration of the tire 30a, will be explained. Figure 2 This is a flowchart illustrating an example of a decision routine executed by ECU 60. This routine is executed at a predetermined time, or whenever the ignition switch is turned on and the vehicle 20 system is started.

[0022] When this routine is executed, the CPU of ECU60 inputs the first data Dt, the second data Dθ, the vehicle position Pv from the navigation system 50, the acceleration αv from the acceleration sensor 70, the humidity Hu from the humidity sensor 74, and the vehicle height H stored in ROM (S100).

[0023] Then, using the pre-created and stored learning model in ROM, and the first data Dt, the second data Dθ, the vehicle position Pv, ​​the acceleration αv, the humidity Hu, and the vehicle height H input in S100, the presence or absence of tire 30a deterioration is determined (S110), and this routine ends. Here, the learning model will be explained.

[0024] Figure 3 This is an explanatory diagram illustrating the method for constructing the learned model. During the construction of the learned model, the experimenter drove vehicle 20 for a specified period of time. The ECU 60 acquired data as input (input variables) to the neural network during the driving of vehicle 20, namely, first data Dt, second data Dθ, vehicle position Pv, ​​acceleration αv, and humidity Hu. For acquiring the first data Dt, there is a correlation between the deterioration of the tire 30a (a rubber component) and its historical temperature. For acquiring the second data Dθ, there is a correlation between the deterioration of the rubber component and the external force acting on the tire 30a due to the turning of vehicle 20. For acquiring the vehicle position Pv, ​​there is a correlation between the deterioration of tire 30a, the deterioration of the rubber component, and the condition of the road surface during driving (whether it is a paved road, the duration of road surface wetness in rainy areas, etc.). For acquiring the acceleration αv, there is a correlation between the deterioration of tire 30a and the external force acting on tire 30a. For acquiring the humidity Hu, there is a correlation between the deterioration of tire 30a and humidity. For obtaining the vehicle height H, there is a correlation between the deterioration of tire 30a and the external force acting on tire 30a according to the vehicle height H.

[0025] If data is obtained in this way and becomes an input variable, the tester stops the vehicle 20, visually confirms the presence or absence of tire deterioration 30a, and uses an electronic device such as a smartphone that can connect to the ECU 60 to input the presence or absence of tire deterioration 30a as an output (output variable) into the ECU 60. At this time, if the tire 30a's tread exhibits hardening, cracking, or splitting, or if the minimum condition that the tire should be maintained to function as a component of the vehicle 20 is not maintained, the tester determines that tire 30a has deteriorated.

[0026] ECU60 repeatedly acquires the following data: first data Dt, second data Dθ, vehicle position Pv, ​​acceleration αv, humidity Hu, vehicle height H, and the presence or absence of tire 30a degradation. Furthermore, if ECU60 acquires the required number of first data Dt, second data Dθ, vehicle position Pv, ​​acceleration αv, humidity Hu, vehicle height H, and tire 30a degradation values ​​for constructing a high-precision neural network, it constructs a learned model using the acquired first data Dt, second data Dθ, vehicle position Pv, ​​acceleration αv, humidity Hu, and vehicle height H as input and the presence or absence of tire 30a degradation as output, and pre-stores it in ROM. In S110, since the presence or absence of tire 30a degradation is determined based on the learned model stored in ROM and the first data Dt, second data Dθ, vehicle position Pv, ​​acceleration αv, humidity Hu, and vehicle height H input in S100, tire 30a degradation can be determined with higher accuracy.

[0027] The vehicle 20 equipped with the degradation determination device of this embodiment described above determines the presence or absence of tire degradation by using a learned model with first data Dt, second data Dθ, vehicle position Pv, ​​acceleration αv, and humidity Hu as inputs and the presence or absence of tire degradation as output, and the first data Dt, second data Dθ, vehicle position Pv, ​​acceleration αv, and humidity Hu to determine the presence or absence of tire degradation, thereby determining tire degradation with higher accuracy.

[0028] In the above embodiment, the first data Dt is set as the time series data of the wheel temperature Tw from the temperature sensor 31a, but it can also be set as the time series data of the temperature of other components located near the tire 30a and reflecting the temperature of the tire 30a, i.e., the temperature response parameter.

[0029] In the above embodiment, the humidity Hu is set to the humidity Hu from the humidity sensor 74, but it can also be set to a humidity response parameter that can indirectly measure the humidity Hu.

[0030] In the above implementation, the inputs used to construct the neural network are set as first data Dt, second data Dθ, vehicle position Pv, ​​acceleration αv, humidity Hu, and vehicle height H. However, it is sufficient to include at least the first data Dt, second data Dθ, vehicle position Pv, ​​and humidity Hu, but it is also possible to exclude acceleration αv and vehicle height H.

[0031] In the above embodiment, the presence or absence of deterioration of tire 30a was determined. However, instead of tire 30a, the presence or absence of deterioration of other rubber components used in the vehicle 20's running system, such as rubber bushings 42a-42d of suspension devices 40a-40d, drive shaft guards, link end guards, lower arm joint guards, etc., can also be determined.

[0032] In the above-described embodiment, the tester visually confirms the presence or absence of deterioration of tire 30a, and the presence or absence of deterioration of tire 30a is input to ECU 60 as an output (output variable) using an electronic device such as a smartphone that can be connected to ECU 60. However, it is also possible to take an image of the appearance of tire 30a with a camera and input it as image data to ECU 60, where ECU 60 uses the image data to determine the presence or absence of deterioration of tire 30a.

[0033] In the above implementation, a neural network is used to construct the fully learned model. However, a fully learned model can also be constructed using machine learning methods different from neural networks.

[0034] The correspondence between the main elements of the implementation method and the main elements of the invention listed in the means for solving the problem will be explained. In the implementation method, tire 30a corresponds to "rubber component" and ECU 60 corresponds to "deterioration determination device".

[0035] The correspondence between the main elements of the implementation method and the main elements of the invention listed in the "Means for Solving the Problem" column is merely an example of how the implementation method carries out the invention listed in the "Means for Solving the Problem" column, and therefore does not limit the elements of the invention listed in the "Means for Solving the Problem" column. In other words, the explanation of the invention listed in the "Means for Solving the Problem" column is based on the description in that column, and the implementation method is simply a specific example of the invention listed in the "Means for Solving the Problem" column.

[0036] The above describes the methods for implementing this disclosure using the embodiments, but this disclosure is not limited to such embodiments. Of course, it can be implemented in various ways without departing from the spirit of this disclosure.

[0037] This disclosure can be used in industries such as manufacturing deterioration detection devices.

[0038] Explanation of reference numerals in the attached figures

[0039] 20 - Vehicle; 30a-30d - Tires; 31a - Temperature sensor; 32 - Power source; 35 - Differential gear; 36 - Drive shaft; 40a-40d - Suspension system; 42a-42d - Rubber bushings; 50 - Navigation system; 60 - Electronic control unit (ECU); 70 - Acceleration sensor; 72 - Steering angle sensor; 74 - Humidity sensor.

Claims

1. A degradation determination device for determining the degradation of rubber components used in the running gear of a vehicle, wherein, The deterioration of the rubber component is determined using a machine learning model obtained by taking first data, second data, the vehicle's location information, and humidity or a humidity response parameter reflecting the humidity as input variables and the presence or absence of deterioration of the rubber component as the output variable. The first data is time series data of the temperature of the rubber component or a temperature response parameter reflecting the temperature of the rubber component, and the second data is time series data of the vehicle's steering angle.

Images