Real-time misfire detection for automobile engines with medium data rate crankshaft sampling

• US 5,862,507 A
• Filed: 04/07/1997
• Issued: 01/19/1999
• Est. Priority Date: 04/07/1997
• Status: Expired due to Term

• Current Assignee: FCA US LLC (Stellantis NV)
• Original Assignee: Chrysler Corporation (Stellantis NV)
• Inventors: Wu, Zhijian James, Lee, Anson
• Primary Examiner(s): Nelli, Raymond A.
• Application Number:
  US08/835,393
• Time in Patent Office: 652 Days
• Field of Search: 701/111, 123/119 A, 123/571, 123/419, 123/425, 123/478, 123/436, 123/414, 73/116, 73/117.3, 364/431.01, 364/431.08
• US Class Current: 701/111
• CPC Class Codes: F02D 2041/288 for performing a transformation into the frequency domain, e.g. Fourier transformation
• CPC Class Codes: F02D 2200/1015 Engines misfires
• CPC Class Codes: F02D 2250/12 Timing of calculation, i.e. specific timing aspects when calculation or updating of engine parameter is performed
• CPC Class Codes: F02D 41/1498 measuring engine roughness
• CPC Class Codes: G01M 15/11 by detecting misfire

Abstract

A method and system for detecting engine misfire in an internal combustion engine. A data signal containing a plurality of groups of data points generated from sampling crankshaft rotational displacement over three predetermined angular sampling windows is generated. Each of the plurality of groups of sampled data points represents a cylinder firing event. The data signal is conditioned, and the plurality of groups of sampled data points are mapped into single misfire detection data points. The mapped single misfire detection points are then compared to a misfire detection dynamic threshold, and a misfire detection signal is output if the misfire detection points fall outside the dynamic threshold. The misfire detection system achieves high degree of accuracy through a multi-stage signal conditioning, multi-rate signal processing and statistical decision technology and a mixed size of window sampling strategy, and is capable of being installed on a large scale in motor vehicles through use of existing onboard automotive microcontroller technology.

Claims

1. A computer-implemented method of detecting misfire of an internal combustion engine, said engine including a plurality of cylinders and a crankshaft, said crankshaft having an angular velocity, comprising the steps of:

sensing for two firings of one of said cylinders the angular velocity of the crankshaft;

generating a data signal containing a plurality of groups of sampled data points, each of said plurality of groups of sampled data points representing a sensed crankshaft angular velocity for a particular cylinder firing event;

mapping each group of said plurality of groups of sampled data points into a single data point based upon a predetermined firing criteria;

determining a misfire detection dynamic threshold based upon firing signature signals of said engine;

comparing said mapped data point to said misfire detection dynamic threshold; and

outputting a misfire detection signal if said mapped data point does not satisfy said dynamic threshold, said outputted misfire detection signal being indicative of a misfire of said engine.

2. The method of claim 1, further comprising the step of enhancing separation between mapped data points representing normal cylinder firing events and mapped data points representing misfire cylinder firing events in said data signal.

3. The method of claim 2, wherein said step of enhancing separation between data points comprises cubing said mapped sampled data points;

and

passing said cubed mapped sampled data points through a low pass filter to filter out higher order harmonics generated by said step of cubing and to filter out data signal noise.

4. The method of claim 1, wherein said step of generating a data signal comprises generating a data signal including sensed data from both an engine speed signal based on crankshaft angular velocity sampled at a medium data rate, and a manifold absolute pressure (MAP) signal.

5. The method of claim 4 further comprises the step of extracting average and fluctuation signal components from said engine speed and MAP signals.

6. The method of claim 4, wherein said step of generating a data signal comprises sampling said crankshaft angular velocity over M mixed size sampling windows per cylinder firing event to generate each of said plurality of groups of sampled data points, where M is an integer greater than 1.

7. The method of claim 6, wherein said step of sampling said crankshaft angular velocity over M mixed size sampling windows comprises:

sampling said crankshaft angular velocity to generate a first data point during a first 40 degree window after top dead center;

sampling said crankshaft angular velocity to generate a second data point during a 60 degree window occurring at about 20 degrees after said 40 degree window; and

sampling said crankshaft angular velocity to generate a third data point during a second 40 degree window occurring immediately after said 60 degree window.

8. The method of claim 1, further comprising the step of filtering said data signal to eliminate signal frequency components at and above an engine firing frequency after said step of conditioning said data signal.

9. The method of claim 8, further comprising the step of filtering noise from said data signal to improve the signal to noise ratio of said data signal.

10. The method of claim 1, wherein said step of mapping each group of said plurality of groups of sample data points comprises mapping every M data points into one data point for signal processing purposes, where said data signal has a medium data rate of sampling of M points per firing, where M is a positive integer greater than one.

11. The method of claim 10, wherein said step of mapping each group of said plurality of groups of sampled data points is incorporated into said step of comparing said mapped data points to a misfire detection dynamic threshold.

12. The method of claim 1, further comprising the step of compensating for a mean of said mapped single data points representing normal cylinder firing events by dynamically adjusting said data signal based on a running mean calculated from previously input mapped sampled data points representing normal cylinder firing events.

13. The method of claim 1, wherein the step of conditioning the data signal comprises compensating for variations in misfire signature magnitude caused by engine operating conditions.

14. The method of claim 1, wherein said step of comparing said mapped sampled data points further comprises the steps of:

feeding said mapped data points into a mean filter and a deviation filter;

estimating mean and deviation values of said mapped sampled data points for both normal cylinder firing events and misfire cylinder firing events;

updating said dynamic threshold based on said estimated mean and deviation values within predetermined upper and lower boundaries; and

determining if values of said mapped sampled data points fall below or above said dynamic threshold.

15. The method of claim 14, further comprising the steps of generating a signal indicative of a normal cylinder firing event if said values of said mapped sampled data points fall below or above said dynamic threshold;

and

generating an engine misfire signal if said values of said mapped single data points fall below said dynamic threshold.

16. The method of claim 1, further comprising the step of compensating for variations in mean values of said mapped sampled data points representing normal cylinder firing events for equalization thereof.

17. A motor vehicle internal combustion engine misfire detection system for detecting misfire of an internal combustion engine, said engine including a plurality of cylinders and a manifold and a crankshaft, said crankshaft having an angular velocity, comprising:

a crankshaft velocity sensor that senses velocity of said vehicle crankshaft based on medium data rate crankshaft sampling and that generates a corresponding engine speed signal;

a manifold sensor that senses manifold absolute pressure (MAP) of said manifold and generates a corresponding manifold absolute pressure signal;

a signal processor that extracts average and fluctuation signal components from said engine speed and MAP signals for detecting both engine normal fire and misfire cylinder firing events;

a decimeter that, for a medium data sampling rate with M points sampled per cylinder firing event, where M is a positive integer greater than one, maps every M points into one mapped data point for misfire detection processing; and

a dynamic threshold generator for generating a dynamic detection threshold based upon firing signature signals of said engine;

a dynamic decision sub-system connected to said threshold generator for determining if said signal represents an engine misfire cylinder firing event or an engine normal fire cylinder firing event based upon a comparison between said mapped data points and said generated dynamic detection threshold and that generates an output signal indicative thereof, said output signal being indicative of a misfire of said engine.

18. The system of claim 17, further comprising a multiplier that enhances separation between said mapped data points representing engine normal fire and misfire cylinder firing events.

19. The system of claim 17, wherein said dynamic decision sub-system comprises a mean filter that generates a mean output of said normal and misfire cylinder firing events, and a deviation filter that generates a deviation output of said normal and misfire cylinder firing events;

and

said dynamic threshold generator updating the dynamic threshold of said sub-system based on outputs of said mean and deviation filters.

20. The system of claim 17, further comprising an equalizer that compensates for variations in engine misfire signature magnitude based on said engine speed fluctuation signal component and said engine speed and manifold absolute pressure average signal components.

21. The system of claim 19, further comprising a filter coupled to said misfire magnitude equalizer that filters out frequency components at or above an engine signal firing frequency.

22. The system of claim 17, further comprising a resonance removal device coupled to said decimeter that detects and removes resonance noise from said data signal;

and

an equalizer that equalizes data samples from said normal fire cylinder firing events in said data signal to compensate for variation in the mean level thereof to enhance said detection of said misfire cylinder firing events.

23. The method of claim 17, further comprising a sampling window having a mixed sampling window size for sampling rotational displacement of a crankshaft sensor wheel operatively associated with said crankshaft during engine cylinder firing events.

24. The method of claim 23, wherein said sampling window size is about a 40 degree angular displacement after top dead center for a first crankshaft sensor wheel data interval, about a 60 degree angular displacement for a second crankshaft sensor wheel data interval occurring after about a 20 degree angular displacement of said crankshaft sensor wheel after said first data interval, and about a 60 degree angular displacement for a third crankshaft sensor wheel data interval occurring immediately after said second data interval.