Combustion knock detection and control through statistical characterization of knock levels

• US 7,415,347 B2
• Filed: 08/25/2006
• Issued: 08/19/2008
• Est. Priority Date: 08/25/2006
• Status: Expired due to Fees

• Current Assignee: Michigan Technological University
• Original Assignee: Michigan Technological University
• Inventors: Naber, Jeffrey D., Rajagopalan, Satheesh Rajh
• Primary Examiner(s): KWON, JOHN
• Application Number:
  US11/467,313
• Publication Number:
  US 20080051981A1
• Time in Patent Office: 725 Days
• Field of Search: 701/101, 701/111, 701/114, 701/115, 73/350.4
• US Class Current: 701/111
• CPC Class Codes: G01L 23/225 circuit arrangements therefor

Abstract

A method of statistically characterizing combustion knock events includes receiving signals from a sensing device such as an accelerometer, estimating at least one parameter of a probability distribution function based on the received signals, and calculating a value indicative of an n^th percentile based on the parameter to predict upcoming combustion knock events.

Claims

1. A method of statistically characterizing combustion events in a combustion engine, the method comprising:

receiving at least one signal indicative of knock level associated with combustion events;

estimating at least one parameter of a probability density function and a cumulative distribution function based on receiving the at least one signal;

estimating a probability distribution of the at least one signal based on estimating the at least one parameter; and

calculating at least one value indicative of an n^th percentile of the at least one signal based on the at least one parameter.

2. The method of claim 1, further comprising estimating at least one of a first alternative value indicative of skewness and a second alternative value indicative of kurtosis.

3. The method of claim 1, wherein receiving the at least one signal includes receiving at least one electric signal from at least one of an accelerometer and a pressure transducer coupled to an engine.

4. The method of claim 1, wherein estimating the probability distribution includes estimating at least one of a log-normal distribution, a beta distribution, a chi-squared distribution, a gamma distribution, an inverse chi-distribution, and an inverse gamma distribution.

5. The method of claim 1, wherein receiving the at least one signal includes receiving the at least one signal under knocking conditions.

6. The method of claim 1, wherein estimating the at least one parameter includes applying the at least one signal to at least one of an IIR filter and a FIR filter.

7. The method of claim 1, wherein calculating the at least one value includes calculating a value between about an 80^th percentile of the at least one signal and about a 99^th percentile of the at least one signal.

8. The method of claim 1, further comprising estimating a second parameter of the probability density function and the cumulative distribution function.

9. The method of claim 8, wherein calculating the at least one value includes calculating the at least one value based on the at least one parameter and the second parameter.

10. The method of claim 9, wherein calculating the at least one value includes calculating a value between about an 80^th percentile of the at least one signal and about a 99^th percentile of the at least one signal.

11. The method of claim 8, wherein estimating the second parameter includes estimating the second parameter based on estimating the at least one parameter.

12. The method of claim 8, wherein estimating the probability distribution includes estimating a log-normal probability distribution.

13. A method of statistically characterizing combustion events in a combustion engine, the method comprising:

receiving at least one signal indicative of knock level associated with combustion events;

estimating at least one parameter of a probability density function and a cumulative distribution function based on receiving the at least one signal; and

estimating a probability distribution of the at least one signal based on estimating the at least one parameter, wherein estimating the at least one parameter includes calculating at least one of a running average and a standard deviation of the logarithm of the at least one signal.

14. The method of claim 13, wherein estimating the at least one parameter includes calculating at least one of the running average and the standard deviation of the natural log of the at least one signal to estimate at least one of a mean value and a standard deviation value.

15. A method of statistically characterizing combustion events in a combustion engine, the method comprising:

receiving at least one signal indicative of knock level associated with combustion events;

estimating at least one parameter of a probability density function and a cumulative distribution function based on receiving the at least one signal; and

estimating a probability distribution of the at least one signal based on estimating the at least one parameter, wherein estimating the at least one parameter includes calculating at least one of a weighted average and a standard deviation of the logarithm of the at least one signal.

16. The method of claim 15, wherein estimating the at least one parameter includes calculating at least on of the weighted average and the standard deviation of the natural log of the at least one signal to estimate at least one of a mean value and a standard deviation value.

17. A method of predicting combustion knock events based on real-time statistical characterization of knock level metrics, the method comprising:

receiving at least one signal indicative of knock level;

estimating a first parameter of a probability density function and a cumulative distribution function based on the at least one signal indicative of knock level;

calculating a first value indicative of an n^th percentile of the at least one signal;

calculating a second value indicative of the n^th percentile of the at least one signal; and

comparing the first value to the second value to predict an upcoming combustion knock event.

18. The method of claim 17, further comprising estimating a probability distribution based on the first parameter.

19. The method of claim 17, wherein estimating the first parameter includes calculating the running average of the logarithm of the at least one signal.

20. The method of claim 19, wherein estimating the first parameter includes calculating the running average of the natural log of the at least one signal to estimate a mean value.

21. The method of claim 17, wherein estimating the first parameter includes calculating the weighted average of the logarithm of the at least one signal.

22. The method of claim 21, wherein estimating the first parameter includes calculating the weighted average of the natural log of the at least one signal to estimate a mean value.

23. The method of claim 17, wherein calculating the first value includes calculating a value between about an 80^th percentile of the at least one signal and about a 99^th percentile of the at least one signal when the at least one signal indicative of knock level is received under knocking conditions of an engine.

24. The method of claim 17, wherein calculating the second value includes calculating a value between about an 80^th percentile and about of the at least one signal and about a 99^th percentile of the at least one signal when the at least one signal indicative of knock level is received under non-knocking conditions of an engine.

25. The method of claim 17, wherein receiving the at least one signal includes receiving at least one electric signal from at least one of an accelerometer and a pressure transducer coupled to an engine.

26. The method of claim 17, further comprising estimating a probability distribution based on the first parameter, wherein estimating the probability distribution includes estimating at least one of a log-normal distribution, a beta distribution, a chi-squared distribution, a gamma distribution, an inverse chi-distribution, and an inverse gamma distribution.

27. The method of claim 17, wherein comparing the first value to the second value includes calculating a ratio of the first value with respect to the second value.

28. The method of claim 17, wherein comparing the first value to the second value includes calculating the difference between the first value and the second value.

29. The method of claim 17, further comprising estimating a second parameter of the probability density function and the cumulative distribution function.

30. The method of claim 29, wherein calculating the first value includes calculating a value between about an 80^th percentile of the at least one signal and about a 99^th percentile of the at least one signal based on the first parameter and the second parameter when the at least one signal indicative of knock level is received under knocking conditions of an engine.

31. The method of claim 29, wherein calculating the second value includes calculating a value between about an 80^th percentile of the at least one signal and about a 99^th percentile of the at least one signal based on the first parameter and the second parameter when the at least one signal indicative of knock level is received under non-knocking conditions of an engine.

32. A method of predicting combustion knock events based on knock level values substantially independent of knock conditions, the method comprising:

receiving at least one signal indicative of knock level;

estimating a first parameter of a probability density function and a cumulative distribution function based on the at least one signal indicative of knock level;

calculating a first value indicative of a first n^th percentile of the at least one signal based on the first parameter;

calculating a second value indicative of a second n^th percentile of the at least one signal based on the first parameter; and

normalizing the first value with respect to the second value to predict an upcoming combustion knock event substantially independent on knock conditions.

33. The method of claim 32, further comprising estimating a probability distribution based on the first parameter.

34. The method of claim 32, wherein receiving the at least one signal includes receiving at least one electric signal from at least one of an accelerometer and a pressure transducer mounted coupled to an engine.

35. The method of claim 32, further comprising estimating a probability distribution based on the first parameter, wherein estimating the probability distribution includes estimating at least one of a log-normal distribution, a beta distribution, a chi-squared distribution, a gamma distribution, an inverse chi-distribution, and an inverse gamma distribution.

36. The method of claim 32, wherein estimating the first parameter includes calculating a weighted average of the logarithm of the at least one signal to estimate a mean value.

37. The method of claim 32, wherein estimating the first parameter includes calculating a running average of the logarithm of the at least one signal to estimate a mean value.

38. The method of claim 32, wherein estimating the first parameter includes applying the at least one signal to at least one of an IIR filter and a FIR filter.

39. The method of claim 32, wherein calculating the first value includes calculating a value between about an 80^th percentile of the at least one signal and about a 99^th percentile of the at least one signal, and wherein calculating the second value includes calculating a value between about a 15^th percentile of the at least one signal and about a 35^th percentile of the at least one signal.

40. A method of substantially eliminating a measurement error related at least to a gain of an accelerometer coupled to an engine for predicting combustion knock events, the method comprising:

generating at least one electric signal with the accelerometer;

calculating at least one knock level based on the at least one electric signal;

estimating a first parameter of a probability density function and a cumulative distribution function based on the at least one knock level;

calculating a first value indicative of a first n^th percentile of the at least one knock level based on the first parameter;

calculating a second value indicative of a second n^th percentile of the at least one knock level based on the first parameter; and

normalizing the first value with respect to the second value for substantially eliminating the measurement error in predicting upcoming combustion knock events.