Measuring and testing – Vibration – By mechanical waves
Reexamination Certificate
2003-06-03
2004-06-08
Williams, Hezron (Department: 2856)
Measuring and testing
Vibration
By mechanical waves
C073S602000, C073S861060, C073S861270, C073S861280, C073S861290
Reexamination Certificate
active
06745632
ABSTRACT:
CROSS-REFERENCE TO RELATED APPLICATIONS
This application claims benefit of no related applications.
TECHNICAL FIELD OF THE INVENTION
This invention generally relates to an improvement in the measurement of the transit time of an ultrasonic wave through a fluid. More particularly, embodiments of the invention relate to an algorithm allowing determination of the transit time in the presence of noise and dispersion.
BACKGROUND OF THE INVENTION
In many fields use has been made of the measurement of the transit time for an ultrasonic wave through a fluid. This measurement determines the properties, flow or position of the fluids (liquid, gas, or particulate) through which the wave travels, or the location of bodies in the fluid distorting the flow of that waveform. The transit time must be measured to a high degree of accuracy and resolution. This has been done either by measuring the phase angle of the received wave compared to the phase angle of the transmitted wave, by measuring the time-of-flight of a wave packet, or by determining the peak of the cross-correlation with the transmitted wave.
The measurement of the phase angle difference between the transmitted and received signals, as by synchronous demodulation, has several drawbacks. The accurate resolution of the phase angle requires the use of a high frequency signal with a short period. It then becomes difficult to account for the number of integral periods between the transmitted and received signals since displacing the signal by one period will give the same response when the phase is measured. For a digital determination of the phase, a measurement significantly higher than the Nyquist frequency is often employed, which increases the required computational power in digital measurements. In addition, if multiple paths for the ultrasonic signal are present, they will combine at the receiver as a single frequency signal with a phase angle shift representing an inseparable contribution from each path. A phase measurement has an advantage in that a single frequency signal is commonly used and the accurate filtering of the received signal allows a large rejection of off-frequency noise.
A time-of-flight measurement modulates the envelope of the transmitted wave into a packet and detects the arrival of the packet at the receiver. The packet represents an increase of the bandwidth of the transmitted wave to roughly the inverse of the packet width. This requires a larger bandwidth in the filter of the received packet and requires an increase in the transmitted packet energy when noise is present due to increased noise passing through the input filter. Often there are multiple waveform paths slightly differing from the unperturbed waveform caused by inhomogeneous media, turbulence in the media or reflections. This uncontrollably broadens a pulse and makes the time-of-flight measurement prone to inaccuracies. Usually the determination of the arrival of the variably attenuated envelope is too imprecise to obtain an accurate transit time, so additional measurements are made on the carrier waveform using the packet arrival as a gating mechanism. This, in turn requires a high rate of data measurements and significant computations. Often these systems average over many measurements in order to improve the accuracy and resolution.
A common problem to both phase measurements and time-of-flight measurements is the nature of noise encountered in fluid measurements. Often the noise is generated by the disruptions to the fluid flow, giving rise to a “whistle” noise generation. This noise is concentrated at particular frequencies that vary with the flow rate and flow conditions as opposed to a wider bandwidth “white” noise. The measured signal can be severely affected when the noise is in the region of the measured frequencies.
In an attempt to reduce the influence of this noise, use has been made of the cross-correlation of the launched waveform and the detected waveform. U.S. Pat. No. 4,576,047 and U.S. Pat. No. 4,576,047 describe attempts to use the peak of the correlation waveform as an indication of the transit time in such a manner.
There are two principle problems encountered in attempts to measure the transit time using the methods of the prior art: the accurate determination of the transit time in the presence of noise, and the effects of dispersion on the measurement. The effects of the noise have been addressed in phase and time-of-flight measurements by increasing the power of the transmitted pulse to improve the signal-to-noise ratio. The noise is more successfully attacked by utilizing the cross-correlation between the received and transmitted waveforms. A properly chosen launched waveform has little cross-correlation with anticipated noise. The effects of dispersion in the measurements have not been generally addressed in the prior art.
There can be several causes of dispersions in the measurement of the transit time. Inhomogeneous media, where there are fluctuations in the composition or pressure of a fluid, or turbulence in the media can cause fluctuations in the velocity of sound across the transmit path causing a transmitted pulse to spread, in effect “smearing” the arrival time. A similar smearing of the arrival time is present if the media velocity is non-uniform in time, e.g. if there are “surges” in the flow. In this case, if the measurement is averaged over longer times to reduce the noise, as is done in phase measurements or cross-correlation measurements, the accelerating or deceleration appears as dispersion in the final measurement. In the case of time-of-flight measurements in such a condition the measurement is a “snapshot” in time and must be repeated often enough to obtain a statistical mean of the transit time. The aforementioned effects can be exacerbated with an increase in flow velocity in fluids, particularly gasses. The effect can be expected to depend on such parameters as average velocity, wall condition, angle to the flow path, and flow profile. There can be a systematic bias in the transit time measurements due to dispersion effects, e.g. the spread of the pulse leading edge can cause a time-of-flight measurement to indicate an early detection of a pulse compared to the pulse center.
SUMMARY OF THE INVENTION
To address these problems this invention proposes that the transmitting ultrasonic transducer be excited with a frequency-modulated signal and the signal from the receiving transducer be appropriately digitized. A reference waveform is created utilizing either the transmitted waveform, a calculation of the expected received waveform, or a measurement of the received waveform under reference conditions. The square of the convolution of the reference waveform with the received waveform is created. This calculation can be further simplified by undersampling the received signal. The squared convolution is normalized to unity area and the time-weighted average of this normalized function is summed over the proper region to give the transit time. Algorithms are presented for this calculation.
The summation can be either over minimums in the squared convolution envelope or between the minimums about the highest peak. In the latter case if the measurements are appropriately spaced a minimum number of calculations will still give a very accurate result. The region over which the summation is made is determined by the symmetry of the squared convolution in the vicinity of the peak to be considered. If there is insufficient symmetry to indicate that the measurement will be accurate, the fundamental frequency of the frequency-modulated signal may be reduced or deconvolution may be used to separate the composite signals. The former case is useful where the dispersion is asymmetric.
REFERENCES:
patent: 4301684 (1981-11-01), Thompson et al.
patent: 4307616 (1981-12-01), Vasile
patent: 4837578 (1989-06-01), Gammell
patent: 5042303 (1991-08-01), Geluk et al.
patent: 5197019 (1993-03-01), Delon-Martin et al.
patent: 5203823 (1993-04-01), Wadaka et al.
patent: 5650571 (1997-07-01), Freud et al.
patent: 5770806 (
Dryer Joseph Ernest
Lambert John David
Miller Rose M.
Williams Hezron
LandOfFree
Method for measuring ultrasonic transit times does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for measuring ultrasonic transit times, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for measuring ultrasonic transit times will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3365489