Liquid metal flow condition detection

• US 6,539,805 B2
• Filed: 03/15/2001
• Issued: 04/01/2003
• Est. Priority Date: 07/19/1994
• Status: Expired due to Fees

• Current Assignee: Vesuvius Crucible Company (Vesuvius Plc)
• Original Assignee: Vesuvius Crucible Company (Vesuvius Plc)
• Inventors: Dorricott, James D., Heaslip, Lawrence J.
• Primary Examiner(s): Larkin, Daniel S.
• Assistant Examiner(s): MILLER, ROSE MARY
• Application Number:
  US09/812,110
• Publication Number:
  US 20010029785A1
• Time in Patent Office: 747 Days
• Field of Search: 736/49, 73/18.3, 73/18.6, 73/645.3, 736/58, 736/59, 164/453, 164/457, 164/4.1, 164/150.1, 340/683, 137/486, 266/99, 702/34, 702/35, 702/39, 702/48, 702/56, 702/103
• US Class Current: 73/649
• CPC Class Codes: B22D 2/001 for the slag appearance in a molten metal stream
• CPC Class Codes: G01F 1/666 by detecting noise and sounds generated by the flowing fluid
• CPC Class Codes: G01N 2291/014 Resonance or resonant frequency
• CPC Class Codes: G01N 2291/0252 Melting, molten solids
• CPC Class Codes: G01N 2291/02836 Flow rate, liquid level
• CPC Class Codes: G01N 29/036 by measuring frequency or resonance of acoustic waves
• CPC Class Codes: G01N 29/14 using acoustic emission techniques echo of particles G01N29/046; measuring mechanical vibrations or acoustic waves in solids in general G01H1/00
• CPC Class Codes: G01N 29/4436 with a reference signal amplitude comparison G01N29/48
• CPC Class Codes: G01N 29/449 Statistical methods not provided for in G01N29/4409, e.g. averaging, smoothing and interpolation
• CPC Class Codes: G01N 29/46 by spectral analysis, e.g. Fourier analysis or wavelet analysis spectral signal processing per se G06F17/14

Abstract

An apparatus and method for detecting the condition of the flow of liquid metal in or from a teeming vessel includes one or more sensors for detecting vibration caused by a flow of liquid metal in or from the teeming vessel and for outputting a sensor signal corresponding to mechanical and acoustic vibrations detected by the sensor. A signal processor receives the sensor signal and compares the sensor signal to a reference calibration signal and outputs a comparison signal. A logic unit receives the comparison signal and outputs a status signal indicative of the condition of the flow of the liquid metal in or from the teeming vessel which can be used to stop the flow of metal. The calibration signal can be static or may be dynamically updated.

Claims

1. A method of detecting the condition of the flow of a liquid metal through an outlet from a teeming vessel, to indicate an undesired flow condition, the method comprising:

a) generating a vibration calibration signal indicative of a desired flow condition of the liquid metal;

b) sensing vibrations in the liquid metal flow;

c) generating a sensor signal embodying information regarding the sensed vibrations;

d) analyzing the sensor signal into a frequency band spectrum comprising multiple frequency bands;

e) comparing at least one of the frequency bands with the calibration signal; and

f) generating a status signal indicating an undesired flow condition in response to unacceptable differences between the sensor signal and the calibration signal.

2. A method according to claim 1 comprising outputting the frequency band spectrum for comparison with the calibration signal as a digital data spectrum.

3. A method according to claim 2 wherein the frequency band spectrum lies in a frequency range of from 0.1 Hz to 20 kHz.

4. A method according to claim 1 performed in a steel-making process, the liquid metal being liquid steel, wherein the frequency band spectrum lies in a frequency range of from 10 Hz to 1 kHz.

5. A method according to claim 1 wherein the at least one frequency band compared with the calibration signal is selected from among the frequency bands available from analysis of the sensor signal as being a frequency band responsive to the undesired flow condition.

6. A method according to claim 1 wherein the sensor signal is an analog signal, the method further comprising converting the sensor signal to a digital data signal prior to comparing at least one of the frequency bands with the calibration signal, employing constant-percentage bandwidth filters to analyze the sensor signal into the frequency band spectrum and outputting the analyzed signal for comparison with the calibration signal as a digital data spectrum of frequencies.

7. A method according to claim 1 wherein the sensor signal is continuously analyzed into the frequency band spectrum by a real-time frequency analyzer permitting rapid frequency analysis of the sensor signal and wherein spectral comparison with the calibration signal is effected simultaneously with frequency analysis to avoid loss of signal data.

8. A method according to claim 1 wherein the liquid metal is steel, the frequency band spectrum is output as a rapidly and continuously generated digital data spectrum, the desired flow condition is a substantially slag-free flow condition and wherein the method further comprises:

g) calculating a standard deviation for the calibration signal corresponding to acceptable differences between the sensor signal and the calibration signal;

h) employing the calculated standard deviation when comparing the digital data spectrum to the calibration spectrum by continually laying the data spectrum over the calibration spectrum to determine when the intensity level of the data spectrum is outside the calculated standard deviation;

and

i) generating the status signal indicating an undesired flow condition when the differences between the data spectrum and the calculated standard deviation are unacceptable.

9. A method according to claim 8 comprising determining the magnitudes of differences in intensity between selected frequency bands of the data spectrum and the calibration spectrum, the frequency bands being selected for ability to indicate undesirable flow conditions, to yield spectral comparison data for generation of the status signal.

10. A method according to claim 1 wherein the calibration signal comprises a data spectrum of frequency intensities and a standard deviation is applied to the calibration signal for comparison with each frequency band whereby frequency band intensities exceeding the standard deviation of the calibration spectrum indicate an undesired flow condition.

11. A method according to claim 10 wherein the standard deviation is determined by sensing vibrations while pouring liquid metal from a teeming vessel, during an undesired flow event, in a preliminary step prior to flow detection.

12. A method according to claim 1 wherein the calibration signal is vessel-specific and is generated for each teeming vessel by sensing vibrations during desirable flow conditions through the teeming vessel, the method comprising using the vessel-specific calibration signal for comparison with the at least one frequency band to generate the status signal.

13. A method according to claim 12 comprising flowing liquid metal from a charged teeming vessel until a status signal indicating an undesired flow condition is generated and generating the vessel-specific calibration signal by sensing vibrations in the flowing liquid prior to occurrence of an undesired flow condition.

14. A method according to claim 13 implemented to monitor a flow of molten steel in a continuous casting process wherein molten steel is poured from successive teeming vessels through an outlet in each teeming vessel and wherein the method is repeated for each successive teeming vessel.

15. A method according to claim 13 comprising applying a standard deviation to the calibration spectrum used for comparison with the at least one frequency band, the standard deviation being determined by sensing vibrations in a liquid metal flow having undesired flow conditions.

16. A method according to claim 15 comprising effecting a sensitivity adjustment by increasing or decreasing the standard deviation.

17. A method according to claim 1 comprising generating spectral comparison data from the comparison of the at least one frequency band with the calibration signal and employing a logic unit to effect generation of the status signal by calculating the differences between the sensed signals and the calibration spectrum and generating the status signal indicating an undesired flow condition when the magnitude of the variance of the spectral comparison data is outside a predetermined confidence interval.

18. A method according to claim 17 comprising employing the logic unit to adjust the confidence interval and the magnitude of the differences between the adjusted confidence interval and the measured spectral data, to allow the adjustment of the sensitivity of the system to the flow condition changes previously defined.

19. A method according to claim 18 wherein individual said frequency bands have independently adjustable confidence intervals, the method comprising independently adjusting the confidence intervals to adjust the sensitivity of the system to different flow conditions.

20. A method according to claim 17 wherein the logic unit receives inputs from a teeming vessel weight sensor and a valve position actuator controlling outflow of liquid metal through the teeming vessel outlet to allow the sensitivity of the detection method to be varied as a function of teeming vessel weight and to enable automatic control and closing of the valve in response to an undesired flow condition.

21. A method according to claim 17 wherein the logic unit generates the status signal comprising employing the status signal to indicate an alarm state of the flow condition or to facilitate operator response or to automatically initiate closure of a valve controlling the liquid metal flow.

22. A method according to claim 1 comprising using an accelerometer to sense the vibrations in the liquid flow, the accelerometer communicating with the outlet through an elastic solid to receive vibrations from the liquid metal flow, the accelerometer being in direct contact with the elastic solid.

23. A method of detecting an undesired flow condition in a flow of molten steel in a continuous casting process wherein molten steel is poured from successive teeming vessels through an outlet in each teeming vessel, the method comprising:

a) commencing pouring of liquid steel from one of the successive teeming vessels through the teeming vessel outlet;

b) generating a vessel-specific vibration calibration signal indicative of a desired flow condition of the liquid metal during the pouring of the liquid steel from the one teeming vessel;

c) sensing vibrations in the liquid steel flow;

d) generating a sensor signal embodying information regarding the sensed vibrations;

e) comparing the sensor signal with the calibration signal;

f) generating a status signal indicating an undesired flow condition in response to differences between the sensor signal and the calibration signal;

g) replacing the one teeming vessel with a new teeming vessel charged with liquid steel; and

h) repeating elements b) through f) with the new teeming vessel.

24. A method according to claim 23 further comprising

i) analyzing the sensor signal into a frequency band spectrum comprising multiple frequency bands; and

j) selecting from among the multiple frequency bands a frequency band or bands responsive to the undesired flow condition and employing the selected frequency band or bands for comparison of the sensor signal with the calibration signal.

25. A method according to claim 23 wherein the calibration signal comprises a data spectrum of frequency intensities and a standard deviation is applied to the calibration signal for comparison with each frequency band whereby frequency band intensities exceeding the standard deviation of the calibration spectrum indicate an undesired flow condition.

26. A method according to claim 25 wherein the standard deviation is determined by sensing vibrations while pouring liquid metal from a teeming vessel, during an undesired flow event, in a preliminary step prior to flow detection.

27. A method of detecting the condition of the flow of a liquid metal, optionally molten steel, through an outlet from a teeming vessel, to indicate an undesired flow condition, the method comprising:

a) generating a vibration calibration signal indicative of a desired flow condition of the liquid metal;

b) applying a standard deviation to the calibration signal whereby vibration intensities exceeding the standard deviation of the calibration signal indicate an undesired flow condition;

c) sensing vibrations in the liquid metal flow;

d) generating a sensor signal embodying information regarding the sensed vibrations;

e) generating a status signal indicating an undesired flow condition in response to unacceptable differences between the sensor signal and the calibration signal.

28. A method according to claim 27 comprising determining the standard deviation by sensing vibrations while pouring liquid metal from a teeming vessel, during an undesired flow event, in a preliminary step prior to flow detection.

29. A method according to claim 28 wherein the calibration signal comprises a data spectrum of frequency intensities and the standard deviation is applied to the calibration signal for comparison with each frequency band whereby frequency band intensities exceeding the standard deviation of the calibration spectrum indicate an undesired flow condition.

30. A method of predicting an undesired condition in a liquid metal flow through an outlet from a molten metal pouring vessel, the method comprising the following elements:

a) sensing flow-related perturbations in a physical parameter indicative of the undesired condition of the liquid metal flow to generate a sensed parameter signal;

b) comparing the sensed parameter signal with a reference signal to generate a difference signal, the reference signal being generated prior to the sensed parameter signal;

c) interpreting the difference signal to predict the presence of the undesired condition in the liquid metal outflow; and

d) updating the reference signal with more recent data from the sensed parameter signal.

31. A method according to claim 30 further comprising

e) repeating elements a) through d) so that the difference signal indicates changes of the sensed parameter with time.

32. A method according to claim 30 wherein elements b), c) and d) are performed by a logic unit and cyclically repeated on each cycle of the logic unit.

33. A method according to claim 30 wherein the sensed parameter comprises natural vibrations sensed by a vibration sensor physically connected with the liquid metal flow through vibration-transmissive solid structure.

34. A method according to claim 32 wherein the reference signal becomes a predicted data spectrum that can be updated on each cycle of the logic unit.

35. A method of predicting an undesired condition in a flow of liquid metal comprising the elements of:

a) obtaining time-sequenced calibration spectra of a physical wave phenomenon indicative of the flow condition;

b) calculating a predicted data spectrum of the physical wave phenomenon from the calibration spectra to indicate a flow condition at a future point in time;

c) obtaining a current data spectrum at the future point in time;

d) comparing the current data spectrum with the predicted data spectrum to provide spectral comparison data;

e) processing the spectral comparison data to provide a status signal;

f) storing the current data spectrum as the most recent entry in the sequence of calibration spectra;

g) repeating elements b) through f).

36. A method according to claim 35 wherein elements b) through f) are performed by a logic unit and cyclically repeated on each cycle of the logic unit.

37. A method according to claim 35 wherein the physical wave phenomenon comprises natural vibrations sensed by a vibration sensor physically connected with the liquid metal flow through vibration-transmissive solid structure.

38. A method according to claim 37 wherein the time sequenced calibration spectra are generated by monitoring desirable flow conditions with the vibration sensor and are stored and wherein the calibration spectra are statistically analyzed and the predicted data spectrum is generated by forward extrapolation of the calibration spectra.

39. A method according to claim 35 wherein an old calibration spectrum is removed from the calibration spectra when the current data spectrum is stored to the sequence of calibration spectra.

40. A method according to claim 35 wherein a standard deviation is applied to the predicted data spectrum to facilitate processing of the spectral comparison data.

41. A method according to claim 35 wherein the time-sequenced calibration spectra and the current data spectrum comprise multiple frequency bands and wherein each frequency band is processed by applying a standard deviation to the respective calibration frequency band and comparing the respective data frequency band to the respective calibration frequency band whereby frequency band intensities exceeding the standard deviation of the calibration spectrum indicate an undesired flow condition.

42. A method according to claim 37 wherein the vibration sensor is mechanically coupled to a teeming vessel outlet to sense vibrations in molten steel flowing through a pouring tube leading from the teeming vessel.

43. A method according to claim 35 wherein flow condition detection system is coupled with a liquid metal flow shut off valve for the valve to be operated automatically in response to a negative status signal.