JP5097503B2 - Ultrasonic flow meter signal processor - Google Patents

Description

  The present invention relates to a signal processing apparatus for an ultrasonic flowmeter that digitally processes a signal.

  As a method of measuring the flow velocity in the pipeline using an ultrasonic beam, the difference between the propagation time of the ultrasonic beam from upstream to downstream and the propagation time of the ultrasonic beam from downstream to upstream is obtained. A time difference method for calculation is often used.

  As a method for obtaining this time difference, Patent Document 1 discloses a method using a zero cross point of a signal, and Patent Document 2 discloses a method using a cross-correlation of signals.

JP 2002-162269 A JP 2002-243514 A

  In an ultrasonic flowmeter, a pulsed ultrasonic beam is transmitted in a fluid with a period of several ms to several tens of ms, a passing signal or a reflected signal is received, and a frequency shift due to the arrival time or the Doppler effect is received. Etc. are measured to determine the fluid velocity, thereby measuring the flow rate. As a method of measuring the arrival time, there are a method of capturing the rising edge of the received signal, a method of capturing the zero cross point of the received signal, and the like.

  In addition, as a method of calculating the time difference between the arrival time from the upstream side to the downstream side of the fluid and the arrival time in the reverse direction, a method using the cross-correlation of the signal or a multiplication part of the calculation of the cross-correlation There is a method of using the one replaced with an absolute value.

  FIG. 3 is a waveform diagram of data obtained from an ultrasonic beam propagating in two directions in the first and second pipes 1. As a method for obtaining a time difference from this signal, there is a method using cross-correlation of signals described in Patent Document 2.

For example, a method for obtaining the arrival time difference using the cross-correlation will be described. The cross-correlation at the corresponding point of the first and second data x and y consisting of the number N of data is obtained by Expression (1).
Rxy [m] = Σ | x [n] * y [n + m] |
(M = 0, 1, 2,..., N-1) (1)

  In Equation (1), Rxy represents a correlation function, and x [n] and y [n + m] are functions (data) for obtaining cross-correlation. The subscript n + m of y uses the remainder divided by N. The value of m that gives the point where the value of the correlation function is maximized gives the time difference between the two data x and y, that is, the time difference of the received ultrasonic beam.

  FIG. 4 is a waveform diagram obtained as a result of calculating cross-correlation for the data of FIG. The time difference between the signals is obtained by the deviation of the maximum value of the waveform from the origin.

The following equation (2) is a calculation method using a signal difference.
Sxy [m] = Σ | x [n] −y [n + m] |
(M = 0,1,2, ..., N-1) (2)

  In equation (2), the multiplication part of equation (1) is replaced with the absolute value of subtraction. The reason for replacing multiplication with subtraction is that subtraction is advantageous in various aspects. For example, the area of the circuit, current consumption, calculation time, and the like, and subtraction is more advantageous than multiplication in these alone or in combination. This is the same whether the processing is performed by hardware or software.

  When performing signal processing to determine the fluid flow velocity by these methods, it is convenient to make the amplitude of the received waveform of the ultrasonic beam as constant as possible. Therefore, normally, an AGC (Automatic Gain Control) loop is formed. The amplitude of the received waveform is controlled to be a constant value. However, since the gain at the time of the next reception is determined based on the current reception data, there is a problem that the gain adjustment cannot follow if a fluctuation occurs in a time shorter than the emission cycle of the ultrasonic beam.

  In order to make the signal amplitude constant in normal AGC, the gain of the variable gain amplifier for reception is adjusted based on the peak value, average value, or effective value of the received signal. As a configuration of the variable gain amplifier for reception, there are a method of switching the feedback resistance of the amplifier with an analog switch, a method of using an analog multiplier, and the like.

  In the method of switching the feedback resistor with an analog switch, the gain cannot be continuously changed. Therefore, when the signal continuously changes, the amplitude cannot be made constant.

  In the method using an analog multiplier, if the gain control voltage can be changed continuously, the gain can also be changed continuously. However, when the control voltage is generated by a DA converter, it is necessary to increase the resolution (number of bits) of the D / A converter in order to perform continuous control. Use of a transducer is difficult.

  Further, in the ultrasonic flowmeter, the amplitude measurement and the gain control of the variable gain amplifier are periodically performed, for example, several ms to several tens of ms, and therefore cannot follow changes faster than this cycle.

  An object of the present invention is to provide a signal processing apparatus for an ultrasonic flowmeter that can solve the above-described problems and can obtain data having a constant amplitude even for a signal that changes faster than the response time of a feedback loop. There is to do.

In order to achieve the above object, an ultrasonic flowmeter signal processing device according to the present invention is alternately arranged in a first direction from the downstream side to the upstream side of the conduit and in a second direction from the upstream side to the downstream side. The ultrasonic beam is transmitted and received, and the obtained signal from the ultrasonic beam is stored as digital data in the memory for each of the data in the first and second directions, and the data is read from the memory for signal processing. In the signal processing device of the ultrasonic flowmeter, as a numerical value representing the characteristics of the data obtained for each of the two types of data in the memory , at least the reciprocal number based on the maximum value of each of the two types of digital data first means for determining the coefficients comprising, by obtaining the difference by multiplying the coefficient for each two data read out from said memory, seeking propagation time difference of the ultrasonic beam by two types of data Characterized by comprising a second means for measuring the amount.

  According to the signal processing device of the ultrasonic flowmeter according to the present invention, it is possible to obtain an accurate ultrasonic beam propagation time difference that is a basis of flow rate measurement in a short time, so that the unit time is larger than in the case of cross-correlation. It is possible to increase the number of times of measurement per unit, and if the number of times of measurement is the same as the conventional one, the operation speed of the circuit can be reduced.

  Even if the AGC control is incomplete, the signal processing circuit can be supplied with signals having the same amplitude. Furthermore, it is possible to more accurately determine whether normal reception has been performed.

The present invention will be described in detail based on the embodiment shown in FIGS.
FIG. 1 is a block circuit diagram of the embodiment. An upstream ultrasonic transducer 2 and a downstream ultrasonic transducer 3 that both transmit and receive ultrasonic beams are arranged on the upstream side and the downstream side of the conduit 1 through which the fluid to be measured flows. These ultrasonic transducers 2 and 3 are connected to a transmission amplifier 5 and a reception variable gain amplifier 6 through a changeover switch 4. The output of the receiving variable gain amplifier 6 is connected to an A / D converter 7 and a data bus 8. The data bus 8 is connected to a peak detection circuit 9, an arithmetic unit 10, a memory 11, a CPU 12 having a RAM and a ROM, and an I / O controller 13.

  The memory 11 has two memories RAMa and RAMb, and their outputs are connected to the arithmetic unit 10 via a multiplier 14. The multiplier 14 also has two multipliers, and the outputs of the RAMa and RAMb are output to the arithmetic unit 10 via the two multipliers. The output of the clock generation circuit 15 is connected to the I / O controller 13, and the output of the I / O controller 13 is connected to the changeover switch 4, the transmission amplifier 5, the variable gain amplifier 6, and the A / D converter 7. Has been.

  In the time difference type ultrasonic flowmeter, as described above, the flow velocity of the fluid is obtained from the difference in arrival time between when the flow is transmitted from the upstream side to the downstream direction and when the flow is transmitted in the reverse direction. The flow rate is measured by multiplying the cross-sectional area of the road. The operation of each circuit at the time of measurement is performed by a program command stored in a ROM built in the CPU 12.

  The steps of the process using the difference in this embodiment will be described in time series.

  [Step S1] The changeover switch 4 is controlled by a command from the CPU 12, the upstream ultrasonic transducer 2 is connected to the output of the transmission amplifier 5, and the downstream ultrasonic transducer 3 is connected to the reception variable gain amplifier 6. The peak detection circuit 9 is initialized.

  [Step S2] In response to a command from the CPU 12, the I / O controller 13 generates a burst signal for driving the upstream ultrasonic transducer 2. This burst signal is converted into an ultrasonic beam by the ultrasonic transducer 2, propagates through the pipe 1, is received by the downstream ultrasonic transducer 3, and is converted into an electrical signal. This electric signal is amplified by the receiving variable gain amplifier 6, converted into a digital signal by the A / D converter 7, and stored in the RAM a of the memory 11 through the data bus 8. Thereby, data as shown in FIG. 3 is obtained, and the stored data is set as the first data x. At the same time, the peak detection circuit 9 detects the maximum value representing the characteristics of the digital data. The maximum value at this time is defined as a first peak value Px.

  The propagation time of the ultrasonic beam is about 60 μs when the fluid is water and the distance between the two ultrasonic vibrators 2 and 3 is 10 cm. On the other hand, if the frequency of the ultrasonic beam is 2 MHz and a signal wave of about 10 cycles can be received, the duration is 5 μs. In order to store this signal, if it is 10 μs with a margin, it is preferable to store 512 points when the clock frequency of AD conversion is 50 MHz.

  [Step S3] The propagation direction of the ultrasonic beam in step S1 is the reverse direction. That is, the selector switch 4 is controlled by a command from the CPU 12, the downstream ultrasonic transducer 3 is connected to the output of the transmission amplifier 5, and the upstream ultrasonic transducer 2 is connected to the reception variable gain amplifier 6 to detect the peak. The circuit 9 is initialized.

  [Step S4] The same processing as step S2 is performed, and the obtained data is stored in the RAMb of the CPU 12. The data obtained here is the second data y, and the maximum value is the second peak value Py.

  Note that the order of steps S1 and S2 and steps S3 and S4 may be reversed.

  The data obtained in this way is obtained as two types of first and second data x and y as shown in FIG.

  [Step S5] In steps S2 and S4, the first and second peaks Px and Py are obtained from the data x and y in FIG. 3, and the coefficients Kx and Ky that are the reciprocals thereof are obtained.

  Since these coefficients Kx and Ky are used in digital signal processing, they have a floating point or fixed point format. The number of bits of the first and second peak values Px and Py is determined by the number of bits of the AD converter 7. If this is assumed to be 8 bits, it is preferable that these 8 bits be a signed integer. In addition, the possible values of the first and second peak values Px and Py are between 0 and 127. If the data is of this order, the reciprocal number is not actually calculated, but is stored as a table in the ROM in the CPU 12, and the reciprocal number is extracted from this table, so that the processing time can be shortened. .

  Further, this reciprocal number is not the reciprocal number with respect to 1, but may be a value obtained by dividing an arbitrary numerical value. When division is performed by integer arithmetic, if the number of bits of the dividend is not larger than the number of bits of the divisor, the number of significant bits of the quotient will decrease, so the number of bits of the dividend needs to be larger than the divisor.

  For example, if the divisor (peak value) is 8 bits with a sign and the result of division is represented by a signed 16-bit integer, the maximum value 32767 of the signed 16-bit integer is divided by the peak value. be able to. As a result of division, a value after the decimal point may occur. In that case, the value after the decimal point may be rounded down or rounded off.

  Note that the number of bits of digital data, reciprocal data, and output data input to the multiplier 14 need not be the same.

[Step S6] Equation (3) is a calculation method for calculation in this embodiment, and is executed in the calculation unit 10.
Sxy [m] = Σ | Kx * x [n] −Ky * y [n + m] |
(M = 0, 1, 2,..., N-1) (3)

  The coefficients Kx and Ky obtained in step S5 are set as the coefficients of the multiplier 14. The multiplier 14 is composed of two multipliers A and B. Each multiplier A and B has two inputs, one input is connected to the aforementioned coefficients Kx and Ky, and the other input is connected to the two RAMa and RAMb of the memory 11. The first and second data x and y can be simultaneously extracted from the RAMa and RAMb of the memory 11, respectively.

  [Step S7] The difference is calculated according to equation (3). First, n = 0 and m = 0. n and n + m are addresses of the RAMa and RAMb of the memory 11 storing the first and second data x and y. In this example, n and m take values of 0 to 511, and when the value of n + m exceeds 512, the remainder is removed by 512. The value of x [n] is extracted from the memory storing the first data x, and the value of y [n + m] is extracted from the memory storing the second data y, and calculation is performed in the arithmetic unit 10. The value of n is changed from 0 to 511, all the obtained values are integrated, and the value is stored in the memory 11 as one data. Next, m is increased by 1, and n is again set to 0 to 511, and the above-described integration operation is performed.

  In this way, by repeating the value of m from 0 to 511, a waveform diagram as shown in FIG. 2 is obtained. Actually, it is not necessary to obtain all the data evenings, and a range in which a time difference corresponding to the range in which the flow velocity changes can be obtained may be calculated.

  In Equation (3), the coefficients Kx and Ky are multiplied, but the same effect can be obtained by changing this to division by the first and second peak values Px and Py. These reciprocal values remain constant until a series of calculations is completed. However, a circuit that performs division tends to be larger in circuit scale and slower in operation than a circuit that performs multiplication, and there are few advantages of a method using division.

  [Step S8] The time difference between the signals is obtained from the deviation of the maximum value of the waveform in FIG. 2 from the origin. The CPU 12 obtains a signal arrival time difference from the data obtained in step S7 based on the deviation from the origin of the data corresponding to the lowest point, and obtains the fluid flow velocity in the pipe line 1 from this time difference.

  According to this embodiment, even when the amplitude of the data digitally converted by the AD converter 7 is not constant, the inverse of the first and second peak values Px and Py is multiplied (resulting in division). Thus, the peak values Px and Py can be made constant, and it becomes possible to perform processing on signals having these constant peak values Px and Py.

  In the calculation of the difference, if the shapes and amplitudes of the two waveforms match, the minimum value can be expected to be zero. Since the minimum value increases as the waveform shape or amplitude deviation between the two signals increases, the minimum value of the difference can be used to determine the quality of the signal.

  When measuring the flow rate of the liquid, if bubbles or fine particles are mixed in the fluid, the progress of the ultrasonic pulse is hindered, the signal waveform is disturbed, the amplitude is changed, and the signal quality is deteriorated. In particular, if the waveform is greatly disturbed, accurate measurement of the flow velocity becomes difficult and causes a malfunction. Therefore, it is important to judge the signal quality.

  However, in the normal AGC control, the amplitudes of the two waveforms cannot be accurately matched. Therefore, when the minimum value is increased, it is determined whether the cause is due to the waveform difference or the AGC control. I can't.

  When this embodiment is not applied, that is, when the multiplier 14 is not used, the minimum value of the difference increases as the amplitude shift increases, and the measurement data at that time is regarded as an error and discarded. However, according to the present embodiment, the peak values Px and Py of the signal to be processed are always constant, so that data having an amplitude deviation in the AD-converted digital data can also be used. It becomes possible to do.

It is a block circuit block diagram. It is the wave form diagram which performed the difference calculation based on the signal of FIG. It is a graph figure of the received waveform of an ultrasonic signal. FIG. 4 is a waveform diagram in which cross-correlation is calculated based on the signal of FIG. 3.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Pipe line 2, 3 Ultrasonic transducer 4 Changeover switch 5 Transmitting amplifier 6 Receiving variable gain amplifier 7 A / D converter 8 Data bus 9 Peak detection circuit 10 Arithmetic unit 11 Memory 12 CPU
13 I / O controller 14 Multiplier 15 Clock generation circuit

Claims (2)

The ultrasonic beam is transmitted and received alternately in the first direction from the downstream side to the upstream side of the pipe and in the second direction from the upstream side to the downstream side, and the obtained signal by the ultrasonic beam is digitally transmitted. In the signal processing device of the ultrasonic flowmeter that stores the data in the memory according to the data in the first and second directions as the data, reads the data from the memory, and performs signal processing, the data is obtained for each of the two types of data in the memory. A first means for obtaining a coefficient that is the reciprocal number based on the maximum value of each of the two types of digital data as a numerical value that represents the characteristics of the obtained data, and for each of the two types of data read from the memory A second means for measuring a flow rate by obtaining a difference in propagation time of the ultrasonic beam by two types of data by multiplying the coefficient and obtaining a difference; Processing apparatus.   The signal processing apparatus for an ultrasonic flowmeter according to claim 1, wherein the coefficient is extracted from a storage table.

Images