JP2022176528A - Vehicle management system, vehicle and traveling management method of vehicle - Google Patents

Abstract

To properly correct a fuel combustion state of an engine even if fuel composition information is not required.SOLUTION: Each of a vehicle 1A and a plurality of vehicles 1B includes sensors 4 for detecting a fuel combustion state of an engine 3. A server 23 includes a memory 232 in which a learnt model related to engine control is stored. In the learnt model, a correction amount for correcting the fuel combustion state of the engine 3 is learnt on the basis of sensor data indicating a detection result of each of the sensors 4 of the plurality of vehicles 1B, and service station data capable of specifying a service station at which each of the plurality of vehicles 1B has received the supply of fuel. A processor 231 calculates the correction amount by imparting the sensor data of the vehicle 1A and the service station data to the learnt model, and transmits the calculated correction amount to the vehicle 1A. The vehicle 1A corrects the fuel combustion state of the engine 3 of the vehicle 1A by using the received correction amount.SELECTED DRAWING: Figure 11

Description

TECHNICAL FIELD The present disclosure relates to a vehicle management system, a vehicle, and a vehicle travel management method, and more particularly to a technique for managing a vehicle equipped with an engine.

Japanese Patent Laying-Open No. 2005-113715 (Patent Document 1) discloses a control device that controls the operation of a power plant (specifically, an engine) that uses a predetermined fuel.

Japanese Patent Application Laid-Open No. 2005-113715

Patent Literature 1 describes acquiring fuel composition information based on vehicle position information when a vehicle is refueled. The engine control device disclosed in Patent Document 1 is configured to correct the operation of the engine based on fuel composition information.

In order for the engine control device disclosed in Patent Document 1 to be effective, that is, to appropriately correct the operation of the engine (fuel combustion state in the engine), it is necessary to obtain information on the composition of the fuel. However, obtaining fuel composition information may require complex processes such as fuel analysis.

The present disclosure has been made to solve the above problems, and an object of the present disclosure is to provide a technology capable of appropriately correcting the fuel combustion state in an engine even if fuel composition information is not required. is.

(1) A vehicle management system according to one aspect of the present disclosure includes a target vehicle and a server configured to communicate with the target vehicle and the plurality of vehicles. Each of the subject vehicle and the plurality of vehicles includes at least one sensor that detects fuel combustion conditions in the engine. The server includes a memory that stores a trained model for engine control, and a processor that executes arithmetic processing using the trained model. In the trained model, based on sensor data indicating the detection results of at least one sensor of each of the plurality of vehicles, and gas station data capable of identifying the gas station where each of the plurality of vehicles received fuel, A correction amount for correcting the fuel combustion state in the engine is learned. The processor provides the learned model with the sensor data of the target vehicle and the gas station data, calculates the correction amount, and transmits the calculated correction amount to the target vehicle. The target vehicle uses the received correction amount to correct the fuel combustion state in the engine of the target vehicle.

(2) The gas station data includes data on the installation location of the gas station and data on the timing of refueling at the gas station.

(3) In the learned model, correction amounts are learned based on sensor data, gas station data, and data on the traveling positions of a plurality of vehicles. In addition to the sensor data and gas station data, the processor provides the learned model with data regarding the travel position of the target vehicle, thereby calculating the correction amount of the target vehicle.

(4) Sensor data includes the cylinder internal pressure of the engine, the rotation speed of the engine, the temperature of the exhaust gas from the engine, the pressure difference before and after the filter that collects particulate matter in the exhaust gas, and the nitrogen in the exhaust gas. including data relating to at least one of the oxide concentrations;

(5) A vehicle according to another aspect of the present disclosure is a vehicle that is configured to communicate with a server that includes a learned model for engine control. In the trained model, the correction amount for correcting the fuel combustion state in the engine is learned based on sensor data and gas station data collected through communication with multiple vehicles each equipped with an engine. . The sensor data is data indicating the detection result of the fuel combustion state in each engine of a plurality of vehicles. Gas station data is data that can identify gas stations where a plurality of vehicles have been supplied with fuel. A vehicle includes a communication device and a control device. The communication device transmits the sensor data of the vehicle and the gas station data to the server, and receives the correction amount calculated by the server giving the sensor data from the vehicle and the gas station data to the learned model. The controller corrects the fuel combustion state in the vehicle engine using the received correction amount.

(6) A vehicle travel management method according to still another aspect of the present disclosure manages vehicle travel using a learned model related to engine control. In the trained model, the correction amount for correcting the fuel combustion state in the engine is learned based on sensor data and gas station data collected through communication with multiple vehicles each equipped with an engine. . The sensor data is data indicating the detection result of the fuel combustion state in each engine of a plurality of vehicles. Gas station data is data that can identify gas stations where a plurality of vehicles have been supplied with fuel. The run management method includes first and second steps. The first step is to calculate a correction amount for correcting the fuel combustion state in the engine of the target vehicle by providing the learned model with sensor data and gas station data from the target vehicle. The second step is a step of correcting the fuel combustion state in the engine of the target vehicle using the correction amount calculated in the first step.

According to the present disclosure, it is possible to appropriately correct the fuel combustion state in the engine even if the fuel composition information is not required.

1 is a diagram schematically showing the overall configuration of a vehicle management system according to an embodiment; FIG. FIG. 4 is a diagram showing the configuration of the vehicle and the control center in more detail; It is a figure which shows an example of a structure of an engine. FIG. 5 is a diagram showing another example of the configuration of the engine; It is a figure which shows an example of a structure of sensors. FIG. 2 is a functional block diagram showing an example of a predictive model regarding the combustion state of an engine according to the present embodiment; FIG. FIG. 4 is a diagram for explaining an example of combustion improvement processing for correcting the operation of the injection system of the engine; FIG. 5 is a diagram for explaining another example of combustion improvement processing for correcting the operation of the injection system of the engine; FIG. 4 is a diagram for explaining an example of combustion improvement processing for correcting the operation of the EGR system of the engine; FIG. 5 is a diagram for explaining an example of combustion improvement processing for correcting the operation of the intake system of the engine; 4 is a flow chart showing a processing procedure of combustion improvement processing in the present embodiment;

Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. The same or corresponding parts in the drawings are denoted by the same reference numerals, and the description thereof will not be repeated.

[Embodiment]
<Overall system configuration>
FIG. 1 is a diagram schematically showing the overall configuration of a vehicle management system according to this embodiment. The vehicle management system 100 includes a vehicle 1A, a plurality of vehicles 1B, and a control center 2. FIG. 1 shows how a vehicle 1A is refueled at a refueling station 9. As shown in FIG. The vehicle 1A corresponds to a "target vehicle" according to the present disclosure.

The vehicle 1A and the plurality of vehicles 1B basically have a common configuration. When the vehicle 1A and the vehicle 1B are not distinguished from each other, they are also described as "vehicle 1". Two-way communication is possible between each vehicle 1 and the control center 2 .

A vehicle 1 is a vehicle in which an engine 3 (see FIG. 3) is mounted, and is specifically a so-called conveyor vehicle, hybrid vehicle, or plug-in hybrid vehicle. The type of fuel combusted by the engine 3 is not particularly limited. The fuel is, for example, gasoline fuel, diesel fuel, biofuel (such as ethanol), or gaseous fuel (such as propane gas). The fuel may also be hydrogen, ammonia, e-fuel (synthetic liquid fuel of carbon dioxide and hydrogen), dimethyl ether (DME), or the like.

FIG. 2 is a diagram showing the configuration of the vehicle 1 and the control center 2 in more detail. The vehicle 1 includes a navigation system 11 , a DCM (Data Communication Module) 12 , an ECU (Electronic Control Unit) 13 , an in-vehicle network 14 , an engine 3 , and sensors 4 .

The navigation system 11 includes a GPS (Global Positioning System) receiver (not shown) and identifies the position of the vehicle 1 based on radio waves from artificial satellites (not shown). The navigation system executes various navigation processes for the vehicle 1 using the position information of the vehicle 1 specified by the GPS receiver.

The DCM 12 is configured to be capable of wireless two-way communication with the control center 2 . Thereby, the control center 2 can manage the travel history of the vehicle 1 by collecting the position information of the vehicle 1 . In addition, below, the positional information on the vehicle 1 is also described as "vehicle positional data."

The ECU 13 includes a processor 131 such as a CPU (Central Processing Unit), a memory 132 such as a ROM (Read Only Memory) and a RAM (Random Access Memory), and an input/output port 133 through which signals are input/output. The ECU 13 controls the engine 3 based on the input of signals from the sensors 4 and the map and programs stored in the memory 132 so that the vehicle 1 is in a desired state.

In-vehicle network 14 is, for example, a wired network such as a CAN (Controller Area Network). An in-vehicle network 14 connects the navigation system 11, the DCM 12, and the ECU 13 to each other.

The engine 3 burns fuel in accordance with a control command from the ECU 13 to output driving force for the vehicle 1 to run. The sensors 4 comprehensively describe various sensors mounted on the vehicle 1 . In this embodiment, the sensors 4 are mainly used for predicting the combustion state of the engine 3 . The configurations of the engine 3 and the sensors 4 will be described in more detail with reference to FIGS. 3 and 5. FIG.

The control center 2 is configured to manage travel of a plurality of vehicles 1 . The control center 2 includes a vehicle database 21 , a travel history database 22 , a server 23 , a communication module 24 and an intranet 25 .

The vehicle database 21 stores identification information, manufacturing information, and the like regarding each of the plurality of vehicles 1 . The identification information includes, for example, the serial number of vehicle 1 and the communication device ID of DCM 12 . The manufacturing information includes, for example, information on the manufacturer and vehicle type (model, grade, model year, etc.) of the vehicle 1 .

The travel history database 22 stores travel histories of each of the plurality of vehicles 1 . The travel history includes, in addition to vehicle position data, data related to the combustion state of fuel in the engine 3 (also referred to as the combustion state of the engine 3). Gas station data, which will be described later, is also stored in the travel history database 22 .

The server 23 includes a processor 231 , a memory 232 and an input/output port 233 like the ECU 13 . Server 23 may be a cloud server. Main processing executed by the server 23 in the present embodiment includes processing for improving the combustion state of the engine 3 . This processing is described as "combustion improvement processing". The combustion improvement process will be described later in detail.

The communication module 24 is configured to be capable of wireless two-way communication with the DCM 12 mounted on the vehicle 1 .

The intranet 25 connects the vehicle database 21, the travel history database 22, the server 23, and the communication module 24 to each other.

<Engine configuration>
FIG. 3 is a diagram showing an example of the configuration of the engine 3. As shown in FIG. The engine 3 is, for example, an in-line four-cylinder spark ignition internal combustion engine. The engine 3 includes an engine body 31 . The engine body 31 includes four cylinders 311-314 arranged in one direction. Since the configurations of the cylinders 311 to 314 are the same, the configuration of the cylinder 311 will be representatively described below.

The cylinder 311 is provided with two intake valves 321 , two exhaust valves 322 , an injector 323 and an ignition plug 324 . An intake passage 33 and an exhaust passage 34 are connected to the cylinder 311 . The intake passage 33 is opened and closed by an intake valve 321 . The exhaust passage 34 is opened and closed by an exhaust valve 322 . By adding fuel to the air supplied to the engine body 31 through the intake passage 33, a mixture of air and fuel is generated. Fuel is injected into cylinder 311 by injector 323 and an air-fuel mixture is generated within cylinder 311 . Then, the ignition plug 324 ignites the air-fuel mixture in the cylinder 311 . The air-fuel mixture is thus combusted in the cylinder 311 . Combustion energy generated when the air-fuel mixture is combusted in cylinder 311 is converted into kinetic energy by a piston (not shown) in cylinder 311 and output. The fuel supply method of the engine 3 is not limited to in-cylinder injection, and may be port injection, or a combination of in-cylinder injection and port injection.

The engine 3 further includes a supercharger 35 , a waste gate valve (WGV) device 36 and an EGR device 37 . The supercharger 35 is configured to supercharge intake air using exhaust energy. More specifically, supercharger 35 includes compressor 351 , turbine 352 and shaft 353 . A compressor 351 is arranged in the intake passage 33 and a turbine 352 is arranged in the exhaust passage 34 . Compressor 351 and turbine 352 are configured to be connected to each other via shaft 353 and rotate integrally. The turbine 352 rotates by receiving the flow of exhaust gas discharged from the engine body 31 . The rotational force of turbine 352 is transmitted to compressor 351 via shaft 353 to rotate compressor 351 . The rotation of the compressor 351 compresses intake air directed toward the engine body 31 and supplies the compressed air to the engine body 31 .

An air flow meter 405 (described later) is provided at a position upstream of the compressor 351 in the intake passage 33 . An intercooler 331 is provided at a position downstream of the compressor 351 in the intake passage 33 . A throttle valve 332 is provided at a position downstream of the intercooler 331 in the intake passage 33 . Air (intake) flowing into the intake passage 33 passes through the compressor 351 , the intercooler 331 and the throttle valve 332 and is supplied to the cylinders 311 to 314 of the engine body 31 . Intercooler 331 cools the intake air compressed by compressor 351 . The throttle valve 332 is configured to be able to adjust the flow rate of intake air flowing through the intake passage 33 .

A catalyst device 341 and an aftertreatment device 342 are provided in the exhaust passage 34 . Catalyst device 341 includes, for example, a three-way catalytic converter. Catalyst device 341 oxidizes unburned components (eg, hydrocarbons (HC) or carbon monoxide (CO)) contained in the exhaust gas discharged from engine 3, and oxidizes components (eg, nitrogen oxides (NOx)). to reduce. The post-treatment device 342 is provided downstream of the catalyst device 341 . The aftertreatment device 342 includes a filter such as a DPF (Diesel Particulate Filter) and collects particulate matter (PM) emitted from the engine 3 .

The WGV device 36 is provided upstream of the catalyst device 341 in the exhaust passage 34 . The WGV device 36 is configured to allow the exhaust gas discharged from the engine body 31 to flow around the turbine 352 and to adjust the amount of bypassed exhaust gas. WGV device 36 includes bypass passage 361 , WGV 362 , and WGV actuator 363 .

A bypass passage 361 is connected to the exhaust passage 34 to flow exhaust around the turbine 352 . Specifically, the bypass passage 361 branches from a portion of the exhaust passage 34 upstream of the turbine 352 (for example, between the engine main body 31 and the turbine 352) and a portion of the exhaust passage 34 downstream of the turbine 352. (eg, between turbine 352 and catalyst device 341).

WGV 362 is arranged in bypass passage 361 . WGV 362 is a negative pressure valve driven by WGV actuator 363 . The WGV 362 is configured to be able to adjust the flow rate of the exhaust gas guided from the engine body 31 to the bypass passage 361 by its opening. As the WGV 362 closes, the flow rate of the exhaust gas introduced from the engine body 31 to the bypass passage 361 decreases, while the flow rate of the exhaust gas flowing into the turbine 352 increases, and the intake pressure (supercharging pressure) increases.

The EGR device 37 is provided in the exhaust passage 34 and causes the exhaust gas to flow into the intake passage 33 . EGR device 37 includes an EGR passage 371 , an EGR valve 372 and an EGR cooler 373 . The EGR passage 371 connects a portion of the exhaust passage 34 between the catalyst device 341 and the aftertreatment device 342 and a portion of the intake passage 33 between the compressor 351 and the air flow meter 405, thereby Part of the exhaust gas is taken out as EGR gas and led to the intake passage 33 . The EGR valve 372 is configured to be able to adjust the flow rate of EGR gas flowing through the EGR passage 371 . The EGR cooler 373 cools the EGR gas flowing through the EGR passage 371 .

FIG. 4 is a diagram showing another example of the configuration of the engine 3A. The engine 3A differs from the engine 3 shown in FIG. 3 in that it includes a variable nozzle device 38 instead of the WGV device 36.

The variable nozzle device 38 includes a plurality of variable nozzles (nozzle vanes) provided on the outer circumference of the turbine wheel. The variable nozzle is configured such that its opening degree (VN opening degree) is changed within a range from the opening degree at the fully closed position to the opening degree at the fully open position according to a control command from the ECU 13 . When the nozzle opening of the variable nozzle is reduced, the opening area of the region through which the exhaust gas passes is reduced, and the flow velocity of the exhaust gas introduced into the turbine 352 is increased. As a result, the energy recovered in the turbine wheel increases and the boost pressure increases. On the other hand, if the variable nozzle opening is increased, the opening area of the region through which the exhaust gas passes increases, and the flow velocity of the exhaust gas decreases. As a result, the energy recovered in the turbine wheel is reduced and the boost pressure is reduced.

<Configuration of sensors>
FIG. 5 is a diagram showing an example of the configuration of the sensors 4. As shown in FIG. The sensors 4 include a vehicle speed sensor 401, an accelerator opening sensor 402, an engine rotation speed sensor 403, a knock sensor 404, an air flow meter 405, a throttle opening sensor 406, a cylinder pressure sensor 407, and a cooling water temperature sensor. 408 , boost pressure sensor 409 , boost temperature sensor 410 , turbine rotation speed sensor 411 , exhaust temperature sensor 412 , air-fuel ratio sensor 413 , NOx sensor 414 , and differential pressure sensor 415 .

Vehicle speed sensor 401 detects the speed of vehicle 1 . The accelerator opening sensor 402 detects the amount of depression (accelerator opening) of an accelerator pedal (not shown) by the driver. An engine rotation speed sensor 403 detects the rotation angle (crank angle) of a crankshaft (not shown) and the rotation speed of the engine 3 . Knock sensor 404 detects the occurrence of knocking in engine 3 (vibration of engine body 31). Airflow meter 405 detects the amount of intake air to intake passage 33 .

A throttle opening sensor 406 detects the opening of the throttle valve 332 (throttle opening). An in-cylinder pressure sensor 407 detects the pressure inside a combustion chamber (not shown). A cooling water temperature sensor 408 detects the temperature of cooling water (cooling water temperature) circulating in a water jacket (not shown) provided in the engine body 31 . The supercharging pressure sensor 409 detects the supercharging pressure by the supercharger 35 (pressure downstream of the compressor 351). Supercharging temperature sensor 410 detects the temperature of intake air cooled by intercooler 331 (supercharging temperature).

A turbine rotation speed sensor 411 detects the rotation speed of the turbine 352 of the supercharger 35 . An exhaust temperature sensor 412 detects the temperature of the exhaust gas. The air-fuel ratio sensor 413 detects oxygen concentration in the exhaust gas. A NOx sensor 414 detects the concentration of nitrogen oxides (NOx) in the exhaust gas. Differential pressure sensor 415 detects the pressure difference across a filter (not shown) included in aftertreatment device 342 . Each sensor outputs a signal indicating the detection result to the ECU 13 .

<Data collection>
It is desirable to improve the combustion state of the engine 3 in order to improve the fuel efficiency and drivability of the vehicle 1A. It is conceivable that the vehicle 1A acquires fuel composition information and improves the combustion state of the engine 3 based on the composition information (see Patent Document 1, for example). However, obtaining fuel composition information may require complex processes such as fuel analysis. Therefore, in the present embodiment, instead of using the fuel composition information, a configuration described below is adopted.

The vehicle 1 (ECU 13) transmits data acquired from signals from each sensor included in the sensors 4 to the control center 2 (server 23). Thereby, the control center 2 can obtain or estimate the combustion state of the fuel in the engine 3 (cylinders 311 to 314). Specifically, the control center 2 can obtain the combustion state of fuel in the cylinders 311 to 314 based on the signal detected by the in-cylinder pressure sensor 407 (the waveform signal of the in-cylinder pressure). Further, the control center 2 can estimate the combustion state of the engine 3 based on the signal detected by the knock sensor 404 (the signal representing the sound or vibration generated in the cylinders 311-314). Furthermore, the control center 2 can also estimate the combustion state of the engine 3 based on signals detected by the exhaust temperature sensor 412, the NOx sensor 414 and/or the differential pressure sensor 415. Hereinafter, the data transmitted from the vehicle 1 to the control center 2 and collected by the control center 2 will be collectively referred to as "sensor data".

It should be noted that the sensors included in the sensors 4 shown in FIG. 5 are only examples. In other words, sensor data is not limited to those obtained from these sensors. The sensor data may be data obtained from some of these sensors, or may include data obtained from other sensors not shown in FIG.

Vehicle 1 (vehicle 1A and vehicle 1B) also transmits to server 23 data for identifying the gas station where vehicle 1 was refueled in addition to the sensor data. Hereinafter, data that can specify a gas station is also referred to as "gas station data". The gas station data specifically includes vehicle position data when the vehicle 1 is refueled. Since the installation location of the gas station is known to the control center 2 from the map information, the gas station can be identified from the vehicle position data at the time of refueling.

Moreover, even at the same gas station, the properties of the fuel to be refueled may change according to the refueling time (for example, the season). Therefore, it is preferable that the gas station data include data regarding the timing of refueling the vehicle 1 in addition to the vehicle position data when the vehicle 1 is refueled. Since the gas station data includes data (for example, time data) related to the timing of refueling, it is possible to associate the property of the refueled fuel with time as well. However, gas station data does not have to include data on refueling timing.

The memory 232 of the server 23 installed in the control center 2 stores prediction models constructed by machine learning and/or data mining regarding combustion states of various fuels. A prediction model is constructed for each vehicle type (model, grade, model year, etc.). The control center 2 preferably updates each prediction model to correspond to the latest fuel properties by continuing learning of each prediction model using sensor data and gas station data collected from the vehicle 1B. When the control center 2 receives the sensor data and the gas station data from the vehicle 1A, the control center 2 inputs the data into a prediction model corresponding to the vehicle type of the vehicle 1A and obtains an output from the prediction model. This makes it possible to improve the combustion state of the engine 3 of the vehicle 1A without using specific fuel composition information.

<Combustion prediction model>
FIG. 6 is a functional block diagram showing an example of a prediction model (combustion prediction model) regarding the combustion state of the engine 3 in this embodiment. Upon receiving the sensor data and the gas station data, the combustion prediction model outputs a correction amount according to the calculation rule obtained by machine learning, using the indices indicated by the sensor data and the gas station data as feature values.

The combustion prediction model is a trained model, such as a neural network model constructed by unsupervised learning or reinforcement learning. A neural network model includes an input layer, a hidden layer, and an output layer. An input variable to the input layer is described as y _i (i is a natural number). An output variable from the hidden layer to the output layer is denoted as θ _j (j is a natural number). An output variable taken out from the output layer is described as _Zp . The weight from the i-th node of the input layer to the j-th node of the hidden layer is denoted as _Wji . We denote the weight from the j-th node of the hidden layer to the node of the output layer as W _j '.

Input variables y _i are sensor data and gas station data collected from vehicle 1 . In the example shown in FIG. 6, the position information (vehicle position data) of the vehicle 1B acquired by the navigation system 11 is also included in the input variable _yi . The output variable _Zp is the amount of correction of the control command to the engine 3 . Machine learning is performed using this correction amount as teacher data.

The correction amount can be classified into, for example, a correction amount of the control command to the injection system, a correction amount of the control command to the intake system, and a correction amount of the control command to the EGR system. More specifically, the correction amount of the control command to the injection system is the amount of change from the reference amount of the fuel injection amount from the injector 323, the amount of change from the reference timing of the fuel injection timing from the injector 323, It is the amount of change from the reference timing of the ignition timing of the spark plug 324 and the like. The correction of the control command to the intake system includes the amount of change from the reference amount of the throttle opening, the amount of change from the reference amount of the opening of the WGV 262 (or the variable nozzle device 38), and the like. The correction amount of the control command to the EGR system is, for example, the amount of change from the reference amount of the opening of the EGR valve 372 . These correction amounts are transmitted to the vehicle 1A by the communication module 24 provided in the control center 2. FIG.

It is desirable to use vehicle position data in addition to sensor data and gas station data for prediction model learning. The vehicle position data reflects various information (weather, temperature, amount of rainfall, gradient, speed limit, presence or absence of traffic congestion, etc.) regarding the traveling point of the vehicle 1B. Therefore, by using the vehicle position data, it becomes possible to construct a prediction model learned with higher accuracy.

Note that the neural network model is merely an example of a technique that can be employed for machine learning of the combustion state of the engine 3, and the machine learning technique is not limited to this. The server 23 may use other machine learning models such as deep learning models, logistic regression models, support vector machines, for example.

Further, the server 23 may analyze the sensor data and gas station data collected by the control center 2 by a data mining technique for so-called big data. As a data mining method, for example, the following known method can be adopted. The server 23 uses, for example, the fuel injection timing from the injector 323 as an objective variable and other parameters as explanatory variables, and finds the relationship between the explanatory variables by, for example, clustering, thereby correcting the fuel injection timing. Select multiple explanatory variables to use. Then, in order to predict how the objective variable (fuel injection timing in this example) will change when the selected explanatory variable changes, the server 23 determines the relationship between the explanatory variable and the objective variable. Extract the regularity that The extracted regularity is expressed as a predictive model or relational expression of the objective variable (statistical modeling). As an example, when performing regression analysis (for example, multiple regression analysis) using a plurality of explanatory variables, the server 23 can calculate the parameters of the prediction model used for regression analysis by parameter fitting (fitting).

<Improvement of combustion state>
FIG. 7 is a diagram for explaining an example of combustion improvement processing for correcting the operation of the injection system of the engine 3. FIG. 7 and FIGS. 8 to 10, which will be described later, the horizontal axis represents the crank angle of the engine 3. As shown in FIG. The vertical axis represents the in-cylinder pressure of the cylinders 311 to 314 of the engine 3 .

Depending on the properties of the fuel, the in-cylinder pressure may deviate from the target in-cylinder pressure (target in-cylinder pressure) (see the in-cylinder pressure before correction in the figure). In such a case, for example, by correcting the operation of the injection system of the engine 3 using the correction amount, the in-cylinder pressure can be brought closer to the target in-cylinder pressure (see corrected in-cylinder pressure). FIG. 7 shows an example of retarding the fuel injection timing. The construction of the combustion prediction model (learning the model) in this example includes intake air volume, exhaust volume (difference between target and actual exhaust volume), unburned components in exhaust gas (hydrocarbons, etc.), ) and/or concentration of nitrogen oxides, ignition timing to the mixture, combustion noise, etc. can be used.

FIG. 8 is a diagram for explaining another example of the combustion improvement process for correcting the operation of the injection system of the engine 3. FIG. The in-cylinder pressure may deviate from the target in-cylinder pressure even when the ignitability of the fuel supplied to the engine 3 is better or worse than the ignitability of typical fuel. While the amount of fuel injected from the injector 323 is adjusted by changing the magnitude of the pulse (pulse height/pulse width) output from the ECU 13 to the injector 323, the magnitude of the pulse is corrected using the correction amount. By doing so, the in-cylinder pressure can be brought close to the target in-cylinder pressure. In multi-injection, it is desirable to correct the fuel injection amount for main combustion.

FIG. 9 is a diagram for explaining an example of combustion improvement processing for correcting the operation of the intake system of the engine 3. FIG. Another factor that causes the in-cylinder pressure to deviate from the target in-cylinder pressure is the amount of air intake to the engine 3 . In such a case, the in-cylinder pressure can be brought closer to the target in-cylinder pressure by correcting the intake air amount by adjusting the opening of the throttle valve 332 and/or the WGV 362 (variable nozzle device 38) using the correction amount.

FIG. 10 is a diagram for explaining an example of combustion improvement processing for correcting the operation of the EGR system of the engine 3. FIG. The amount of EGR gas can also be a factor causing the in-cylinder pressure to deviate from the target in-cylinder pressure. In such a case, the in-cylinder pressure can be brought close to the target in-cylinder pressure by correcting the EGR gas amount by adjusting the opening degree of the EGR valve 372 using the correction amount.

7 to 10, an example in which the in-cylinder pressure is brought close to the target in-cylinder pressure has been described, but the correction amount is used to bring other parameters (for example, the heat release rate of the engine 3) closer to the target value. Combustion conditions may be improved.

<Processing flow>
FIG. 11 is a flow chart showing the procedure of the combustion improvement process in this embodiment. This flowchart is executed when a predetermined condition is satisfied (for example, every predetermined control period while the engine 3 is operating such as when the vehicle 1A is running). In the figure, the left side shows the processing executed in the vehicle 1A, and the right side shows the processing executed in the control center 2. As shown in FIG. Each step is realized by software processing by vehicle 1A (ECU 13) or control center 2 (server 23), but may be realized by hardware (electric circuit) arranged in ECU 13 or server 23. A step is abbreviated as S.

In this example, the ECU 13 uses flags to control transmission of sensor data, vehicle position data, and gas station data to the control center 2 . This flag is described as a "data transmission flag".

In S11, the ECU 13 determines whether or not the vehicle 1A has been refueled. The ECU 13 can determine that refueling has been performed when the fuel gauge detects an increase in the amount of fuel in the fuel tank (not shown). If refueling has been performed (YES in S11), the ECU 23 turns on the data transmission flag (S12). Then, the ECU 23 transmits the gas station data to the control center 2 (S13). Gas station data is stored in the travel history database 22 . When vehicle 1A is not being refueled (NO in S11), the processes of S12 and S13 are skipped and the data transmission flag is kept off.

In S14, the ECU 13 determines whether the data transmission flag is ON. If the data transmission flag is ON (YES in S14), ECU 13 transmits sensor data and vehicle position data to control center 2 (S15). Since the processing of the example on the left side of the figure is executed each time a predetermined condition is satisfied (for example, every control cycle), the processing of S15 is repeatedly executed. That is, sensor data and vehicle position data are continuously transmitted until the required amount of data has been collected (described below). These sensor data and vehicle position data are stored in the travel history database 22 in association with the gas station data 13 collected in S13. When the data transmission flag is off (NO in S14), transmission of sensor data and vehicle position data from vehicle 1A to control center 2 is not performed.

In S21, the server 23 determines whether or not the amount of data required to calculate the correction amount using the combustion prediction model has been collected. When the collection of necessary data is completed (YES in S21), server 23 applies the collected data (sensor data, vehicle position data and gas station data) to the combustion prediction model to calculate a correction amount (S22). . The server 23 then transmits the calculated correction amount to the vehicle 1A (S23). It should be noted that if the correction amount is out of the normally assumed range, there is a possibility that inappropriate fuel (for example, non-standard fuel) is supplied. In this case, the server 23 may transmit a warning to the vehicle 1A instead of or in addition to the correction amount.

In S<b>16 , the ECU 13 determines whether or not the correction amount has been received from the control center 2 . When the correction amount is received (YES in S16), the ECU 13 switches the data transmission flag from on to off. Further, the ECU 13 corrects the operation of the injection system, EGR system and/or intake system of the engine 3 using the correction amount, as described with reference to FIGS. 7 to 10 (S18). As a result, the fuel condition of the engine 3 is improved.

As described above, in the present embodiment, by using the combustion prediction model, which is a prediction model regarding the combustion state of fuel, the engine 3 of the vehicle 1A can be determined from the sensor data, the vehicle position data, and the gas station data of the vehicle 1A. A correction amount for correcting the combustion state is calculated. The gas station data is data for specifying the gas station, and is not data relating to the analysis result of the composition of the supplied fuel. According to this embodiment, the combustion state of the engine 3 can be appropriately corrected even if the fuel composition information is not obtained. In addition, the occurrence of combustion problems (for example, misfires or abnormal combustion noises) in the engine 3 can be suppressed.

The properties of the fuel in the fuel tank may change each time refueling is performed. However, the combustion prediction model is constructed using big data representing detection results of combustion states of various fuels collected from a large number of vehicles 1B. Therefore, according to this embodiment, it is possible to appropriately correct the combustion state of the engine 3 not only for a specific fuel but also for a wide variety of fuels.

Also, in general, the computational processing capability of an in-vehicle ECU is limited. By using the server 23 installed in the control center 2, it is possible to easily ensure a high level of computational processing capability and reduce the computational load of the ECU 13. FIG. Further, it is possible to reduce the calculation load of the server 23 by learning the correction amount by the server 23 .

The embodiments disclosed this time should be considered as examples and not restrictive in all respects. The scope of the present disclosure is indicated by the scope of claims rather than the description of the above-described embodiments, and is intended to include all modifications within the scope and meaning equivalent to the scope of the claims.

100 vehicle management system 1 (1A, 1B) vehicle 11 navigation system 111 GPS receiver 13 ECU 131 processor 132 memory 133 input/output port 14 in-vehicle network 2 control center 21 vehicle database 22 traveling history database, 23 server, 231 processor, 232 memory, 233 input/output port, 24 communication module, 25 intranet, 3, 3A engine, 31 engine body, 311 to 314 cylinders, 321 intake valve, 322 exhaust valve, 323 injector, 324 Spark plug, 33 intake passage, 331 intercooler, 332 throttle valve, 34 exhaust passage, 341 catalyst device, 342 aftertreatment device, 35 supercharger, 351 compressor, 352 turbine, 353 shaft, 36 WGV device, 361 bypass passage, 362 WGV, 363 WGV actuator, 37 EGR device, 371 EGR passage, 372 EGR valve, 373 EGR cooler, 38 variable noise device, 4 sensors, 401 vehicle speed sensor, 402 accelerator opening sensor, 403 engine rotation speed sensor, 404 knock Sensor 405 Airflow meter 406 Throttle opening sensor 407 In-cylinder pressure sensor 408 Cooling water temperature sensor 409 Supercharging pressure sensor 410 Supercharging temperature sensor 411 Turbine rotation speed sensor 412 Exhaust gas temperature sensor 413 Air-fuel ratio sensor 414 sensor, 415 differential pressure sensor.

Claims (6)

target vehicle and
a server configured to communicate with the target vehicle and a plurality of vehicles;
each of the target vehicle and the plurality of vehicles includes at least one sensor that detects a combustion state of fuel in an engine;
The server is
a memory in which a learned model related to engine control is stored;
a processor that executes arithmetic processing using the trained model,
The trained model includes sensor data indicating detection results of the at least one sensor of each of the plurality of vehicles, and gas station data capable of identifying a gas station where each of the plurality of vehicles is supplied with fuel. A correction amount for correcting the fuel combustion state in the engine is learned based on
The processor calculates the correction amount by giving the sensor data and the gas station data of the target vehicle to the learned model, transmits the calculated correction amount to the target vehicle,
The vehicle management system, wherein the target vehicle corrects a combustion state of fuel in the engine of the target vehicle using the received correction amount. 2. The vehicle management system according to claim 1, wherein said gas station data includes data regarding installation locations of said gas stations and data regarding refueling times at said gas stations. In the learned model, the correction amount is learned based on the sensor data, the gas station data, and data relating to the traveling positions of the plurality of vehicles,
3. The processor calculates the correction amount of the target vehicle by providing the learned model with data relating to the traveling position of the target vehicle in addition to the sensor data and the gas station data. The vehicle management system described in . The sensor data includes the in-cylinder pressure of the engine, the rotation speed of the engine, the temperature of exhaust gas from the engine, the pressure difference before and after a filter that collects particulate matter in the exhaust gas, and the exhaust gas Vehicle management system according to any one of claims 1 to 3, comprising data relating to at least one of nitrogen oxide concentrations in the vehicle. A vehicle configured to communicate with a server containing a trained model for engine control,
In the learned model, a correction amount for correcting the fuel combustion state in the engine is learned based on sensor data and gas station data collected through communication with a plurality of vehicles each having an engine mounted thereon. and
The sensor data is data indicating a detection result of a fuel combustion state in the engine of each of the plurality of vehicles,
The gas station data is data capable of identifying a gas station where fuel is supplied to the plurality of vehicles,
The vehicle is
The sensor data of the vehicle and the gas station data are transmitted to the server, and the correction amount calculated by the server giving the sensor data and the gas station data from the vehicle to the learned model is calculated. a receiving communicator;
and a control device that corrects a combustion state of fuel in the engine of the vehicle using the received correction amount. A vehicle driving management method for managing driving of a vehicle using a learned model related to engine control,
In the learned model, a correction amount for correcting the fuel combustion state in the engine is learned based on sensor data and gas station data collected through communication with a plurality of vehicles each having an engine mounted thereon. and
The sensor data is data indicating a detection result of a fuel combustion state in the engine of each of the plurality of vehicles,
The gas station data is data capable of identifying a gas station where fuel is supplied to the plurality of vehicles,
The running management method includes:
a step of calculating the correction amount for correcting the combustion state of fuel in the engine of the target vehicle by providing the sensor data and the gas station data from the target vehicle to the learned model;
and a step of correcting a combustion state of fuel in the engine of the target vehicle using the correction amount.

Images