Temperature estimation device for throttle valve

Abstract

To provide a temperature estimation device capable of accurately estimating a temperature of a throttle valve.SOLUTION: Mapping data that defines mapping based on a parameter learned through machine learning is stored in a storage device 46. The mapping includes an intake air variable involving heat quantity that moves between intake air passing through a throttle valve 21 and the throttle valve 21 as an input variable and time series data of the input variable, and includes a temperature of the throttle valve 21 as an output variable. A CPU 42 executes acquisition processing for acquiring a value of the input variable and calculation processing for calculating a value of the output variable by inputting the value of the input variable acquired through the acquisition processing to the mapping.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a throttle valve temperature estimating device. [Background technology]

[0002] Various techniques have been proposed for estimating the temperature of a throttle valve provided in an intake passage of an internal combustion engine. For example, an internal combustion engine described in Patent Document 1 is configured to determine whether or not the throttle valve is frozen based on the temperature of the cooling water, which is related to the temperature of the throttle valve. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2020-2843 A Summary of the Invention [Problem to be solved by the invention]

[0004] When estimating the temperature of the throttle valve, it is desirable to estimate the temperature with high accuracy. [Means for solving the problem]

[0005] A throttle valve temperature estimation device that solves the above problem is a device that estimates the temperature of a throttle valve provided in an intake passage of an internal combustion engine. This temperature estimation device includes an execution device and a storage device. The storage device stores mapping data that defines a mapping based on parameters learned by machine learning, and the mapping includes, as input variables, intake variables related to the amount of heat transferred between the intake air passing through the throttle valve and the throttle valve, and time-series data of the input variables, and includes the temperature as an output variable. The execution device executes an acquisition process that acquires values ​​of the input variables, and a calculation process that calculates values ​​of the output variables by inputting the values ​​of the input variables acquired by the acquisition process into the mapping. Effect of the Invention

[0006] This throttle valve temperature estimating device can estimate the throttle valve temperature with high accuracy. [Brief description of the drawings]

[0007] [Figure 1] FIG. 1 is a diagram showing a configuration of a throttle valve temperature estimation device according to one embodiment. [Diagram 2] FIG. 2 is a flowchart showing a procedure of a process executed by the control device according to the embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0008] Hereinafter, an embodiment of the throttle valve temperature estimating device will be described with reference to FIGS. <About the internal combustion engine configuration> As shown in FIG. 1, an internal combustion engine 10 mounted on a vehicle is provided with an intake passage 12 through which intake air flows to be introduced into a combustion chamber.

[0009] A throttle body 20 is provided in the intake passage 12. The throttle body 20 is provided with a throttle valve 21 that adjusts the amount of intake air flowing through the intake passage 12. The throttle valve 21 is biased in the closing direction by a spring (not shown). The throttle body 20 is provided with an electric motor 16 that drives the throttle valve 21. A cooling water passage 30 for circulating the cooling water of the internal combustion engine 10 within the throttle body 20 is connected to the throttle body 20. The flow rate of the cooling water flowing through the cooling water passage 30 is adjusted by an electric water pump provided in the cooling system of the internal combustion engine 10.

[0010] The control device 40 controls the internal combustion engine 10 as a control target, and controls the drive of the fuel injection valves of the internal combustion engine 10, the throttle valve 21, etc., in order to control the torque, exhaust component ratio, etc., which are control variables.

[0011] When performing various controls, the control device 40 refers to the outside air temperature THout detected by the outside air temperature sensor 51, the intake air temperature THA detected by the intake air temperature sensor 52, and the intake air amount GA detected by the air flow meter 53. The control device 40 also refers to the output signal Scr of the crank angle sensor 54 that detects the rotation angle of the crankshaft of the internal combustion engine 10, and the coolant temperature THW that is the temperature of the coolant detected by the water temperature sensor 55. The control device 40 also refers to the oil temperature THO that is the temperature of the lubricating oil of the internal combustion engine 10 that is detected by the oil temperature sensor 56, and the vehicle speed SP that is detected by the vehicle speed sensor 57. The control device 40 also refers to the output signal Swp of the rotation speed sensor 58 that detects the rotation speed of the electric water pump. The control device 40 calculates the engine rotation speed NE based on the output signal Scr of the crank angle sensor 54. The control device 40 also calculates the engine load factor KL based on the engine rotation speed NE and the intake air amount GA. The engine load factor KL represents the ratio of the current cylinder inflow air amount to the cylinder inflow air amount when the internal combustion engine 10 is operated steadily under full load. The cylinder inflow air amount is the amount of air that flows into the cylinder during the intake stroke. The control device 40 also calculates the cooling water flow rate Q, which is the flow rate of the cooling water flowing through the throttle body 20, based on the output signal Swp of the rotation speed sensor 58.

[0012] The control device 40 includes a CPU 42, a ROM 44, and a storage device 46 which is an electrically rewritable non-volatile memory, which are capable of communicating with each other. The control device 40 executes various controls by the CPU 42 executing a program stored in the ROM 44. The CPU 42 and the ROM 44 constitute an execution device.

[0013] <Throttle valve temperature estimation process> The control device 40 executes a temperature estimation process to estimate the temperature THv of the throttle valve 21. The control device 40 executing this temperature estimation process constitutes a temperature estimation device for the throttle valve 21.

[0014] Fig. 2 shows the procedure of the temperature estimation process executed by the control device 40. The process shown in Fig. 2 is realized by the CPU 42 repeatedly executing a program stored in the ROM 44, for example, at a predetermined interval. In the following, the step number of each process is represented by a number preceded by "S."

[0015] 2, the CPU 42 first executes an acquisition process to acquire various values ​​(S10). Specifically, the CPU 42 acquires the outside air temperature THout, the intake air temperature THA, the intake air amount GA, the engine speed NE, the engine load factor KL, the cooling water temperature THW, the cooling water flow rate Q, the oil temperature THO, and the vehicle speed SP. The CPU 42 also acquires the previous outside air temperature THoutp, the previous intake air temperature THAp, the previous intake air amount GAp, the previous engine speed NEp, the previous engine load factor KLp, the previous cooling water temperature THWp, the previous cooling water flow rate Qp, the previous oil temperature THOp, and the previous vehicle speed SPp.

[0016] The previous outside air temperature THoutp is the value of the outside air temperature THout acquired by the CPU 42 in the present process executed before the preset time Tp. The previous intake air temperature THAp is the value of the intake air temperature THA acquired by the CPU 42 in the present process executed before the preset time Tp. The previous intake air amount GAp is the value of the intake air amount GA acquired by the CPU 42 in the present process executed before the preset time Tp. The previous engine speed NEp is the value of the engine speed NE acquired by the CPU 42 in the present process executed before the preset time Tp. The previous engine load factor KLp is the value of the engine load factor KL acquired by the CPU 42 in the present process executed before the preset time Tp. The previous cooling water temperature THWp is the value of the cooling water temperature THW acquired by the CPU 42 in the present process executed before the preset time Tp. The previous cooling water flow rate Qp is the value of the cooling water flow rate Q acquired by the CPU 42 in the present process executed before the preset time Tp. The previous oil temperature THOp is the value of the oil temperature THO acquired by the CPU 42 in this process executed before the predetermined time Tp. The previous vehicle speed SPp is the value of the vehicle speed SP acquired by the CPU 42 in this process executed before the predetermined time Tp. Note that the predetermined time Tp is a time required to improve the estimation accuracy of the temperature THv and is set in advance.

[0017] Next, the CPU 42 assigns each value acquired in the process of S10 to an input variable for the mapping defined by the mapping data DM stored in the storage device 46 (S20). That is, the CPU 42 assigns the outside air temperature THout to input variable x(1), the intake air temperature THA to input variable x(2), the intake air amount GA to input variable x(3), the engine speed NE to input variable x(4), and the engine load factor KL to input variable x(5).The CPU 42 also assigns the cooling water temperature THW to input variable x(6), the cooling water flow rate Q to x(7), the oil temperature THO to x(8), and the vehicle speed SP to x(9).

[0018] The CPU 42 also assigns the previous outside air temperature THoutp to input variable x(10), the previous intake air temperature THAp to input variable x(11), and the previous intake air amount GAp to input variable x(12). The CPU 42 also assigns the previous engine speed NEp to input variable x(13), and the previous engine load factor KLp to input variable x(14). The CPU 42 also assigns the previous cooling water temperature THWp to input variable x(15), the previous cooling water flow rate Qp to x(16), the previous oil temperature THOp to x(17), and the previous vehicle speed SPp to x(18).

[0019] In this embodiment, the above-mentioned input variables into which the outside air temperature THout, the engine speed NE, the intake air volume GA, and the engine load factor KL are substituted are intake variables related to the amount of heat transferred between the intake air passing through the throttle valve 21 and the throttle valve 21.

[0020] The input variables into which the coolant temperature THW and the coolant flow rate Q are substituted are coolant variables related to the amount of heat transferred between the coolant flowing through the throttle body 20 and the throttle valve 21.

[0021] The input variables into which the oil temperature THO and the vehicle speed SP are substituted are atmospheric variables related to the amount of heat transferred from the air in the engine compartment in which the internal combustion engine 10 is housed to the throttle valve 21 via the throttle body 20.

[0022] Furthermore, when the engine load ratio KL is large, the opening of the throttle valve 21 is increased. Therefore, the torque of the electric motor 16 required to drive the throttle valve 21, which is biased by a spring to the closing side, to the opening side increases, and the amount of heat generated by the electric motor 16 increases. Therefore, the input variable into which the engine load ratio KL is substituted is a motor variable related to the amount of heat transferred from the electric motor 16 to the throttle valve 21 via the throttle body 20.

[0023] Each variable from input variable x(10) to input variable x(18) is time-series data indicating the past value of the input variable. Next, the CPU 42 executes a calculation process to calculate the temperature THv, which is an output variable, by substituting each value of the input variable x(1) to the input variable x(18) into the mapping defined by the mapping data DM (S30). Then, the CPU 42 temporarily ends this process.

[0024] In this embodiment, a function approximator is exemplified as the mapping, and more specifically, a fully connected forward propagation type neural network with one intermediate layer is exemplified. Specifically, the input variables x(1) to x(18) and the bias parameter x(0) to which values ​​are assigned by the process of S30 are converted into "m" values ​​by a linear mapping defined by coefficients wFjk (j=1 to m, k=0 to 18). Then, the value of the intermediate layer node is determined by substituting each of the converted "m" values ​​into the activation function f. In addition, the value of the temperature THv, which is the output variable, is determined by substituting the value into the activation function g, which is obtained by converting each of the values ​​of the intermediate layer nodes by the linear mapping defined by the coefficients wS1j. In this embodiment, the activation function f is exemplified by a hyperbolic tangent, and the activation function g is exemplified by a ReLU function.

[0025] The mapping data DM is a trained model trained using, for example, an internal combustion engine 10 mounted on a vehicle. That is, various input variables are acquired, and the temperature THv at that time is measured. In this way, training data is generated, and the mapping data DM is trained based on the training data. Then, the values ​​of the coefficients wFjk and wS1j are trained so as to reduce the difference between the value output by the mapping data DM and the value of the training data with respect to the temperature THv, which is an output variable. Specifically, for example, the values ​​of the coefficients wFjk and wS1j may be trained so as to minimize the tolerance entropy, which is one of the error functions. In this way, the mapping data DM can be trained using machine learning. Then, when such training is completed, the mapping data DM, which is the trained model, is stored in the storage device 46.

[0026] <About action and effects> The operation and effects of this embodiment will be described. (1) The amount of heat transferred between the intake air passing through the throttle valve 21 and the throttle valve 21 is used as the input variable, and the temperature THv of the throttle valve 21 is calculated by inputting the input variable into a mapping defined by the mapping data DM. Here, there is a time delay between the temperature THv of the throttle valve 21 calculated based on the input variable and the actual temperature of the throttle valve 21. Therefore, in this embodiment, the above-mentioned input variable includes time series data of the input variable. Therefore, the temperature THv, which is the output variable, is calculated taking into account such a time delay, so that the estimation accuracy of the temperature THv of the throttle valve 21 is improved.

[0027] (2) The input variables include the coolant variables related to the amount of heat transferred between the coolant flowing through the throttle body 20 and the throttle valve 21. Therefore, the estimation accuracy of the temperature THv of the throttle valve 21 is improved compared to a case where the coolant variables are not included in the input variables.

[0028] (3) The input variables include the atmospheric variables related to the amount of heat transferred from the air in the engine compartment housing the internal combustion engine 10 to the throttle valve 21 via the throttle body 20. Therefore, the estimation accuracy of the temperature THv of the throttle valve 21 is improved compared to a case in which the atmospheric variables are not included in the input variables.

[0029] (4) The input variables include the motor variables related to the amount of heat transferred from the electric motor 16 to the throttle valve 21 via the throttle body 20. Therefore, the estimation accuracy of the temperature THv of the throttle valve 21 is improved compared to a case where the motor variables are not included in the input variables.

[0030] <Examples of changes> The above embodiment can be modified as follows: The above embodiment and the following modifications can be combined with each other to the extent that no technical contradiction occurs.

[0031] The values ​​substituted for the input variables as intake variables may be changed as appropriate. The values ​​substituted for the input variables as cooling water variables may be changed as appropriate. The values ​​substituted for the input variables as atmosphere variables may be changed as appropriate. The values ​​substituted for the input variables as motor variables may be changed as appropriate.

[0032] At least one of the cooling water variables, the atmosphere variables, and the motor variables may be omitted from the input variables. At least one of the input variables x(10) to x(18) may be input to the mapping as time series data indicating the past values ​​of the input variables.

[0033] Although the data from the past one time is used as the time series data of the input variables, the data from the past two or more times may be used as the time series data of the input variables. Increasing the amount of time series data in this way can further improve the estimation accuracy of the temperature THv.

[0034] The activation function of the above mapping is an example, other functions may be used. Although the neural network in the above embodiment has one hidden layer, the number of hidden layers may be two or more.

[0035] Although a fully connected forward propagation type neural network has been exemplified as the neural network, the present invention is not limited to this. For example, a recurrent connection type neural network may be adopted as the neural network.

[0036] The function approximator as the mapping may be a regression equation. This corresponds to the neural network not having an intermediate layer. [Explanation of symbols]

[0037] 10... internal combustion engine, 12... intake passage, 16... electric motor, 20... throttle body, 21... throttle valve, 30... cooling water passage, 40... control device, 42... CPU, 44... ROM, 46... storage device, 51... outside air temperature sensor, 52... intake air temperature sensor, 54... crank angle sensor, 55... water temperature sensor, 56... oil temperature sensor, 57... vehicle speed sensor, 58... rotational speed sensor

Claims

[Claim 1] A device for estimating a temperature of a throttle valve provided in an intake passage of an internal combustion engine, comprising: an execution device and a storage device, The storage device stores mapping data that defines a mapping based on parameters learned by machine learning, the mapping includes, as input variables, intake variables relating to the amount of heat transferred between the intake air passing through the throttle valve and the throttle valve, and time series data of the input variables, and includes the temperature as an output variable; The execution device is An acquisition process for acquiring values ​​of the input variables; a calculation process for calculating a value of the output variable by inputting the value of the input variable acquired by the acquisition process into the mapping; Throttle valve temperature estimator.

Images