EP4133171A1 - System and method for predicting high frequency emission information of an engine - Google Patents

System and method for predicting high frequency emission information of an engine

Info

Publication number
EP4133171A1
EP4133171A1 EP20718275.9A EP20718275A EP4133171A1 EP 4133171 A1 EP4133171 A1 EP 4133171A1 EP 20718275 A EP20718275 A EP 20718275A EP 4133171 A1 EP4133171 A1 EP 4133171A1
Authority
EP
European Patent Office
Prior art keywords
engine
emission
high frequency
emission information
model
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
EP20718275.9A
Other languages
German (de)
French (fr)
Inventor
Camila CHOVET
Julie LE LOUVETEL-POILLY
Shota NAGANO
François LAFOSSAS
Rémi LOSERO
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Toyota Motor Corp
Original Assignee
Toyota Motor Europe NV SA
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Toyota Motor Europe NV SA filed Critical Toyota Motor Europe NV SA
Publication of EP4133171A1 publication Critical patent/EP4133171A1/en
Pending legal-status Critical Current

Links

Classifications

    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F02—COMBUSTION ENGINES; HOT-GAS OR COMBUSTION-PRODUCT ENGINE PLANTS
    • F02D—CONTROLLING COMBUSTION ENGINES
    • F02D41/00—Electrical control of supply of combustible mixture or its constituents
    • F02D41/02—Circuit arrangements for generating control signals
    • F02D41/14—Introducing closed-loop corrections
    • F02D41/1438—Introducing closed-loop corrections using means for determining characteristics of the combustion gases; Sensors therefor
    • F02D41/1444—Introducing closed-loop corrections using means for determining characteristics of the combustion gases; Sensors therefor characterised by the characteristics of the combustion gases
    • F02D41/1452—Introducing closed-loop corrections using means for determining characteristics of the combustion gases; Sensors therefor characterised by the characteristics of the combustion gases the characteristics being a COx content or concentration
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F02—COMBUSTION ENGINES; HOT-GAS OR COMBUSTION-PRODUCT ENGINE PLANTS
    • F02D—CONTROLLING COMBUSTION ENGINES
    • F02D41/00—Electrical control of supply of combustible mixture or its constituents
    • F02D41/0097—Electrical control of supply of combustible mixture or its constituents using means for generating speed signals
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F02—COMBUSTION ENGINES; HOT-GAS OR COMBUSTION-PRODUCT ENGINE PLANTS
    • F02D—CONTROLLING COMBUSTION ENGINES
    • F02D41/00—Electrical control of supply of combustible mixture or its constituents
    • F02D41/02—Circuit arrangements for generating control signals
    • F02D41/04—Introducing corrections for particular operating conditions
    • F02D41/06—Introducing corrections for particular operating conditions for engine starting or warming up
    • F02D41/062—Introducing corrections for particular operating conditions for engine starting or warming up for starting
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F02—COMBUSTION ENGINES; HOT-GAS OR COMBUSTION-PRODUCT ENGINE PLANTS
    • F02D—CONTROLLING COMBUSTION ENGINES
    • F02D41/00—Electrical control of supply of combustible mixture or its constituents
    • F02D41/02—Circuit arrangements for generating control signals
    • F02D41/14—Introducing closed-loop corrections
    • F02D41/1438—Introducing closed-loop corrections using means for determining characteristics of the combustion gases; Sensors therefor
    • F02D41/1444—Introducing closed-loop corrections using means for determining characteristics of the combustion gases; Sensors therefor characterised by the characteristics of the combustion gases
    • F02D41/1452—Introducing closed-loop corrections using means for determining characteristics of the combustion gases; Sensors therefor characterised by the characteristics of the combustion gases the characteristics being a COx content or concentration
    • F02D41/1453—Introducing closed-loop corrections using means for determining characteristics of the combustion gases; Sensors therefor characterised by the characteristics of the combustion gases the characteristics being a COx content or concentration the characteristics being a CO content or concentration
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F02—COMBUSTION ENGINES; HOT-GAS OR COMBUSTION-PRODUCT ENGINE PLANTS
    • F02D—CONTROLLING COMBUSTION ENGINES
    • F02D41/00—Electrical control of supply of combustible mixture or its constituents
    • F02D41/02—Circuit arrangements for generating control signals
    • F02D41/14—Introducing closed-loop corrections
    • F02D41/1438—Introducing closed-loop corrections using means for determining characteristics of the combustion gases; Sensors therefor
    • F02D41/1444—Introducing closed-loop corrections using means for determining characteristics of the combustion gases; Sensors therefor characterised by the characteristics of the combustion gases
    • F02D41/1459—Introducing closed-loop corrections using means for determining characteristics of the combustion gases; Sensors therefor characterised by the characteristics of the combustion gases the characteristics being a hydrocarbon content or concentration
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F02—COMBUSTION ENGINES; HOT-GAS OR COMBUSTION-PRODUCT ENGINE PLANTS
    • F02D—CONTROLLING COMBUSTION ENGINES
    • F02D41/00—Electrical control of supply of combustible mixture or its constituents
    • F02D41/02—Circuit arrangements for generating control signals
    • F02D41/14—Introducing closed-loop corrections
    • F02D41/1438—Introducing closed-loop corrections using means for determining characteristics of the combustion gases; Sensors therefor
    • F02D41/1444—Introducing closed-loop corrections using means for determining characteristics of the combustion gases; Sensors therefor characterised by the characteristics of the combustion gases
    • F02D41/146—Introducing closed-loop corrections using means for determining characteristics of the combustion gases; Sensors therefor characterised by the characteristics of the combustion gases the characteristics being an NOx content or concentration
    • F02D41/1461—Introducing closed-loop corrections using means for determining characteristics of the combustion gases; Sensors therefor characterised by the characteristics of the combustion gases the characteristics being an NOx content or concentration of the exhaust gases emitted by the engine
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F02—COMBUSTION ENGINES; HOT-GAS OR COMBUSTION-PRODUCT ENGINE PLANTS
    • F02D—CONTROLLING COMBUSTION ENGINES
    • F02D41/00—Electrical control of supply of combustible mixture or its constituents
    • F02D41/02—Circuit arrangements for generating control signals
    • F02D41/14—Introducing closed-loop corrections
    • F02D41/1438—Introducing closed-loop corrections using means for determining characteristics of the combustion gases; Sensors therefor
    • F02D41/1444—Introducing closed-loop corrections using means for determining characteristics of the combustion gases; Sensors therefor characterised by the characteristics of the combustion gases
    • F02D41/146—Introducing closed-loop corrections using means for determining characteristics of the combustion gases; Sensors therefor characterised by the characteristics of the combustion gases the characteristics being an NOx content or concentration
    • F02D41/1461—Introducing closed-loop corrections using means for determining characteristics of the combustion gases; Sensors therefor characterised by the characteristics of the combustion gases the characteristics being an NOx content or concentration of the exhaust gases emitted by the engine
    • F02D41/1462—Introducing closed-loop corrections using means for determining characteristics of the combustion gases; Sensors therefor characterised by the characteristics of the combustion gases the characteristics being an NOx content or concentration of the exhaust gases emitted by the engine with determination means using an estimation
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F02—COMBUSTION ENGINES; HOT-GAS OR COMBUSTION-PRODUCT ENGINE PLANTS
    • F02D—CONTROLLING COMBUSTION ENGINES
    • F02D41/00—Electrical control of supply of combustible mixture or its constituents
    • F02D41/02—Circuit arrangements for generating control signals
    • F02D41/14—Introducing closed-loop corrections
    • F02D41/1401—Introducing closed-loop corrections characterised by the control or regulation method
    • F02D2041/1412—Introducing closed-loop corrections characterised by the control or regulation method using a predictive controller
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F02—COMBUSTION ENGINES; HOT-GAS OR COMBUSTION-PRODUCT ENGINE PLANTS
    • F02D—CONTROLLING COMBUSTION ENGINES
    • F02D41/00—Electrical control of supply of combustible mixture or its constituents
    • F02D41/02—Circuit arrangements for generating control signals
    • F02D41/14—Introducing closed-loop corrections
    • F02D41/1401—Introducing closed-loop corrections characterised by the control or regulation method
    • F02D2041/1433—Introducing closed-loop corrections characterised by the control or regulation method using a model or simulation of the system
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F02—COMBUSTION ENGINES; HOT-GAS OR COMBUSTION-PRODUCT ENGINE PLANTS
    • F02D—CONTROLLING COMBUSTION ENGINES
    • F02D2250/00—Engine control related to specific problems or objectives
    • F02D2250/12—Timing of calculation, i.e. specific timing aspects when calculation or updating of engine parameter is performed
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F02—COMBUSTION ENGINES; HOT-GAS OR COMBUSTION-PRODUCT ENGINE PLANTS
    • F02D—CONTROLLING COMBUSTION ENGINES
    • F02D2250/00—Engine control related to specific problems or objectives
    • F02D2250/14—Timing of measurement, e.g. synchronisation of measurements to the engine cycle

Definitions

  • the present disclosure is related to a system and method for predicting high frequency emission information of an engine, in particular as a function of an emission measurement having a low sampling frequency.
  • Fast emission analyser devices are better representing engine behaviour slow emission analyser devices. However, since slow emission analyser devices are significantly less expensive, they are often used instead.
  • the characteristics "fast” and “slow” thereby refer to the accuracy of the analyser device, in particular to accurately measure high-frequency emission dynamics, ,or in particular the capability of the emission analyser device to obtain said emission dynamics. For example said accuracy can be represented by the sampling frequency of the respective emission measurement device.
  • CA2363378A1 From CA2363378A1 a method for analyzing the exhaust emissions of large industrial engines is known. Emission information is produced in real time and permits the generation of test results immediately after an emission test is conducted. The method includes the real time calculation of exhaust volumetric flow rate from fuel gas flowrate and the use of real time intake manifold conditions to determine engine load from an engine load curve.
  • WO9910728A2 relates to an IR analyze apparatus of the most important essential to environment substance such as CO, HC and NO comprised by vehicle emission based on the (ultra-red) gaseous absorption principle.
  • a system for predicting high frequency emission information of an engine comprises: a control device being configured to receive low frequency emission information from an emission measurement device, said emission measurement device having a first sampling frequency equal to or below a first predetermined sampling threshold configured to the output low frequency emission information having the first sampling frequency.
  • the control device is further configured to execute a model. Said model, when executed by the control device, predicts high frequency emission information as a function of the low frequency emission information and at least one predetermined high frequency engine input parameter, the predicted high frequency emission information having at least a second sampling frequency above the first predetermined sampling threshold.
  • the system is desirably able to predict high frequency (e.g. 100Hz) emissions from slower analyser measurements, which are strongly linked to engine behavior.
  • the system may use non-linear and dynamic machine-learning based models to predict e.g. CO, HC and NOx predictions.
  • Results may be coupled with an engine out emission model giving high frequency results. Also, a deep understanding of engine out emissions during engine restart is possible, in order to support engine development and calibration. The system may therefore replace fast analyser measurements.
  • Low frequency emission information refers to e.g. around l-2Hz measurements. Also, the frequency emission information could be referred to as lower as emission dynamics, which means that measurements cannot fully catch emission dynamics.
  • the system may further comprise a speed sensor configured to measure the engine speed, wherein the model predicts high frequency emission information additionally as a function of the engine speed.
  • the system may further comprises an air-fuel ratio sensor configured to measure the air-fuel ratio of the engine, wherein the model predicts high frequency emission information additionally as a function of the air-fuel ratio.
  • the system may further be configured to receive the at least one engine input parameter or comprises a sensor configured to measure the at least one engine input parameter, wherein the engine input parameter comprises at least one of: volumetric efficiency, engine speed, and air mass flow.
  • the measured low frequency emission information may comprise at least one of: carbon oxide (CO) emission, nitrogen oxide (NOx) emission and hydrocarbon (HC) emission.
  • the predicted high frequency emission information may comprise respectively at least one of: carbon oxide (CO) emission, nitrogen oxide (NOx) emission and hydrocarbon (HC) emission.
  • the model may comprise a machine learning model, in particular a regression model and/or a linear model.
  • the first predetermined sampling threshold may be between 2Hz and 10Hz.
  • the first sampling frequency may be between 1 and 2Hz.
  • the predicted high frequency emission information may have a third sampling frequency above a second predetermined sampling threshold which is above the first predetermined sampling threshold.
  • the second predetermined sampling threshold may be 100Hz.
  • the present disclosure further relates to a method of predicting high frequency emission information of an engine, comprising the steps of: providing a system according to any one of the preceding claims, providing a test vehicle comprising an engine, in particular a hybrid vehicle comprising an engine and at least one electrical motor, operating the vehicle by restarting the engine a plurality of times for limited time periods, in particular less than 10s, predicting high frequency emission information of the engine by means of the system.
  • the present disclosure further relates to a method of training a model for predicting high frequency emission information of an engine, comprising the steps of: providing a training data set of an emission pattern of a test vehicle, said data set comprising low frequency emission information having a first sampling frequency of the emission pattern equal to or below the first predetermined sampling threshold, and high frequency emission information having at least a second sampling frequency of the emission pattern above the first predetermined sampling threshold, and train the model (in a supervised fashion) by using the low frequency emission information as input for the model and the high frequency emission information as target output.
  • the emission pattern of a test vehicle may hence comprise low frequency emission information and the corresponding high frequency emission information, i.e. which are identical except their frequency resolution.
  • FIG. 1 shows a block diagram of a system with a control device according to embodiments of the present disclosure
  • FIG. 2 shows a schematic pipeline for training the model according to embodiments of the present disclosure
  • Fig. 3 shows an exemplary schematic diagram of the difference between high frequency emission measurements and respective low frequency emission measurements both representing the same measured high frequency emission information (e.g. CO emissions of a test vehicle) according to embodiments of the present disclosure, and
  • Fig. 4 shows an exemplary schematic diagram of the high frequency and low frequency emission measurements and a predicted high frequency emission according to embodiments of the present disclosure.
  • Fig. 1 shows a block diagram of a system with a control device according to embodiments of the present disclosure.
  • the system 10 is configured to predict high frequency emission information of an engine 20 (e.g. of a vehicle which is external to the system).
  • the system comprises a control device 1 being configured to receive low frequency emission information from an emission measurement device 3.
  • Said emission measurement device is preferably also comprises by the system or is connected to the system.
  • the emission measurement device 3 takes measurements of the engine 30 with a relatively low first sampling frequency equal to or below a first predetermined sampling threshold.
  • the emission measurement device 3 is further configured to output low frequency emission information having the first sampling frequency.
  • the control device is further configured to execute a trained model.
  • Said model predicts high frequency emission information as a function of the low frequency emission information and at least one predetermined high frequency engine input parameter.
  • the predicted high frequency emission information has at least a second sampling frequency above the first predetermined sampling threshold.
  • the high frequency emission information may have a frequency of 100Hz or more.
  • the control device 1 is connected to or comprises a data storage 2.
  • Said data storage may be used to store e.g. the trained model, in particular a combined regression and linear model, and/or an algorithm for calculating the predicted high frequency emission information.
  • the system may also be connected to an external server 20 which provides the model and/or training data to train the model.
  • the control device 1 may comprise an electronic circuit, a processor (shared, dedicated, or group), a combinational logic circuit, a memory that executes one or more software programs, and/or other suitable components that provide the described functionality.
  • the emission measurement device 3 measures e.g. at least one of carbon oxide (CO) emission, nitrogen oxide (NOx) emission and hydrocarbon (HC) emission of the engine.
  • the emission measurement device 3 may be a relatively inexpensive product with a reduced sampling frequency of e.g. a value between only 1 to 2 Hz.
  • the control device 1 may additionally carry out further monitoring functions of the engine and/or the vehicle.
  • the system may e.g. comprise a speed sensor 4 configured to measure the engine speed, wherein the model predicts high frequency emission information additionally as a function of the engine speed.
  • the system may further comprise an air-fuel ratio sensor 5 configured to measure the air-fuel ratio of the engine, wherein the model predicts high frequency emission information additionally as a function of the air-fuel ratio.
  • the emission measurement device 3 may be a conventional slow analyser, e.g. MEXA Horiba. Such analysers are not able to accurately measure the restart emission up to 5s, while fast analyser (e.g.: Cambustion) correctly measures emissions for all engine states. However the latter, is extremely expensive.
  • the present disclosure leverages dynamic machine learning based and non-linear models to predict fast analyser behaviour from slow analyser measurements. This model could thus replace fast analyser measurements.
  • Fig. 2 shows a schematic pipeline for training the model according to embodiments of the present disclosure.
  • the model of the present disclosure may be referred to as a slow to fast analyser model.
  • Said model is trained by inputting engine inputs, i.e. at least one predetermined high frequency engine input parameter, and slow measurements, i.e. low frequency emission information.
  • the model may be trained in a supervised fashion by using respective fast emission measurements, i.e. high frequency emission information as target output for the model. The model correspondingly learns to predict such high frequency emission information.
  • a fast to slow analyser model may (as also shown in fig. 2) be used to validate the predictions of the slow to fast analyser model with slow analyser measurements.
  • the slow to fast analyser could also be integrated into engine model estimation models or engine restart models.
  • high frequency emission measurements REF_Fast e.g. with 100 Hz or more
  • the main inputs of the model are slow analyser measurements and optionally engine speed and/or air fuel ratio sensor value.
  • the engine speed can be used as an input to identify engine operation.
  • CO emissions are very sensitive to Air fuel ratio value, considering A/F (air-fuel) as input can support better prediction of CO high frequency behavior.
  • the model is trained using measurement data.
  • the predicted outputs of the model are high frequency CO, NOx and HC emissions.
  • Fig. 4 shows an exemplary schematic diagram of the measured high frequency emission REF_Fast (first dashed line) and low frequency emission measurements REF_Slow (second dashed line) and predicted high frequency emission SIM (continuous line) according to embodiments of the present disclosure.
  • the y-axis in fig. 3 and 4 may correspond to the CO emission (e.g. in g/s).
  • the predicted high frequency emission SIM is close to the actually measured frequency emission measurements REF_Fast.
  • this model may provide dynamic behavior of emission predictions within e.g. ⁇ 7% deviation for Worldwide Flarmonized Light Vehicle Test cycle (WLTC) and Real Driving Emissions compliant cycles.
  • WLTC Worldwide Flarmonized Light Vehicle Test cycle
  • the simulation time of this model is about e.g. 0.02 times the real time.
  • This model can be thus used to deeply understand the behavior of engine out emission at high frequency.
  • This model could therefore replace fast analyser measurements. Having a predicted high frequency response can provide a deep understanding of engine out emissions during engine restart. Such findings may be of interest e.g. for engine development and calibration support.

Landscapes

  • Engineering & Computer Science (AREA)
  • Chemical & Material Sciences (AREA)
  • Combustion & Propulsion (AREA)
  • Mechanical Engineering (AREA)
  • General Engineering & Computer Science (AREA)
  • Combined Controls Of Internal Combustion Engines (AREA)

Abstract

The invention relates to a system for predicting high frequency emission information of an engine is provided. The system comprises a control device being configured to receive low frequency emission information from an emission measurement device, said emission measurement device having a first sampling frequency equal to or below a first predetermined sampling threshold configured to the output low frequency emission information having the first sampling frequency. The control device is further configured to execute a model. Said model, when executed by the control device, predicts high frequency emission information as a function of the low frequency emission information and at least one predetermined high frequency engine input parameter, the predicted high frequency emission information having at least a second sampling frequency above the first predetermined sampling threshold. The invention further relates to a corresponding prediction method and a training method.

Description

System and method for predicting high frequency emission information of an engine
FIELD OF THE DISCLOSURE [0001] The present disclosure is related to a system and method for predicting high frequency emission information of an engine, in particular as a function of an emission measurement having a low sampling frequency.
BACKGROUND OF THE DISCLOSURE
[0002] During development, testing (or even use) of a vehicle it can be desired to continuously measure the emission caused by the vehicle, i.e. by its internal combustion engine. In this context, it is in particular interesting to simulate real driving conditions during the measurements. [0003] However, under real driving conditions the number of short restarts of a vehicle power train (i.e. engine restarts) is very high, in particular in case of a hybrid vehicle. For this reason, an accurate prediction of the restarts emission is important. Hybrid vehicles may have in conventional use about 50% of engine ON events which are shorter than 9s. To have an accurate and reliable estimation of engine emissions and tailpipe emissions (e.g. during development or testing of a vehicle), the estimation of the restart emissions should be accurate. Fast emission analyser devices are better representing engine behaviour slow emission analyser devices. However, since slow emission analyser devices are significantly less expensive, they are often used instead. [0004] The characteristics "fast" and "slow" thereby refer to the accuracy of the analyser device, in particular to accurately measure high-frequency emission dynamics, ,or in particular the capability of the emission analyser device to obtain said emission dynamics. For example said accuracy can be represented by the sampling frequency of the respective emission measurement device.
[0005] From CA2363378A1 a method for analyzing the exhaust emissions of large industrial engines is known. Emission information is produced in real time and permits the generation of test results immediately after an emission test is conducted. The method includes the real time calculation of exhaust volumetric flow rate from fuel gas flowrate and the use of real time intake manifold conditions to determine engine load from an engine load curve.
[0006] WO9910728A2 relates to an IR analyze apparatus of the most important essential to environment substance such as CO, HC and NO comprised by vehicle emission based on the (ultra-red) gaseous absorption principle.
SUMMARY OF THE DISCLOSURE
[0007] Currently, it remains desirable to provide a system and method for reliably predicting high frequency emission information of an engine but at reduced costs. More in particular, it remains desirable to provide a system and method for reliably for predicting high frequency emission information of an engine by using an emission measurement device having a relatively low sampling frequency. [0008] The present disclosure resolves these issues desirably by using dynamic machine learning based and/or non-linear models to predict fast analyser behavior from slow analyser measurements, in order to have a deeper understanding on high frequency emission.
[0009] Therefore, according to the embodiments of the present disclosure, a system for predicting high frequency emission information of an engine is provided. The system comprises: a control device being configured to receive low frequency emission information from an emission measurement device, said emission measurement device having a first sampling frequency equal to or below a first predetermined sampling threshold configured to the output low frequency emission information having the first sampling frequency. The control device is further configured to execute a model. Said model, when executed by the control device, predicts high frequency emission information as a function of the low frequency emission information and at least one predetermined high frequency engine input parameter, the predicted high frequency emission information having at least a second sampling frequency above the first predetermined sampling threshold.
[0010] Accordingly, the system is desirably able to predict high frequency (e.g. 100Hz) emissions from slower analyser measurements, which are strongly linked to engine behavior. The system may use non-linear and dynamic machine-learning based models to predict e.g. CO, HC and NOx predictions.
By providing such a system is becomes possible to provide high frequency emission estimations (e.g. 100Hz or more). Results may be coupled with an engine out emission model giving high frequency results. Also, a deep understanding of engine out emissions during engine restart is possible, in order to support engine development and calibration. The system may therefore replace fast analyser measurements.
[0011] Low frequency emission information refers to e.g. around l-2Hz measurements. Also, the frequency emission information could be referred to as lower as emission dynamics, which means that measurements cannot fully catch emission dynamics.
[0012] The system may further comprise a speed sensor configured to measure the engine speed, wherein the model predicts high frequency emission information additionally as a function of the engine speed.
[0013] The system may further comprises an air-fuel ratio sensor configured to measure the air-fuel ratio of the engine, wherein the model predicts high frequency emission information additionally as a function of the air-fuel ratio.
[0014] The system may further be configured to receive the at least one engine input parameter or comprises a sensor configured to measure the at least one engine input parameter, wherein the engine input parameter comprises at least one of: volumetric efficiency, engine speed, and air mass flow.
[0015] The measured low frequency emission information may comprise at least one of: carbon oxide (CO) emission, nitrogen oxide (NOx) emission and hydrocarbon (HC) emission.
[0016] Also the predicted high frequency emission information may comprise respectively at least one of: carbon oxide (CO) emission, nitrogen oxide (NOx) emission and hydrocarbon (HC) emission.
[0017] The model may comprise a machine learning model, in particular a regression model and/or a linear model.
[0018] The first predetermined sampling threshold may be between 2Hz and 10Hz.
[0019] The first sampling frequency may be between 1 and 2Hz. [0020] The predicted high frequency emission information may have a third sampling frequency above a second predetermined sampling threshold which is above the first predetermined sampling threshold.
[0021] The second predetermined sampling threshold may be 100Hz.
[0022] The present disclosure further relates to a method of predicting high frequency emission information of an engine, comprising the steps of: providing a system according to any one of the preceding claims, providing a test vehicle comprising an engine, in particular a hybrid vehicle comprising an engine and at least one electrical motor, operating the vehicle by restarting the engine a plurality of times for limited time periods, in particular less than 10s, predicting high frequency emission information of the engine by means of the system.
[0023] The present disclosure further relates to a method of training a model for predicting high frequency emission information of an engine, comprising the steps of: providing a training data set of an emission pattern of a test vehicle, said data set comprising low frequency emission information having a first sampling frequency of the emission pattern equal to or below the first predetermined sampling threshold, and high frequency emission information having at least a second sampling frequency of the emission pattern above the first predetermined sampling threshold, and train the model (in a supervised fashion) by using the low frequency emission information as input for the model and the high frequency emission information as target output.
[0024] The emission pattern of a test vehicle may hence comprise low frequency emission information and the corresponding high frequency emission information, i.e. which are identical except their frequency resolution.
[0025] For training the model one may use an existing dataset (e.g. NEDYC RDE, WLTC) containing both low frequency measurement system and at the same time with an high frequency measurement system.
[0026] It is intended that combinations of the above-described elements and those within the specification may be made, except where otherwise contradictory. [0027] It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure, as claimed.
[0028] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments of the disclosure and together with the description, and serve to explain the principles thereof.
BRIEF DESCRIPTION OF THE DRAWINGS
[0029] Fig. 1 shows a block diagram of a system with a control device according to embodiments of the present disclosure;
[0030] Fig. 2 shows a schematic pipeline for training the model according to embodiments of the present disclosure;
[0031] Fig. 3 shows an exemplary schematic diagram of the difference between high frequency emission measurements and respective low frequency emission measurements both representing the same measured high frequency emission information (e.g. CO emissions of a test vehicle) according to embodiments of the present disclosure, and
[0032] Fig. 4 shows an exemplary schematic diagram of the high frequency and low frequency emission measurements and a predicted high frequency emission according to embodiments of the present disclosure.
DESCRIPTION OF THE EMBODIMENTS
[0033] Reference will now be made in detail to exemplary embodiments of the disclosure, examples of which are illustrated in the accompanying drawings. Wherever possible, the same reference numbers will be used throughout the drawings to refer to the same or like parts.
[0034] Fig. 1 shows a block diagram of a system with a control device according to embodiments of the present disclosure.
[0035] The system 10 is configured to predict high frequency emission information of an engine 20 (e.g. of a vehicle which is external to the system). For this purpose the system comprises a control device 1 being configured to receive low frequency emission information from an emission measurement device 3. Said emission measurement device is preferably also comprises by the system or is connected to the system. The emission measurement device 3 takes measurements of the engine 30 with a relatively low first sampling frequency equal to or below a first predetermined sampling threshold. [0036] The emission measurement device 3 is further configured to output low frequency emission information having the first sampling frequency. The control device is further configured to execute a trained model. Said model predicts high frequency emission information as a function of the low frequency emission information and at least one predetermined high frequency engine input parameter. The predicted high frequency emission information has at least a second sampling frequency above the first predetermined sampling threshold. Hence, the high frequency emission information may have a frequency of 100Hz or more.
[0037] The control device 1 is connected to or comprises a data storage 2. Said data storage may be used to store e.g. the trained model, in particular a combined regression and linear model, and/or an algorithm for calculating the predicted high frequency emission information. The system may also be connected to an external server 20 which provides the model and/or training data to train the model.
[0038] The control device 1 may comprise an electronic circuit, a processor (shared, dedicated, or group), a combinational logic circuit, a memory that executes one or more software programs, and/or other suitable components that provide the described functionality.
[0039] The emission measurement device 3 measures e.g. at least one of carbon oxide (CO) emission, nitrogen oxide (NOx) emission and hydrocarbon (HC) emission of the engine. The emission measurement device 3 may be a relatively inexpensive product with a reduced sampling frequency of e.g. a value between only 1 to 2 Hz.
[0040] The control device 1 may additionally carry out further monitoring functions of the engine and/or the vehicle.
[0041] In addition to the emission measurement device 3 the system may e.g. comprise a speed sensor 4 configured to measure the engine speed, wherein the model predicts high frequency emission information additionally as a function of the engine speed.
[0042] Additionally the system may further comprise an air-fuel ratio sensor 5 configured to measure the air-fuel ratio of the engine, wherein the model predicts high frequency emission information additionally as a function of the air-fuel ratio. [0043] The emission measurement device 3 may be a conventional slow analyser, e.g. MEXA Horiba. Such analysers are not able to accurately measure the restart emission up to 5s, while fast analyser (e.g.: Cambustion) correctly measures emissions for all engine states. However the latter, is extremely expensive.
[0044] However, since under a real driving condition, the number of short restarts is very high, an accurate prediction of the restarts emission is important. The system of the present disclosure is configured to generate such predictions, even if only a slow analyser is used.
[0045] In other words, it is proposed to use dynamic machine learning based and non-linear models to predict fast analyser behavior from slow analyser measurements, in order to have a deeper understanding on high frequency emission.
[0046] The present disclosure leverages dynamic machine learning based and non-linear models to predict fast analyser behaviour from slow analyser measurements. This model could thus replace fast analyser measurements.
[0047] Fig. 2 shows a schematic pipeline for training the model according to embodiments of the present disclosure.
[0048] The model of the present disclosure may be referred to as a slow to fast analyser model. Said model is trained by inputting engine inputs, i.e. at least one predetermined high frequency engine input parameter, and slow measurements, i.e. low frequency emission information. In the engine model training process the model may be trained in a supervised fashion by using respective fast emission measurements, i.e. high frequency emission information as target output for the model. The model correspondingly learns to predict such high frequency emission information.
[0049] A fast to slow analyser model may (as also shown in fig. 2) be used to validate the predictions of the slow to fast analyser model with slow analyser measurements.
[0050] The slow to fast analyser could also be integrated into engine model estimation models or engine restart models.
[0051] Fig. 3 shows an exemplary schematic diagram of the difference between high frequency emission measurements REF_Fast (e.g. with 100 Hz or more) and respective low frequency emission measurements REF_SLOW (with <=10Hz) both representing the same measured high frequency emission information (e.g. CO emissions of a test vehicle) together with the engine speed Eng_spd according to embodiments of the present disclosure.
[0052] Accordingly, the model is used to predict high frequency (e.g. 100Hz or more) emissions from slow analyser (< = 10Hz) measurements. The main inputs of the model are slow analyser measurements and optionally engine speed and/or air fuel ratio sensor value. The engine speed can be used as an input to identify engine operation. As CO emissions are very sensitive to Air fuel ratio value, considering A/F (air-fuel) as input can support better prediction of CO high frequency behavior.
[0053] The model is trained using measurement data. In particular, the inputs of the models are slow analyser measurements (<=10Hz) and engine input parameters, for example, volumetric efficiency, engine speed, air mass flow, air by fuel ratio. The predicted outputs of the model are high frequency CO, NOx and HC emissions. Two different implementations of the model are possible, machine learning models and non-linear models. The first requires a learning process targeting a function (f) that best maps inputs variables (x) to an output variable (y): y=f(x). The second describes nonlinear relationships between experimental data. Both models need Input and Output data for creation (commonly known as training for machine learning approaches).
[0054] For model creation, a set of existing driving conditions can be used - NEDC, RDE, Custom. Measurements are done including cold start condition. The amount of necessary data will depend on the approach. For non-linear models, a RDE cycle is sufficient for training. For machine learning, a minimum of 3 different driving cycles is suitable. In general, the training domain has to be wide enough in terms of operating conditions and transient conditions (gradient of all inputs) in order to be predictive for any driving conditions. Software parameters are tuned to obtain the optimal model. Parameters, such as sampling time, feedback time lag, dimensional reduction, may be taken into consideration for model creation and development. Final models may be first evaluated in a WLTC (Worldwide Flarmonized Light Vehicle Test cycle) on the same engine. A second verification may be done in another engine for different driving cycles: RDE (Real Driving Emissions), NEDC (New European Driving Cycle), WLTC. Final models are independent of engines characteristics.
[0055] Fig. 4 shows an exemplary schematic diagram of the measured high frequency emission REF_Fast (first dashed line) and low frequency emission measurements REF_Slow (second dashed line) and predicted high frequency emission SIM (continuous line) according to embodiments of the present disclosure. The y-axis in fig. 3 and 4 may correspond to the CO emission (e.g. in g/s). As it can be seen in fig. 4, the predicted high frequency emission SIM is close to the actually measured frequency emission measurements REF_Fast. In particular, this model may provide dynamic behavior of emission predictions within e.g. ±7% deviation for Worldwide Flarmonized Light Vehicle Test cycle (WLTC) and Real Driving Emissions compliant cycles. The simulation time of this model is about e.g. 0.02 times the real time. This model can be thus used to deeply understand the behavior of engine out emission at high frequency. This model could therefore replace fast analyser measurements. Having a predicted high frequency response can provide a deep understanding of engine out emissions during engine restart. Such findings may be of interest e.g. for engine development and calibration support.
[0056] Throughout the description, including the claims, the term "comprising a" should be understood as being synonymous with "comprising at least one" unless otherwise stated. In addition, any range set forth in the description, including the claims should be understood as including its end value(s) unless otherwise stated. Specific values for described elements should be understood to be within accepted manufacturing or industry tolerances known to one of skill in the art, and any use of the terms "substantially" and/or "approximately" and/or "generally" should be understood to mean falling within such accepted tolerances.
[0057] Although the present disclosure herein has been described with reference to particular embodiments, it is to be understood that these embodiments are merely illustrative of the principles and applications of the present disclosure.
[0058] It is intended that the specification and examples be considered as exemplary only, with a true scope of the disclosure being indicated by the following claims.

Claims

1. A system for predicting high frequency emission information of an engine, comprising: a control device being configured to receive low frequency emission information from an emission measurement device, said emission measurement device having a first sampling frequency equal to or below a first predetermined sampling threshold configured to the output low frequency emission information having the first sampling frequency, and the control device being further configured to execute a model, wherein the model, when executed by the control device, predicts high frequency emission information as a function of the low frequency emission information and at least one predetermined high frequency engine input parameter, the predicted high frequency emission information having at least a second sampling frequency above the first predetermined sampling threshold.
2. The system according to any one of the preceding claims 1, wherein the system further comprises a speed sensor configured to measure the engine speed, wherein the model predicts high frequency emission information additionally as a function of the engine speed.
3. The system according to any one of the preceding claims 1 and 2, wherein the system further comprises an air-fuel ratio sensor configured to measure the air-fuel ratio of the engine, wherein the model predicts high frequency emission information additionally as a function of the air-fuel ratio.
4. The system according to any one of the preceding claims, wherein the system is further configured to receive the at least one engine input parameter or comprises a sensor configured to measure the at least one engine input parameter, wherein the engine input parameter comprises at least one of: volumetric efficiency, engine speed, and air mass flow.
5. The system according to any one of the preceding claims, wherein the measured low frequency emission information comprises at least one of: carbon oxide (CO) emission, nitrogen oxide (NOx) emission and hydrocarbon (HC) emission, and/or the predicted high frequency emission information comprises respectively at least one of: carbon oxide (CO) emission, nitrogen oxide (NOx) emission and hydrocarbon (HC) emission.
6. The system according to any one of the preceding claims, wherein the model comprises a machine learning model, in particular a regression model and/or a linear model.
7. The system according to any one of the preceding claims, wherein the first predetermined sampling threshold is between lHz and 10Hz, more in particular between 1 and 2Hz, and/or the first sampling frequency is between 1 and 2Hz.
8. The system according to any one of the preceding claims, wherein the predicted high frequency emission information has a third sampling frequency above a second predetermined sampling threshold which is above the first predetermined sampling threshold.
9. The system according to any one of the preceding claims, wherein the second predetermined sampling threshold is 100Hz.
10. The system according to any one of the preceding claims, further comprising the emission measurement device.
11. A method of predicting high frequency emission information of an engine, comprising the steps of: providing a system according to any one of the preceding claims, providing a test vehicle comprising an engine, in particular a hybrid vehicle comprising an engine and at least one electrical motor, operating the vehicle by restarting the engine a plurality of times for limited time periods, in particular less than 10s, predicting high frequency emission information of the engine by means of the system.
12. A method of training a model for predicting high frequency emission information of an engine, comprising the steps of: providing a training data set of an emission pattern of a test vehicle, said data set comprising low frequency emission information having a first sampling frequency of the emission pattern equal to or below the first predetermined sampling threshold, and high frequency emission information having at least a second sampling frequency of the emission pattern above the first predetermined sampling threshold, and train the model by using the low frequency emission information as input for the model and the high frequency emission information as target output.
EP20718275.9A 2020-04-06 2020-04-06 System and method for predicting high frequency emission information of an engine Pending EP4133171A1 (en)

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/EP2020/059816 WO2021204355A1 (en) 2020-04-06 2020-04-06 System and method for predicting high frequency emission information of an engine

Publications (1)

Publication Number Publication Date
EP4133171A1 true EP4133171A1 (en) 2023-02-15

Family

ID=70277378

Family Applications (1)

Application Number Title Priority Date Filing Date
EP20718275.9A Pending EP4133171A1 (en) 2020-04-06 2020-04-06 System and method for predicting high frequency emission information of an engine

Country Status (2)

Country Link
EP (1) EP4133171A1 (en)
WO (1) WO2021204355A1 (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20240095591A1 (en) * 2022-09-21 2024-03-21 Unitedhealth Group Incorporated Processing different timescale data utilizing a model

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2339152B1 (en) * 2009-12-23 2013-06-19 FPT Motorenforschung AG Method and device for controlling an EGR system in a combustion engines

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5625750A (en) * 1994-06-29 1997-04-29 Ford Motor Company Catalyst monitor with direct prediction of hydrocarbon conversion efficiency by dynamic neural networks
RU2180107C2 (en) 1997-08-25 2002-02-27 Виссеншафтлихе Веркштатт Фюр Умвельтместехник Гмбх Automobile exhaust-gas analyzer
DE19741973C1 (en) * 1997-09-23 1999-04-22 Daimler Chrysler Ag Method of determining the soot conc. of self-igniting internal combustion engines
DE102008004221A1 (en) * 2008-01-14 2009-07-16 Robert Bosch Gmbh Determining an occurring during the operation of an internal combustion engine NOx and soot emission
DE102011081949B4 (en) * 2011-09-01 2021-06-10 Robert Bosch Gmbh Method and device for implementing a control, in particular for use in a motor vehicle

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2339152B1 (en) * 2009-12-23 2013-06-19 FPT Motorenforschung AG Method and device for controlling an EGR system in a combustion engines

Also Published As

Publication number Publication date
WO2021204355A1 (en) 2021-10-14

Similar Documents

Publication Publication Date Title
CN110863919B (en) Engine virtual test environment system and engine management system mapping method
US8793004B2 (en) Virtual sensor system and method for generating output parameters
Singh et al. Hybrid physics-infused 1D-CNN based deep learning framework for diesel engine fault diagnostics
JP6139896B2 (en) Engine test apparatus and method
CN112443414B (en) Internal combustion engine state judging device, internal combustion engine state judging system, data analysis device, and internal combustion engine control device
JP2017110630A (en) Method and device for monitoring the operation of a device for variably adjusting the cylinder compression ratio in a piston-internal combustion engine
JP2005090353A (en) Transient engine performance adaptation method and system
Mentink et al. Development and application of a virtual NOx sensor for robust heavy duty diesel engine emission control
Fischer Transient NOx estimation using artificial neural networks
Hanzevack et al. Virtual sensors for spark ignition engines using neural networks
US20210062747A1 (en) Internal combustion engine condition determination apparatus, internal combustion engine condition determination system, data analyzing apparatus, and internal combustion engine control apparatus
Soliman et al. Diagnosis of an automotive emission control system using fuzzy inference
JP2013144975A (en) Model based calibration of inferential sensing
WO2021204355A1 (en) System and method for predicting high frequency emission information of an engine
Rizzoni et al. Diagnosis and prognosis of automotive systems: motivations, history and some results
JP6151473B2 (en) Method for diagnosing and / or adjusting at least one system of devices
JP4235116B2 (en) Apparatus and method for testing transient characteristics of internal combustion engine
Arsie et al. Real-time estimation of engine NOx emissions via recurrent neural networks
Smits et al. Excitation signal design and modeling benchmark of nox emissions of a diesel engine
Togun et al. Nonlinear identification of a spark ignition engine torque based on ANFIS with NARX method
Sundaram et al. Modeling of transient gasoline engine emissions using data-driven modeling techniques
US20250005218A1 (en) Method, Device, Vehicle and Computer Program for Modeling and Monitoring a Warming-Up Behavior of a Vehicle Component
Gurel et al. Multi-objective optimization of transient air-fuel ratio limitation of a diesel engine using DoE based Pareto-optimal approach
Cavina et al. Individual cylinder air-fuel ratio control for engines with unevenly spaced firing order
Wang et al. A detection and isolation of faults technique in automotive engines using a data-driven and model-based approach

Legal Events

Date Code Title Description
STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: UNKNOWN

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE INTERNATIONAL PUBLICATION HAS BEEN MADE

PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: REQUEST FOR EXAMINATION WAS MADE

17P Request for examination filed

Effective date: 20220913

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

DAV Request for validation of the european patent (deleted)
DAX Request for extension of the european patent (deleted)
STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: EXAMINATION IS IN PROGRESS

17Q First examination report despatched

Effective date: 20240808

RAP1 Party data changed (applicant data changed or rights of an application transferred)

Owner name: TOYOTA JIDOSHA KABUSHIKI KAISHA