JP2019109194A - Flow rate measuring device - Google Patents

Abstract

To provide a flow rate measuring device for measuring a flow rate by reducing an influence of an installation condition of a sensor and a piping state in passive type flow rate measurement.SOLUTION: A flow rate measuring device includes: a signal level calculation part for calculating a relationship between a frequency and a signal level on the basis of sensor data for indicating vibration of fluid from a sensor provided in piping; a band signal level calculation part for calculating a band signal level for a plurality of frequency bands on the basis of a signal level; a flow rate calculation part for calculating a flow rate for each frequency band for a plurality of frequency bands on the basis of a band signal level, and a model for expressing a relationship between a flow rate calculated for each frequency band and the band signal level; and an estimated result determination part for outputting an estimation result for a flow rate on the basis of a plurality of calculated flow rates.SELECTED DRAWING: Figure 1

Description

  The present invention relates to technology for measuring the flow rate of piping, and more particularly to an apparatus for measuring the flow rate of fluid from the outside of piping.

  Generally, in order to measure the fluid flow rate in the piping, a piping component to which a flow meter is added is inserted into a part of the piping that becomes a path, and measurement is performed. In this method, since the fluid and the sensor are in direct contact, reliable calibration is possible. However, in the case of new piping, it is necessary to cut the existing piping network and insert new parts, and it is not possible to send fluid during construction. This method is not always the best solution depending on the factory equipment.

  Therefore, technology has been developed for detecting the flow rate from outside the pipe without processing the pipe. One is the so-called active type, and Patent Document 1 includes an oscillation unit installed in contact with the pipe and emitting longitudinal ultrasonic waves in the thickness direction of the pipe, and the longitudinal wave transmitted in the pipe. A receiving unit that receives ultrasonic waves, and a flow rate calculating unit that calculates the flow rate of gas based on the reception result by the receiving unit. The flow rate calculating unit calculates the ultrasonic wave reception time and the ultrasonic wave amplitude in the receiving unit. It is a flow rate measuring device which computes the flow rate of gas by performing spectrum analysis etc. for every predetermined unit time about the function which shows a relation.

  On the other hand, Non-Patent Document 1 proposes a so-called passive type flow rate measurement technology for directly measuring vibrations generated by the interaction between piping and fluid. This technique utilizes the correlation between the flow energy and the magnitude of pipe vibration energy generated by fluid movement.

JP, 2017-75834, A

Nakata et al., "Method of measuring liquid flow rate using AE sensor", Japan Fluid Power System Society, May 2013

  In the conventional active type as shown in Patent Document 1, since it is easily affected by the difference in acoustic impedance between the pipe and the fluid, it is necessary to select the type of fluid to be measured. In particular, in order to measure accurately with a gas such as air whose impedance is different from that of the piping, it is necessary to precisely create a jig for fitting the sensor installation conditions and the transducers to the piping, making the device expensive. Become. Or, I had no choice but to exclude from the measurement.

  On the other hand, in the conventional passive type as shown in Non-Patent Document 1, the information obtained from the flow rate correlation of vibration is the relative amount of flow rate. The correlation between the flow rate obtained by the minimum flowmeter (for example, the outlet of the compressor) installed in the piping network and the individual vibrations obtained by a large number of sensors arranged in the branch piping is measured in advance. It is a use method of estimating the flow rate of each piping.

  However, it is expected that the correlation coefficient easily fluctuates depending on the installation conditions and the flow rate. In particular, depending on the state of the piping structure, resonance occurs at a specific flow rate or frequency, and the previously determined correlation can not be used. In this case, the problem is that if the magnitude of the vibration is obtained and if it is different from the normal time, the correlation is normal and the flow is simply changed, or the correlation is changed and the flow is changed. It is indistinguishable whether it can not be caught correctly.

  An object of the present invention is to provide a flow rate measuring device which measures the flow rate by reducing the influence of the installation condition of the sensor and the piping state in the passive type flow rate measurement.

  A preferred example of the present invention is a flow rate measuring device for measuring a fluid flow rate in a pipe by vibration of the pipe, and a relationship between frequency and signal level based on sensor data indicating vibration of fluid from a sensor provided in the pipe. And a band signal level calculator for calculating band signal levels for a plurality of frequency bands based on the signal level, a band signal level for each band, and A flow rate calculating unit which calculates the flow rate for each of the frequency bands based on a flow rate and a model representing a relationship between the band signal level, and the plurality of flow rates which are calculated for the plurality of frequency bands; It is a flow measurement device which has the presumed result judgment part which outputs the presumed result about the above-mentioned flow.

  According to the present invention, in the passive type flow rate measurement, it is possible to obtain a flow rate measuring device which measures the flow rate by reducing the influence of the installation condition of the sensor and the piping state.

(A) is an example of the whole flow rate measuring device of fluid in Example 1, (B) is an example of a relation of a sensor of a flow rate measuring device and piping, (C) is an example of an installation situation of a sensor of a flow rate measuring device Figure showing. FIG. 6 is a view showing an example of a pipe vibration spectrum with respect to the flow rate in the first embodiment. (A) shows the correlation between the average oscillating voltage and the flow rate within a certain frequency band in Example 1, and (B) shows another correlating the average oscillating voltage and the flow rate within a certain frequency range Showing the correlation in time. (A) is a figure explaining the correlation of the frequency band of a sensor and flow volume in the condition without resonance in Example 1, (B) is the correlation of the frequency band of a sensor and flow volume in the condition with resonance Explaining the nature. (A) is an example of the whole flow rate measuring device of the fluid in Example 4, (B) is a figure which shows an example of the installation condition of the sensor of the flow rate measuring device of a fluid. (A) is a figure explaining the correlation of the frequency band of a sensor and flow volume in the conditions without resonance in Example 4, (B) is the correlation of the frequency band of a sensor and flow volume in the condition with resonance Explaining the nature. (A) is a figure explaining the correlation of the frequency band of a sensor and flow volume in the conditions without resonance in Example 7, (B) is the frequency band of the sensor in the condition with resonance in Example 7 The figure explaining the correlation of and flow volume. (A) is a figure which shows the flow at the time of preparation in Example 1, (B) shows the flow at the time of the calibration for every sensor. FIG. 6 is a diagram showing a flow at the time of flow rate calculation in the first embodiment. FIG. 16 is a diagram showing a flow at the time of flow rate calculation in the second embodiment. FIG. 18 is a diagram showing a flow of calibration in the third embodiment. FIG. 18 is a diagram showing a flow at the time of flow rate calculation in Example 3. (A) is a figure which shows the flow at the time of preparation in Example 4, (B) is a figure which shows the flow at the time of calibration in Example 4. FIG. FIG. 16 is a diagram showing a flow at the time of flow rate calculation in Example 4. FIG. 18 is a diagram showing a flow at the time of flow rate calculation in Example 5. FIG. 18 is a diagram showing a flow of calibration in the sixth embodiment. FIG. 18 is a diagram showing a flow at the time of flow rate calculation in Example 6. (A) is a figure which shows the flow at the time of preparation in Example 7, (B) is a figure which shows the flow at the time of calibration in Example 7. FIG. FIG. 18 is a diagram showing a flow at the time of flow rate calculation in Example 7;

  Below, an Example is described in detail using a drawing.

  FIG. 1 is an example of a block diagram of a flow rate measuring apparatus according to a first embodiment. FIG. 1A shows the entire system including a compressor and a piping network. A plurality of branch pipes and one sensor 104 per branch pipe are arranged in the flow rate sensor 102 arranged in the pipe connected to the compressor 101 and the pipe network 103. The sensors 104a, 104b,... 104N send their respective sensor data to the computer 120 via the network, and the computer 120 processes the sensor data based on the sensor data. Terminal devices 105a, 105b... 105N are connected to the ends of the pipes.

  The structure of the sensor in Example 1 is shown in FIG. 1 (B) and FIG. 1 (C). In this example, a vibration sensor is assumed as the sensor 104. The vibration of the piping generated by the turbulent flow of compressed air, which is a fluid, passing through the piping is converted to a voltage by the sensor. As the sensor 104, an acceleration sensor or an AE sensor may be used. The detailed installation situation is shown in FIG. 1 (C). The sensor 111 is fixed to the mounting substrate 112, and the mounting substrate 112 is further fixed to the pipe 110 by the clamp fixture 113. In order to fix the sensor 111, the mounting substrate 112, and the pipe 110, the clamp fixture 113 is used, and although not shown, it is configured to be tightened with a screw or the like.

  The computer 120 comprises an I / O interface 121 for receiving sensor data indicative of fluid vibrations, an arithmetic unit 122, a console 123 and several storages 124 to 126 for recording data. The arithmetic device 122 is, for example, a CPU. In order to execute each processing unit in the processing flow to be described later, the arithmetic unit 122 calls a program stored in a recording apparatus such as the storage 124 to 126 and executes various processes.

  In the passive measurement which measures the fluid flow rate in piping by the vibration of the piping, the average value (voltage root mean square) within a certain frequency band of the flow rate and the output voltage of each sensor monotonically increases with the flowmeter installed at the beginning Find in advance the correlation. After obtaining this relationship in advance, the actual amount is estimated by actual compressed air operation. Here, voltage is used, but if correlation can be found, it may be power.

FIG. 2 shows an example of a diagram (spectrum) of frequency versus voltage at 0 to 50 kHz in the sensor in a state where the flow rate of compressed air under a pressure of 0.4 MPa in the steel pipe 20A is changed. The vertical axis in FIG. 2 is the voltage level (relative value), and the horizontal axis is the frequency (Hz). The numerical value at the upper right of the figure is the flow rate (L / min), and indicates the spectrum with each flow rate as a parameter.
This spectrum changes with the flow rate of compressed air. Depending on the environment of the piping network, by selecting an appropriate frequency band, the average voltage in the frequency band exhibits a monotonically increasing (approximately direct proportion) relationship to the flow rate.

  As an example, FIG. 3A shows the relationship between the flow rate and the average voltage at a frequency bandwidth of 20 kHz to 30 kHz at time A. The vertical axis in FIG. 3A is the 10 kHz in-band voltage average value (au), and the horizontal axis is the flow rate (L / min). After obtaining this relationship in advance, the actual amount is estimated by actual compressed air operation.

  The vertical axis in FIG. 3B is the 10 kHz in-band voltage average value (au), and the horizontal axis is the flow rate (L / min). Compared with FIG. 3 (A), despite the same arrangement condition, depending on the time (time B), resonance may occur with the structure of the pipe in the vicinity of the sensor, and a peak may occur in a specific flow rate and frequency band, The exact flow may not be known. For example, it can not be discerned whether the flow rate is 125 L / min or 175 L / min at an average voltage level of 0.4 obtained.

  The cause is, for example, a change in the open / close state of the valve around the pipe or a change in the operation state of the apparatus along with the change in the flow rate. In addition, there is a possibility that the inclination may change due to peeling of the sensor from the pipe. Only by evaluating the voltage in one frequency band, it can not be distinguished whether the flow rate has actually changed or the correlation between the flow rate and the output voltage of the sensor has changed when the estimated flow rate value has changed. Further, in FIG. 3B, there is a protruding portion at a flow rate of 125 L / min, and a recessed portion is observed at a flow rate exceeding 200 L / min.

  Therefore, in passive measurement that measures the fluid flow rate in a pipe by the vibration of the pipe, the relationship between the flow rate and the voltage average value in a certain frequency band monotonously increases is determined in advance, and then an average value is obtained for one sensor. A plurality of frequency bands (hereinafter, frequency bands are abbreviated as bands) are prepared. For example, 15 kHz to 20 kHz, 25 to 30 kHz, 45 to 50 kHz, and the like.

  The diagrams on the left of FIG. 4A and FIG. 4B are diagrams showing the relationship between the signal level (average voltage) and the frequency, the vertical axis indicates the signal level, and the horizontal axis indicates the frequency. The diagrams on the right side of FIG. 4A and FIG. 4B are diagrams showing the relationship between the signal level (average voltage) and the flow rate, the vertical axis is the signal level, and the horizontal axis is the flow rate. In a normal correlation having no resonance peak as shown in FIG. 4A, the correlations (inclination of slope A, band B, and band C with respect to changes in flow rate, for example, flow rate 1, flow rate 2 and flow rate 3) Shows an average voltage change according to.

  However, when a resonance peak as shown in FIG. 4B occurs in, for example, the band B, the change in voltage of only the band B deviates from the predetermined value. In other words, in the band B with resonance although the same flow rate is estimated between the band A, the band C, and the band B, the average voltage contradicts with the other bands. After excluding the zone causing the contradiction, the flow rate can be estimated by obtaining an average voltage in the remaining zones.

  In this example, the specific determination of the resonance band and the non-resonance band is performed by determining the average voltage in each band, and determining the final flow rate based on the majority of the flow rate obtained by the correlation obtained in advance. The small number of bands may have changed in slope due to resonance, etc., and are excluded from the flow rate estimation. More specifically, when the change in voltage with respect to the previously determined flow rate is different, it is considered that the installation situation of the sensor and the pipe has changed as well as the resonance, and therefore, it can be excluded. This example is effective in the situation where the flow rate does not change at the time of evaluation.

  FIG. 8A shows a flow at the time of preparation of the first embodiment. The operation of the compressor 101 is started to generate compressed air (S1001). The flow meter installed in the pipe connected to the compressor 101 is operated to continuously acquire flow data (S1002). The end pipe 1 (for example, a pipe connected to the end device 105a of FIG. 1A) is opened (S1011). The sensor is operated to acquire voltage data of the sensor at several different flow rates (S1012). The data is recorded in the calibration data storage 124 as time series data. The rate to be acquired may be twice the frequency required for the fast Fourier transform (hereinafter referred to as FFT) analysis to be performed later, and the period may be adjusted to the necessary frequency resolution. Close the end pipe 1 (S1013).

  The series of operations in S1011 to S1013 are repeatedly performed as many as the number of end pipes (80). Finally, the flow meter and the compressor 101 are stopped (S1003 and S1004, respectively), and the process ends. In each step in FIG. 8A, the arithmetic device 122 may control the opening and closing of the valve of the compressor 101 and each pipe, control acquisition of flow rate data of the flow meter, recording, etc. May perform a part.

  FIG. 8B shows a calibration flow per sensor. FIG. 8B is a process performed by the arithmetic device 122. The signal level calculation unit performs FFT on the sensor voltage (recorded time series data) acquired in S1012 and the like, and records the result as frequency series data in the calibration data storage 124 (S1051). In general, the preprocessing (band limitation, window function processing) required for the FFT is not limited here.

  The band signal level calculation unit obtains the in-band average voltage using the recorded frequency sequence data (S1052). In order to obtain the average voltage in the band, data converted by FFT is divided by several kHz (for example, 5 kHz), and individual voltage data in the band are once squared and integrated. By taking that を, the average voltage in that band can be obtained. The correlation calculation unit creates a correlation table of flow rate and average voltage for each band, and records this in the correlation data storage 125 (S1053).

  The model calculation unit generates a correlation model of the flow rate and the average voltage by the least square method from the obtained correlation table (S1054). Steps S1052 to S1054 are repeatedly executed to cover the measurement band of the sensor 1 (81). The results are recorded. If the measurement frequency band is 0 to 50 KHz, 10 bands can be taken at 5 KHz band.

  The band candidate selection unit calculates the divergence rate in a plurality of bands, and selects a plurality of band candidates suitable for the direct proportional model from the divergence rate with the model. (S1055). In the selection of band candidates, data may be output and bands may be selected visually.

The correlation model described above can be expressed by the following (1) and (2) which are simple direct proportional relationships.
Flow rate = α × average voltage (1),
Alternatively, if the pipe diameter is large, the flow rate = α × average voltage + β (2) in consideration of the dead flow rate
When using power as an index, a quadratic function is used.

The divergence rate is, for example,
√ (Σ (flow rate during calibration-calculated flow rate even with model) ^ 2 / number of samples) (3)
Defined in, and reflected in the error of estimation.

  Finally, a plurality of bands used for flow rate calculation for each sensor and α and β of the above equation applied to the model are registered in the correlation data storage 125 as data related to the correlation model (S1056). And it ends.

  In the following description, the process proceeds on the basis of model (1), but the same applies to model 2. In the actual flow rate calculation, the flow rate is determined by driving only the sensor and registering a voltage signal obtained from the sensor or calculating an average voltage in a band and calculating it from the mathematical expression model (1). The results are registered in the estimated data storage 126.

  FIG. 9 shows a flow at the time of flow rate calculation of the first embodiment. The arithmetic device 122 executes the following processes of the flow of FIG. 9. The data acquisition unit acquires the sensor voltage received by the I / O interface 121 from the sensor 1 via the network (time-series data) (S1101).

  The signal level calculation unit subjects the acquired sensor voltage 1 to FFT to calculate the relationship between the sensor voltage and the frequency, which is an example of the signal level as shown in FIG. 2 (S1102). In general, the preprocessing (band limitation, window function processing) required for the FFT is not limited here.

  The band signal level calculation unit obtains an average voltage in the band 1 which is a candidate for flow rate calculation in advance (S1111). The method of determining the average voltage is the same as that of S1052. The flow rate calculating unit calculates the flow rate based on the correlation model of the band 1 recorded in S1056 and the calculated average voltage in the band 1 (S1112).

  This series of operations (S1111-1112) is performed by the number of selected bands (82). The estimation result judging unit compares the flow rates calculated in the respective bands (S1113). If they match (Y in S1113), the value is output as an estimation result (S1103), and the process ends.

  If they do not match (N in S1113), the estimation result judgment unit takes majority with different flow rate values (S1104). Then, the estimation result determination unit determines whether there is a flow rate value that occupies a large number of values of the estimated flow rate (S1105). If there is a flow rate value that occupies a large number (YES in S1105), the estimation result determination unit outputs it as an estimation result (S1106) and ends.

  If the result is different and the majority can not be decided (NO in S1105), it is assumed that the flow rate sensing by this sensor 1 becomes impossible, and the estimation result judgment unit causes the display device or the like to display that (S1107) . And it ends.

  According to the first embodiment, it is possible to estimate the flow rate more accurately. Alternatively, it is possible to obtain a trigger for finding the correlation between the flow rate and the sensor output voltage again.

  In the second embodiment, another implementation method of the method for determining the resonance band and the non-resonance band described in the first embodiment will be described. In this example, the change of the average value is obtained for each band with respect to the change of the flow rate, and even if one is different from the inclination examined in advance, it is excluded from the calculation candidates as the possibility near the resonance peak.

  For example, in an environment where the flow rate is changing, the flow rate is calculated in each zone at time A1, and the flow rate is calculated in each zone at time A2. Changes in the calculated flow rate in each zone from time A1 to time A2 are stored. If the correlation between voltage and flow rate is maintained, the change in flow rate calculated in any zone shows the same value. However, when the same correlation as that at the time of calibration is not maintained in a specific band due to resonance or the like, the amount of change shows a value different from that in the other bands. Calculated flow rate in this zone is excluded.

  More specifically, when the change in voltage with respect to the previously determined flow rate is different, it is considered that the installation situation of the sensor and the pipe has changed as well as the resonance, and therefore, it can be excluded. This example is effective in the situation where the flow rate changes at the time of evaluation.

  The process flow at the time of flow rate calculation in Example 2 is shown in FIG. The processing of FIG. 10 is executed by the arithmetic unit 122. The data acquisition unit acquires the voltage (time-series data) of the sensor 1 at time 1 (S1201). The signal level calculation unit performs FFT on the voltage (time-series data) of the sensor 1 at time 1 (S1202).

  The band signal level calculation unit obtains an average voltage in the band 1 which is a candidate for calculating the flow rate in advance (S1203). The method of obtaining is the same as that of S1052. The flow rate calculating unit calculates the flow rate based on the calculated average voltage and the correlation model of band 1 (S1204). This series of operations (S1203 to S1210) is executed by the number of the plurality of selected bands (83).

  Next, the data acquisition unit acquires the voltage (time-series data) of the sensor 1 at time 2 (S1205). The signal level calculation unit subjects the voltage (time series data) of the sensor 1 at time 2 to FFT (S1206). The band signal level calculation unit obtains an average voltage in the band 1 selected in advance as a band for calculating the flow rate (S1207). This is the same band as time 1. The method of determining the average voltage is the same as that of S1052.

  The flow rate calculating unit calculates the flow rate based on the calculated average voltage and the correlation model of band 1 (S1208). This series of operations (S1205-S1208) is performed by the number of the plurality of selected bands (83). For each zone, the flow rate calculating unit obtains a difference which is a change in the flow rate calculated at time 1 and time 2 (S1210).

  Next, the estimation result judging unit compares the differential flow rates of the respective bands (S1211). S1210 and S1222 are also performed by the number of the plurality of selected bands (83). As a result of comparison, if the differential flow rates in all the bands are the same (Y in S1211), the flow rate at either time 1 or time 2 may be output as an estimation result, or time 1 and time 2 The estimated flow rate in both of

  In addition to the time, the estimated flow rate of the time may be output (S1212). When the estimation result is output, the process ends. If at least one difference flow rate in the calculated band does not match (N in S1211), the estimation result judgment unit takes majority with different flow rate values (S1204). Then, the estimation result determination unit determines whether there is a flow rate value that occupies a large number (S1214). If there is a flow rate value that occupies a large number (YES in S1214), the estimation result determination unit causes the flow rate value that occupies the large number to be output as an estimation result (S1206), and the process ends.

  If the results vary and the majority decision can not be made (NO in S1214), the estimation result judgment unit displays the fact that the flow rate sensing by this sensor has become impossible (S1206). And it ends. According to the second embodiment, when evaluating at the time of calibration or calculation, it becomes effective in the situation where the flow rate changes.

  In the third embodiment, another implementation method of the method for determining the resonance band and the non-resonance band described in the first embodiment will be described. In this example, the band determined in the calibration stage, that is, the frequency range in which the average voltage therein is in direct correlation with the flow rate, and the correlation between the average voltage and the flow rate in further divided bands Use

  At the calibration stage, the correlation between the average voltage and the flow rate in the divided bands is approximately directly proportional using the least squares method etc. (here, α is named “provisional slope”. This relationship is used only to confirm the resonance within the band, not to calculate the flow rate, and if the same correlation is maintained in each band as at the time of calibration, the frequency characteristics within the band (shape of the curve) Also, the correlation between the divided bands, "temporary slope", does not change either.

  Therefore, there is no contradiction in the "provisionally calculated flow rate" obtained. However, if the same correlation is not maintained due to resonance or the like in a specific band at the time of calculation from the time of calibration, the frequency characteristics (the shape of the curve) in the band are deformed, and thus the correlation between divided bands. The "provisional slope" also changes. For this reason, contradiction also arises between the temporarily calculated flow rates obtained between the divided bands. Therefore, the calculated flow rate in this zone is excluded.

  More specifically, when the change in voltage with respect to the previously determined flow rate is different, it is considered that the installation situation of the sensor and the pipe has changed as well as the resonance, and therefore, it can be excluded. The present embodiment is effective in the case where a large number of values as in the first and second embodiments can not be obtained because the deformation of the frequency characteristic (curve) in one band can be detected.

  The specific flow in this example is shown in FIG. 11 and FIG. FIG. 11 is a flow of calibration. Steps S1051 to S1056 are the same as in the first embodiment, that is, FIG. 8B. Further, as in FIG. 8B, the process is repeatedly performed so as to cover the measurement band (84). In this example, each band is further divided into, for example, a first half band and a second half band centering on the middle of the bands from the result of the FFT, and the average value of the voltages is determined (S1307 and S1310).

  Then, the correlation table creation unit creates a correlation table of flow rate and average voltage in each divided band (S1308 and S1311). Furthermore, a temporary slope (temporary coefficient) is obtained using the least squares method or the like, and the model calculation unit models and records the temporary slope as a parameter (S1309 and S1312).

  The flow at the time of calculation is shown in FIG. Steps S1101 to S1102 are the same as in the first embodiment, that is, FIG. In this example, each band is further divided into the first half and the second half, the band signal level calculation unit obtains the average value of the voltage in S1353 and S1355, respectively, and the flow rate calculation unit calculates the provisional flow rate in S1354 and S1356. calculate.

  The estimation result determination unit compares the calculated first half band and the two provisional flow rates in the second half band in S1357. If the comparison results do not match (N in S1357), band 1 is excluded from the calculated band (S1358). This series of operations (S1353 to S1357) is performed on all selected bands (85). If a match is found in S1357 (Y in S1357), the estimation result determination unit determines whether a matched band remains (S1359).

  If a coincident band remains (Yes at S1359), the flow rate is actually calculated in that band, and the result is output (S1361). If there is no remaining band (No at S1359), it is determined that the flow rate sensing by this sensor has become impossible, and the estimation result judgment unit displays that effect (S1360). And it ends.

  According to the third embodiment, it is possible to detect the deformation of the frequency characteristic (curve) in one band, and in the first and second embodiments, even if the calculated flow rate can not obtain a large number of values, a large number of values can be obtained. You can increase the chances of getting it.

  FIG. 5 is an example of a configuration diagram of a fluid flow rate measuring device according to a fourth embodiment. FIG. 5A shows the entire system including the compressor and the piping network. One flow rate sensor 102 is disposed in a pipe connected to the compressor 101. In the piping network 103, a plurality of branch pipes and a plurality of sensors 104 are installed in each of the branch pipes.

  Each sensor, such as sensors 104-1a, 104-1b, 104-2a, 104-2b,... 104-Na, 104-Nb, sends respective sensor data to the computer 120 via the network, and the computer 120 Processing is performed based on sensor data. End devices 105-1 and 105-N are connected to the ends of the pipes.

  The configuration of the sensor in the fourth embodiment is shown in FIG. 5 (B). In this example, a vibration sensor is assumed as the sensor, and the vibration of the pipe accompanying the movement of the fluid is picked up by the sensor. Two sensors 111A and 111B are symmetrically arranged for one pipe. Also, the sensors 111A and 111B are fixed to the mounting boards 112A and 112B, respectively.

  Although not shown, the mounting boards 112A and 112B and the sensors 111A and 111B are fixed to the pipe 110 by being clamped with a clamp. The computer 120 is the same as the configuration of FIG. Three or more sensors may be fixed to one pipe.

  In the passive measurement which measures the fluid flow rate in piping by the vibration of the piping, the average value (voltage root mean square) within a certain zone of the flow rate and the output voltage of each sensor monotonically increases with the flowmeter installed at the beginning Find the relationship in advance. One band is prepared in one sensor. The same is done with multiple sensors. After obtaining this relationship in advance, the actual flow rate of compressed air is calculated.

  In this example, by arranging a plurality of sensors in the same pipe, the sensor of the resonated condition is detected and excluded from the flow rate calculation. FIG. 13A shows a flow at the time of preparation of the fourth embodiment. S1001, S1002, S1011, S1013, S1003, and S1004 are the same as in FIG. Then, a plurality of sensors (such as sensors 1 and 2) are operated to acquire voltage data of the sensors at several different flow rates (S1412-1, S1412-2, ...).

  This operation is performed as many as the number of sensors per end pipe (86). The voltage data of the sensor is recorded in the calibration data storage 124 as time series data. The acquisition rate may be adjusted to twice the frequency required for the later FFT analysis, and the period may be adjusted to the required frequency resolution. A series of operations (S1411-S1013) are repeatedly performed as many as the number of end pipes (87). Finally, the flow meter and the compressor 101 are stopped (S1003 and 1004, respectively), and the process ends.

  In each step of FIG. 13A, as in FIG. 8A, the arithmetic unit 122 controls the opening and closing of the valve of the compressor 101 and each pipe, and controls acquisition of flow rate data of the flowmeter, recording, and the like. You may

  Next, the calibration flow in this example is shown in FIG. FIG. 13B is a process performed by the arithmetic device 122. The contents are substantially the same as the steps in FIG. 8B in the first embodiment. The difference is that there is a plurality of sensor data in one pipe, and one band is prepared by one sensor. In S1455, one band is selected for calculating the flow rate in each sensor.

  Further, the steps S1051 to S1056 are repeated (88) by the number of all sensors. In actual calculation, the flow rate is determined by driving only the sensor and calculating from the average voltage calculation obtained from the sensor and the mathematical expression model (1). The flow rate and the average voltage value for each sensor, that is, the correlation data are recorded in the correlation data storage 125, and the calculation data is recorded in the estimation data storage 126.

  The diagrams on the left side of FIGS. 6A and 6B are diagrams showing the relationship between the signal level (average voltage) and the frequency, the vertical axis indicates the signal level, and the horizontal axis indicates the frequency. The diagrams on the right side of FIGS. 6A and 6B are diagrams showing the relationship between the signal level (average voltage) and the flow rate, the vertical axis is the signal level, and the horizontal axis is the flow rate. The normal correlation without resonance peak as shown in FIG. 6A shows an average voltage change according to the correlation obtained in advance for the sensor A, the sensor B and the sensor C with respect to the change of the flow rate. The same inclination may be used as illustrated, or may be different depending on the arrangement.

  However, when a resonance peak as shown in FIG. 6B occurs in, for example, the sensor B, the change in voltage of only the sensor B deviates from the planned value. More specifically, in spite of the change in the same flow rate between the sensor A, the sensor C and the sensor B, a contradiction arises in the corresponding change in voltage. From this, it is detected that a phenomenon such as resonance or the like in which a predetermined correlation is broken occurs in any one or a plurality of sensors of the sensor A, the sensor B, and the sensor C. Therefore, the flow rate may be calculated after excluding it in the subsequent steps.

  In this example, the method of discriminating between the sensor having resonance and the sensor not having determined the final flow rate by determining the average voltage in each zone with each sensor and then determining the calculated flow rate obtained by the correlation obtained in advance. Do. The small number of bands may have changed in inclination due to resonance or the like, and are excluded from the flow rate calculation. More specifically, when the change in voltage with respect to the previously determined flow rate is different, it is considered that the installation situation of the sensor and the pipe has changed as well as the resonance, and therefore, it can be excluded. This example is effective in the situation where the flow rate does not change at the time of evaluation.

  The process flow at the time of calculation of the fourth embodiment is shown in FIG. The arithmetic device 122 executes each process of FIG. Steps S1101 to S1112 are similar to the processing of FIG. The difference is that the series of operations (S1101-S1112) are repeatedly executed by the number of sensors (89). Further, the estimation result judging unit compares the flow rates estimated by the respective sensors (S1505). If all the calculated flow rates match (Y in S1505), the value is output as an estimation result (S1103), and the process is ended.

  If they do not match (N in S1505), the estimation result judging unit takes a majority decision (S1104). If there is a flow rate value that occupies a large number (Yes in S1105), the flow rate value that occupies the large number is output as an estimation result (S1106), and the process ends. If the results vary and it is not possible to determine the majority (No in S1105), it is determined that the flow rate sensing of the flow rate of the pipe has become impossible, and the estimation result determination unit displays the effect (S1510) and ends.

  According to the fourth embodiment, by comparing the estimated flow rates of a plurality of sensors, it is possible to estimate the flow rate more accurately. Alternatively, it is possible to obtain a trigger for determining the correlation (calibration) of the flow rate and the output voltage of the sensor again.

  In the fifth embodiment, another implementation method of the method for determining the resonance band and the non-resonance band described in the fourth embodiment will be described. In this example, for a slight change in the flow rate, a change in the voltage average value is obtained for each sensor, and if any one is different from the previously examined inclination, it is excluded from the calculation as a possibility near the resonance peak. For example, in an environment where the flow rate is changing, the flow rate is calculated in each zone at time A1, and the flow rate is calculated by each sensor at time A2.

  Changes in the calculated flow rate in each zone from time A1 to time A2 are stored. If the correlation between voltage and flow rate is maintained, the change in flow rate calculated by any sensor shows the same value. However, when the same correlation as that at the time of calibration is not maintained due to resonance or the like in a particular sensor, the amount of change shows a value different from that of the other sensors. Calculated flow rate with this sensor is excluded. More specifically, when the change in voltage with respect to the previously determined flow rate is different, it is considered that the installation situation of the sensor and the pipe has changed as well as the resonance, and therefore, it can be excluded. This example is effective in the situation where the flow rate changes at the time of evaluation.

  FIG. 15 is a process flow at the time of calculation in the fifth embodiment. Each process of FIG. 15 is the same as that of FIG. However, this series of operations (S1201 to S1210) are repeatedly executed by the number of sensors (90).

  According to the fifth embodiment, in the case where estimated flow rates from a plurality of sensors are compared, it is effective in a situation where the flow rate changes when making an evaluation such as at the time of calibration or at the time of flow rate calculation.

  In the sixth embodiment, another implementation method of the method for determining the resonance band and the non-resonance band described in the second embodiment will be described. In this example, the sensor-specific band determined at the calibration stage, that is, the average voltage and the flow rate in the band further divided into the frequency range in which the average voltage therein is directly proportional to the flow rate Use correlations of At the calibration stage, the correlation between the average voltage and the flow rate in the divided bands is approximately directly proportional using the least squares method etc. (here, α is named “provisional slope”. .

  This relationship is used only to check the resonance in the band and not used to calculate the flow rate. When the same correlation as in the calibration is maintained in each band, the frequency characteristic (shape of the curve) in the band does not change either, so the correlation between the divided bands and the "provisional slope" also do not change. Therefore, there is no contradiction in the "provisionally calculated flow rate" obtained. However, if the same correlation is not maintained due to resonance or the like in a specific band at the time of calculation from the time of calibration, the frequency characteristics (the shape of the curve) in the band are deformed, and thus the correlation between divided bands. The "provisional slope" also changes.

For this reason, contradiction also arises between the temporarily calculated flow rates obtained between the divided bands. Therefore, the calculated flow rate in this zone is excluded.
More specifically, when the change in voltage with respect to the previously determined flow rate is different, it is considered that the installation situation of the sensor and the pipe has changed as well as the resonance, and therefore, it can be excluded. The present example can detect the deformation of the frequency characteristic (curve) in the band of one sensor, and is thus effective when a large number of values can not be obtained between the sensors as in the fourth and fifth embodiments.

  The specific flow in this example is shown in FIG. FIG. 16 is a flow at the time of calibration. Steps S1051 to S1056 are the same as in FIG. 8B. However, S1051 to S1056 are repeatedly performed by the number of sensors (91). Further, in the sixth embodiment, there is one band per sensor.

  In this example, each band is further divided into the first half and the second half, and steps S1307 to S1312 are executed. The steps are the same as in FIG.

  FIG. 17 shows a process flow at the time of calculation in the sixth embodiment. The main processing steps S1101 and S1102 are the same as in FIG. Moreover, S1353 to S1361 are the same as FIG. However, this series of operations (S1101 to S1357) are performed on all the sensors (92).

  According to the sixth embodiment, the calculated flow rates from the plurality of sensors are compared, and in the fourth and fifth embodiments, even if the multiple values of the calculated flow rates can not be obtained, the multiple values are obtained. The possibilities can be increased.

The configuration diagram of the fluid flow rate measuring device in the seventh embodiment is the same as that of FIG.
In this example, unlike the fourth embodiment, a plurality of sensors are arranged in the same pipe, and one sensor uses a plurality of bands to be used for calculation, thereby detecting a sensor in a resonated condition and calculating the flow rate. Exclude from

  FIG. 18A shows a flow of preparation of the seventh embodiment. This figure is similar to FIG. 13 (A). However, the operation of obtaining sensor voltages such as S1412-1 is repeatedly performed as many as the number of sensors disposed per one end pipe (93). Moreover, a series of operation (S1011-S1013) is repeatedly implemented by the number of piping (94).

  Next, the calibration flow in this example is shown in FIG. FIG. 18B is substantially the same as the flow of FIG. However, S1051 to S1056 are repeatedly executed by the number of sensors (95). Furthermore, it is to have the step (S1855) of selecting a plurality of band candidates for one sensor. In this embodiment, one pipe has a plurality of sensor data, and one sensor prepares a plurality of bands.

  In actual calculation, the flow rate is determined by driving only the sensor and calculating from the average voltage calculation obtained from the sensor and the mathematical expression model (1). The flow rate and the average voltage value for each sensor, that is, the correlation data are recorded in the correlation data storage 125, and the estimation data is recorded in the estimation data storage 126.

  However, in spite of the same arrangement condition, depending on the time, resonance may occur with the structure of the piping near the sensor, and a peak may appear in a specific flow rate and frequency vs. range, and the accurate flow rate may not be known.

  Therefore, in the passive measurement in which the fluid flow rate in the pipe is measured by the vibration of the pipe, the relationship between the flow rate and the voltage average value in a certain band is monotonously increased is previously obtained. Then, for a slight change in flow rate, the change (slope) of the average value is determined for each sensor. As a result of finding the change (inclination) of the average value, the same band is compared with each other, and even if one of them is different from the inclination examined in advance, it is excluded from the calculation as the possibility near the resonance peak.

The diagrams on the left of FIG. 7A and FIG. 7B are diagrams showing the relationship between the signal level (average voltage) and the frequency, the vertical axis indicates the signal level, and the horizontal axis indicates the frequency. The diagrams on the right side of FIGS. 7A and 7B show the relationship between the signal level (average voltage) and the flow rate, the vertical axis indicates the signal level, and the horizontal axis indicates the flow rate.
In the normal correlation having no resonance peak as shown in FIG. 7A, the sensor A, the sensor B, and the sensor C all have an average according to the correlation determined in advance in each of the bands A, B and C with respect to changes in the flow rate. Indicates a change in voltage.

  However, when a resonance peak as shown in FIG. 7B occurs, for example, in the band B of the sensor B, the change in voltage of only the band B of the sensor B deviates from the planned value. More specifically, in the sensor B, in spite of the change of the same flow rate between the band A and the band C and the band B, contradiction occurs in the corresponding voltage change width. Furthermore, even in the same zone B, in spite of the change in the same flow rate between the sensor A, the sensor C and the sensor B, a contradiction arises in the corresponding voltage change width.

  From this, it is detected that a phenomenon such as resonance that is broken in advance has occurred in any of the sensor A, the sensor B, and the sensor C, or a plurality of bands. Therefore, the flow rate may be calculated after eliminating it in the subsequent steps.

In this embodiment, the method of discriminating between the sensor having resonance and the sensor not determining the average voltage in each band, determines the final flow rate to be calculated based on the majority of the calculated flow rate obtained by the correlation obtained in advance. There is a possibility that the slope has changed due to resonance or the like, and it is excluded from the flow rate calculation target.
More specifically, when the change in voltage with respect to the previously determined flow rate is different, not only the resonance but also the installation situation of the sensor and the pipe is considered to have changed, so that it can be excluded. This example is effective in the situation where the flow rate does not change at the time of evaluation.

  The process flow at the time of flow rate calculation of Example 7 is shown in FIG. The process in FIG. 19 is similar to the process in FIG. However, this series of operations (S1111-S1112) are executed as many as the number of candidate bands (96). Further, this series of operations (S1101-S1112) are executed by the number of sensors installed in the same pipe (97).

  According to the seventh embodiment, since a plurality of sensors are used and the flow rate calculation results in a plurality of frequency bands are used for one sensor, the flow rate can be calculated more accurately. Alternatively, it is possible to make it easier to obtain a trigger for determining the correlation (calibration) of the flow rate and the output voltage of the sensor again.

101: compressor, 102: flow sensor, 104: sensor, 110: piping, 120: computer, 121: I / O interface, 122: arithmetic device

Claims (12)

A flow rate measuring device for measuring a fluid flow rate in a pipe by vibration of the pipe,
A signal level calculation unit that calculates the relationship between frequency and signal level based on sensor data indicating vibration of fluid from a sensor provided in the pipe;
A band signal level calculator configured to calculate band signal levels for a plurality of frequency bands based on the signal levels;
The flow rate for each of the frequency bands is calculated for a plurality of the frequency bands based on the band signal level, and a model representing the relationship between the flow rate calculated for each frequency band and the band signal level. A flow rate calculation unit,
A flow rate measuring apparatus comprising: an estimation result judging unit which judges to output an estimation result of the flow rate based on a plurality of calculated flow rates. The flow rate measurement apparatus according to claim 1, wherein the estimation result determination unit compares the flow rates of the plurality of frequency bands, and estimates the flow rate of any one of the plurality of flow rates based on the comparison result. A flow rate measuring device characterized by judging to output as a result. The flow rate measuring apparatus according to claim 2, wherein the estimation result judging unit outputs the flow rate as an estimation result when the compared plurality of flow rates are all the same, and the comparison results do not match In the case, the flow rate measuring device characterized by outputting the flow rate which became many as an estimation result. The flow rate measuring device according to claim 1, wherein the flow rate calculating unit calculates the flow rate at a first time and a second time for each of the frequency bands, and the flow rate at the first time and the flow rate Calculating the difference of the flow rate at the second time,
The estimation result judgment unit
The calculated differences are compared for a plurality of the frequency bands, and it is determined based on the comparison result that one of the plurality of flow rates is output as an estimation result. Flow measurement device. The flow rate measuring device according to claim 1, wherein the flow rate calculating unit calculates the flow rate for each divided band obtained by further dividing the frequency band,
The estimation result judgment unit
The flow rates of the divided bands are compared, and the comparison is repeated for a plurality of the frequency bands, and it is determined that the flow rate of any one of the plurality of flow rates is output as an estimation result based on the comparison result. A flow rate measuring device characterized by The flow rate measuring apparatus according to claim 1, wherein a plurality of the sensors are disposed in one of the pipes, and the sensor is assigned a specific one of the frequency bands,
The flow rate calculation unit
After the flow rate is calculated for each of the sensors based on the sensor data from the sensors and the flow rates of the plurality of sensors are calculated,
The estimation result judgment unit
A flow rate measuring device, which is determined to output the flow rate of any one of a plurality of flow rates as an estimation result based on a plurality of the flow rates. The flow rate measuring device according to claim 6, wherein the flow rate calculating unit calculates the flow rate at a first time and a second time for the sensor, and the flow rate at the first time and the second time After calculating the difference of the flow rate at the time of and calculating the difference for a plurality of the sensors,
The estimation result judgment unit
The flow rate measurement is characterized by comparing the differences calculated for each of the plurality of sensors, and based on a result of the comparison, outputting any one of the flow rates among the plurality of flow rates as an estimation result. apparatus. The flow rate measuring device according to claim 6, wherein the flow rate calculating unit calculates the flow rate for each divided band obtained by further dividing the frequency band allocated to the sensor,
The estimation result judgment unit
The flow rates of the divided zones are compared, and the flow rates of the divided zones are repeated for a plurality of sensors, and the flow rate of any one of the plurality of flow rates is estimated as a result of the comparison. A flow rate measuring device characterized by judging to output. The flow rate measuring device according to claim 1, wherein a plurality of the sensors are disposed in one of the pipes, and a plurality of the frequency bands are selected for the sensors, respectively.
The flow rate calculation unit
After the flow rate is calculated for the plurality of frequency bands selected for each of the plurality of sensors based on the sensor data from the sensors,
The estimation result judgment unit
A flow rate measuring device, which is determined to output the flow rate of any one of a plurality of flow rates as an estimation result based on a plurality of the flow rates. The flow rate measurement apparatus according to claim 1, wherein the estimation result determination unit compares the flow rates of a plurality of the frequency bands, and the comparison results do not match, and there is no large number of flow rates. Is a flow rate measuring device characterized by outputting that flow rate sensing is impossible. The flow rate measuring apparatus according to claim 1, wherein a flowmeter for acquiring flow rate data of the pipe is disposed, and the acquired flow rate data and the band signal level indicating an average voltage in the frequency band are used. A flow rate measuring apparatus comprising a band candidate selection unit for selecting the frequency band for calculating a band signal level. The flow rate measurement apparatus according to claim 11, wherein the correlation calculation unit calculates a correlation between the flow rate and the band signal level for each frequency band based on the flow rate data and the band signal level, and A flow rate measuring apparatus comprising: a model calculation unit that calculates the model based on a correlation.

Images