JP2002257874A - Measuring method for a specified cycle - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a measuring method for a specified cycle, using a simple circuit structure that can suppress power consumption and accurately measure a signal cycle of a specified frequency for a measuring method for the specified cycle needed, for example, in a method measuring the average value of a pulse which contains higher harmonics. SOLUTION: The autocorrelation function of a corresponding input measuring signal is determined from sampling the data that the input measuring signal is sampled, and the signal cycle of the specified frequency is determined in a predicted range, based on the autocorrelation function. After triggering to the input-measuring signal under prescribed trigger conditions, the timing for a next trigger is limited in the predictive range and a time between the trigger and the next trigger is then pound as the period.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for measuring a specific period required in a method for measuring an average value of pulsations including, for example, harmonics.

[0002]

2. Description of the Related Art A pulsation including a harmonic appears in, for example, a flow rate and a pressure of a fluid flowing in a pipe, or a current and an applied voltage supplied from a power supply. For example, a fluid such as a gas or a liquid flowing in a pipe includes pulsation due to vibration transmitted from the outside to the pipe system. Examples of the vibration transmitted from the external vibration source to the piping path include, for example, vibrations (5 to 20) generated by a gas engine heat pump (hereinafter, referred to as GHP), a gas engine of a gas engine generator, a compressor, or the like.
0 Hz). The GHP has a gas engine that burns gas, reciprocates a piston in a combustion chamber, and converts it into rotational movement.

[0003] The vibration generated from such a GHP has a fundamental frequency based on the reciprocating motion of the piston and the rotational motion of the rotating system and its harmonic components. When the vibration of the fundamental frequency and its harmonic components is transmitted to the piping system, the fluid flowing in the pipe becomes a pulsating flow having a vibration waveform obtained by synthesizing the respective frequency components.

As a method for measuring the average value of the flow rate of a fluid flowing through a pipe with a pulsating flow having a vibration waveform in which a plurality of harmonics are superimposed, a patent application filed by the present applicant (Japanese Patent Application No. 200
No. 0-387905). According to the proposed method, it is assumed that the rotation speed of the GHP gas engine during the data sampling period is constant (that is, the fundamental frequency of the vibration component transmitted during the sampling period is constant). By setting the sampling parameters so that the sum of the vibration amplitude components becomes zero, an average flow rate can be obtained only by arithmetic averaging without performing complicated calculations.

[0005] This method is based on the premise that the fundamental frequency of the vibration component transmitted during the sampling period is known. However, since the rotation speed of the GHP fluctuates according to the use mode, in order to implement the above average flow rate measuring method, the actual fundamental frequency (basic period) of the pulsation on which the harmonic component is superimposed is accurately determined in advance. Although it is necessary to measure it, the above application does not disclose any method for measuring the fundamental period.

As a conventional method for measuring a specific period (specific frequency) such as a basic period, a measurement signal is passed through a low-pass filter (low-pass filter) for cutting off harmonic components of the specific frequency to remove the harmonic components. A method for measuring a fundamental frequency is known. For example, if the input circuit is AC
(AC) coupling (coupling) and D of input measurement signal
The C (direct current) component is removed, and then passed through a low-pass filter to remove the harmonic components of the input measurement signal. Next, a trigger is applied under a predetermined trigger condition, and the period from the time when the trigger is applied to the time when the next trigger is applied is measured, thereby measuring the period of the waveform of the fundamental frequency.
The trigger condition is, for example, a trigger level = 0 V and a trigger slope = “+” (0 V passes from a negative (−) level to a positive (+) level).

[0007]

However, there is only 2 octaves (oct) away from the fundamental frequency.
If the amplitude level of the second harmonic or higher harmonics is large, the roll-off characteristic becomes 6 dB / oct or 12 dB /
With a low-pass filter of about oct, these harmonic components cannot be completely removed, resulting in a distorted waveform in which the harmonics are superimposed on the time waveform of the fundamental frequency. If the amplitude level of the superimposed harmonic component is large, the distortion of the waveform becomes remarkable, and an unnecessary zero cross point where the signal passes the cell level within a specific period to be obtained is generated. Therefore, if a trigger is applied to the measurement signal that has passed through the low-pass filter under the above-described trigger condition, the next trigger is applied at an unnecessary zero-crossing point, and an accurate specific cycle may not be measured.

To improve this, it is conceivable to lower the cut-off frequency of the low-pass filter or to sharpen the roll-off characteristic to improve the frequency cutoff characteristic of the filter. However, this reduces the cost required for measurement. It is unrealistic to raise it.

As another conventional method for measuring a specific period, a method is known in which a sliding search method is applied to obtain an autocorrelation function of a measurement signal, and a specific period is obtained based on the function.

In this method, an analog measurement signal is sampled at a sampling frequency that is at least twice the fundamental frequency to be measured, and data is collected at a sampling time that is at least twice the period of the fundamental frequency signal to be measured. I do. Then, a sliding search method is applied to the collected sampling data to obtain an autocorrelation function. The time point at which the correlation coefficient becomes maximum in the obtained autocorrelation function is determined, and the time from the reference time point to the time point is defined as the period of the specific frequency signal.

However, in order to accurately obtain a specific period by the method of measuring a specific period using the autocorrelation function, it is necessary to increase the time resolution of the measurement as much as possible. For this reason, a sampling frequency that is twice or more the specific frequency to be measured as described above is not enough, and it is required to perform sampling at a sufficiently higher sampling frequency (for example, ten times or more). For example, in order to determine the period of the fundamental frequency signal in the range of 5 Hz to 50 Hz, a sampling frequency fs = about 500 Hz is required.
For this reason, a high-speed A / D (analog / digital) converter is required, which causes a problem that a circuit becomes complicated and power consumption increases.

Furthermore, in order to apply the sliding search method, data must be collected for a sampling time at least twice as long as the period of the fundamental frequency signal to be measured. Therefore, when the sampling frequency fs is increased, a large data buffer is required. Therefore, there is a problem that the circuit becomes complicated and the power consumption increases. Therefore, if an accurate measurement is to be performed, the measurement cost is increased by a specific period measurement method using an autocorrelation function.

An object of the present invention is to provide a method for measuring a specific period, which can suppress the power consumption with a simple circuit configuration and accurately measure the period of a signal of a specific frequency.

[0014]

An object of the present invention is to obtain an autocorrelation function of an input measurement signal from sampling data obtained by sampling the input measurement signal, and to determine a period of a signal of a specific frequency based on the autocorrelation function. Obtained in the prediction range, after triggering the input measurement signal under a predetermined trigger condition, limit the time when the next trigger is applied within the prediction range, and set the time between the trigger and the next trigger as the cycle. This is achieved by a method of measuring a specific period, which is characterized in that it is determined.

In the method for measuring a specific period of the present invention, sampling is performed after removing a DC component of the input measurement signal.

In the above-described method of measuring a specific period, the sampling is performed after removing a harmonic component contained in the input measurement signal to some extent.

In the method for measuring a specific period according to the present invention, the input measurement signal may be at least twice as large as the specific frequency.
Sampling at a sampling frequency of twice or less, and
The method is characterized in that data is collected at a sampling time that is at least twice the period of the signal of the specific frequency.

In the above-described method of measuring a specific period according to the present invention, the autocorrelation function is obtained by applying a sliding search method to the collected sampling data.

In the method for measuring a specific period according to the present invention, the time t at which the correlation coefficient becomes the maximum from the autocorrelation function.
1 is obtained, and the prediction range t of the cycle of the signal of the specific frequency is set to t1−s / 2 ≦ t ≦ t1 + s / 2 (where s is the time resolution of sampling and the reciprocal of the sampling frequency). I do.

[0020]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A method for measuring a specific period according to an embodiment of the present invention will be described with reference to FIGS. First, a schematic configuration of a measuring apparatus for realizing the measuring method of a specific cycle according to the present embodiment will be described with reference to FIG. The measuring apparatus 1 has an input stage (not shown) to which an analog measurement signal is input, and a low-pass filter 10 arranged next to the input stage. The input stage may be AC coupled to remove the DC component of the measurement signal or D
It can be switched to one of the C couplings. The low-pass filter 10 is 5H in this example.
The low-pass filter is provided for the purpose of removing some harmonic components of the specific frequency f in the range of z to 50 Hz to a certain extent, and a low-pass filter having filter characteristics comparable to those of the related art is sufficient.

A trigger circuit 20 and a trigger condition setting circuit 50 are arranged next to the low-pass filter 10.
The trigger circuit 20 can apply a trigger to an input measurement signal under predetermined trigger conditions based on control from the trigger condition setting circuit 50.

The trigger condition setting circuit 50 has a trigger time limit circuit (not shown) for limiting the trigger time in addition to the setting of the trigger conditions such as the trigger level and the trigger slope. The trigger timing limiting circuit determines a timing at which the next trigger is applied after the first trigger is applied by the trigger circuit 20 based on a clock signal supplied from a clock device (not shown) that synchronizes the circuits of the measuring device 1. Has a function to restrict. For example, the first trigger is time t
Assuming that 0 = 0, the next trigger is A ≦ t1 ≦ B
It does not take until it enters the range. As the values A and B, for example, A = −s / 2 and B = + s / 2 are set using the value of the sampling time resolution s, which is one of the sampling parameters in the A / D converter 30. Further, the trigger condition setting circuit 50 measures the time from the first trigger to the next trigger as the period T of the signal of the specific frequency f based on the clock signal, and uses the measured period T as a predetermined processing system. (Not shown).

An A / D converter 3 is provided next to the trigger circuit 20.
0 is arranged, and the analog measurement signal after the start of the trigger is sampled and converted into digital data. The A / D converter 30 has an input dynamic range of, for example, several bits, and outputs a measurement signal of, for example, 5 Hz.
Sampling frequency more than twice the specific frequency f in the range of 50 Hz (in this example, fs = 100 Hz)
And the maximum period Tmax = 1 / fmin = 1/5 = 0.2 s of the signal of the specific frequency f that can be taken.
sampling time Ts (Ts in this example)
= 0.4 sec) to collect data. The sampling time resolution s is s = 1 / fs
= 10 msec. Number of sampling points collected N
Is N = Ts / s = 40. As described above, the A / D converter 30 according to the present embodiment can use an A / D converter with a relatively low sampling rate and low cost and low power consumption. The A / D converter 30 can output at least the value of the sampling time resolution s among the above-described sampling parameters to the trigger condition setting circuit 50.

At the next stage of the A / D converter 30, a correlator 40 to which the sampling data obtained by the A / D converter 30 is inputted is provided. The correlator 40 calculates an autocorrelation function by applying a sliding search method to the sampled data. The correlator 40 obtains the time point at which the correlation coefficient becomes maximum from the obtained autocorrelation function, and outputs the time t from the reference time point to the time point to the trigger condition setting circuit 50.

Next, a method of measuring a specific period according to the present embodiment will be specifically described with reference to FIGS. 2 and 3 in addition to FIG. (1) As a premise, the range of the specific frequency (fundamental frequency to be measured) f included in the input measurement signal is 5 Hz to 50 H
z.

(2) First, an input circuit (not shown) of an input stage is set to AC coupling to input a measurement signal, and a DC component of the input measurement signal is cut.

(3) Next, the input measurement signal is passed through the low-pass filter 10 to remove some of the harmonic components contained in the signal.

(4) Sampling frequency fs = 100H twice the fundamental frequency f in the range of 5 Hz to 50 Hz for the input measurement signal from which the harmonic components have been removed to some extent.
sampling at z and the maximum period Tmax = 1 / fmin = 1/5 = 0.2 of the signal of the fundamental frequency f
Data is collected at a sampling time Ts = 0.4 sec twice as long as the second. The sampling time resolution s is 10
msec. The number of sampling points N collected is 40
Individual.

(5) An autocorrelation function is obtained by applying a sliding search method to the collected sampling data. FIG. 2 shows an example of the obtained autocorrelation function. The horizontal axis represents the time t from the reference time, and the vertical axis represents the correlation coefficient. In FIG. 2, a waveform having a maximum correlation coefficient at t = 16 s (s is the time resolution) is obtained.

(6) FIG. 3 is an enlarged view of the vicinity of t = 16 s where the correlation coefficient becomes maximum in FIG. The solid line shown in FIG. 3 indicates a continuous correlation function curve, and indicates that the actual maximum value of the correlation coefficient is at tx = 15.6 s. As shown in FIG. 3, in the discretized autocorrelation function, correlation coefficient data is arranged for each sampling time resolution s, and therefore the actual maximum value of the correlation coefficient is t = 16s ± s / 2.
Will be within the range. Therefore, the period T of the signal of the specific frequency f to be obtained is 16 s−s / 2 ≦ t ≦ 16
s + s / 2.

(7) Therefore, as trigger conditions, in addition to the trigger level = 0 V and the trigger slope = “+”, after the first trigger is applied, the next trigger is 16 s−s / 2.
Do not start until it falls within the range of ≦ t ≦ 16s + s / 2.

(8) When a trigger is applied to the input measurement signal using this trigger condition, the input measurement signal crosses the level zero point many times within the actual basic period T with the slope = “+”. Even if it is turned off, the next trigger will not be activated from the time when the first trigger is activated until it enters the range of 16 s−s / 2 ≦ t ≦ 16 s + s / 2. And 16s
-S / 2 ≦ t ≦ 16s + s / 2, the time t =
Since the next trigger is applied at 15.6 s, the period T of the waveform of the specific frequency f can be accurately measured. In addition, 1
If there are a plurality of zero cross points in the range of 6s−s / 2 ≦ t ≦ 16s + s / 2, the next trigger is activated at the first zero cross point in the range, but −s / 2 ≦ t ≦ + s /
The range of 2 is s = 10 msec, which is sufficiently within the measurement accuracy required for the measurement of the specific period T and can be ignored.

As described above, according to the present embodiment, the amplitude level of the second or higher harmonic component superimposed on the time waveform of the fundamental frequency is large and cannot be completely removed by the low-pass filter 10. Even if a number of unnecessary zero-cross points appear in the specific period to be obtained, the specific period can be accurately measured.

Therefore, it is not necessary to lower the cut-off frequency of the low-pass filter or to sharpen the roll-off characteristic to improve the frequency cut-off characteristic of the filter, so that the cost required for the specific period measurement can be suppressed. .

Further, according to the present embodiment, the high sampling time resolution required in the measuring method of the specific period for obtaining the autocorrelation function using the conventional sliding search method is unnecessary. For example, in order to obtain the period of the signal of the fundamental frequency f included in the range of 5 Hz to 50 Hz, the sampling frequency fs in the above-described conventional measuring method of the specific period using the autocorrelation function needs to be about 500 Hz. For this reason, a high-speed A / D converter is required, which causes a problem that a circuit becomes complicated and power consumption increases. However, the A / D converter 3 according to the present embodiment has a problem.
0 means that the sampling frequency fs is at most fs = 100 Hz
Is enough.

When the specific frequency f is in the range of 5 Hz to 50 Hz, a specific comparison will be made. In this embodiment, the sampling frequency fs = 100 Hz (time resolution s = 10 msec) and the sampling time ts = 0.4.
Although the number of sampling points is 40, the number of sampling points is 40, but in the above-described conventional method of measuring a specific cycle using the autocorrelation function, the sampling frequency fs = 500 Hz (time resolution s = 2 msec) and the sampling time ts = 0.4 sec.
c, 200 sampling points are required. Thus, according to the present embodiment, the clock frequency of the operation clock used in the A / D converter and other circuits can be reduced to about 1/5 of the conventional one.
About 1/5 of the power consumption of the related art can be realized.

Further, in order to apply the sliding search method in the conventional method, it is necessary to prepare a correlator having a large data buffer, which has the disadvantage that the circuit becomes complicated and the power consumption increases. But
In the present embodiment, such a problem does not occur and the measurement cost can be suppressed.

On the other hand, if the number of sampling points can be increased, the sampling frequency fs is set to 200 to 5
00Hz or less (4 to 10 times or less the desired fundamental frequency)
Then, the time resolution can be further improved and highly accurate measurement can be performed.

As described above, according to the present embodiment, it is possible to realize a method for measuring a specific cycle that can accurately measure the cycle of a signal of a specific frequency while suppressing power consumption with a simple circuit configuration.

The present invention is not limited to the above embodiment, but can be variously modified. For example, in the above embodiment, as the trigger conditions, the trigger level = 0 V, the trigger slope =
Although "+" is employed, the present invention is not limited to this. For example, a trigger level other than 0 V can be set, or the trigger slope may be set to "-" (a negative gradient passes the trigger level).

In the above embodiment, the case where the specific frequency is included in the range of 5 to 50 Hz has been described as an example. However, the present invention is not limited to this, and can be applied to any other frequency (band). It is.

[0042]

As described above, according to the present invention, power consumption can be suppressed with a simple circuit configuration, and the period of a signal of a specific frequency can be accurately measured.

[Brief description of the drawings]

FIG. 1 is a diagram showing a schematic configuration of a measuring apparatus for realizing a measuring method of a specific cycle according to an embodiment of the present invention.

FIG. 2 is a diagram illustrating an example of an autocorrelation function obtained by applying a sliding search method in a specific period measurement method according to an embodiment of the present invention.

FIG. 3 is t = 16s at which the correlation coefficient becomes maximum in FIG.
It is the figure which expanded and displayed the vicinity.

[Explanation of symbols]

 Reference Signs List 10 low-pass filter 20 trigger circuit 30 A / D converter 40 correlator 50 trigger condition setting circuit

 ────────────────────────────────────────────────── ─── Continuing on the front page (72) Inventor Kenchi Kobayashi F-term (reference) 2F035 GA02 2G029 AA01 AC08 AG04 AH08 3G084 AA05 DA04 DA13 EA01 EA05 EA06 EC04 FA07 FA11

Claims (6)

[The claims] An autocorrelation function of an input measurement signal is obtained from sampling data obtained by sampling the input measurement signal, and a period of a signal of a specific frequency is obtained in a predetermined prediction range based on the autocorrelation function. After applying a trigger under a predetermined trigger condition, the time when the next trigger is applied is limited within the prediction range, and the time between the trigger and the next trigger is obtained as the cycle. Measuring method. 2. The method for measuring a specific period according to claim 1, wherein the sampling is performed after removing a DC component of the input measurement signal. 3. The method for measuring a specific period according to claim 1, wherein the sampling is performed after removing a harmonic component included in the input measurement signal to some extent. 4. The method of measuring a specific period according to claim 1, wherein the input measurement signal is sampled at a sampling frequency of 2 to 10 times the specific frequency, and A method for measuring a specific period, wherein data is collected at a sampling time that is at least twice the period of a signal of a specific frequency. 5. The method for measuring a specific period according to claim 1, wherein the autocorrelation function is obtained by applying a sliding search method to the collected sampling data. Measuring method. 6. The method for measuring a specific period according to claim 5, wherein a time t1 at which a correlation coefficient becomes maximum is obtained from the autocorrelation function, and the prediction range t of the period of the signal of the specific frequency is obtained.
Is given by t1−s / 2 ≦ t ≦ t1 + s / 2 (where s is the sampling time resolution and the reciprocal of the sampling frequency)
A method for measuring a specific period, characterized in that: