Using an artificial neural network for combustion phasing control in a spark ignited internal combustion engine

Patent metadata

Field	Value
Publication number	US-10788396-B2
Application number	US-201816224988-A
Country	US
Kind code	B2
Filing date	Dec 19, 2018
Priority date	Dec 19, 2018
Publication date	Sep 29, 2020
Grant date	Sep 29, 2020

Abstract

A calibration system and method for a spark ignition engine of a vehicle involve artificially weighting engine dynamometer data in high engine load regions and using it to generate training data for an artificial neutral network (ANN). A plurality of ANNs are trained using the training data and the plurality of ANNs are then filtered based on their maximum error to obtain a filtered set of trained ANNs. A statistical analysis is performed on each of the filtered set of trained ANNs including determining a set of statistical metrics for each of the filtered set of trained ANNs and then one of the filtered set of trained ANNs having a best combination of error at high engine loads and the set of statistical error metrics is then selected. Finally, an ANN calibration is generated using the selected one of the filtered set of trained ANNs.

First claim

What is claimed is:

1. A calibration system for a spark ignition (SI) engine of a vehicle, the calibration system comprising:
dynamometer sensors configured to measure dynamometer data for the SI engine; and
a computing system configured to:
receive the dynamometer data for the SI engine,
artificially weight the dynamometer data in high engine load regions to obtain weighted engine dynamometer data,
generate training data for an artificial neural network (ANN) using the weighted engine dynamometer data,
train a plurality of ANNs using the generated training data,
filter the plurality of trained ANNs based on their maximum error to obtain a filtered set of trained ANNs,
perform a statistical analysis on each of the filtered set of trained ANNs including determining a set of statistical metrics for each of the filtered set of trained ANNs,
select one of the filtered set of trained ANNs having a best combination of error at high engine loads and the set of statistical metrics, and
generate an ANN calibration using the selected one of the filtered set of trained ANNs.

2. The calibration system of claim 1 , wherein the computing system is further configured to output the generated ANN calibration to a controller of the SI engine.

3. The calibration system of claim 2 , wherein receipt of the generated ANN calibration causes the controller to store the generated ANN calibration and use the generated ANN calibration for combustion phasing control to at least one of decrease engine knock and increase fuel economy.

4. The calibration system of claim 3 , wherein the use of the generated ANN calibration involves the controller:
receiving, from camshaft position sensors, measured positions of intake and exhaust camshafts of the SI engine;
receiving, from an air charge sensor, a measured air charge provided to each cylinder of the SI engine;
receiving, from an engine speed sensor, a measured engine speed indicative of a rotational speed of a crankshaft of the SI engine;
using the generated ANN calibration, determining at least one of a desired base spark timing and a desired maximum brake torque (MBT) spark timing based on the measured intake and exhaust camshaft positions, the measured air charge, and the measured engine speed; and
controlling a spark system of the SI engine based on at least one of the desired base and MBT spark timings.

5. The calibration system of claim 4 , wherein the controller does not generate and output empirically calibrated surfaces for determining the base or MBT spark timings.

6. The calibration system of claim 1 , wherein generating the training data includes using the weighted engine dynamometer data and a mathematical method, wherein the generated training data includes more data points than the weighted engine dynamometer data.

7. The calibration system of claim 1 , wherein the set of statistical metrics comprises at least one of a sum of squares due to error (SSE) and a root-mean-square error (RMSE).

8. The calibration system of claim 1 , wherein each of the plurality of trained ANNs defines two hidden layers, twelve neurons per hidden layer, a Bayesian regularization backpropagation training function, and a hyperbolic tangent sigmoid transfer function.

9. A calibration method for a spark ignition (SI) engine of a vehicle, the method comprising:
receiving, by a calibration system, dynamometer data for the SI engine;
artificially weighting by the calibration system, the dynamometer data in high engine load regions to obtain weighted engine dynamometer data;
generating, by the calibration system, training data for an artificial neural network (ANN) using the weighted engine dynamometer data;
training, by the calibration system, a plurality of ANNs using the generated training data;
filtering, by the calibration system, the plurality of trained ANNs based on their maximum error to obtain a filtered set of trained ANNs;
performing, by the calibration system, a statistical analysis on each of the filtered set of trained ANNs including determining a set of statistical metrics for each of the filtered set of trained ANNs;
selecting, by the calibration system, one of the filtered set of trained ANNs having a best combination of error at high engine loads and the set of statistical metrics; and
generating, by the calibration system, an ANN calibration using the selected one of the filtered set of trained ANNs.

10. The calibration method of claim 9 , further comprising outputting, by the calibration system and to a controller of the SI engine, the generated ANN calibration.

11. The calibration method of claim 10 , wherein receipt of the generated ANN calibration causes the controller to store the generated ANN calibration and use the generated ANN calibration for combustion phasing control to at least one of decrease engine knock and increase fuel economy.

12. The calibration method of claim 11 , wherein the use of the generated ANN calibration involves the controller:
receiving, from camshaft position sensors, measured positions of intake and exhaust camshafts of the SI engine;
receiving, from an air charge sensor, a measured air charge provided to each cylinder of the SI engine;
receiving, from an engine speed sensor, a measured engine speed indicative of a rotational speed of a crankshaft of the SI engine;
using the generated ANN calibration, determining at least one of a desired base spark timing and a desired maximum brake torque (MBT) spark timing based on the measured intake and exhaust camshaft positions, the measured air charge, and the measured engine speed; and
controlling a spark system of the SI engine based on at least one of the desired base and MBT spark timings.

13. The calibration method of claim 12 , wherein the controller does not generate and output empirically calibrated surfaces for determining the base or MBT spark timings.

14. The calibration method of claim 9 , wherein generating the training data includes using the weighted engine dynamometer data and a mathematical method, wherein the generated training data includes more data points than the weighted engine dynamometer data.

15. The calibration method of claim 9 , wherein the set of statistical metrics comprises at least one of a sum of squares due to error (SSE) and a root-mean-square error (RMSE).

16. The calibration method of claim 9 , wherein each of the plurality of trained ANNs defines two hidden layers, twelve neurons per hidden layer, a Bayesian regularization backpropagation training function, and a hyperbolic tangent sigmoid transfer function.

Assignees

• Attard William P
• Wang Shu
• Xu Shuonan
• Tutton Tyler
• Gopinath Srinath
• Lawrence David A
• Fca Us Llc

Inventors

• Attard William P
• Wang Shu
• Xu Shuonan
• Tutton Tyler
• Gopinath Srinath
• Lawrence David A

Classifications

• G01M15/04 Primary

  Testing internal-combustion engines · CPC title

• F02P5/152

  dependent on pinking (detecting or indicating knocks in internal-combustion engines G01L23/22) · CPC title

• Y02T10/40

  Engine management systems · CPC title

• F02D2200/101

  Engine speed · CPC title

• F02P5/1502

  using one central computing unit · CPC title