ANOMALITY DETERMINATION DEVICE FOR A POWER TRANSFER DEVICE

Abstract

Anomaly detection device (90) for a power transmission device, wherein the anomaly detection device (90) is used for a vehicle (VC) that has the power transmission device (40) and includes an oil temperature sensor (103), wherein the Power transmission device (40) includes a friction engagement element (75) and is configured to transmit a force delivered by a power source (10) of the vehicle (VC) to a drive wheel (60), and wherein the oil temperature sensor (103) is configured to detect an oil temperature which is a temperature of oil circulating in the power transmission device (40), wherein the anomaly detection device (90) comprises: a processing circuit (91, 92); and a storage device (93) wherein the storage device (93) stores a characteristic map data entry (DM1, DM2, DM3) which specifies a characteristic map and includes data learned through machine learning, an oil temperature sensor reading (103) is an oil temperature reading (Toil), If oil temperature relationship data (RDToil), which are data corresponding to time series data of the oil temperature sensing value (Toil), are entered into the map as an input variable, the map outputs a variable that determines whether the friction engagement element (75) has an anomaly, and the processing circuit (91, 92) is set up to to execute a procurement process that procures the input variable, and to perform an anomaly determination process which, based on the output variable which is output by the characteristic map as a result of the input variables obtained in the procurement process into the characteristic map, determines whether the friction engagement element (75) has an anomaly.

Description

1. Technical field

[0001] The following description refers to an anomaly detection device for a power transmission device with a friction engagement element, which determines whether the friction engagement element has an anomaly. 2. State of the art

[0002] JP 2011-058510A discloses an example of a power transmission device containing a frictional engagement element such as a clutch or a brake. If an anomaly occurs in the frictional engagement element of such a power transmission device, and the deterioration of the oil circulating in the device progresses, the oil emits a characteristic odor. As the deterioration of the oil circulating in the power transmission device progresses due to the anomaly of the frictional engagement element, the odor components of the oil change compared to those present when the frictional engagement element is not anomalous and the oil has not deteriorated.

[0003] CN 1 01 012 883 A relates to a device and a method for determining anomalies in a vehicle with a multi-stage automatic transmission. The device comprises an anomaly assessment unit, a storage device, and a transmission setup device for reassessment. When the vehicle restarts and the storage device contains information about a previously detected anomaly, the transmission is shifted into the appropriate gear to recheck whether the anomaly persists.

[0004] This allows anomalies that would otherwise go undetected to be identified before the journey begins, which can, for example, prevent switching to a higher gear and thus improve driving performance.

[0005] JP 2009-079716A concerns a shift control system for an automatic transmission designed to reduce the thermal stress on the friction elements while simultaneously improving driving performance. The shift control system includes a calculation of the current thermal stress, a prediction of heat generation before the start of a phase, a prediction of the thermal stress at the end of a phase, and a determination of whether the phase is permissible based on the predicted thermal stress. Heat generation is predicted based on the average value of the transmission torque and the relative rotational speed of the friction elements.

[0006] In the patent application described above, an odor sensor is arranged in an oil pan in which the oil is stored, in order to detect the odor components emitted by the oil. Thus, an anomaly in the power transmission device is predicted based on a detection value from the sensor.

[0007] To predict an anomaly in the power transmission device using the above method, a detection value from the odor sensor is required. Therefore, there is a need for a technique that detects an anomaly in a friction engagement element without using a detection value from an odor sensor. Summary of Revelation

[0008] This summary serves to present, in simplified form, a selection of concepts that are described in detail below. This summary is not intended to identify key features or essential characteristics of the claimed subject matter, nor is it intended to be used as an aid in determining the scope of protection of the claimed subject matter.

[0009] Aspects of the present disclosure, its functioning and benefits are as follows.

[0010] Aspect 1. An anomaly detection device for a power transmission device is used for a vehicle, comprising the power transmission device and an oil temperature sensor. The power transmission device includes a friction engagement element and is configured to transmit the power delivered by a power source of the vehicle to a drive wheel. The oil temperature sensor is configured to detect an oil temperature, which is the temperature of the oil circulating in the power transmission device. The anomaly detection device includes a processing circuit and a storage device. The storage device stores a map data entry (map data entry) that specifies a map (a map) and contains machine learning-learned data. A detection value from the oil temperature sensor is an oil temperature detection value.When oil temperature relationship data, which are time series data corresponding to the oil temperature measurement value, are entered as an input variable into the characteristic map (the figure), the characteristic map (the figure) outputs a variable that determines whether the friction engagement element exhibits an anomaly. The processing circuit is configured to execute a procurement process that retrieves the input variable and an anomaly detection process that, based on the output variable provided by the characteristic map as a result of the input variables retrieved during the procurement process, determines whether the friction engagement element exhibits an anomaly.

[0011] If heat is generated in the friction engagement element during the operation of the power transmission device, the heat can be transferred to the oil circulating in the power transmission device and change the oil temperature, i.e. the temperature of the oil.

[0012] The amount of heat generated by the friction element differs between cases where the friction element exhibits an anomaly and cases where it does not. Changes in the amount of heat generated cause the oil temperature to change differently. Therefore, whether the friction element exhibits anomalies is predicted by analyzing the oil temperature relationship data, which corresponds to the time series data of the oil temperature acquisition value.

[0013] In the configuration described above, the storage device stores a map data entry that specifies a characteristic map. The characteristic map receives the oil temperature relationship data as an input variable and outputs a variable that determines whether the friction engagement element exhibits an anomaly. During operation of the power transmission device, an output variable, which is generated by the characteristic map as a result of a provided input variable being fed into the map, is used to determine whether the friction engagement element exhibits an anomaly. This configuration allows for the determination of whether the friction engagement element exhibits an anomaly without using a reading from an odor sensor.

[0014] Aspect 2. In the anomaly detection device according to Aspect 1, the procurement process comprises a data acquisition process and a relationship data generation process. The data acquisition process receives time-series data of the oil temperature data acquisition, which contains a multitude of oil temperature data acquisitions taken in each acquisition cycle within a predetermined measurement period. The relationship data generation process normalizes the multitude of oil temperature data acquisitions contained in the time-series data of the oil temperature data to generate the oil temperature relationship data.

[0015] Time series data of the oil temperature reading when the friction element exhibits an anomaly differs from time series data of the oil temperature reading when the friction element does not exhibit anomalies. The degree of difference can vary depending on whether the oil temperature reading is relatively large or relatively small. If time series data of the oil temperature reading is used as an input variable for the characteristic map and the degree of difference is relatively small, the accuracy of the determination is likely to be lower than if the degree of difference is relatively large. In other words, the accuracy of the determination can vary depending on the magnitude of the oil temperature reading.

[0016] In this respect, in the configuration described above, oil temperature relationship data is inputted to the characteristic map as an input variable. The oil temperature relationship data is normalized data from the time series data of the oil temperature measurement value. Therefore, the degree of difference in the oil temperature relationship data between the case where the friction engagement element exhibits an anomaly and the case where the friction engagement element does not exhibit anomalies changes only slightly when the oil temperature measurement value is relatively large and when the oil temperature measurement value is relatively small. Thus, using the oil temperature relationship data as the input variable of the characteristic map reduces variations in the determination accuracy caused by the magnitude of the oil temperature measurement value.

[0017] Aspect 3. In the anomaly detection device according to Aspect 2, a value obtained by normalizing the oil temperature acquisition value is a normalized oil temperature acquisition value. In the relationship data generation process, the processing circuit is set up to normalize the multitude of oil temperature acquisition values ​​contained in the time series data of the oil temperature acquisition value in order to derive time series data of the normalized oil temperature acquisition value containing a multitude of normalized oil temperature acquisition values, and to generate as the oil temperature relationship data data that show a distribution of the numerical magnitude of the multitude of normalized oil temperature acquisition values ​​contained in the time series data of the normalized oil temperature acquisition value.

[0018] In the configuration described above, the oil temperature relationship data show the distribution of the numerical magnitude of the normalized oil temperature sensing values ​​contained in the time series data of the normalized oil temperature sensing value. Variations in the magnitude of the normalized oil temperature sensing values ​​can differ between the case where the friction engagement element has no anomaly and the case where the friction engagement element has an anomaly. Therefore, the oil temperature relationship data is used as an input variable of the characteristic map to increase the accuracy of the determination.

[0019] Aspect 4. In the anomaly detection device according to one of aspects 1 to 3, the input variable includes a vehicle speed.

[0020] The operating parameters of the power transmission device at relatively high vehicle speeds differ from those at relatively low vehicle speeds. As the operating parameters of the power transmission device change, so does the temperature of the oil circulating within it. Therefore, in the configuration described above, vehicle speed is used as an input variable for the characteristic map. This means the map outputs a variable that takes the vehicle speed into account. Using such an output variable increases the accuracy of the determination.

[0021] Aspect 5. In the anomaly detection device according to one of aspects 1 to 4, the power transmission device includes a clutch as a friction engagement element. The input variable includes at least one rotational speed of an input-side element of the clutch, a rotational speed of an output-side element of the clutch, or a rotational speed difference between the input-side element and the output-side element.

[0022] For example, when the clutch is engaged, the amount of heat generated by the clutch can vary depending on the rotational speed of the input element, the rotational speed of the output element, and the speed difference between the input and output elements. In this context, in the configuration described above, at least one of these values—the rotational speed of the input element, the rotational speed of the output element, or the speed difference—is used as an input variable for the characteristic map. That is, the characteristic map outputs a variable that takes into account at least one of these values. Using such an output variable increases the accuracy of the determination.

[0023] Aspect 6. In the anomaly detection device according to any of aspects 1 to 5, an input-output speed differential is a speed difference between an input section that inputs a torque into the frictional engagement element and an output section that receives a torque output by the frictional engagement element. The input variable includes a calculated value of a heat generation quantity of the frictional engagement element, which is calculated based on the product of the torque input into the frictional engagement element and the input-output speed differential.

[0024] The calculated value of the heat generated by the friction element is determined under the assumption that the friction element is free of anomalies. Therefore, the relationship between the calculated heat generation value and the changes in the oil temperature reading can differ between cases where the friction element exhibits anomalies and cases where it does not. In the configuration described above, the calculated heat generation value of the friction element is used as an input variable to the characteristic map. This means the characteristic map outputs a variable that incorporates the calculated heat generation value. Using such an output variable improves the accuracy of the determination.

[0025] Aspect 7. In the anomaly detection device according to one of aspects 1 to 6, the input variable includes an engagement force of the friction engagement element.

[0026] The amount of heat generated by the frictional engagement element can differ, for example, between a case where the engagement force of the frictional engagement element is relatively large and a case where the engagement force is relatively small. For instance, if the engagement force is relatively small and the frictional engagement element is in sliding engagement, the frictional engagement element generates a greater amount of heat than if the engagement force is relatively large and the frictional engagement element is in full engagement. In this context, in the configuration described above, the engagement force of the frictional engagement element is used as an input variable of the characteristic map. That is, the characteristic map outputs a variable that takes the engagement force of the frictional engagement element into account. Using such an output variable increases the accuracy of the determination.

[0027] Aspect 8. In the anomaly detection device according to one of aspects 1 to 7, the input variable includes a packing clearance of the friction engagement element.

[0028] In general, the engagement force required for the friction element will likely increase as the packing clearance of the friction element increases. This means that the amount of heat generated when engaging the friction element can vary in proportion to the packing clearance. Therefore, in the configuration described above, the packing clearance of the friction element is used as an input variable for the characteristic map. This means the characteristic map outputs a variable that takes the packing clearance of the friction element into account. Using such an output variable improves the accuracy of the determination.

[0029] Aspect 9. In the anomaly detection device according to one of aspects 1 to 8, the input variable includes a detection value from an acceleration sensor attached to the vehicle.

[0030] When the friction engagement element is engaged or disengaged, its operation generates a vibration in the power transmission device. This vibration is detected by the vehicle-mounted accelerometer. The accelerometer reading can differ between cases where the friction engagement element exhibits an anomaly and cases where it does not. Therefore, in the configuration described above, an accelerometer reading is used as an input variable for the characteristic map. This means the characteristic map outputs a variable that incorporates the accelerometer reading. Using such an output variable improves the accuracy of the determination.

[0031] Aspect 10. In the anomaly detection device according to one of aspects 1 to 9, the characteristic map outputs a variable that determines whether cold welding has occurred in the friction engagement element. In the anomaly detection process, the processing circuit is configured to determine, based on the output variable provided by the characteristic map as a result of inputting the input variables obtained during the procurement process, whether cold welding has occurred in the friction engagement element.

[0032] When the friction element is engaged and cold welding occurs within it, the amount of heat generated by the friction element changes compared to the state where no cold welding has occurred. This means the oil temperature reading changes differently. Therefore, the output variable provided by the characteristic map is used to determine whether cold welding has occurred within the friction element.

[0033] Aspect 11. In the anomaly detection device according to one of aspects 1 to 10, the characteristic map outputs a variable that determines whether the friction engagement element is in faulty engagement. In the anomaly detection process, the processing circuit is configured to determine, based on the output variable provided by the characteristic map as a result of inputting the input variables obtained during the procurement process, whether the friction engagement element is in faulty engagement.

[0034] When the friction engagement element is engaged and a faulty engagement occurs, the amount of heat generated by the friction engagement element changes compared to the case where no faulty engagement occurs. This means the oil temperature reading changes differently. Therefore, the output variable provided by the characteristic map is used to determine whether the friction engagement element is in a faulty engagement state.

[0035] Aspect 12. In the anomaly detection device according to one of aspects 1 to 11, the characteristic map outputs a variable that determines whether a seizing has occurred in the friction engagement element. In the anomaly detection process, the processing circuit is configured to determine, based on the output variable provided by the characteristic map as a result of inputting the input variables obtained in the procurement process into the characteristic map, whether a seizing has occurred in the friction engagement element.

[0036] If the friction element seizes, it cannot be disengaged. Therefore, the amount of heat generated by the friction element differs depending on whether it seizes or not. This means the oil temperature reading changes differently. Thus, the output variable provided by the characteristic map is used to determine whether the friction element has seized.

[0037] Aspect 13. In the anomaly detection device according to one of aspects 1 to 9, the storage device stores map data entries. The map data entries comprise a first map data entry, a second map data entry, and a third map data entry. The first map data entry specifies a map that, upon input of the input variables, outputs a variable determining whether cold welding has occurred in the friction engagement element. The second map data entry specifies a map that, upon input of the input variables, outputs a variable determining whether the friction engagement element is in faulty engagement. The third map data entry specifies a map that, upon input of the input variables, outputs a variable determining whether seizing has occurred in the friction engagement element.

[0038] In this configuration, the first map data entry is learned through machine learning specifically to determine whether cold welding has occurred in the friction engagement element. The second map data entry is learned through machine learning specifically to determine whether the friction engagement element is in faulty engagement. The third map data entry is learned through machine learning specifically to determine whether seizing has occurred in the friction engagement element. In this case, when an input variable is fed into a map specified by the first map data entry, the output variable provided by the map is used to determine whether the friction engagement element exhibits an anomaly caused by cold welding.Even if an input variable is entered into a map specified by the second map data entry, the output variable provided by the map is used to determine whether the friction engagement element exhibits an anomaly caused by faulty engagement. Similarly, if an input variable is entered into a map specified by the third map data entry, the output variable provided by the map is used to determine whether the friction engagement element exhibits an anomaly caused by seizing.

[0039] Aspect 14. In the anomaly detection device according to one of Aspects 1 to 12, the storage device stores map data entries separately according to each operating state of the friction engagement element. The map data entries include a first map data entry and a second map data entry. The first map data entry specifies a map that outputs a variable determining whether the friction engagement element has an anomaly when the oil temperature relationship data corresponding to the operating state of the friction engagement element (the first operating state) is input into the map as an input variable.The second map data entry specifies a map that outputs a variable determining whether the friction engagement element exhibits an anomaly when the oil temperature relationship data corresponding to the operating state of the friction engagement element (a second operating state distinct from the first) is input into the map. The processing circuit is configured to perform a data selection process that chooses the map data entry corresponding to the operating state of the friction engagement element from the map data entries stored in the memory device.In the anomaly determination process, the processing circuit is set up to determine, based on the output variable output from the characteristic map specified by the characteristic map data entry selected in the data selection process as a result of inputting the input variables obtained in the procurement process into the characteristic map, whether the friction engagement element has an anomaly.

[0040] Even if the friction control element exhibits an anomaly, the oil temperature sensing value can change differently depending on the friction control element's operating state. In the configuration described above, the first map data entry is specifically learned by machine learning when the friction control element's operating state is the first. The second map data entry is specifically learned by machine learning when the friction control element's operating state is the second. From these map data entries, one corresponding to the current operating state is selected, and an input variable is fed into a map defined by the selected map data entry.Subsequently, based on the output variable derived from the characteristic map, it is determined whether the friction engagement element exhibits an anomaly. Thus, the accuracy of the determination is increased by separately using the characteristic map data entries according to the operating state.

[0041] Aspect 15. In the anomaly detection device according to one of aspects 1 to 12, the storage device stores characteristic map data entries corresponding to a degree of deterioration of the power transmission device's characteristic. The processing circuit is configured to execute a data selection process that selects from the characteristic map data entries stored in the storage device the one corresponding to the degree of deterioration of the power transmission device's characteristic. In the anomaly detection process, the processing circuit is configured to determine, based on the output variable specified by the characteristic map data entry selected in the data selection process, whether the friction engagement element exhibits an anomaly. This output variable is generated as a result of inputting the input variables obtained in the procurement process into the characteristic map.

[0042] For example, even if the power transmission device does not exhibit anomalies, the amount of heat generated by the power transmission device can change in accordance with the degree of deterioration of its characteristics. In the configuration described above, the characteristic map data entries are used separately according to the degree of deterioration of the power transmission device's characteristics. More precisely, a characteristic map data entry corresponding to the degree of deterioration of the power transmission device's characteristics is selected, and an input variable is fed into a characteristic map specified by the selected characteristic map data entry. Then, based on the output variable provided by the characteristic map, it is determined whether the frictional engagement element exhibits anomalies.This increases the accuracy of the determination by using the characteristic map data entries separately according to the degree of deterioration of the characteristics.

[0043] Further features and aspects will be evident from the following detailed description, drawings, and claims. Brief description of the characters Fig. Figure 1 is a diagram showing a control unit and a drive system of a vehicle controlled by the control unit in a first embodiment. Fig. Figure 2 is a schematic cross-sectional view showing part of a torque converter when a lock-up clutch is in a disengaged state. Fig. Figure 3 is a schematic cross-sectional view showing part of the torque converter when the lock-up clutch is in an engaged state. Fig. 4 is the first half of a flowchart showing a series of processes performed by the controller. Fig. 5 is the second half of the flowchart, which shows the series of processes that are carried out by the controller. Fig. Figure 6 is a diagram showing changes in the oil temperature detection value. Fig. Figure 7 is a diagram showing a histogram of the oil temperature relationship data. Fig. Figure 8 is a diagram showing a histogram of oil temperature relationship data. Fig. Figure 9 is a block diagram showing a storage device of a control unit in a second embodiment. Fig. 10 is the first half of a flowchart showing a series of processes performed by the control unit. Fig. 11 is the second half of a flowchart showing a series of processes performed by the control unit. Fig. Figure 12 is a block diagram showing a storage device of a control unit in a third embodiment. Fig. 13 is the first half of a flowchart showing a series of processes performed by the control unit. Fig. 14 is the second half of the flowchart, which shows a series of processes performed by the control unit.

[0044] In the drawings and the detailed description, the same reference numbers refer to the same elements. The drawings may not be to scale, and the relative size, proportions, and representation of elements in the drawings may be exaggerated for clarity, illustration, and convenience. Detailed description

[0045] This description provides a comprehensive understanding of the described methods, devices, and / or systems. Modifications and equivalents of the described methods, devices, and / or systems are obvious to a person skilled in the art. The sequence of operations is exemplary and may be modified by a person skilled in the art, except for operations that necessarily follow a specific order. Descriptions of functions and designs that are well known to a person skilled in the art may be omitted.

[0046] Exemplary embodiments can take different forms and are not limited to the examples described. However, the examples described are detailed and complete and convey to a person skilled in the art the full scope of protection of the disclosure. First embodiment

[0047] A first embodiment of an anomaly detection device for a power transmission device is now described with reference to the Fig. 1, Fig. 2, Fig. 3, Fig. 4, Fig. 5, Fig. 6, Fig. 7 to Fig. 8 described.

[0048] First, a schematic configuration of a vehicle containing the anomaly detection device is described.

[0049] As in Fig. As shown in Figure 1, a vehicle VC includes an internal combustion engine 10, a transmission 40, and drive wheels 60. The internal combustion engine 10 includes a crankshaft 11, which is coupled to a torque converter 70 of the transmission 40. The torque converter 70 is coupled to an input shaft 81 of a transmission mechanism 80. The drive wheels 60 are coupled via a differential (not shown) to an output shaft 82 of the transmission mechanism 80.

[0050] The torque converter 70 comprises a front cover 71, a pump impeller 72, a turbine 73, and a stator 74. The front cover 71 is coupled to the crankshaft 11 of the internal combustion engine 10. The pump impeller 72 rotates integrally with the front cover 71. The turbine 73 rotates integrally with the input shaft 81 of the transmission 40. The stator 74 amplifies the torque between the pump impeller 72 and the turbine 73. When the pump impeller 72 rotates in accordance with the engine operation, its rotation is transmitted to the turbine 73 via oil in the torque converter 70. As a result, the output torque of the internal combustion engine 10 is fed into the transmission mechanism 80.

[0051] The torque converter 70 includes a lock-up clutch 75. When the lock-up clutch 75 is engaged, it mechanically connects the pump impeller 72 and the turbine 73. Thus, when the lock-up clutch 75 is engaged, the output torque of the internal combustion engine 10 is transmitted from the front cover 71 through the lock-up clutch 75 to the turbine 73 and fed into the transmission mechanism 80.

[0052] As in the Fig. 2 and Fig. As shown in Figure 3, the bridging clutch 75 comprises a carrier 76 that rotates integrally with the turbine 73, output-side elements 77 that are rotatably mounted integrally with the carrier 76, and input-side elements 78 that are supported by the front cover 71. When the direction in which the input shaft 81 of the transmission mechanism 80 extends is the lateral direction in the Fig. 2 and Fig. 3 corresponds to an axial direction AD, the inlet-side elements 78 are displaceable relative to the front cover 71 in the axial direction AD. However, a stopper 79 limits the displacement of an inlet-side element 78A, which is closest to the turbine 73, in the direction of the turbine 73 (i.e., in the Fig. 2 and Fig. 3 to the right) in the axial direction AD.

[0053] The outlet-side elements 77 are supported on the carrier 76 and are displaceable in the axial direction AD. Friction elements 77a are attached to opposite surfaces of each outlet-side element 77. The outlet-side elements 77 are arranged such that each outlet-side element 77 is located in the axial direction AD between adjacent input-side elements 78.

[0054] The lock-up clutch 75 is switched between the engaged and disengaged states by adjusting the hydraulic pressure of a pressure adjustment range 70a in the torque converter 70. More precisely, when the hydraulic pressure of the pressure adjustment range 70a in the Fig. When the condition shown in Figure 2 is increased, the outlet-side elements 77 and the input-side elements 78, with the exception of the input-side element 78A, slide in the axial direction AD towards the turbine 73 (to the right). Fig. 2) This means, as in Fig. Figure 3 shows the operating state of the bridging clutch 75 shifted into the engaged state, and axially AD adjacent elements of the input-side elements 78 and the output-side elements 77 are pressed against each other. When the hydraulic pressure of the pressure adjustment range 70a is in the Fig. When the state shown in Figure 3 is reduced, adjacent elements of the input-side elements 78 and the output-side elements 77 are separated from each other. As a result, as shown in Figure 3, the following occurs: Fig. Figure 2 shows the operating state of the bridging clutch 75 shifted to the disengaged state.

[0055] Among the input-side elements 78, the input-side element 78 that is furthest from the turbine 73 in the axial direction AD is designated as "input-side element 78B". If the operating state of the bypass coupling 75 differs from that described in Fig. 2 shown disengaged state in the in Fig. When the coupling clutch 75 transitions from the disengaged state shown in Figure 3, the sliding amount of the input-side element 78B is greater than the sliding amount of the other input-side elements 78 and the output-side elements 77. In the present embodiment, the sliding amount of the input-side element 78B when the operating state of the coupling clutch 75 is switched from the disengaged state to the engaged state is referred to as the "packing clearance PCtc of the coupling clutch 75".

[0056] As in Fig. As shown in Figure 1, the transmission mechanism 80 includes a first clutch C1, a second clutch C2, a brake mechanism B1, and a one-way clutch F1. The gear stage of the transmission 40 is changed according to a combination of the engaged and disengaged states in the first clutch C1, the second clutch C2, and the brake mechanism B1, and a combination of a restricted state and a released state of the one-way clutch F1.

[0057] The vehicle VC includes an oil supply unit 50, which supplies the transmission 40 with oil. The oil supply unit 50 includes an oil pan 51, in which the oil is stored, and a mechanically driven oil pump 52. The oil pump 52 includes a driven shaft 52a, which is coupled to the crankshaft 11 of the internal combustion engine 10. The oil pump 52 draws oil from the oil pan 51 and delivers the oil to the transmission 40. The pressure of the oil delivered by the oil pump 52 is regulated by a hydraulic pressure control circuit 41 of the transmission 40. The hydraulic pressure control circuit 41 includes solenoid valves 41a. The hydraulic pressure control circuit 41 energizes each solenoid valve 41a to control the oil flow and pressure.

[0058] The internal combustion engine 10 is controlled by a control unit 90, which operates various operating units of the internal combustion engine 10 to control the torque, an exhaust component, and other control aspects of the internal combustion engine 10. The control unit 90 also controls the transmission 40 and actuates the solenoid valves 41a of the hydraulic pressure control circuit 41.

[0059] In controlling the control aspects described above, the control unit 90 refers to an output signal Scr from a crank angle sensor 101 and an output signal Sin from an input shaft rotation angle sensor 102, which detects a rotation angle of the input shaft 81 of the transmission 40. The control unit 90 also refers to an oil temperature detection value Toil, which is an oil temperature detected by an oil temperature sensor 103, a vehicle speed SPD, which is a vehicle speed VC detected by a vehicle speed sensor 104, and a vehicle acceleration G, which is an acceleration of the vehicle VC detected by an acceleration sensor 105.

[0060] The control unit 90 comprises a central processing unit (CPU) 91, a read-only memory (ROM) 92, a storage device 93, which is an electrically rewritable non-volatile memory, and a peripheral circuit 94, configured to communicate with each other via a local network 95. The peripheral circuit 94 includes, for example, a circuit that generates a clock signal that controls internal operation, a power supply circuit, and a reset circuit. The control unit 90 controls the control aspects by having the CPU 91 execute programs stored in the ROM 92.

[0061] Storage device 93 stores map data entries DM1, DM2, and DM3. Each of the map data entries DM1, DM2, and DM3 contains data that specifies a map which outputs a variable corresponding to one of the various input variables (described below) when the input variable is applied. The data is learned through machine learning.

[0062] During operation of the transmission 40, an anomaly may occur in the lock-up clutch 75 of the transmission 40. Possible anomalies of the lock-up clutch 75 include the following: cold welding of the lock-up clutch 75; faulty engagement of the lock-up clutch 75; and seizing of the lock-up clutch 75.

[0063] When the lock-up clutch 75 is engaged, the friction elements 77a of the output elements 77 are pressed against the input elements 78. Since both the input and output elements 77 are made of metal, the friction elements 77a are pressed against the metal (input elements 78). As the friction elements 77a wear, the surfaces of the output elements 77 are exposed. In this state, when the lock-up clutch 75 is engaged, the output elements 77 directly contact the input elements 78. That is, the friction elements are not pressed against the metal (input elements 78), but rather the metal (output elements 77) is pressed against the metal (input elements 78). This can lead to cold welding in the lock-up clutch 75.The amount of heat generated by the bridging clutch 75 differs between pressing metal onto metal and pressing the friction elements onto metal.

[0064] The hydraulic pressure of pressure adjustment range 70a in the torque converter 70 is set when the operating state of the lock-up clutch 75 changes between the engaged and disengaged states. If, for example, an anomaly occurs in the hydraulic pressure control circuit 41, the hydraulic pressure of pressure adjustment range 70a cannot be set correctly. This can lead to the operating state of the lock-up clutch 75 not being properly controlled. More precisely, when switching the operating state of the lock-up clutch 75 to the engaged state, if the hydraulic pressure of pressure adjustment range 70a cannot be increased sufficiently, faulty engagement can occur, resulting in an insufficient force pressing the output elements 77 against the input elements 78.The faulty engagement causes the output elements 77 to slide on the input elements 78, reducing the efficiency of the torque transmission of the lock-up clutch 75. Increasing the engagement force restricts the sliding of the output elements 77 on the input elements 78. When the output elements 77 slide on the input elements 78, the lock-up clutch 75 generates a greater amount of heat than when the output elements 77 do not slide on the input elements 78.

[0065] If, during the switching of the operating state of the bypass clutch 75 from the engaged state to the disengaged state, the hydraulic pressure of the pressure adjustment range 70a cannot be reduced due to an anomaly in the hydraulic pressure control circuit 41, the bypass clutch 75 may remain in the engaged state. A condition in which the bypass clutch 75 remains in the engaged state, regardless of any attempt to change its operating state to the disengaged state, is referred to as the bypass clutch 75 seizing. When the bypass clutch 75 is seized, it generates a greater amount of heat than when the bypass clutch 75 is switched normally to the disengaged state.

[0066] If the lock-up clutch 75 exhibits an anomaly such as those described above, the oil temperature, i.e., the temperature of the oil circulating in the transmission 40, changes differently than if the lock-up clutch 75 does not exhibit an anomaly. In the present embodiment, the control unit 90 determines whether the lock-up clutch 75 exhibits an anomaly based on changes in the oil temperature sensing value Toil. The control unit 90 uses the map data entries DM1, DM2, and DM3 stored in the memory device 93 for this purpose.

[0067] In the present embodiment, the characteristic map entry DM1 specifies a characteristic map that outputs a variable Y(1) which determines whether cold welding has occurred in the lock-up clutch 75. The characteristic map entry DM2 specifies a characteristic map that outputs a variable Y(2) which determines whether the lock-up clutch 75 is in faulty engagement. The characteristic map entry DM3 specifies a characteristic map that outputs a variable Y(3) which determines whether the lock-up clutch 75 has seized.

[0068] The sequence of a series of processes executed by the control unit 90 to determine whether the bypass clutch 75 has an anomaly is now described with reference to the Fig. 4 and Fig. 5 described. The ones in the Fig. 4 and Fig. The sequence of processes shown in Figure 5 is implemented by the CPU 91 when executing the programs stored in ROM 92. The sequence of processes is executed repeatedly in a predetermined cycle. More precisely, the CPU 91 restarts the execution of the sequence of processes when the time elapsed since the temporary end of the sequence reaches the time corresponding to the predetermined cycle.

[0069] In step S11, the CPU 91 sets a coefficient z to one. Next, in step S13, the CPU 91 receives the current oil temperature acquisition value Toil as an oil temperature acquisition value Toil(z). Then, in step S15, the CPU 91 increments the coefficient z by one. In step S17, the CPU 91 determines whether the coefficient z is greater than a coefficient determination value zTh. In the present embodiment, time-series data of oil temperature acquisition values ​​Toil are used to determine whether the bypass coupling 75 has an anomaly. The time-series data of the oil temperature acquisition values ​​Toil include oil temperature acquisition values ​​Toil that follow each other in chronological order. The coefficient determination value zTh is set as a determination reference that determines whether the required number of oil temperature acquisition values ​​Toil for the determination has been obtained.If the coefficient z is less than or equal to the coefficient determination value zTh (S17: NO), CPU 91 proceeds to step S13. This means that the oil temperature measurement value Toil continues to be determined. If the coefficient z is greater than the coefficient determination value zTh (S17: YES), time series data of oil temperature measurement values ​​Toil, including "z", have been obtained, and CPU 91 proceeds to step S19.

[0070] In step S19, the CPU 91 normalizes the time-series data of the oil temperature acquisition values ​​Toil. For example, among the oil temperature acquisition values ​​Toil(1), Toil(2), ..., and Toil(z) contained in the time-series data of the oil temperature acquisition values ​​Toil, the CPU 91 sets the largest value as a reference oil temperature acquisition value ToilB. The CPU 91 divides each of the oil temperature acquisition values ​​Toil(1), Toil(2), ..., and Toil(z) by the reference oil temperature acquisition value ToilB to normalize the oil temperature acquisition values ​​Toil(1), Toil(2), ..., and Toil(z). The normalized oil temperature acquisition values ​​Toil(1), Toil(2), ..., and Toil(z) are referred to as normalized oil temperature acquisition values ​​ToilN(1), ToilN(2), ..., and ToilN(z). The normalized oil temperature measurement value ToilN(1) is determined, for example, by dividing the oil temperature measurement value Toil(1) by the reference oil temperature measurement value ToilB.Data containing the normalized oil temperature measurement values ​​ToilN(1), ToilN(2), ..., and ToilN(z) can also be referred to as "time series data of the normalized oil temperature measurement values ​​ToilN".

[0071] In step S21, the CPU 91 generates oil temperature relationship data RDToil based on the time series data of the normalized oil temperature measurement values ​​ToilN. In the present embodiment, each of the normalized oil temperature measurement values ​​ToilN(1), ToilN(2), ..., and ToilN(z) is greater than zero and less than or equal to one. A range of numerical values ​​from zero to one is divided into subranges. For example, the range of numerical values ​​from zero to one is divided into intervals of 0.2. For each subdivided range, the CPU 91 counts the number of normalized oil temperature measurement values ​​ToilN contained within the subdivided range. For example, if the normalized oil temperature measurement values ​​ToilN(1), ToilN(2), ...If ToilN(z) contains four normalized oil temperature acquisition values ​​ToilN that are greater than 0.4 and less than or equal to 0.6, the CPU 91 determines that the number of normalized oil temperature acquisition values ​​ToilN contained within a partitioned range of 0.4 to 0.6 is four. The CPU 91 calculates the count result for each partitioned range, obtained as described above, as the oil temperature relationship data RDToil. More precisely, the oil temperature relationship data RDToil shows the distribution of the numerical magnitude of the normalized oil temperature acquisition values ​​ToilN(1), ToilN(2), ..., and ToilN(z) contained within the time series data of the normalized oil temperature acquisition values ​​ToilN.

[0072] For example, the CPU 91 uses the number of normalized oil temperature (ToilN) values ​​in the range of 0 to 0.2 as count value Cnt(1) and the number of normalized oil temperature (ToilN) values ​​in the range of 0.2 to 0.4 as count value Cnt(2). Furthermore, the CPU 91 uses the number of normalized oil temperature (ToilN) values ​​in the range of 0.4 to 0.6 as count value Cnt(3) and the number of normalized oil temperature (ToilN) values ​​in the range of 0.6 to 0.8 as count value Cnt(4). The CPU 91 uses the number of normalized oil temperature (ToilN) values ​​in the range of 0.8 to 1 as count value Cnt(5). This means that the oil temperature relationship data RDToil includes the count values ​​Cnt(1), Cnt(2), Cnt(3), Cnt(4) and Cnt(5).

[0073] In Fig. Figure 6 shows the solid line and the dashed line respectively, time series data of oil temperature recording values ​​Toil. Fig. 7 is a histogram of the oil temperature relationship data RDToil, based on the time series data of the oil temperature recording values ​​Toil, represented by the dashed line in Fig. The 6 are shown, and it is generated. Fig. Figure 8 is a histogram of the oil temperature relationship data RDToil, generated based on the time series data of the oil temperature acquisition values ​​Toil, which are represented by the solid line in Fig. Figure 6 shows the time series data of the oil temperature measurement values ​​Toil, which are represented by the dashed line in Fig. Figure 6 shows that the oil temperature detection value Toil increases slowly at an essentially constant rate. Therefore, the values ​​shown in Fig. The oil temperature relationship data shown in Figure 7 (RDToil) exhibit small variations in the count values ​​Cnt(1) to Cnt(5). The time series data of the oil temperature measurement values ​​(Toil), represented by the solid line in Figure 7, show small variations in the count values ​​Cnt(1) to Cnt(5). Fig. Figure 6 shows that the rate of increase of the oil temperature measurement value Toil changes at an intermediate point. Therefore, the figures in Fig. The oil temperature relationship data shown in the 8 RDToil data show large variations in the count values ​​Cnt(1) to Cnt(5).

[0074] Going back to the Fig. 4 and Fig. In step S23, the CPU 91 receives the vehicle speed SPD, the engine speed NE, the input shaft speed Nat, a speed difference ΔNtc, a calculated heat generation quantity CVtc of the lock-up clutch 75, the engagement force EFtc of the lock-up clutch 75, the packing clearance PCtc of the lock-up clutch 75, and the vehicle acceleration G. The engine speed NE is the speed of the crankshaft 11, calculated based on the output signal Scr of the crank angle sensor 101, and is simultaneously the speed of the input-side elements 78 of the lock-up clutch 75. The input shaft speed Nat is the speed of the input shaft 81 of the transmission 80, calculated from the output signal Sin of the input shaft rotation angle sensor 102, and is simultaneously the speed of the output-side elements 77 of the lock-up clutch 75.The speed difference ΔNtc is a difference between the engine speed NE and the input shaft speed Nat, and is also a speed difference between the input elements 78 and the output elements 77 of the lock-up clutch 75. Furthermore, the speed difference ΔNtc can also refer to a speed difference between the front cover 71, which introduces the torque into the input elements 78, and the input shaft 81, which receives the torque delivered by the output elements 77. When the lock-up clutch 75 is engaged, the calculated heat generation quantity CVtc of the lock-up clutch 75 is calculated based on the product of the speed difference ΔNtc and an input torque of the lock-up clutch 75. The input torque of the lock-up clutch 75 is the torque supplied to the torque converter 70 by the internal combustion engine 10.When the lock-up clutch 75 is in the disengaged state, the calculated heat generation quantity value CVtc is zero. The engagement force EFtc of the lock-up clutch 75 is the force that presses the output elements 77 against the input elements 78 and can be derived from the hydraulic pressure of the pressure adjustment range 70a. The packing clearance PCtc is a value that is measured during a clearance check of the gearbox 40 and pre-stored in the storage device 93.

[0075] In step S25, the CPU 91 assigns the oil temperature relationship data RDToil generated in step S21 and the data obtained in step S23 to the input variables x(1) to x(13) of the characteristic maps, which determine whether the bypass clutch 75 exhibits an anomaly. More precisely, the CPU 91 assigns the count value Cnt(1) of the oil temperature relationship data RDToil to the input variable x(1), assigns the count value Cnt(2) to the input variable x(2), and assigns the count value Cnt(3) to the input variable x(3). The CPU 91 assigns the count value Cnt(4) to the input variable x(4) and assigns the count value Cnt(5) to the input variable x(5). The CPU 91 assigns the vehicle speed SPD to the input variable x(6), the engine speed NE to the input variable x(7), and the input shaft speed Nat to the input variable x(8).CPU 91 assigns the rotational speed difference ΔNtc to input variable x(9), the calculated heat generation quantity CVtc to input variable x(10), and the engagement force EFtc to input variable x(11). CPU 91 assigns the packing clearance PCtc to input variable x(12) and the vehicle acceleration G to input variable x(13).

[0076] In step S27, the CPU 91 sets a determination coefficient MP to one. In step S29, the CPU 91 selects a map data entry corresponding to the determination coefficient MP from the map data entries DM1, DM2, and DM3 stored in the memory device 93. For example, the CPU 91 selects the map data entry DM1 if the determination coefficient MP is one, selects the map data entry DM2 if the determination coefficient MP is two, and selects the map data entry DM3 if the determination coefficient MP is three.

[0077] In step S31, the CPU 91 inputs the input variables x(1) to x(13) into a map specified by the selected map data entry in order to calculate an output variable Y(MP).

[0078] In the present embodiment, the characteristic map comprises a fully linked feedforward neural network with a single intermediate layer. The neural network includes an activation function h(x), which is used as an input nonlinear characteristic map. This input nonlinear characteristic map performs a nonlinear transformation on an input coefficient wFjk (j=0 to n, k=0 to 13) and outputs a linear characteristic map specified by the input coefficient wFjk. In the present embodiment, a hyperbolic tangent tanh(x) is used as an example of the activation function h(x). The neural network also includes an activation function f(x), which is used as an output nonlinear characteristic map.The output nonlinear characteristic map performs a nonlinear conversion on an output coefficient wSj (j = 0 to n) and an output linear characteristic map, which is a linear characteristic map specified by the output coefficient wSj. In the present embodiment, a hyperbolic tangent tanh(x) is used as the activation function f(x). A value n specifies the dimension of an intermediate layer. In the present embodiment, the value n is less than thirteen, i.e., the dimension of the input variable x. The input coefficient wFj0 ​​is a bias parameter and is the coefficient of an input variable x(0). The input variable x(0) is defined as one. The output coefficient wS0 is a bias parameter.

[0079] The map data entry DM1 is a learned model that is trained using a vehicle with the same specifications as vehicle VC before being loaded onto vehicle VC. To learn the map data entry DM1, training data, including monitored data and input data, is obtained in advance. Specifically, the time-series data of the oil temperature detection values ​​Toil are obtained while the vehicle is actually driving. When the time-series data of the oil temperature detection values ​​Toil undergo the same processes as in steps S19 and S21, the oil temperature relationship data RDToil is obtained as input data. Additionally, at this point, the vehicle speed SPD, engine speed NE, input shaft speed Nat, speed difference ΔNtc, the calculated heat generation quantity value CVtc of the lock-up clutch 75, the engagement force EFtc of the lock-up clutch 75, and the vehicle acceleration G are obtained as input data.Furthermore, information on the occurrence of cold welding, i.e., information indicating whether cold welding has occurred in the lock-up clutch 75, is obtained as monitored data. For example, the information on the occurrence of cold welding can be zero if cold welding has occurred, and one if no cold welding has occurred. The packing clearance PCtc of the lock-up clutch 75 loaded on the vehicle is also obtained as input data before the vehicle starts moving.

[0080] Training data entries are generated during vehicle operation under various conditions. For example, a lock-up clutch, which is susceptible to cold welding when engaged, is installed on a vehicle, and the vehicle is driven. If no cold welding occurs during the drive, various types of input data entries are collected, and information indicating that no cold welding occurred is also collected as monitored data. If cold welding occurs during the drive, various types of input data entries are collected, and information indicating that cold welding occurred is also collected.

[0081] Such training data entries are used to learn the characteristic map data entry DM1. More precisely, an input variable and an output variable are adjusted so that the difference between the output variable, which is output by a characteristic map into which input data is fed, and the actual information about the occurrence of cold welding converges to a value that is less than or equal to a predefined value.

[0082] Furthermore, the map data entry DM2 is a learned model, which is learned using a vehicle with the same specifications as vehicle VC before it is loaded onto vehicle VC. To learn the map data entry DM2, training data, including monitored data and input data, is obtained in advance. More precisely, as described above, various types of input data entries are obtained while the vehicle is actually driving. At this point, information about the occurrence of faulty engagement, i.e., information indicating whether the lock-up clutch 75 is faultily engaged, is obtained as monitored data. For example, the information about the occurrence of faulty engagement can be zero if the lock-up clutch 75 is faultily engaged, and it can be one if the lock-up clutch 75 is not faultily engaged.

[0083] Training data entries, containing monitored data and input data, are generated when the vehicle is driven in various situations. For example, a lock-up clutch, which is capable of faulty engagement, is installed on a vehicle, and the vehicle is driven. If the lock-up clutch is engaged while the vehicle is driving, and no faulty engagement occurs, various types of input data are received indicating that no faulty engagement occurred, and information about the occurrence of a faulty engagement, indicating that no faulty engagement occurred, is also received as monitored data.When the lock-up clutch is engaged while the vehicle is in motion, if a faulty intervention has occurred, various types of input data are received regarding when a faulty intervention occurred, and information about the occurrence of a faulty intervention, indicating that a faulty intervention has occurred, is also received as monitored data.

[0084] Such training data entries are used to learn the characteristic map data entry DM2. More precisely, an input-side variable and an output-side variable are adjusted so that the difference between the output variable, which is output by a characteristic map into which input data is fed, and the actual information about the occurrence of a faulty intervention converges to a value that is less than or equal to a predetermined value.

[0085] Furthermore, the map data entry DM3 is a learned model that is trained using a vehicle with the same specifications as vehicle VC before it is loaded onto vehicle VC. To learn the map data entry DM3, training data, including monitored data and input data, is obtained in advance. More precisely, various types of input data entries are obtained while the vehicle is actually driving. At this point, information about the occurrence of seizing, i.e., information indicating whether the lock-up clutch 75 is stuck, is obtained as monitored data. For example, the information about the occurrence of seizing can be zero if a seizing has occurred and can be one if no seizing has occurred.

[0086] Training data entries, containing monitored data and input data, are generated during vehicle operation under various conditions. For example, a lock-up clutch, which is prone to seizing, is installed on a vehicle, and the vehicle is being driven. If the lock-up clutch is disengaged during the vehicle's operation and no seizure occurs, various types of input data are collected for the event that no seizure occurred, and information indicating that no seizure occurred is collected as monitored data.If the lock-up clutch is disengaged while the vehicle is in motion and a seizure has occurred, various types of input data are received regarding when a seizure occurred, and information about the occurrence of seizures, indicating that a seizure has occurred, is received as monitored data.

[0087] Such training data entries are used to learn the characteristic map data entry DM3. More precisely, an input-side variable and an output-side variable are set such that the difference between the output variable, which is output by a characteristic map into which input data is fed, and the actual information about the occurrence of locking converges to a value that is less than or equal to a predetermined value.

[0088] After the output variable Y(MP) is calculated in step S31, the CPU 91 proceeds to step S33 and evaluates the output variable Y(MP) calculated in step S31. More precisely, the CPU 91 determines, based on the output variable Y(MP), whether the bypass coupling 75 exhibits an anomaly. For example, if the determination coefficient MP is equal to one and the output variable Y(1) is less than or equal to an anomaly determination value, the CPU 91 determines that cold welding has occurred in the bypass coupling 75. If the output variable Y(1) is greater than the anomaly determination value, the CPU 91 determines that no cold welding has occurred. For example, if the determination coefficient MP is equal to two and the output variable Y(2) is less than or equal to the anomaly determination value, the CPU 91 determines that the bypass coupling 75 is improperly engaged.If the output variable Y(2) is greater than the anomaly determination value, CPU 91 determines that no faulty intervention has occurred. For example, if the determination coefficient MP is equal to 3 and the output variable Y(3) is less than or equal to the anomaly determination value, CPU 91 determines that the bypass coupling 75 is stuck. If the output variable Y(3) is greater than the anomaly determination value, CPU 91 determines that no sticking has occurred. If the output variable Y is less than or equal to the anomaly determination value, it can be determined, as described above, that an anomaly in the bypass coupling 75 has occurred with a high probability. Thus, it is determined that an anomaly has occurred.

[0089] If, in step S35, the evaluation result determines that an anomaly has occurred in the bypass coupling 75 (YES), the CPU 91 proceeds to step S37. In step S37, the CPU 91 stores information indicating that the anomaly has occurred in memory device 93. For example, if the output variable Y(1) is less than or equal to the anomaly determination value, the CPU 91 stores information indicating that cold welding has occurred in memory device 93. For example, if the output variable Y(2) is less than or equal to the anomaly determination value, the CPU 91 stores information indicating a faulty engagement in memory device 93. For example, if the output variable Y(3) is less than or equal to the anomaly determination value, the CPU 91 stores information indicating that a seizing has occurred in memory device 93.Then the CPU 91 continues with step S39.

[0090] In step S35, the CPU 91 proceeds to step S39 if, based on the evaluation result, it is determined that no anomaly has occurred in the bridging coupling 75 (NO).

[0091] In step S39, CPU 91 determines whether the determination coefficient MP is greater than or equal to three. If the determination coefficient MP is greater than or equal to three, all three anomaly determinations relating to the bypass coupling 75 have been completed. If the determination coefficient MP is less than three, one or more of the three anomaly determinations relating to the bypass coupling 75 have not yet been completed. Therefore, if the determination coefficient MP is less than three (S39: NO), CPU 91 proceeds to step S41. In step S41, CPU 91 increments the determination coefficient MP by one and then proceeds to step S29. In step S39, if the determination coefficient MP is greater than or equal to three (YES), CPU 91 temporarily terminates the sequence of processes.

[0092] The operation of the present embodiment will now be described.

[0093] In Fig. Figure 6 shows the time series data of the oil temperature sensing values ​​Toil, indicated by the dashed line, changes in the oil temperature sensing value Toil when the lock-up clutch 75 is engaged and exhibits no anomaly. Fig. Figure 6 shows the time series data of the oil temperature sensing values ​​Toil, represented by the solid line, changes in the oil temperature sensing value Toil when the lock-up clutch 75 is switched to the engaged state and cold welding has occurred. In the present embodiment, when time series data of oil temperature sensing values ​​Toil are obtained, oil temperature relationship data RDToil, as described in the Fig. 7 and Fig. Figure 8 shows the data generated based on the time series data of oil temperature acquisition values ​​(Toil). This means that the generated oil temperature relationship data (RDToil) differs depending on whether cold welding occurred during the acquisition of the time series data. More precisely, the numerical size of the count values ​​Cnt(1) to Cnt(5) in the oil temperature relationship data (RDToil) varies.

[0094] Furthermore, the variations of the count values ​​Cnt(1) to Cnt(5) differ between the oil temperature relationship data RDToil, which is based on time series data of oil temperature acquisition values ​​Toil obtained when the bridging clutch 75 is in the engaged state and a faulty engagement has occurred, and the oil temperature relationship data RDToil, which is based on time series data of oil temperature acquisition values ​​Toil obtained when no anomaly has occurred.

[0095] Furthermore, the count values ​​Cnt(1) to Cnt(5) differ between the oil temperature relationship data RDToil, which is based on time series data of oil temperature acquisition values ​​Toil obtained when the lock-up clutch 75 is stuck and switched to the disengaged state, and the oil temperature relationship data RDToil, which is based on time series data of oil temperature acquisition values ​​Toil obtained when no anomaly has occurred.

[0096] When the count values ​​Cnt(1) to Cnt(5) of the oil temperature relationship data RDToil are entered into the map specified by the map data entries DM1, DM2, and DM3, the maps output the output variable Y(MP) corresponding to the oil temperature relationship data RDToil. Subsequently, the output variable Y(MP) is used to determine whether an anomaly has occurred.

[0097] The present embodiment has the following advantages. (1-1) In the present embodiment, the storage device 93 stores the characteristic map data entries. Each characteristic map data entry specifies a characteristic map that receives the oil temperature relationship data RDToil as an input variable and outputs a variable Y that determines whether the lock-up clutch 75 has an anomaly. During operation of the transmission 40, the output variable Y, which is output by the characteristic map as a result of the input of a received input variable into the characteristic map, is used to determine whether the lock-up clutch 75 has an anomaly. In this way, it is possible to determine whether the lock-up clutch 75 has an anomaly without using a detection value from an odor sensor.

[0098] Even if a vehicle does not have an odor sensor like the vehicle VC, it will be determined whether the vehicle's jump-start clutch 75 has an anomaly.

[0099] When anomaly detection is performed using a detection value from an odor sensor, the occurrence of an anomaly can only be detected when the components of the odor emitted by the oil change as a result of oil deterioration. In the present embodiment, the occurrence of an anomaly is detected when the oil temperature detection value, Toil, has changed differently. Thus, the anomaly is detected quickly.

[0100] (1-2) In the present embodiment, the time series data of the oil temperature sensing values ​​Toil are normalized to obtain time series data of the normalized oil temperature sensing values ​​ToilN, which contain several normalized oil temperature sensing values ​​ToilN. Then, the oil temperature relationship data RDToil are generated based on the time series data of the normalized oil temperature sensing values ​​ToilN. Therefore, the degree of difference in the oil temperature relationship data RDToil between the time at which the lock-up clutch 75 exhibits an anomaly and the time at which the lock-up clutch 75 does not exhibit anomalies varies slightly between a case where the oil temperature sensing value Toil is relatively large and a case where the oil temperature sensing value Toil is relatively small.Thus, using the oil temperature relationship data RDToil as an input variable of the characteristic map reduces variations in the determination accuracy caused by the size of the oil temperature measurement values ​​Toil.

[0101] (1-3) The accuracy of the determination increases with an increasing number of data entries in the time series data of the normalized oil temperature measurement values ​​ToilN. In the present embodiment, the oil temperature relationship data RDToil, which is an input variable of a characteristic map, is a histogram of time series data of normalized oil temperature measurement values ​​ToilN. Therefore, the data volume of the oil temperature relationship data RDToil does not increase significantly, even if the number of data entries in the time series data increases. The determination is performed with high accuracy even when a small data volume is used.

[0102] (1-4) The operating parameters of the transmission 40 at a relatively high vehicle speed SPD differ from those at a relatively low vehicle speed SPD. When the operating parameters of the transmission 40 change, the temperature of the oil circulating in the transmission 40 also changes. In this context, the vehicle speed SPD is used as an input variable of a characteristic map in the present embodiment. That is, the characteristic map outputs a variable Y taking the vehicle speed SPD into account. When such an output variable Y is used to determine whether the lock-up clutch 75 has an anomaly, the accuracy of the determination is increased.

[0103] (1-5) If the lock-up clutch 75 exhibits no anomaly and is in the engaged state, changes in oil temperature can be estimated to a certain extent. Changes in oil temperature when the lock-up clutch 75 is in the disengaged state, changes in oil temperature when the operating state is switched from the disengaged state to the engaged state, and changes in oil temperature when the operating state is switched from the engaged state to the disengaged state can also be estimated to a certain extent.

[0104] The vehicle speed SPD determines the point in time at which the lock-up clutch 75 switches from the disengaged state to the engaged state, and the point in time at which the lock-up clutch 75 switches from the engaged state to the disengaged state. That is, the operating state of the lock-up clutch 75 is controlled as a function of the vehicle speed SPD. Therefore, whether the lock-up clutch 75 exhibits anomalies is determined by comparing changes in oil temperature, estimated from the operating state of the lock-up clutch 75 as determined by the vehicle speed SPD, with changes in the oil temperature sensing value Toil.

[0105] As in the present embodiment, the accuracy of determining whether the bridging clutch 75 has an anomaly is increased when the vehicle speed SPD is added to the input variable of a characteristic map.

[0106] (1-6) For example, when the bypass clutch 75 is in the engaged state, the amount of heat generated by the bypass clutch 75 can vary according to the rotational speed of the input elements 78, the rotational speed of the output elements 77, and the speed difference between the input elements 78 and the output elements 77. In the present embodiment, the motor speed NE, which corresponds to the rotational speed of the input elements 78, the input shaft speed Nat, which corresponds to the rotational speed of the output elements 77, and the speed difference ΔNtc are used as input variables of a characteristic map. That is, the characteristic map outputs a variable Y taking into account the motor speed NE, the input shaft speed Nat, and the speed difference ΔNtc.

[0107] Using output variables Y to determine whether the bridging clutch 75 has an anomaly increases the accuracy of the determination.

[0108] (1-7) The calculated heat generation quantity value CVtc of the bypass coupling 75 is calculated under the assumption that the bypass coupling 75 has no anomaly. Therefore, the relationship between the calculated heat generation quantity value CVtc and the changes in the oil temperature sensing value Toil may differ when the bypass coupling 75 has an anomaly and when the bypass coupling 75 does not. In this respect, in the present embodiment, the calculated heat generation quantity value CVtc is used as an input variable of a characteristic map. That is, the characteristic map outputs a variable Y taking into account the calculated heat generation quantity value CVtc. Using such an output variable Y to determine whether an anomaly exists increases the accuracy of the determination.

[0109] (1-8) When the bypass clutch 75 is in the engaged state, the amount of heat generated by the bypass clutch 75 can differ between a case where the engagement force EFtc is relatively large and a case where the engagement force EFtc is relatively small. For example, if the engagement force EFtc is relatively small and the bypass clutch 75 is in sliding engagement, the bypass clutch 75 generates a greater amount of heat than if the engagement force EFtc is relatively large and the bypass clutch 75 is in full engagement. In this respect, the engagement force EFtc is used as an input variable of a characteristic map in the present embodiment. That is, the characteristic map outputs a variable Y taking the engagement force EFtc into account. Using such an output variable Y to determine whether the bypass clutch 75 has an anomaly increases the accuracy of the determination.

[0110] (1-9) If the packing clearance PCtc of the bypass clutch 75 is increased in magnitude, the engagement force that pushes the bypass clutch 75 into the engaged state tends to increase. That is, when the bypass clutch 75 is in the engaged state, the amount of heat generated by the bypass clutch 75 can vary in accordance with the size of the packing clearance PCtc. In this context, in the present embodiment, the packing clearance PCtc is used as an input variable of a characteristic map. That is, the characteristic map outputs a variable Y taking the packing clearance PCtc into account. If such an output variable Y is used to determine whether the bypass clutch 75 has an anomaly, the accuracy of the determination is increased.

[0111] (1-10) When the lock-up clutch 75 is switched from the disengaged state to the engaged state, or vice versa, changes in the operating state of the lock-up clutch 75 generate vibrations in the transmission 40. These vibrations can be detected by the acceleration sensor 105. The vehicle acceleration G, which is a detection value of the acceleration sensor 105, can differ between a case in which the lock-up clutch 75 exhibits an anomaly and a case in which the lock-up clutch 75 does not exhibit an anomaly. In the present embodiment, the vehicle acceleration G is used as an input variable of a characteristic map. That is, the characteristic map outputs a variable Y taking the vehicle acceleration G into account.Using such an output variable Y to determine whether an anomaly exists increases the accuracy of the determination.

[0112] (1-11) The amount of heat generated by the lock-up clutch 75 differs between a case where cold welding has occurred in the lock-up clutch 75 and a case where cold welding has not occurred in the lock-up clutch 75. That is, the oil temperature sensing value Toil changes differently. Therefore, when an input variable is fed into a map specified by the map data entry DM1, the output variable Y provided by the map is used to determine whether cold welding has occurred.

[0113] (1-12) When the lock-up clutch 75 is switched to the engaged state, the amount of heat generated by the lock-up clutch 75 differs between a case where the lock-up clutch 75 is in faulty engagement and a case where the lock-up clutch 75 is not in faulty engagement. That is, the oil temperature sensing value Toil changes differently. Therefore, when an input variable is fed into a map specified by the map data entry DM2, the output variable Y provided by the map is used to determine whether faulty engagement is present.

[0114] (1-13) If the lock-up clutch 75 is stuck, it cannot be switched to the disengaged state. The amount of heat generated by the lock-up clutch 75 differs between the case where the lock-up clutch 75 has seized and the case where it has not seized. That is, the oil temperature sensing value Toil changes differently. Therefore, when an input variable is fed into a map specified by the map data entry DM3, the output variable Y provided by the map is used to determine whether a seizure has occurred.

[0115] (1-14) In the present embodiment, the map data entries DM1, DM2, and DM3 are prepared for each type of anomaly of the lock-up clutch 75. The map data entry DM1 is used to determine whether cold welding has occurred. The map data entry DM2 is used to determine whether there is a faulty engagement. The map data entry DM3 is used to determine whether a seizing has occurred. Thus, the accuracy of each determination is increased by using the map data entries DM1, DM2, and DM3 separately according to the type of determination. Second embodiment

[0116] A second embodiment is described below with reference to the drawings. The main focus is on the differences compared to the first embodiment.

[0117] As in Fig. As shown in Figure 9, the storage device 93 in the present embodiment stores map data entries DM11, DM12, DM13, and DM14, each corresponding separately to a specific operating state of the lock-up clutch 75. Map data entry DM11 is learned by machine learning specifically for when the operating state of the lock-up clutch 75 is a disengaged state. Map data entry DM12 is learned by machine learning specifically for when the operating state of the lock-up clutch 75 is a transitional-engagement state. Map data entry DM13 is learned by machine learning specifically for when the operating state of the lock-up clutch 75 is an engaged state. Map data entry DM14 is learned by machine learning specifically for when the operating state of the lock-up clutch 75 is a transitional-disengagement state.The transitional engagement state refers to an operating state of the bypass clutch 75 when switching from the disengaged state to the engaged state. The transitional disengagement state refers to an operating state of the bypass clutch 75 when switching from the engaged state to the disengaged state.

[0118] The sequence of a series of processes executed by the control unit 90 to determine whether the bypass clutch 75 has an anomaly is now described with reference to the Fig. 10 and Fig. 11 described. The ones in the Fig. 10 and Fig. The sequence of processes shown in Figure 11 is implemented by the CPU 91 when executing the programs stored in ROM 92. This sequence of processes is executed repeatedly in a predetermined cycle. More precisely, the CPU 91 restarts the execution of the sequence of processes when the time elapsed since the temporary end of the sequence reaches the time corresponding to the predetermined cycle.

[0119] In step S51, the CPU 91 sets a coefficient z to one. Next, in step S53, the CPU 91 receives the current oil temperature measurement value Toil as oil temperature measurement value Toil(z). Then, in step S55, the CPU 91 increments the coefficient z by one. In step S57, the CPU 91 determines whether the coefficient z is greater than a coefficient determination value zTh. If the coefficient z is less than or equal to the coefficient determination value zTh (S57: NO), the CPU 91 proceeds to step S53. If the coefficient z is greater than the coefficient determination value zTh (S57: YES), the CPU 91 proceeds to step S59.

[0120] In step S59, the CPU 91 receives time-series data of normalized oil temperature acquisition values ​​ToilN, including several normalized oil temperature acquisition values ​​ToilN(1), ToilN(2), ..., and ToilN(z), in the same manner as in step S19. In step S61, the CPU 91 generates oil temperature relationship data RDToil based on the time-series data of the normalized oil temperature acquisition values ​​ToilN, in the same manner as in step S21. In step S63, the CPU 91 receives the vehicle speed SPD, the engine speed NE, the input shaft speed Nat, a speed difference ΔNtc, a calculated heat generation quantity value CVtc of the lock-up clutch 75, the engagement force EFtc of the lock-up clutch 75, the packing clearance PCtc of the lock-up clutch 75, and the vehicle acceleration G, in the same manner as in step S23.In step S65, the CPU 91 assigns the oil temperature relationship data RDToil calculated in step S61 and the data obtained in step S63 to the input variables x(1) to x(13) of the characteristic maps that determine whether the bridging clutch 75 has an anomaly, in the same way as in step S25.

[0121] In step S67, the CPU 91 receives the operating state of the bypass clutch 75. More precisely, the CPU 91 selects the current operating state of the bypass clutch 75 from the disengaged state, the transition-to-engagement state, the engaged state, and the transition-to-disengagement state. "The current operating state of the bypass clutch 75" refers to an operating state that the CPU 91 achieves based on the control system. Therefore, if the bypass clutch 75 has an anomaly, the received operating state may differ from the actual operating state.

[0122] In step S69, the CPU 91 sets a state coefficient SV to a value corresponding to the operating state obtained in step S67. For example, if the operating state is the disengaged state, the CPU 91 sets the state coefficient SV to one. If the operating state is the transition-engagement state, the CPU 91 sets the state coefficient SV to two. If the operating state is, for example, the engaged state, the CPU 91 sets the state coefficient SV to three. If the operating state is the transition-disengagement state, the CPU 91 sets the state coefficient SV to four.

[0123] In step S71, the CPU 91 selects a map data entry corresponding to the state coefficient SV from the map data entries DM11, DM12, DM13, and DM14, which are stored in the memory device 93. If the state coefficient SV is, for example, one, the CPU 91 selects the map data entry DM11. If the state coefficient SV is equal to two, the CPU 91 selects the map data entry DM12. If the state coefficient SV is equal to three, the CPU 91 selects the map data entry DM13. If the state coefficient SV is four, the CPU 91 selects the map data entry DM14.

[0124] In step S73, the CPU 91 inputs the input variables x(1) to x(13) into a map specified by the selected map data entry in order to calculate an output variable Y(SV).

[0125] The map data entry DM11 is a learned model that is trained on a vehicle with the same specifications as vehicle VC before being loaded onto vehicle VC. To learn the map data entry DM11, training data, including monitored data and input data, is obtained in advance. Specifically, various types of input data entries are obtained when the lock-up clutch is disengaged and the vehicle is being driven. Furthermore, information about the occurrence of anomalies, i.e., information indicating whether the lock-up clutch has an anomaly, is obtained as monitored data. In the present embodiment, the types of anomalies, i.e., cold welding, faulty engagement, and seizing, are not differentiated.For example, if an anomaly has occurred in the jump-start coupling, the information about the occurrence of an anomaly can be zero, and if no anomaly has occurred in the jump-start coupling, the information about the determination of the anomaly can be one.

[0126] Training data entries are generated during the vehicle's operation in various situations. For example, a jump-start clutch, which is prone to anomalies, is installed on a vehicle, and the vehicle is driven. If no anomaly occurs during the vehicle's operation, various types of input data entries are collected, and information indicating that no anomaly occurred is also collected as monitored data. If an anomaly occurs during the vehicle's operation, various types of input data entries are collected, and information indicating that the anomaly occurred is also collected as monitored data.

[0127] Such training data entries are used to learn the DM11 map data entry. More precisely, an input variable and an output variable are set so that the difference between an output variable, which is output by a map into which input data is fed, and the actual anomaly determination information converges to a value that is less than or equal to a predetermined value.

[0128] Furthermore, each of the map data entries DM12, DM13, and DM14 is a learned model, which is learned using a vehicle with the same specifications as vehicle VC before being loaded onto vehicle VC. To learn map data entry DM12, input data is obtained based on various types of data acquired while the vehicle is driving and the lock-up clutch is in the transition-engagement state, and the corresponding information about the occurrence of anomalies is obtained as monitored data. As described above, training data entries are generated. These training data entries are used to learn map data entry DM12.More precisely, an input variable and an output variable are set such that a difference between the output variable, which is output by a characteristic map into which input data is entered, and the actual anomaly determination information converges to a value that is less than or equal to a predetermined value.

[0129] To learn the DM13 map data entry, input data is obtained based on various types of data acquired when the vehicle is driving and the lock-up clutch is engaged. The corresponding information about the occurrence of anomalies is also collected as monitored data. As described above, training data entries are generated. These training data entries are used to learn the DM13 map data entry. More precisely, an input variable and an output variable are adjusted so that the difference between an output variable, provided by a map into which input data is fed, and the actual anomaly detection information converges to a predetermined value.

[0130] To learn the DM14 map data entry, input data is obtained based on various types of data acquired while the vehicle is in motion and the lock-up clutch is in the transition-disengagement state. The corresponding information about the occurrence of anomalies is also collected as monitored data. As described above, training data entries are generated. These training data entries are used to learn the DM14 map data entry. More specifically, an input variable and an output variable are adjusted so that the difference between the output variable, which is provided by a map into which input data is fed, and the actual anomaly detection information converges to a value less than or equal to a predetermined value.

[0131] In step S75, CPU 91 evaluates the output variable Y(SV) calculated in step S73. More precisely, CPU 91 determines whether the output variable Y(SV) is less than or equal to the anomaly determination value. If the output variable Y(SV) is less than or equal to the anomaly determination value, CPU 91 determines that an anomaly has occurred in the bypass coupling 75. If the output variable Y(SV) is greater than the anomaly determination value, CPU 91 determines that no anomaly has occurred in the bypass coupling 75. If, based on the evaluation result in step S77, it is determined that an anomaly has occurred in the bypass coupling 75 (YES), CPU 91 proceeds to step S79. In step S79, CPU 91 stores information indicating that the anomaly has occurred in memory device 93. CPU 91 then temporarily terminates the sequence of processes.

[0132] In step S77, if the evaluation result determines that no anomaly has occurred in the bypass coupling 75 (NO), CPU 91 temporarily terminates the sequence of processes. That is, if, for example, the output variable Y(SV) is greater than the anomaly detection value, CPU 91 does not execute step S79 and temporarily terminates the sequence of processes.

[0133] In addition to the advantages described above (1-1) to (1-10), the present embodiment has the following advantage.

[0134] (2-1) Even if the lock-up clutch 75 exhibits an anomaly, the changes in the oil temperature sensing value Toil may differ depending on the operating state of the lock-up clutch 75. In this respect, in the present embodiment, the map data entries DM11, DM12, DM13, and DM14, which correspond separately to each operating state of the lock-up clutch 75, are prepared in advance. From the map data entries DM11, DM12, DM13, and DM14, a map data entry corresponding to the current operating state is selected, and an input variable is fed into a map defined by the selected map data entry. Subsequently, the output variable Y(SV) output from the map is used to determine whether the lock-up clutch 75 exhibits anomaly. Thus, the accuracy of the determination is increased by using the map data entries separately according to the operating state. Third embodiment

[0135] A third embodiment is described below with reference to the drawings. The main focus is on the differences between this and the first and second embodiments.

[0136] As in Fig. As shown in Figure 12, the storage device 93 in the present embodiment stores map data entries DM21, DM22, DM23, etc., which correspond to a degree of deterioration of the properties of the transmission 40. The map data entry DM11 corresponds to the lowest degree of property deterioration. The map data entry DM12 corresponds to the second lowest degree of property deterioration. The map data entry DM13 corresponds to the third lowest degree of property deterioration.

[0137] The sequence of a series of processes executed by the control unit 90 to determine whether the bypass clutch 75 has an anomaly is now described with reference to the Fig. 13 and Fig. 14 described. The ones in the Fig. 13 and Fig. The sequence of processes shown in Figure 14 is implemented by the CPU 91 when executing the programs stored in ROM 92. This sequence of processes is executed repeatedly in a predetermined cycle. More precisely, the CPU 91 restarts the execution of the sequence of processes when the time elapsed since the temporary end of the sequence reaches the time corresponding to the predetermined cycle.

[0138] In step S91, CPU 91 sets a coefficient z to one. Next, in step S93, CPU 91 receives the current oil temperature measurement value Toil as oil temperature measurement value Toil(z). Then, in step S95, CPU 91 increments the coefficient z by one. In step S97, CPU 91 determines whether the coefficient z is greater than a coefficient determination value zTh. If the coefficient z is less than or equal to the coefficient determination value zTh (S97: NO), CPU 91 proceeds to step S93. If the coefficient z is greater than the coefficient determination value zTh (S97: YES), CPU 91 proceeds to step S99.

[0139] In step S99, the CPU 91 receives time-series data of normalized oil temperature acquisition values ​​ToilN, including several normalized oil temperature acquisition values ​​ToilN(1), ToilN(2), ... and ToilN(z), in the same way as in step S19. In step S101, the CPU 91 generates oil temperature relationship data RDToil based on the time-series data of the normalized oil temperature acquisition values ​​ToilN, in the same way as in step S21. In step S103, the CPU 91 receives the vehicle speed SPD, the engine speed NE, the input shaft speed Nat, a speed difference ΔNtc, a calculated heat generation quantity value CVtc of the lock-up clutch 75, the engagement force EFtc of the lock-up clutch 75, the packing clearance PCtc of the lock-up clutch 75, and the vehicle acceleration G, in the same way as in step S23.In step S105, the CPU 91 assigns the oil temperature relationship data RDToil calculated in step S101 and the data obtained in step S103 to the input variables x(1) to x(13) of the characteristic maps that determine whether the bridging clutch 75 has an anomaly, in the same way as in step S25.

[0140] In step S107, CPU 91 receives a deterioration coefficient AGI, which is a coefficient corresponding to the deterioration of the transmission 40's properties. For example, CPU 91 receives the deterioration coefficient AGI based on the distance traveled by vehicle VC. In this case, CPU 91 can set the deterioration coefficient AGI to a value that increases as the distance traveled by vehicle VC increases.

[0141] In step S109, the CPU 91 selects a map data entry corresponding to the deterioration coefficient AGI from the map data entries DM21, DM22, DM23, ... stored in the memory device 93. If the deterioration coefficient AGI is, for example, one, the CPU 91 selects the map data entry DM21. If the deterioration coefficient AGI is two, the CPU 91 selects the map data entry DM22.

[0142] In step S111, the CPU 91 inputs the input variables x(1) to x(13) into a map specified by the selected map data entry in order to calculate an output variable Y(AGI).

[0143] The map data entry DM21 is a learned model that is trained on a vehicle with the same specifications as vehicle VC before being loaded onto vehicle VC. To learn the map data entry DM21, training data, including monitored data and input data, is obtained in advance. More precisely, various types of input data entries are obtained by actually driving a vehicle whose distance traveled corresponds to a deterioration coefficient AGI of one. Furthermore, information about the occurrence of anomalies, i.e., information indicating whether the lock-up clutch has an anomaly, is obtained as monitored data. In the present embodiment, the types of anomalies, i.e., cold welding, faulty engagement, and seizing, are not differentiated.For example, if an anomaly has occurred in the jump-start coupling, the information about the occurrence of an anomaly can be zero, and if no anomaly has occurred in the jump-start coupling, the information about the determination of the anomaly can be one.

[0144] Training data entries are generated during the vehicle's operation in various situations. For example, a jump-start clutch, which is prone to anomalies, is installed on a vehicle, and the vehicle is driven. If no anomaly occurs during the vehicle's operation, various types of input data entries are collected, and information indicating that no anomaly occurred is also collected as monitored data. If an anomaly occurs during the vehicle's operation, various types of input data entries are collected, and information indicating that the anomaly occurred is also collected as monitored data.

[0145] Such training data entries are used to learn the DM21 map data entry. More precisely, an input variable and an output variable are set so that the difference between the output variable, which is output by a map into which input data is fed, and the actual anomaly determination information converges to a value that is less than or equal to a predetermined value.

[0146] Furthermore, each of the map data entries DM22, DM23, ... is a learned model, which is trained using a vehicle with the same specifications as vehicle VC before being loaded onto vehicle VC. To learn the map data entry DM22, various types of input data are obtained by actually driving a vehicle whose distance traveled corresponds to a deterioration coefficient AGI of two. The corresponding information about the occurrence of anomalies is also obtained as monitored data. As a result, training data entries are generated, containing the input data and the monitored data. These training data entries are used to learn the map data entry DM22.More precisely, an input variable and an output variable are adjusted so that a difference between the output variable, which is output by a characteristic map into which input data is entered, and the actual anomaly determination information converges to a value that is less than or equal to a predetermined value.

[0147] To learn the DM23 map data entry, various types of input data are obtained by actually driving a vehicle whose distance traveled corresponds to a deterioration coefficient (AGI) of three. The corresponding information about the occurrence of anomalies is also obtained as monitored data. As a result, training data entries are generated that contain the input data and the monitored data. These training data entries are used to learn the DM23 map data entry. More precisely, an input variable and an output variable are adjusted so that the difference between the output variable, which is output by a map into which input data is fed, and the actual anomaly determination information converges to a value that is less than or equal to a predetermined value.

[0148] In step S113, CPU 91 evaluates the output variable Y(AGI) calculated in step S111. More precisely, CPU 91 determines whether the output variable Y(AGI) is less than or equal to the anomaly determination value. If the output variable Y(AGI) is less than or equal to the anomaly determination value, CPU 91 determines that an anomaly has occurred in the bypass coupling 75. If the output variable Y(AGI) is greater than the anomaly determination value, CPU 91 determines that no anomaly has occurred in the bypass coupling 75. If, based on the evaluation result in step S115, it is determined that an anomaly has occurred in the bypass coupling 75 (YES), CPU 91 proceeds to step S117. In step S117, CPU 91 stores information indicating that the anomaly has occurred in memory device 93. CPU 91 then temporarily terminates the sequence of processes.

[0149] If, based on the evaluation result in step S115, it is determined that no anomaly has occurred in the bypass coupling 75 (NO), CPU 91 temporarily terminates the sequence of processes. That is, if, for example, the output variable Y(AGI) is greater than the anomaly determination value, CPU 91 does not execute step S117 and temporarily terminates the sequence of processes.

[0150] In addition to the advantages described above (1-1) to (1-10), the present embodiment has the following advantage.

[0151] (3-1) Even if the transmission 40 does not exhibit any anomaly, the amount of heat generated by the transmission 40 can change in accordance with the degree of deterioration of its properties. In this respect, in the present embodiment, the map data entries DM21, DM22, DM23, ..., which correspond to the degree of deterioration of the transmission 40's properties, are prepared in advance. From the map data entries DM21, DM22, DM23, ..., a map data entry corresponding to the current degree of deterioration is selected, and an input variable is fed into a map defined by the selected map data entry. Subsequently, the output variable Y(AGI) output from the map is used to determine whether the lock-up clutch 75 exhibits an anomaly.This increases the accuracy of the determination by using the characteristic map data entries separately according to the degree of deterioration of the properties. Correspondence relationship

[0152] The correspondence between the elements in the embodiments described above and the elements described in the summary is as follows. The correspondence is shown below with each reference numeral of the aspects described in the summary. [1] The anomaly detection device corresponds to the control unit 90. The vehicle's energy source corresponds to the internal combustion engine 10. The drive wheel corresponds to the drive wheel 60. The power transmission device corresponds to the gearbox 40. The friction engagement element corresponds to the lock-up clutch 75. The oil temperature sensor corresponds to the oil temperature sensor 103. The vehicle corresponds to the vehicle VC. An execution device, i.e., the processing circuit, corresponds to the CPU 91 and the ROM 92. The storage device corresponds to the storage device 93. The oil temperature acquisition value corresponds to the oil temperature acquisition value Toil. The oil temperature relationship data corresponds to the oil temperature relationship data RDToil. The map data entries correspond to those in Fig. 1 shown characteristic map data entries DM1, DM2 and DM3, which are in Fig. The 9 shown map data entries DM11, DM12, DM13 and DM14 and those in Fig. The 12 shown characteristic map data entries DM21, DM22, DM23, .... The procurement process corresponds to every process described in the Fig. 4 and Fig. The 5 steps shown, S13 to S23, each process that is in the Fig. 10 and Fig. 11 shown steps S53 to S63 and each process in the Fig. 13 and Fig. The 14 steps shown, S93 to S103, correspond to each process described in the... Fig. 4 and Fig. 5 shown steps S27 to S41, each process of which is in the Fig. 10 and Fig. 11 shown steps S73 to S79 ​​and each process in the Fig. 13 and Fig. 14 steps shown S111 to S117. [2] The acquisition value procurement process corresponds to any process that is in Fig. The 4 steps shown, S13 to S17, each process of which is in Fig. 10 and Fig. 11 shown steps S53 to S57 and each process of the in Fig. 13 and Fig. The 14 steps shown, S93 to S97, correspond to the relationship data generation process, which is the same as each process described in the... Fig. 4 and Fig. 5 shown steps S19 and S21, each process of which is in the Fig. 10 and Fig. 11 shown steps S59 and S61 and each process in the Fig. 13 and Fig. 14 steps shown S99 and S101. [3] The normalized oil temperature measurement value corresponds to the normalized oil temperature measurement value ToilN. [4] The vehicle speed corresponds to the vehicle speed SPD. [5] The clutch corresponds to the bypass clutch 75. The input-side element corresponds to the input-side element 78. The output-side element corresponds to the output-side element 77. The speed of the input-side element corresponds to the engine speed NE. The speed of the output-side element corresponds to the input shaft speed Nat. The speed difference between the input-side element and the output-side element corresponds to the speed difference ΔNtc. [6] The input-side part corresponds to the front cover 71. The output-side part corresponds to the input shaft 81. The input-output speed difference corresponds to the speed difference ΔNtc. The calculated value of the heat generation quantity corresponds to the calculated heat generation quantity value CVtc. [7] The intervention force corresponds to the intervention force EFtc. [8] The packing game is the same as the PCtc packing game. [9] The acceleration sensor corresponds to the acceleration sensor 105. The detection value of the acceleration sensor corresponds to the vehicle acceleration G.

[10] The anomaly detection process corresponds to any process of the in Fig. 4 and Fig. 5 shown steps S31 to S37, if the coefficient of determination MP is equal to one.

[11] The anomaly detection process corresponds to any process that is described in the Fig. 4 and Fig. 5 shown steps S31 to S37, if the coefficient of determination is equal to MP two.

[12] The anomaly detection process corresponds to any process that is described in the Fig. 4 and Fig. 5 shown steps S31 to S37, if the coefficient of determination MP is equal to three.

[13] The first map data entry corresponds to map data entry DM1. The second map data entry corresponds to map data entry DM2. The third map data entry corresponds to map data entry DM3.

[14] The first map data entry corresponds to one of the map data entries DM11, DM12, DM13 and DM14. The second map data entry corresponds to one of the map data entries DM11, DM12, DM13 and DM14, with the exception of the first map data entry. The data selection process corresponds to that described in Fig. 10 and Fig. Step S71, shown in section 11, is the anomaly detection process. This corresponds to every process described in the... Fig. 10 and Fig. 11 steps shown S73 to S79.

[15] The map data entries correspond to the map data entries DM21, DM22, DM23, .... The data selection process corresponds to that described in the Fig. 13 and Fig. Step S109, shown in section 14. The anomaly detection process corresponds to each process described in the Fig. 13 and Fig. 14 steps shown S111 to S117. Modified examples

[0153] The embodiments can be modified as follows. The embodiments and the following modified examples can be combined as long as the combined modified examples remain technically consistent with each other. Characteristic map

[0154] In these embodiments, the activation function of a characteristic map is illustrative and therefore not restrictive. For example, a logical sigmoid function can be used as the activation function of a characteristic map.

[0155] In these embodiments, a neural network with a single intermediate layer is used. However, the neural network can also contain two or more intermediate layers.

[0156] In the embodiments, a fully linked feedforward neural network is used. Alternatively, a recurrent connection neural network can be used. As described, a recurrent connection neural network can be used if, instead of the oil temperature relationship data RDToil, time series data of normalized oil temperature acquisition values ​​ToilN and time series data of oil temperature acquisition values ​​Toil are used as input variables of a characteristic map. Key map data entry

[0157] In the second embodiment, the characteristic map specified by each of the characteristic map data entries DM11, DM12, DM13 and DM14 can be configured to output a variable that distinguishes the type of anomaly of the bypass clutch 75. In this case, the value of the output variable corresponds to the type of anomaly, i.e., occurrence of cold welding, faulty engagement and seizing.

[0158] In the third embodiment, the characteristic map specified by each of the characteristic map data entries DM21, DM22, DM23, ... can be configured to output a variable that distinguishes the type of anomaly of the bypass clutch 75. In this case, the value of the output variable corresponds to the type of anomaly, i.e., occurrence of cold welding, faulty engagement, and seizing.

[0159] In this case, information relating to the probability of cold welding occurring, information relating to the probability of a faulty intervention occurring, and information relating to the probability of seizing occurring are input into an output layer of the characteristic map. The output layer outputs a variable Y that corresponds to each of the input pieces of information.

[0160] In the first embodiment, the storage device 93 can store a single map data entry learned by machine learning, so that each of the determinations—whether cold welding has occurred in the lock-up clutch 75, whether a faulty engagement has occurred in the lock-up clutch 75, and whether the lock-up clutch 75 has seized—can be performed. This configuration eliminates the need to switch the map data entries according to the type of determination.

[0161] Furthermore, if the storage device 93 stores only a single map data entry, as described above, the map data entry can be learned in such a way that the type of anomaly is not distinguished.

[0162] In the first embodiment, as long as the characteristic map data entry DM1 is stored in the storage device 93, the characteristic map data entries DM2 and DM3 do not need to be stored in the storage device 93. In this case as well, it is determined whether cold welding has occurred in the bridging clutch 75.

[0163] In the first embodiment, as long as the characteristic map data entry DM2 is stored in the storage device 93, the characteristic map data entries DM1 and DM3 do not need to be stored in the storage device 93. In this case as well, it is determined whether the bridging clutch 75 is in faulty engagement.

[0164] As long as the characteristic map data entry DM3 is stored in the storage device 93 in the first embodiment, the characteristic map data entries DM1 and DM2 do not need to be stored in the storage device 93. In this case as well, it is determined whether a seizure has occurred in the bridging clutch 75. Input variable

[0165] The input variable does not need to contain the vehicle acceleration G.

[0166] The input variable does not need to contain the PCtc package game.

[0167] The input variable does not need to include the intervention force EFtc.

[0168] If the gearbox 40 contains a sensor that is set up to detect the engagement force EFtc or a correlation value of the engagement force EFtc, a detection value from the sensor can be used as the engagement force EFtc.

[0169] The input variable does not need to contain the calculated heat generation quantity value CVtc.

[0170] If the gearbox 40 contains a sensor that is set up to detect the heat generation quantity of the lock-up clutch 75 or a correlation value of the heat generation quantity, a detection value of the sensor can be used as the input variable.

[0171] If the motor speed NE, i.e., the speed of the input-side elements 78, is used as the input variable, the input variable need not include the input shaft speed Nat, i.e., the speed of the output-side elements 77. Likewise, the input variable does not need to include the speed difference ΔNtc.

[0172] If the input shaft speed Nat, i.e., the speed of the output-side elements 77, is used as the input variable, the input variable need not include the motor speed NE, i.e., the speed of the input-side elements 78. Nor does the input variable need to include the speed difference ΔNtc.

[0173] If the speed difference ΔNtc is used as the input variable, the input variable need not contain the motor speed NE, i.e., the speed of the input-side elements 78. Likewise, the input variable does not need to contain the input shaft speed Nat, i.e., the speed of the output-side elements 77.

[0174] Instead of the speed difference ΔNtc, a speed ratio, i.e. a ratio of the input shaft speed Nat to the motor speed NE, can also be used as an input variable.

[0175] The input variable does not include the motor speed NE, i.e., the speed of the input-side elements 78, the input shaft speed Nat, i.e., the speed of the output-side elements 77, and the speed difference ΔNtc.

[0176] The input variable does not need to contain the vehicle speed SPD. Anomaly detection process

[0177] In the first embodiment, the storage device 93 can be configured to store information indicating that an anomaly has occurred in the bypass clutch 75, if cold welding has occurred, if faulty engagement has occurred, and if seizing has occurred. More precisely, as long as the storage device 93 stores information indicating that an anomaly has occurred, the type of anomaly need not be stored in the storage device 93. Oil temperature relationship data

[0178] In these embodiments, the oil temperature relationship data RDToil includes the five count values ​​Cnt(1) to Cnt(5). However, the number of count values ​​is not limited to five. For example, if the range of numerical values ​​from zero to one is divided into intervals of 0.1, the oil temperature relationship data RDToil can contain ten count values ​​Cnt(1) to Cnt(10).

[0179] Oil temperature relationship data used as input variables for a characteristic map need not be the oil temperature relationship data RDToil. For example, the oil temperature relationship data could be time series data from normalized oil temperature measurement values ​​ToilN.

[0180] The oil temperature relationship data can be time series data from oil temperature recording values. Degree of deterioration

[0181] In the third embodiment, the degree of deterioration of the properties of the transmission 40 is estimated based on the distance traveled by the vehicle VC. Alternatively, the degree of deterioration of the properties of the transmission 40 can be estimated, for example, based on the number of actuations of the lock-up clutch 75 or the total time that the lock-up clutch 75 is in the engaged state. Execution device

[0182] The execution device is not limited to a device containing the CPU 91 and the ROM 92 that executes the software processes. For example, a dedicated hardware circuit may be provided to perform at least part of the software processing carried out in each of the embodiments described above. An example of a dedicated hardware circuit is an ASIC. ASIC is the abbreviation for application-specific integrated circuit. More precisely, the execution device may have one of the following configurations (a) to (c). Configuration (a) includes a processor that executes all the processes described above in accordance with programs, and a program storage device, such as a ROM, that stores the programs.Configuration (b) includes a processor that executes some of the processes described above in accordance with programs, a program storage device, and a dedicated hardware circuit that executes the remaining processes. Configuration (c) includes a dedicated hardware circuit that executes all of the processes described above. Multiple software execution devices, each with a processor and a program storage device, and multiple dedicated hardware circuits may be provided. That is, the processes described above may be executed by a processing circuit that includes at least one of one or more software execution devices or one or more dedicated hardware circuits. The program storage device, i.e., a computer-readable medium, includes any medium accessible by a general-purpose computer or a dedicated computer. Power transmission device

[0183] The power transmission device is not limited to a stepped gearbox and can be any device that contains a frictional engagement element. For example, the power transmission device can be a continuously variable transmission (CVT). Friction engagement element

[0184] The friction engagement element is not limited to the bridging clutch 75. The friction engagement element can be, for example, the first clutch C1, the second clutch C2, or the brake mechanism B1 of the transmission mechanism 80.

[0185] The friction engagement element is not limited to an element that is switched between engaged and disengaged states by adjusting the hydraulic pressure. The friction engagement element can, for example, be switched between engaged and disengaged states by adjusting the electromagnetic force or by being driven by an electric motor. vehicle

[0186] The vehicle can be a hybrid vehicle. The vehicle can be a vehicle that contains a motor-generator but no internal combustion engine. In this case, the motor-generator serves as the vehicle's energy source.

[0187] Various modifications in form and details may be made to the preceding examples without departing from the spirit and scope of protection of the claims and their equivalents. The examples serve only for description and not for limitation. Descriptions of features in each example are to be considered applicable to similar features or aspects in other examples. Suitable results may be obtained if processes are carried out in a different sequence and / or if components in a described system, architecture, device, or circuit are combined differently and / or replaced or supplemented by other components or their equivalents. The scope of disclosure is defined not by the detailed description but by the claims and their equivalents. All variations within the scope of protection of the claims and their equivalents are included in the disclosure.

Claims

[1] Anomaly detection device (90) for a power transmission device, wherein the anomaly detection device (90) is used for a vehicle (VC) which has the power transmission device (40) and includes an oil temperature sensor (103), wherein the Power transmission device (40) includes a friction engagement element (75) and is configured to transmit a force delivered by a power source (10) of the vehicle (VC) to a drive wheel (60), and wherein the oil temperature sensor (103) is configured to detect an oil temperature which is a temperature of oil circulating in the power transmission device (40), wherein the anomaly detection device (90) comprises: a processing circuit (91, 92); and a storage device (93) wherein the storage device (93) stores a characteristic map data entry (DM1, DM2, DM3) which specifies a characteristic map and includes data learned through machine learning, an oil temperature sensor reading (103) is an oil temperature reading (Toil), If oil temperature relationship data (RDToil), which are data corresponding to time series data of the oil temperature sensing value (Toil), are entered into the map as an input variable, the map outputs a variable that determines whether the friction engagement element (75) has an anomaly, and the processing circuit (91, 92) is set up to to execute a procurement process that procures the input variable, and to perform an anomaly determination process which, based on the output variable which is output by the characteristic map as a result of the input variables obtained in the procurement process into the characteristic map, determines whether the friction engagement element (75) has an anomaly. [2] Anomaly detection device (90) according to claim 1, wherein the procurement process includes: a data acquisition process that procures time series data of the oil temperature data acquisition value (Toil) including a multitude of oil temperature data acquisition values ​​(Toil) that are acquired in each acquisition cycle within a predetermined measurement period, and a relationship data generation process that normalizes the multitude of oil temperature acquisition values ​​(Toil) contained in the time series data of the oil temperature acquisition value (Toil) in order to generate the oil temperature relationship data. [3] Anomaly detection device (90) according to claim 2, wherein a value obtained by normalizing the oil temperature measurement value (ToiIN) is a normalized oil temperature measurement value (ToiIN), and In the relationship data generation process, the processing circuit (91, 92) is set up to normalize the multitude of oil temperature acquisition values ​​(Toil) contained in the time series data of the oil temperature acquisition value (Toil), in order to derive time series data of the normalized oil temperature acquisition value (ToiIN) containing a multitude of the normalized oil temperature acquisition values ​​(ToiIN), and to generate as the oil temperature relationship data data showing a distribution of the numerical size of the multitude of normalized oil temperature acquisition values ​​(ToiIN) contained in the time series data of the normalized oil temperature acquisition value (ToiIN). [4] Anomaly detection device (90) according to one of claims 1 to 3, wherein the input variable includes a vehicle speed (SPD). [5] Anomaly detection device (90) according to any one of claims 1 to 4, wherein the power transmission device (40) includes a clutch as a friction engagement element (75), and the input variable includes at least a rotational speed of an input-side element (78) of the clutch, a rotational speed of an output-side element (77) of the clutch or a rotational speed difference (ΔNtc) between the input-side element (78) and the output-side element (77). [6] Anomaly detection device (90) according to any one of claims 1 to 5, wherein an input-output speed difference (ΔNtc) is a speed difference between an input section that inputs a torque into the friction engagement element (75) and an output section that receives a torque output from the friction engagement element (75), and the input variable includes a calculated value of a heat generation quantity of the friction engagement element (75), which is calculated based on the product of the torque input into the friction engagement element (75) and the input-output speed difference (ΔNtc). [7] Anomaly detection device (90) according to one of claims 1 to 6, wherein the input variable includes an engagement force (EFtc) of the friction engagement element (75). [8] Anomaly detection device (90) according to one of claims 1 to 7, wherein the input variable includes a packing clearance (PCTc) of the friction engagement element (75). [9] Anomaly detection device (90) according to one of claims 1 to 8, wherein the input variable includes a detection value of an acceleration sensor (105) attached to the vehicle (VC). [10] Anomaly detection device (90) according to any one of claims 1 to 9, wherein the characteristic map outputs a variable that determines whether cold welding has occurred in the friction engagement element (75), and In the anomaly determination process, the processing circuit (91, 92) is set up to determine, based on the output variable that is output by the characteristic map as a result of the input variables obtained in the procurement process into the characteristic map, whether cold welding has occurred in the friction engagement element (75). [11] Anomaly detection device (90) according to any one of claims 1 to 10, wherein the characteristic map outputs a variable that determines whether the friction engagement element (75) is in faulty engagement, and In the anomaly determination process, the processing circuit (91, 92) is set up to determine, based on the output variable that is output by the characteristic map as a result of the input variables obtained in the procurement process into the characteristic map, whether the friction engagement element (75) is in faulty engagement. [12] Anomaly detection device (90) according to any one of claims 1 to 11, wherein the characteristic map outputs a variable that determines whether seizing has occurred in the friction engagement element (75), and In the anomaly determination process, the processing circuit (91, 92) is set up to determine, based on the output variable that is output by the characteristic map as a result of the input variables obtained in the procurement process into the characteristic map, whether seizing has occurred in the friction engagement element (75). [13] Anomaly detection device (90) according to any one of claims 1 to 9, wherein the storage device (93) stores characteristic map data entries (DM11, DM12, DM13, DM14), the map data entries include a first map data entry (DM11, DM12, DM13, DM14), a second map data entry (DM11, DM12, DM13, DM14) and a third map data entry (DM11, DM12, DM13, DM14), the first map data entry (DM11, DM12, DM13, DM14) specifies a map that outputs a variable that determines whether cold welding has occurred in the friction engagement element (75) when the input variable is entered, the second map data entry (DM11, DM12, DM13, DM14) specifies a map that outputs a variable that determines whether the friction engagement element (75) is in faulty engagement when the input variable is entered, and The third map data entry (DM11, DM12, DM13, DM14) specifies a map that outputs a variable that determines whether seizing has occurred in the friction engagement element (75) when the input variable is entered. [14] Anomaly detection device (90) according to any one of claims 1 to 12, wherein the storage device (93) separately stores characteristic map data entries (DM11, DM12, DM13, DM14) corresponding to each operating state of the friction engagement element (75), the map data entries (DM11, DM12, DM13, DM14) include a first map data entry (DM11, DM12, DM13, DM14) and a second map data entry (DM11, DM12, DM13, DM14), The first map data entry (DM11, DM12, DM13, DM14) specifies a map that outputs a variable that determines whether the friction engagement element (75) has an anomaly when the oil temperature relationship data (RDToil) corresponding to the operating state of the friction engagement element (75), which is a first operating state, are entered into the map as an input variable. The second map data entry (DM11, DM12, DM13, DM14) specifies a map that outputs a variable that determines whether the friction engagement element (75) has an anomaly when the oil temperature relationship data (RDToil), corresponding to the operating state of the friction engagement element (75), which is a second operating state that differs from the first operating state, is entered into the map as an input variable. the processing circuit (91, 92) is set up to perform a data selection process that selects the characteristic map data entry (DM11, DM12, DM13, DM14) corresponding to the operating state of the friction engagement element (75) from the characteristic map data entries (DM11, DM12, DM13, DM14) stored in the storage device (93), and the processing circuit (91, 92) is set up to determine, in the anomaly determination process, based on the output variable that is output by the characteristic map specified by the characteristic map data entry (DM11, DM12, DM13, DM14) selected in the data selection process, as a result of the input of the input variables obtained in the procurement process into the characteristic map, whether the friction engagement element (75) has an anomaly. [15] Anomaly detection device (90) according to any one of claims 1 to 12, wherein the storage device (93) stores characteristic map data entries (DM21, DM22, DM23) which correspond to a degree of deterioration of the characteristics of the power transmission device (40), the processing circuit (91, 92) is set up to perform a data selection process that selects the characteristic map data entry (DM21, DM22, DM23) corresponding to the degree of deterioration of the characteristics of the power transmission device (40) from the characteristic map data entries (DM21, DM22, DM23) stored in the storage device (93), and the processing circuit (91, 92) is set up to determine, in the anomaly determination process, based on the output variable that is output by the characteristic map specified by the characteristic map data entry (DM21, DM22, DM23) selected in the data selection process, as a result of the input of the input variables obtained in the procurement process into the characteristic map, whether the friction engagement element (75) has an anomaly.

Images