JP7027744B2 - Ultrasonic flow meter - Google Patents

Description

The present invention relates to an ultrasonic flow meter that measures the flow rate of a fluid by an ultrasonic method.

In the ultrasonic flow meter, an ultrasonic sensor for transmitting and receiving ultrasonic waves is arranged on the outer periphery of the conduit for the fluid flow. Then, the flow rate of the fluid moving in the pipeline can be measured by transmitting and receiving ultrasonic waves in the pipeline.

For example, in an ultrasonic flow meter called a propagation time difference method, ultrasonic waves are reciprocated in a path that diagonally crosses a fluid moving inside a pipeline. Then, the flow rate of the fluid is obtained from the difference in the time required for the ultrasonic wave to propagate. At this time, as the difference in propagation time, a difference of about several nanoseconds may be measured, and it is important to suppress noise mixed in the signal waveform of ultrasonic waves as much as possible in order to improve the accuracy of flow rate measurement.

Generally, noise includes noise that propagates directly in a pipeline and arrives before ultrasonic waves pass through a fluid and arrive, and so-called reverberation noise (stationary wave) in which ultrasonic waves are repeatedly reflected in a pipe. ), Etc., various noises are generated.

According to the invention described in Patent Document 1, a structure in which an ultrasonic absorber that attenuates ultrasonic waves is arranged between an attachment portion between an ultrasonic sensor and a conduit is disclosed.

Further, according to the invention described in Patent Document 2, a structure in which an ultrasonic absorber is fixed to the outer periphery of a pipeline is disclosed.

In this way, by using the ultrasonic absorber, it is possible to attenuate the line propagating wave and improve the S / N ratio.

Japanese Unexamined Patent Publication No. 2005-337911 Japanese Unexamined Patent Publication No. 2014-157129

However, in the configuration using the ultrasonic absorber as in the inventions described in Patent Document 1 and Patent Document 2, the signal of the ultrasonic wave itself is attenuated to some extent, and the S / N ratio is effectively improved. Can't. In addition, since an extra ultrasonic absorber is required, the mechanism is complicated and the cost is high.

The present invention has been made in view of such a problem, and an object thereof is an ultrasonic flow rate capable of effectively reducing noise mixed in an ultrasonic signal and measuring a flow rate with high accuracy. To provide a total.

The ultrasonic flow meter of the present invention is an ultrasonic flow meter that measures the flow rate of a fluid passing through a pipeline by using ultrasonic waves, and has an ultrasonic transmission unit that transmits the ultrasonic waves to the fluid and the ultrasonic waves. An ultrasonic receiving unit that receives the ultrasonic signal, an averaging filter that reduces noise from the ultrasonic signal received from the ultrasonic receiving unit, a flow rate calculation unit that calculates the flow rate of the fluid based on the ultrasonic signal, and the above. It has a non-linear filter using a non-linear function located between the averaging filter and the flow rate calculation unit, and the non -linear filter is an ultrasonic signal among the output waveforms input from the averaging filter. A noise analysis unit that detects noise amplitude based on the waveform of the timing before is generated, a threshold processing unit that determines a threshold based on the noise amplitude, and an ultrasonic signal superimposed on the ultrasonic signal based on the threshold. It is characterized by including a noise removing unit for reducing the generated noise .

With this configuration, noise mixed in the ultrasonic signal can be effectively reduced, and the flow rate can be measured with high accuracy.

According to the present invention, by providing a non-linear filter using a non-linear function between the averaging filter and the flow rate calculation unit, noise that overlaps with the ultrasonic signal in frequency can be effectively reduced, and the flow rate can be achieved with high accuracy. Can be measured.

It is a functional block diagram which shows the signal processing process of the ultrasonic flow meter which concerns on this embodiment, and is a schematic diagram which shows the state which the ultrasonic sensor constituting an ultrasonic flow meter is installed in a conduit. It is a block diagram which shows the structure of the nonlinear filter which concerns on this embodiment. The nonlinear function used in this embodiment is shown. FIG. 4A shows a signal waveform (output waveform of the averaging filter) before being input to the non-linear filter, FIG. 4B shows the output waveform of the non-linear filter in the present embodiment, and FIG. 4C shows the present embodiment. The output waveform when passing through a general linear filter instead of the non-linear filter in is shown. It is a block diagram which shows the structure of another embodiment different from the nonlinear filter shown in FIG. It is an example of the output waveform of each part of the nonlinear filter shown in FIG.

Hereinafter, the ultrasonic flow meter according to the present embodiment will be described in detail. In the following description, similar components are designated by the same reference numerals in all the drawings, and the description thereof will be omitted as appropriate.

FIG. 1 is a functional block diagram showing a signal processing process of the ultrasonic flow meter according to the present embodiment, and a schematic diagram showing a state in which an ultrasonic sensor constituting the ultrasonic flow meter is installed in a pipeline. As shown in FIG. 1, the ultrasonic flow meter of the present embodiment includes an ultrasonic transmission unit 1, an ultrasonic reception unit 2, an averaging filter 3, a nonlinear filter 4, and a flow rate calculation unit 5. It is composed of having.

As shown in FIG. 1, a pair of ultrasonic sensors 6 are installed on the outer peripheral portion of a pipeline 7 through which a fluid flows. One of the two ultrasonic sensors 6 for transmitting and receiving ultrasonic signals shown in FIG. 1 is arranged on the upstream side in the fluid flow direction, and the other is arranged on the downstream side in the fluid flow direction.

Each of the ultrasonic sensors 6 includes a transducer capable of transmitting and receiving ultrasonic waves, and the ultrasonic sensors 6 are alternately switched between the ultrasonic transmitting unit 1 and the ultrasonic receiving unit 2 to obtain ultrasonic waves. Sound waves are sent and received. That is, in FIG. 1, one ultrasonic sensor 6 is shown as an ultrasonic transmitter 1 and the other ultrasonic sensor 6 is shown as an ultrasonic receiver 2, but transmission / reception can be switched between the ultrasonic sensors 6. .. Then, the flow rate of the fluid is calculated based on the propagation time by measuring the propagation time when the ultrasonic wave is propagated in the flow direction and the opposite direction of the fluid.

The ultrasonic wave transmitted from the ultrasonic wave transmitting unit 1 to the ultrasonic wave receiving unit 2 passes through the fluid in the conduit 7 and propagates directly in the conduit 7, and the noise reverberating in the conduit 7. The voltage is converted into an ultrasonic signal by the ultrasonic receiving unit 2 under the influence of such factors. The ultrasonic signal converted into this voltage signal is amplified by an amplifier on a circuit (not shown) and passed through an A / D converter.

In the present embodiment, the flow rate of the fluid is calculated based on the ultrasonic signal after removing various noises from the ultrasonic signal including noise by filters 3, 4, and the like.

The averaging filter 3 reduces random noise from the input ultrasonic signal and lowers the level of the entire noise. By such a filter process, the threshold value ε can be stably detected by the nonlinear filter 4 described later. The ultrasonic signal (averaged output) from which random noise and the like have been removed in the averaging filter 3 is, for example, as shown in FIG. 2, a non-linear filter 4 via a bandpass filter 8 that passes only the band of the ultrasonic signal. Is entered in. By using the bandpass filter 8 as the filtering process of the ultrasonic signal, noise can be removed to some extent. However, the noise contains various frequency components, and the noise cannot be completely removed by a linear filter such as the bandpass filter 8.

Therefore, in the present embodiment, the non-linear filter 4 using the non-linear function effectively reduces the noise that overlaps with the ultrasonic signal in terms of frequency by using not only the frequency domain but also the amplitude information of the noise. ..

As shown in FIG. 2, the nonlinear filter 4 includes a noise analysis unit 4a, a threshold value processing unit 4b, and a noise removal unit 4c.

The noise analysis unit 4a analyzes the noise amplitude. By the way, the timing at which the ultrasonic waves reach the ultrasonic wave receiving unit 2 from the ultrasonic wave transmitting unit 1 should be predicted in advance to some extent by the distance between the ultrasonic sensors 6, the configuration of the conduit 7, and the speed of sound of the fluid. Can be done. Utilizing this, the noise analysis unit 4a can detect the noise amplitude based on the waveform before the ultrasonic signal is input. The noise detected by the noise analysis unit 4a is hereinafter referred to as small amplitude noise.

The threshold value processing unit 4b sets the threshold value ε based on the amplitude of the small amplitude noise detected by the noise analysis unit 4a. The noise removing unit 4c uses the threshold value ε set by the threshold value processing unit 4b to set the small amplitude noise to the threshold value ε or less. This makes it possible to remove small amplitude noise from the ultrasonic signal. In the present embodiment, by setting the threshold value ε to 1 time or more and 2 times or less the amplitude of the small amplitude noise, the non-linear filter 4 reliably filters the noise of the ultrasonic signal while suppressing it to the threshold value ε or less. can do. As a result, the flow rate can be measured with high accuracy by the flow rate calculation unit 5 based on the output of the nonlinear filter 4.

As the noise reduction unit 4c, for example, an ε filter can be used. The ε filter can effectively remove small-amplitude noise superimposed on a signal with a relatively large change without impairing the steepness of the signal. The input / output equation of the ε filter can be expressed as the equation (1), where x (n) is the input signal on which the small amplitude noise is superimposed and y (n) is the output signal of the ε filter.

は、非巡回型デジタルフィルタ（ＦＩＲ）である。ａ _ｉ は非巡回型デジタルフィルタのフィルタ係数であり、フィルタ係数の直流ゲインを１とするために、その総和は式（３）のように示される。

Is a non-circulating digital filter (FIR). _ai is the filter coefficient of the non-circular digital filter, and the total sum is expressed by the equation (3) in order to set the DC gain of the filter coefficient to 1.

Further, F (x) is a nonlinear function that limits the absolute value to ε, and is expressed as in Eq. (4).

FIG. 3 shows the nonlinear function F (x) used. In this way, the ε filter can be derived by desynchronizing the non-circular digital filter, and the waveform can be generated while suppressing the difference between the input signal x (n) and the output signal y (n) to the threshold value ε or less. Can be smoothed. The ε filter can remove small-amplitude noise that cannot be removed by a linear filter such as an analog circuit filter. The ε filter is a non-circular digital filter, and its phase characteristics do not change with frequency. That is, since the phase constant (phase linearity) is guaranteed, even if a steep frequency characteristic is provided, the distortion of the filter output waveform as used in a general linear filter does not occur.

The output signal of the noise reduction unit 4c of the nonlinear filter 4 is sent to the flow rate calculation unit 5. The flow rate calculation unit 5 calculates the flow rate of the fluid flowing through the pipeline 7. The noise analysis unit 4a of the non-linear filter 4 can easily detect the amplitude of the small amplitude noise by detecting the noise at the timing before the ultrasonic signal waveform is generated. In the threshold value processing unit 4b, the noise removal unit 4c suppresses the small amplitude noise to the threshold value ε or less by using the threshold value ε determined from the amplitude of the small amplitude noise, so that the small amplitude noise superimposed on the ultrasonic signal is effective. Can be reduced. Therefore, the flow rate calculation unit 5 can calculate the flow rate of the fluid with a high S / N ratio from the ultrasonic signal waveform in which the small amplitude noise is reduced.

As described above, in the present embodiment, it is not necessary to use the ultrasonic absorber shown in the patent document, a highly accurate ultrasonic flow meter can be realized with a simple configuration, and the cost is reduced as compared with the conventional case. be able to.

FIG. 4A shows a signal waveform (output waveform of the averaging filter 3) before being input to the nonlinear filter. FIG. 4B shows the output waveform of the nonlinear filter 4 in the present embodiment. FIG. 4C shows an output waveform when a general linear filter is passed instead of the non-linear filter 4 in the present embodiment.

As shown in FIG. 4A, in the output waveform of the averaging filter 3 (corresponding to the “input to the nonlinear filter” waveform shown in FIG. 2), noise having almost the same frequency is superimposed on the ultrasonic signal. Recognize. When the nonlinear filter 4 using the nonlinear function is applied to the output waveform of the averaging filter 3, not only the frequency but also the amplitude information of the small amplitude noise can be added. That is, for example, by using the nonlinear function F (x) as shown in FIG. 3, it is possible to remove the small amplitude noise according to the noise amplitude. Therefore, as shown in the output waveform of the nonlinear filter 4 shown in FIG. 4B, only the small amplitude noise can be appropriately removed, and a high S / N ratio can be obtained.

On the other hand, when a general linear filter is applied, the frequency bands of the ultrasonic signal and the noise overlap. Therefore, as shown in FIG. 4C, when the noise is reduced from the output waveform of FIG. 4A, the ultrasonic signal becomes The amplitude is also greatly reduced. Therefore, a high S / N ratio cannot be obtained in a configuration using a general linear filter.

By the way, when relatively large-amplitude noise (for example, spike noise) is superimposed on the received ultrasonic signal, it is difficult for the above-mentioned averaging filter 3 to remove the large-amplitude noise.

Therefore, it is possible to remove the large-amplitude noise by using the non-linear filter 14 shown in FIG. 5 different from that shown in FIG. An example of the output waveform of each part of the nonlinear filter of FIG. 5 will also be described with reference to FIG.

The nonlinear filter 14 shown in FIG. 5 includes a first difference calculation unit 14a, a second difference calculation unit 14b, and a noise reduction unit 14c.

A filter signal A (see FIG. 6A) that has passed through the bandpass filter 8 and an ultrasonic signal B that was previously acquired through the nonlinear filter 14 are input to the nonlinear filter 14. The previously acquired ultrasonic signal B (see FIG. 6B) is stored in advance in the storage unit 10 shown in FIG.

The filter signal A and the previously acquired ultrasonic signal B are differentially calculated by the first difference calculation unit 14a of the nonlinear filter 14, and the difference signal C shown in FIG. 6C is acquired. As a result, the signal component of the previously acquired ultrasonic signal B is removed from the filter signal A. The obtained difference signal C (remaining signal component) is dominated by the noise component.

As shown in FIG. 6C, large-amplitude noise and small-amplitude noise remain in the difference signal C. Large-amplitude noise is an abnormal noise component whose amplitude is suddenly larger than that of small-amplitude noise. The noise reduction unit 14c removes small amplitude noise from the difference signal C. In order to remove the small amplitude noise, the threshold value is set to a value slightly larger than the small amplitude noise. As a result, the noise removing unit 14c can remove the small amplitude noise and obtain the output signal D (see FIG. 6D) in which the large amplitude noise remains.

Before the noise removing unit 14c removes the small amplitude noise, the noise analysis unit analyzes the amplitude component of the noise, and the threshold value processing unit sets the threshold value to an amplitude slightly larger than the small amplitude noise. Such a noise analysis unit and a threshold value processing unit are provided between the first difference calculation unit 14a and the noise reduction unit 14c, or their functions are integrated in the noise reduction unit 14c. In FIG. 5, the functions of the noise analysis unit and the threshold value processing unit are integrated in the noise reduction unit 14c.

Then, the filter signal A and the output signal D acquired from the noise reduction unit 14c are differentiated by the second difference calculation unit 14b to obtain an output signal E from which the large amplitude noise is removed from the filter signal A. Can be done (see Figure 6E). As shown in FIG. 6E, small amplitude noise remains in the output signal E. As the noise of the output signal E, there is no large-amplitude noise and only small-amplitude noise is present as in the output signal shown in FIG. 4A.

The output signal E from which the large-amplitude noise has been removed is passed through the nonlinear filter 4 of FIG. As described above, the small amplitude noise is removed by passing through the nonlinear filter 4 of FIG.

The present invention is not limited to the above embodiment, and can be modified in various ways. In the above embodiment, the configuration shown in the attached drawings is not limited to this, and can be appropriately changed within the range in which the effect of the present invention is exhibited. In addition, it can be appropriately modified and implemented as long as it does not deviate from the scope of the object of the present invention.

For example, in the above embodiment, the filtering by the bandpass filter 8 may be before the averaging filter 3, may be incorporated in the averaging filter 3, or may be incorporated in the nonlinear filter 4. It may be. Further, the nonlinear function shown in FIG. 3 is also an example and is not limited to this.

Finally, the feature points in the above-described embodiment are summarized.

In an ultrasonic flow meter that measures the flow rate of a fluid passing through a conduit 7 according to the above embodiment by using ultrasonic waves, an ultrasonic transmission unit 1 that transmits ultrasonic waves to the fluid and ultrasonic waves are received. The ultrasonic receiving unit 2, the averaging filter 3 that reduces noise from the ultrasonic signal received from the ultrasonic receiving unit 2, the flow rate calculation unit 5 that calculates the flow rate of the fluid based on the ultrasonic signal, and the averaging filter. It has a non-linear filter 4 using a non-linear function located between 3 and the flow rate calculation unit 5, and an ultrasonic wave output from the averaging filter 3 based on a threshold set by the non-linear filter 4. It is characterized by reducing signal noise.

As described above, by providing the nonlinear filter 4 using the nonlinear function between the averaging filter 3 and the flow rate calculation unit 5, it is possible to appropriately reduce the small amplitude noise from the output signal of the averaging filter 3. Therefore, the noise that overlaps with the ultrasonic signal in frequency can be effectively reduced, and the flow rate can be measured with high accuracy.

Further, in the ultrasonic flow meter according to the above embodiment, the nonlinear filter 4 has the noise analysis unit 4a for detecting the noise amplitude and the noise amplitude based on the waveform at the timing before the ultrasonic signal is generated. It is preferable to include a threshold processing unit 4b that determines the threshold based on the threshold, and a noise removing unit 4c that removes noise of the ultrasonic signal based on the threshold.

As described above, in the present embodiment, the noise analysis unit 4a can detect the noise amplitude in the waveform at the timing before the ultrasonic signal is generated. This makes it possible to detect the noise amplitude easily and accurately. In addition, noise amplitude can be detected by appropriately absorbing noise fluctuations due to the environment. Then, a threshold value can be set based on this noise amplitude, and in the noise removing unit 4c, the small amplitude noise is determined according to the magnitude of the noise amplitude of the small amplitude noise superimposed on the ultrasonic signal based on the threshold value. It is possible to properly remove the noise.

Further, in the ultrasonic flow meter according to the above embodiment, the threshold value is preferably set to twice or more the noise amplitude. This makes it possible to effectively remove small-amplitude noise from the ultrasonic signal.

Further, in the ultrasonic flow meter according to the above embodiment, the noise reduction unit 4c can be an ε filter.

The ε filter can effectively remove small-amplitude noise that cannot be removed by a linear filter such as an analog circuit filter from an ultrasonic signal without impairing the steepness of the signal.

Further, another nonlinear filter 14 shown in FIG. 5 can be arranged in front of the nonlinear filter 4. The nonlinear filter 14 shown in FIG. 5 includes a first difference calculation unit 14a, a noise reduction unit 14c, and a second difference calculation unit 14b. The first difference calculation unit 14a performs a difference calculation between the filter signal A from the averaging filter and the ultrasonic signal B obtained through the nonlinear filter 14 last time. The noise reduction unit 14c removes signals having a noise amplitude equal to or lower than the noise amplitude set in the threshold value. The second difference calculation unit 14b performs a difference calculation between the filter signal A and the output signal D from the noise reduction unit 14c.

As a result, even when an abnormally large-amplitude noise such as spike noise is superimposed, the large-amplitude noise can be appropriately removed.

Further, in the ultrasonic flow meter according to the above embodiment, the noise removing unit 14c provided in another nonlinear filter 14 analyzes the amplitude component based on the difference signal and sets the noise amplitude as the threshold value. However, it is preferable to remove signals having a noise amplitude or less set.

In this way, not only the noise reduction function but also the noise analysis function and the threshold value processing function can be integrated in the noise reduction unit 14c. As a result, the non-linear filter 14 can be made compact and the processing capacity can be increased.

The present invention described above is useful for an ultrasonic flow meter that measures the flow rate of a fluid from the difference in the propagation time of the ultrasonic waves when the ultrasonic waves are propagated diagonally to the flow of a fluid such as gas or liquid. Is.

1: Ultrasonic transmitter 2: Ultrasonic receiver 3: Averaging filter 4, 14: Non-linear filter 4a: Noise analysis unit 4b: Threshold processing unit 4c, 14c: Noise removal unit 5: Flow calculation unit 6: Ultrasonic sensor 7: Pipeline 8: Band path filter 10: Storage unit 14a: First difference calculation unit 14b: Second difference calculation unit A: Filter signal B: Ultrasonic signal C: Difference signal D: Output signal E: Output signal F: Non-linear function x: Input signal y: Output signal ε: Filter

Claims (4)

In an ultrasonic flow meter that measures the flow rate of fluid passing through a pipeline using ultrasonic waves,
An ultrasonic transmission unit that transmits the ultrasonic waves to the fluid, and
The ultrasonic receiving unit that receives the ultrasonic waves and
An averaging filter that reduces noise from the ultrasonic signal received from the ultrasonic receiver,
A flow rate calculation unit that calculates the flow rate of the fluid based on the ultrasonic signal,
A non-linear filter using a non-linear function located between the averaging filter and the flow rate calculation unit,
Have,
The nonlinear filter includes a noise analysis unit that detects a noise amplitude based on a waveform at a timing before the ultrasonic signal is generated among the output waveforms input from the averaging filter.
A threshold processing unit that determines a threshold based on the noise amplitude,
A noise reducing unit that reduces noise superimposed on the ultrasonic signal based on the threshold value, and
An ultrasonic flow meter characterized by being equipped with. The ultrasonic flow meter according to claim 1, wherein the threshold value is set to twice or more the noise amplitude. The ultrasonic flow meter according to claim 1 or 2, wherein the noise removing unit is an ε filter. The difference between the first difference calculation unit and the first difference calculation unit that perform a difference calculation between the filter signal from the averaging filter and the ultrasonic signal obtained through the nonlinear filter last time before the non-linear filter. A second unit that performs a difference calculation between the noise removing unit that removes noise from the difference signal, the filter signal, and the output signal from the noise removing unit using a threshold value set based on the noise amplitude of the signal. It has another non-linear filter with a difference calculator and
The ultrasonic filter according to any one of claims 1 to 3, wherein an output waveform from the second difference calculation unit is input to the nonlinear filter located after the other nonlinear filter. Sonic flow meter.

Images