Method and apparatus for dynamical system analysis

• US 6,208,949 B1
• Filed: 07/01/1998
• Issued: 03/27/2001
• Est. Priority Date: 07/01/1998
• Status: Expired due to Fees

• Current Assignee: Adaptive Audio, Inc.
• Original Assignee: Adaptive Audio, Inc.
• Inventors: Eatwell, Graham P.
• Primary Examiner(s): Heckler, Thomas M.
• Application Number:
  US09/108,253
• Time in Patent Office: 1,000 Days
• Field of Search: 700/108, 700/174, 702/108-113, 702/117, 702/182-185, 702/189-193, 327/39, 327/45, 327/50, 327/60, 327/46
• US Class Current: 702/189
• CPC Class Codes: G05B 23/0256 injecting test signals and analyzing monitored process response, e.g. injecting the test signal while interrupting the normal operation of the monitored system; superimposing the test signal onto a control signal during normal operation of the monitored system
• CPC Class Codes: G05B 5/01 electric

Abstract

A method and apparatus are provided for simultaneously characterizing the disturbance of a dynamical system and the dynamic response model of the system. The disturbance is characterized by a sum of reference signals weighted by disturbance parameters. The dynamic response model is similarly characterized by a set of response parameters, which model the relationship between a set of test signals and the response to these test signals. The disturbance parameters and the response parameters are estimated jointly to provide a characterization of the dynamical system. This estimation is achieved by sensing the system at one or more locations to obtain a sensed signal, obtaining reference signals related to the disturbances, synthesizing an estimate of the sensed signal from the reference signals, the test signals, the disturbance parameters and the response parameters, and jointly adjusting the disturbance parameters and the response parameters according to the difference between the synthesized signal and the sensed signal. The characterization may be incorporated in systems or methods for control or monitoring of dynamical systems.

Claims

1. A method for estimating and updating parameters which characterize the state of a dynamical system subject to one or more disturbances comprising the steps of

obtaining one or more first signals related in time or amplitude to the disturbances,

sensing the system at one or more locations to obtain sensed signals indicative of the dynamical state of the system,

generating predicted parameters from previously estimated parameters,

generating a synthesized signal as a weighted sum of products of said first signals and said predicted parameters, and

adjusting the predicted parameters according to an error signal which is the difference between the synthesized signal and the sensed signal to obtain new estimated parameters.

2. A method as in claim 1 in which

the state of the system includes the disturbance of the system,

said parameters include disturbance parameters characterizing the disturbance of the dynamical system,

said disturbance is characterized as a sum of said first signals each weighted by a disturbance parameter, and

at least one component of the synthesized signal is a synthesized disturbance signal.

3. A method as in claim 1 in which said first signals include reference signals obtained from sensors responsive to the disturbance of the dynamic system.

4. A method as in claim 1 in which said predicted parameters are obtained by passing previously estimated parameters through a low pass filter.

5. A method as in claim 1 in which the adjustment of the parameters is dependent upon said first signals.

6. A method as in claim 1 in which the adjustment of the parameters is proportional to said error signal and to a gain which is dependent upon an estimate of the variance of the error signal.

7. A method as in claim 1 in which one or more of said disturbances is substantially cyclic in time and which said first signals include one or more cyclic reference signals which are time or phase related to the cyclic disturbance or to the source of the cyclic disturbance.

8. A method as in claim 7 in which said cyclic reference signals are sine and cosine signals at multiples of the period of the cyclic disturbance.

9. A method as in claim 8 in which the disturbance is synthesized as $\sum _{}^{}$

m=0M-1

∑

k=0K

(b2

k,m

cos

(k

 



ϕ

m

(n))+b2

k+1,m

sin

(k

 



ϕ

m

(n))),

where m is the index of the disturbance, k is a harmonic number, b_k,m are disturbance parameters and φ

_m(n) is a signal which phase related to the m^th disturbance source at sample n, and in which the state of the system is characterized by the complex amplitudes B_k,m=b_2k,m+i.b_2k+1,m.

10. A method as in claim 7 in which said cyclic reference signals are combinations of sine and cosine signals at multiples of the period of the cyclic disturbance.

11. A method as in claim 1 in which said parameters include response parameters characterizing the dynamic response of the system, said method including the steps of

generating a test signal so as to produce a controlled disturbance in said dynamical system, and

synthesizing a synthesized response signal from current and past values of the test signal and from the response parameters,

wherein said first signals include said test signal.

12. A method as in claim 11 and including the step of adjusting said response parameters according to a time domain algorithm in which the adjustment to the parameters is proportional to the product of the error signal and a gain, said gain being dependent upon the reference signals.

13. A method as in claim 11 and including the step of adjusting said response parameters according to a frequency domain algorithm in which the Fourier transform of the adjustment to the parameters is proportional to the product of the Fourier transform of the error signal and the complex conjugate of the Fourier transform of the reference signals.

14. A method as in claim 1 in which the adjustment of one or more of the estimated parameters is dependent on the values of other estimated parameters.

15. A method as in claim 1 in which said dynamic system is a machine, said method including the step of monitoring the changes in said predicted or estimated parameters over a period of time so as to detect changes in the condition of said machine.

16. A method as in claim 15 in which said error signal is also monitored.

17. A method as in claim 15 in which variations in the synthesized signal are also monitored.

18. A method as in claim 15 in which said parameters include response parameters and disturbance parameters which are estimated jointly.

19. A method for controlling the state of a dynamical system subject to one or more disturbances, where the state is characterized by a set of parameters, comprising the steps of

selecting initial values for the estimated parameters and then repeating the steps of

sensing the system at one or more locations to obtain sensed signals indicative of the state of the system,

obtaining one or more first signals related in time or amplitude to the disturbances,

generating predicted parameters from said previously estimated parameters,

generating a synthesized signal as a weighted sum of the first signals and the predicted parameters,

adjusting the predicted parameters according to an error signal which is the difference between the synthesized signal and the sensed signal to obtain new estimated parameters,

generating a control signal dependent upon the estimated parameters, and

supplying said control signal to an actuator so as to generate controlled disturbances in said dynamic system and thereby control the state of said system.

20. A method as in claim 19 in which

the state of the system includes the disturbance of the system,

said parameters include said disturbance parameters,

said first signals include one or more reference signals from sensors responsive to the disturbance or the source of the disturbance, and

said disturbance is characterized as sum of said reference signals each weighted by a disturbance parameter.

21. A method as in claim 19 in which

said parameters include response parameters

said first signals include a test signal which is added to said control signal, and

the state of the system includes a dynamic response model of the system which is characterized by said response parameters.

22. A method as in claim 19 in which said control signal is also dependent upon said synthesized signal.

23. A method as in claim 19 in which said control signal is also dependent upon said error signal.

24. A system for analyzing the state of a dynamical system subject to one or more disturbances, in which the state of the system includes the disturbance of the system and the disturbance is characterized by a set of disturbance parameters, comprising:

sensors at one or more locations producing sensed signals,

reference signal generating means for generating reference signals related to said disturbances,

disturbance signal synthesizer means for combining said reference signals with the set of disturbance parameters to produce a synthesized disturbance signal,

signal subtraction means for producing an error signal corresponding to the difference between said sensed signal and said synthesized signal, and

disturbance parameter adjustment means for adjusting said parameters in response to said error signal.

25. A system as in claim 24 in which the disturbance at sample n is synthesized as $\sum _{}^{}$

m=0M-1

∑

k=0K

(b2

k,m

cos

(k

 



ϕ

m

(n))+b2

k+1,m

sin

(k

 



ϕ

m

(n))),

where m is the index of the disturbance, k is a harmonic number, b_k,m are disturbance parameters and φ

_m(n) is a signal which phase related to the m^th disturbance source at sample n, and in which the state of the system is characterized by the complex amplitudes B_k,m=b_2k,m+i.b_2k+1,m.

26. A system as in claim 24 in which the disturbance at sample n is synthesized as $\begin{array}{c}\sum _{}^{}\end{array}$

m=0M-1

∑

k=0K

bk

 



m

xm

(n-k),(90)

where m is the index of the reference signal, k is a parameter index, b_km are disturbance parameters and x_m(n) is the m^th reference signal at sample n.

27. A system as in claim 26 in which said disturbance parameter adjustment means operates according to a time domain algorithm in which the adjustment to the disturbance parameters is proportional to the product of the error signal and a gain, said gain being dependent upon the reference signals.

28. A system as in claim 26 in which said disturbance parameter adjustment means operates according to a frequency domain algorithm in which the Fourier transform of the adjustment to the disturbance parameters is proportional to the product of the Fourier transform of the error signal and the complex conjugate of the Fourier transform of the reference signal.

29. A system as in claim 24, in which the state of the system includes the dynamic response of the system and the dynamic response is characterized by a set of response parameters, said system further including:

test signal generating means for generating test signals,

actuator means responsive to said test signals for generating controlled disturbances in said dynamical system

response signal synthesizer means for combining said test signals with a set of response parameters to produce a synthesized response signal,

response parameter adjustment means for adjusting said response parameters in accordance with said error signal,

wherein said error signal corresponds to the said sensed signal minus said synthesized disturbance signal and minus said synthesized response signal.

30. A system as in claim 29 in which said response parameter adjustment means operates according to a time domain algorithm in which the adjustment to the response parameters is proportional to the product of the error signal and a gain, said gain being dependent upon the test signals.

31. A system as in claim 29 in which said response parameter adjustment means operates according to a frequency domain algorithm in which the Fourier transform of the adjustment to the response parameters is proportional to the product of the Fourier transform of the error signal and the complex conjugate of the Fourier transform of the reference signal.

32. A system as in claim 29 in which said dynamical system is a machine and including

machinery monitoring means for monitoring the time variations in said disturbance parameters and said response parameters,

wherein said machinery monitoring means periodically monitors said parameters so as to determine the condition of the machine.

33. A system as in claim 24 in which said dynamical system is a machine and including

machinery monitoring means for monitoring the time variations in said disturbance parameters and periodically comparing said parameters with those expected from known operating or fault conditions so as to determine the condition of the machine.

34. A system as in claim 33 in which said machinery monitoring means also monitors said error signal.

35. A system as in claim 33 in which said machinery monitoring means also monitors said synthesized disturbance signal.

36. A system as in claim 24 and including signal compression means which is responsive to said disturbance parameters and produces a compressed data signal.

37. A system as in claim 36 and including transmission means for transmitting said compressed data signal to a remote location for further processing or analysis.

38. A system as in claim 36 and including storage means for storing said compressed data signal.

39. A system as in claim 36 in which said signal compression means is also responsive to said reference signals, or to a signal from which said reference signals are derived.

40. A system for controlling the state of a dynamical system subject to one or more disturbances in which the state of the system includes the disturbance of the system, characterized by a set of disturbance parameters and also includes the dynamic response of the system, characterized by a set of response parameters, comprising

sensor means at one or more locations producing sensed signals indicative of the state of said dynamical system,

reference signal generating means for generating reference signals related to said disturbances,

disturbance signal synthesizer means for combining said reference signals with a set of disturbance parameters to produce a synthesized disturbance signal,

test signal generating means for generating test signals,

response signal synthesizer means for combining said test signals with a set of response parameters to produce a synthesized response signal,

signal subtraction means for producing an error signal corresponding to said sensed signal minus said synthesized disturbance signal and minus said synthesized response signal,

disturbance parameter adjustment means for adjusting said disturbance parameters in response to said error signal,

response parameter adjustment means for adjusting said response parameters in response to said error signal,

controller means responsive to said response parameters and said reference signals and generating a control signal, and

actuator means responsive to said control signal for generating controlled disturbances in said dynamical system, thereby controlling the state of said system.

41. A system as in claim 40 in which said controller means is also responsive to said synthesized disturbance signal and said error signal or a combination thereof.

42. A system as in claim 40 in which said controller means is also responsive to said disturbance parameters.

43. A system as in claim 40 in which said reference signals are sine and cosine signals and in which the control signal is dependent upon the Fourier transform of the response parameters.

44. A control system as in claim 40 which includes multiple sensors and multiple actuators.

45. A control system as in claim 40 in which said sensor means is a microphone which is responsive to speech in the vicinity of the microphone and in which said error signal is used to communicate the speech to a remote location.