A vortex flowmeter signal processing method and a vortex flowmeter

Abstract

The present application relates to the technical field of signal processing, in particular to a vortex flowmeter signal processing method and a vortex flowmeter. The method acquires the original analog signal after preamplification of the vortex probe, extracts the zero-crossing time of the signal, records adjacent zero-crossing times as a half cycle, and calculates the reliability of each half cycle according to the signal characteristics of the half cycle; according to the reliability, a stable core segment is screened, and the median value of the half cycle frequency in the core segment is taken as the reference frequency; the zero-crossing time is matched from both sides of the core segment by half cycle to obtain the protected zero-crossing point; according to the reference frequency, the candidate search interval is demarcated, the candidate frequency component is extracted, and the mechanical similarity representing the degree of interference is calculated by combining the characteristic differences thereof and the reference frequency; the protected zero-crossing point is taken as the boundary constraint, the mechanical similarity is combined to implement interference suppression on the candidate frequency component, and the processed signal is obtained, thereby converting and outputting the flowmeter measurement flow, thereby improving the accuracy of the flowmeter measurement flow.

Description

Technical Field

[0001] This invention relates to the field of signal processing technology, specifically to a signal processing method for a vortex flowmeter and a vortex flowmeter. Background Technology

[0002] Vortex flow meters are widely used in industrial gas metering. Their working principle is that when fluid flows through the vortex generator in the pipe, Karman vortex streets are alternately generated downstream. The vortex shedding frequency is usually proportional to the fluid velocity. Therefore, the fluid flow rate can be calculated by detecting the vortex shedding frequency.

[0003] However, in industrial operating conditions, vortex flowmeters are susceptible to multi-source interference. For example, mechanical vibrations generated by the air supply pump can be transmitted to the flowmeter probe via pipes and flanges; downstream load changes cause fluid pressure pulsations, which are transmitted to the vortex generator through the fluid medium; and electrical transient interference from pump start-up and shutdown can also interfere with the flowmeter probe signal. In summary, the above interferences act on the flowmeter probe through multiple paths: solid conduction, fluid medium, and electrical coupling. Especially under low flow rate conditions, the original analog signal output by the flowmeter probe includes vortex signal components generated by the shedding of the Karman vortex street, as well as various interference components superimposed on it. The amplitude of these interference components can far exceed that of the vortex signal components, and the interference frequencies of these components are very likely to overlap with the fundamental frequency of the vortex signal components, causing the flowmeter reading to jump, and in severe cases, directly causing measurement inaccuracies. Summary of the Invention

[0004] To address the technical challenge of accurately identifying and effectively suppressing the true vortex flow meter signal under complex operating conditions involving multi-path transmission of combined interference (including solid-borne vibration, fluid pressure pulsation, and electrical noise), thereby improving the measurement accuracy of the flow meter, this invention provides a vortex flow meter signal processing method and a vortex flow meter. The specific technical solution adopted is as follows: This invention proposes a signal processing method for a vortex flowmeter, the method comprising: Acquire the original analog signal from the vortex shear probe after pre-amplification; Extract the zero-crossing times of the original analog signal, define two adjacent zero-crossing times as a half-cycle, and calculate the reliability of each half-cycle based on the signal characteristics of each half-cycle. Stable core segments are obtained by screening based on the reliability of each half-cycle. The median value of the frequency values ​​of each half-cycle within the stable core segment is taken as the reference frequency. Zero-crossing time matching is performed by extending from the stable core segment to both sides half-cycle at a time. The zero-crossing time that passes the matching is recorded as the protection zero-crossing point. Candidate search intervals are defined based on the reference frequency. Candidate frequency components are extracted within the candidate search intervals. Based on the characteristic differences between the candidate frequency components and the reference frequency, the mechanical similarity used to characterize the degree of interference is calculated. Using the protection of the zero-crossing point as a boundary constraint, interference suppression is performed on the candidate frequency components based on mechanical similarity to obtain the suppressed signal. The suppressed signal is then converted into flow rate, and the flow rate is output as the measurement result of the vortex flow meter.

[0005] Furthermore, the original analog signal is an analog voltage signal output by the vortex shear probe, which has been pre-amplified but has not yet undergone zero-shaping and flow conversion. The original analog signal is collected at a preset sampling frequency, and a voltage-time series is formed with the sampling time as the horizontal axis and the voltage value as the vertical axis. The extraction of the zero-crossing time of the original analog signal, defining two adjacent zero-crossing times as a half-cycle, includes: The original analog signal is front-end processed to remove isolated spikes and subtract baseline drift, resulting in a processed analog signal. For each pair of adjacent sampling times in the processed analog signal, if the voltage values ​​of the pair of adjacent sampling times have opposite signs, it is determined that there is a zero-crossing point between the pair of adjacent sampling times. Using linear interpolation, the time when the voltage value between the pair of adjacent sampling times is equal to zero is calculated based on the sampling time and voltage value of each pair of adjacent sampling times, and this time is taken as the zero-crossing time. Arrange all detected zero-crossing moments in chronological order to obtain a zero-crossing moment sequence; Form a half-cycle pair by any two adjacent zero-crossing moments in the zero-crossing moment sequence; take the earliest zero-crossing moment in the half-cycle pair as the start of the half-cycle and the latest zero-crossing moment as the end of the half-cycle; extract the signal segment between the end of the half-cycle and the start of the half-cycle pair from the processed analog signal, and take it as the half-cycle signal segment corresponding to the half-cycle.

[0006] Furthermore, the process of obtaining the credibility includes: For each half-cycle, the integral value of the signal segment of that half-cycle is obtained as the first integral value; the integral value of the signal segment of the next half-cycle adjacent to that half-cycle is obtained as the second integral value; the absolute value of the difference between the first integral value and the second integral value is calculated as the absolute value of the difference; the maximum value between the first integral value and the second integral value is taken as the reference value; based on the absolute value of the difference and the reference value, the difference ratio is obtained; the difference between the value 1 and the difference ratio is calculated as the first evaluation index of that half-cycle in the positive and negative symmetry dimension. Obtain the frequency value of the half-cycle; obtain the frequency values ​​of the first preset number of half-cycles before and the first preset number of half-cycles after the half-cycle, and combine them with the frequency value of the half-cycle itself to calculate the median value of all frequency values ​​as the reference frequency; calculate the absolute value of the difference between the frequency value of the half-cycle and the reference frequency as the absolute value of the deviation; divide the absolute value of the deviation by the reference frequency to obtain the deviation ratio; normalize the difference between the value 1 and the deviation ratio to obtain the second evaluation index of the half-cycle in the frequency continuity dimension. Obtain the integral value and frequency value of the signal segment of the previous half-cycle; calculate the rate of change of the integral value of the current half-cycle relative to the previous half-cycle as the area change rate; calculate the rate of change of the frequency of the current half-cycle relative to the previous half-cycle as the frequency change rate; calculate the absolute value of the difference between the area change rate and the frequency change rate as the absolute value of the rate of change difference; calculate the sum of the absolute values ​​of the area change rate and the absolute values ​​of the frequency change rate as the sum of the absolute values ​​of the rate of change; based on the absolute values ​​of the rate of change difference and the sum of the absolute values ​​of the rate of change, obtain the misscheduling; calculate the difference between the value 1 and the misscheduling as the third evaluation index of the current half-cycle in the amplitude-frequency coordination dimension. Multiply the first evaluation index, the second evaluation index, and the third evaluation index to obtain the first product value; take the cube root of the first product value to obtain the credibility of this half-cycle.

[0007] Furthermore, the process of obtaining the stable core segment includes: Within a preset first processing window, a preset second number of consecutive half-cycles are extracted as candidate segments, starting from each half-cycle within the window, to obtain multiple candidate segments; the median and standard deviation of the confidence of all half-cycles within each candidate segment are calculated to obtain the confidence median and confidence standard deviation. The highest median confidence score is used as the primary screening criterion, and the lowest standard deviation of confidence score is used as the secondary screening criterion. One of the candidate segments is selected as the stable core segment. The median of the frequency values ​​of all half-cycles within the stable core segment is used as the reference frequency.

[0008] Furthermore, the step of performing zero-crossing time matching from the stable core segment outwards in half-cycles, and recording the successful zero-crossing time as the protection zero-crossing point, includes: Based on the reference frequency, the time length of a complete cycle is calculated, and then divided by two to obtain the reference half-cycle length; Starting from the leftmost zero-crossing moment of the stable core segment, the prediction is extrapolated to the left half-cycle according to the baseline half-cycle duration to obtain the predicted zero-crossing moments on the left. At the same time, starting from the rightmost zero-crossing moment of the stable core segment, the prediction is extrapolated to the right half-cycle according to the baseline half-cycle duration to obtain the predicted zero-crossing moments on the right. For each predicted zero-crossing moment, search the sequence of zero-crossing moments obtained by arranging all zero-crossing moments in time to see if there is an actual zero-crossing moment whose time interval between the predicted zero-crossing moment and the actual zero-crossing moment is within the preset sampling interval. If it exists, the match is successful, and the actual zero-crossing time is included in the protected zero-crossing point set. The next zero-crossing time is then predicted to both sides from the actual zero-crossing time as the new starting point. If it does not exist, the match fails, and the extension operation stops.

[0009] Furthermore, the extraction process of the candidate frequency components includes: Establish a frequency range centered on the reference frequency and with a preset deviation as the radius, as the candidate search range; Time-frequency analysis is performed on the processed analog signal to obtain the energy distribution of the processed analog signal in the time-frequency domain. The energy distribution includes the energy value corresponding to each sampling time and each frequency. In the time-frequency domain, within the candidate search interval, the continuous change trajectory of the frequency with the largest energy value at each sampling moment over time is tracked to obtain the candidate ridge line; The candidate frequency range is determined based on the frequency corresponding to each sampling time in the candidate ridge line; The signal components within the candidate frequency range are reconstructed from the processed analog signal and used as candidate frequency components; where the candidate frequency components are time series signals with time as the horizontal axis and voltage value as the vertical axis.

[0010] Furthermore, the process of obtaining the mechanical similarity includes: Perform Hilbert transform on the candidate frequency components to obtain the instantaneous frequency information, amplitude envelope information, and instantaneous phase information of the candidate frequency components; Taking each half-cycle corresponding to the zero-crossing point of the protection as the object, the integral and frequency of each half-cycle are calculated separately to obtain the integral value and frequency value of each half-cycle, which are used as the half-cycle integral value and half-cycle frequency value corresponding to the zero-crossing point of the protection. The predicted zero-crossing time corresponding to each protection zero-crossing point is obtained by extrapolating the reference frequency from the stable core segment to both sides half-cycle, and the theoretical phase is calculated based on the time interval between each predicted zero-crossing time and the actual zero-crossing time. The evaluation index for frequency stability is determined based on the difference between instantaneous frequency information and reference frequency. The evaluation index for amplitude stability is determined based on the difference between the amplitude envelope information and the integral values ​​of each half-cycle corresponding to the protection zero crossing point. Based on the correlation between the amplitude envelope information and the frequency values ​​of each half-cycle corresponding to the zero-crossing point of the protection, the evaluation index of the amplitude-frequency correlation dimension is determined. The evaluation index for phase stability is determined based on the difference between instantaneous phase information and theoretical phase. The mechanical similarity of candidate frequency components is calculated by comprehensively considering evaluation metrics from at least two of the following dimensions: frequency stability, amplitude stability, amplitude-frequency correlation, and phase stability.

[0011] Furthermore, the process of obtaining the suppressed signal includes: Based on the mechanical similarity of candidate frequency components, the suppression weight of the candidate frequency components is determined; wherein, if the mechanical similarity is greater than or equal to a first preset threshold, the suppression weight is set to a first weight value; if the mechanical similarity is less than the first preset threshold but greater than or equal to a second preset threshold, the mechanical similarity of the candidate frequency components is directly used as the suppression weight; if the mechanical similarity is less than the second preset threshold, the suppression weight is set to a second weight value, and the second weight value is less than the first weight value. The suppression weight is multiplied by the voltage value of the candidate frequency component at each sampling time to obtain the suppression component; Using the protected zero-crossing point as the boundary constraint and the signal value corresponding to the protected zero-crossing point as the constraint condition, the suppressed component is removed from the processed analog signal to obtain the suppressed signal.

[0012] Furthermore, the step of removing the suppressed component from the simplified analog signal using the protected zero-crossing point as a boundary constraint and without changing the signal value corresponding to the protected zero-crossing point as a constraint condition, to obtain the suppressed signal, includes: Obtain the voltage values ​​corresponding to each protection zero-crossing point on the processed analog signal; The voltage values ​​at the corresponding times of each protection zero-crossing point in the suppression component are compared with the voltage values ​​at the corresponding times of each protection zero-crossing point on the rectified analog signal. If the absolute value of the voltage value of the suppression component at the corresponding time of the protection zero-crossing point is greater than the absolute value of the voltage value of the protection zero-crossing point on the rectified analog signal, it is determined to be an abnormal protection zero-crossing point; otherwise, it is determined to be a normal protection zero-crossing point. For abnormal protection zero crossings, the absolute value of the voltage value on the rectified analog signal at the abnormal protection zero crossing is used as the limiting value, and the suppression component in the preset neighborhood of the abnormal protection zero crossing is limited. For normal protection zero crossings, the suppression component within the preset neighborhood of the normal protection zero crossing is retained; After traversing all protection zero-crossing points, the suppression components that have been limited or retained in the neighborhood of each protection zero-crossing point are combined to obtain the limited suppression components. The suppressed signal is obtained by subtracting the amplitude-limited suppressed component from the rectified analog signal.

[0013] A vortex flow meter includes: a signal acquisition module for acquiring the original analog signal of the vortex probe after pre-amplification; a signal processing module for executing a vortex flow meter signal processing method; and a flow output module for outputting the flow rate as the measurement result of the vortex flow meter.

[0014] The present invention has the following beneficial effects: The original analog signal after preamplification of the flowmeter is directly acquired for refined waveform analysis, avoiding the loss of waveform features caused by traditional shaping processing, and providing a data foundation for the reliability of each half-cycle in subsequent analysis. The half-cycle is divided using the zero-crossing moment, and the reliability of each waveform segment is quantified by combining the signal characteristics of the half-cycle, so as to accurately select the stable core segment with the best signal-to-noise ratio. The median value of its frequency is used as the reference frequency to ensure the authenticity and stability of the reference frequency, providing a unified and stable basis for subsequent candidate search interval delineation and mechanical similarity judgment, avoiding distortion of global frequency statistics due to mechanical interference (such as interference from the air supply pump and interference caused by pump start-up and shutdown). The protective zero-crossing point is obtained by bidirectional extension matching of the stable core segment, locking the effective waveform boundary, and extracting candidate frequency components based on the reference frequency. The mechanical similarity is calculated by combining the feature differences with the reference frequency, accurately identifying superimposed mechanical interference and real vortex signal components. Relying on the protective zero-crossing point as a boundary constraint and mechanical similarity to suppress interference, the problem of multi-path composite interference and co-frequency interference in industrial sites is effectively overcome, and the measurement accuracy of the vortex flowmeter under low flow velocity and complex working conditions is significantly improved. Attached Figure Description

[0015] Figure 1 This is a flowchart of a vortex flowmeter signal processing method according to an embodiment of the present invention; Figure 2 An example diagram illustrating the process of obtaining credibility provided in one embodiment of the present invention. Detailed Implementation

[0016] The following description, in conjunction with the accompanying drawings, details a signal processing method for a vortex flowmeter and a specific scheme for the vortex flowmeter provided by the present invention.

[0017] Please see Figure 1 The diagram illustrates a flowchart of a vortex flowmeter signal processing method according to an embodiment of the present invention, the method comprising: S101: Acquire the original analog signal from the vortex shear probe after pre-amplification.

[0018] It should be noted that, in actual use, the original analog signal output by the probe of a vortex flow meter is not a pure vortex signal.

[0019] For example, taking a gas supply system with frequent start-stop of the air supply pump as an example, interference signals are coupled to the original analog signal of the probe through at least the following three paths: one type of interference signal is solid mechanical vibration transmitted through pipes and flanges: the vibration generated by the operation of the air supply pump is transmitted to the flow meter body through metal pipes and flanges, directly acting on the piezoelectric probe and generating a periodic electrical signal in the same frequency band as the vortex street signal; another type of interference signal is pressure pulsation transmitted through the fluid medium: the reciprocating motion of the air supply pump piston and the fluctuation of downstream load cause periodic changes in the gas pressure in the pipe, which then propagates to the vortex generator region through the fluid medium, changing the stability of the wake flow field, resulting in the amplitude and frequency of the original analog signal output by the probe being modulated; yet another type of interference signal is electrical transient coupled through the signal link: the high-frequency electromagnetic interference generated when the air supply pump starts and stops is coupled to the probe signal link through the power line or spatial radiation, superimposing spike pulses on the signal.

[0020] In summary, the superposition of the aforementioned multipath interference with the real vortex shear signal components will cause problems such as zero-crossing time shift, amplitude envelope distortion, and frequency trajectory breakage in the original analog signal output by the probe, resulting in the real vortex shear signal components being submerged or distorted.

[0021] It is important to understand that vortex shear probes typically use piezoelectric crystals as sensing elements. When a fluid vortex acts on the probe, the piezoelectric crystal deforms under pressure, generating a charge signal. This signal is usually weak and cannot be directly converted from analog to digital or further digitally processed. It needs to be converted into a voltage signal and amplified by a preamplifier, that is, the weak charge signal is converted into an analog voltage signal that can be sampled. Therefore, preamplification is an essential step before the signal enters digital processing. Thus, this invention acquires the original analog signal after preamplification. At this stage, the original analog signal has not undergone any zero-crossing shaping or waveform transformation and still retains complete information such as zero-crossing position, amplitude envelope, and phase relationship.

[0022] The original analog signal is an analog voltage signal output by a vortex shear probe, which has been pre-amplified but has not yet undergone zero-shaping and flow conversion. The original analog signal is collected at a preset sampling frequency, and a voltage-time series is formed with the sampling time as the horizontal axis and the voltage value as the vertical axis.

[0023] It should be noted that the specific value of the preset sampling frequency must satisfy the Nyquist sampling theorem, that is, the sampling frequency must be at least twice the highest effective frequency of the signal. The highest effective frequency can usually be extracted from the effective frequency range of the vortex flowmeter. The effective frequency range of the vortex flowmeter is determined by the instrument diameter and upper limit of the flow range of the vortex flowmeter itself, which is common knowledge in the field and will not be specifically limited here. For example, assuming the highest vortex frequency is 1000Hz, the preset sampling frequency can be set to 4000Hz to 10000Hz.

[0024] Furthermore, it is important to note that zero-crossing shaping and flow conversion processes convert the original analog signal into a square wave pulse or directly output frequency values. During this process, waveform details such as the amplitude envelope, half-cycle integral value, instantaneous frequency changes, and harmonic phase relationships of the original analog signal are lost. These waveform details precisely reflect the different physical characteristics of the vortex shedding signal components. For example, the amplitude envelope can be used to distinguish between amplitude abrupt changes caused by mechanical interference and actual flow rate changes; the half-cycle integral value can reflect the intensity variation pattern of the vortex shedding signal components; the instantaneous frequency changes can be used to identify electromagnetic interference during pump start-up; and the harmonic phase relationships can be used to distinguish between actual vortex shedding and random mechanical impact. In summary, the aforementioned waveform details need to be retained so that they can be used to distinguish between the actual vortex shedding signal (i.e., the actual vortex shedding signal components contained in the original analog signal) and multipath interference signals, i.e., to identify signals under multipath composite interference.

[0025] S102: Extract the zero-crossing times of the original analog signal, define two adjacent zero-crossing times as a half-cycle, and calculate the reliability of each half-cycle based on the signal characteristics of each half-cycle.

[0026] It is important to understand that in a real vortex shear signal, the frequency of adjacent half-cycles should change continuously and slowly, without abrupt changes. In other words, the most essential characteristic of a real vortex shear signal lies in the alternating pattern of adjacent positive and negative half-cycles. Generally speaking, whether the amplitudes of the positive and negative half-cycles are symmetrical reflects the stability of vortex shedding, whether the frequency changes continuously reflects the smoothness of fluid flow, and whether the amplitude and frequency change synchronously reflects the physical consistency of the real vortex shear signal. These characteristics can only be accurately captured by using half-cycles as the unit of analysis; otherwise, the differences between positive and negative half-cycles will be masked. Therefore, this invention defines adjacent zero-crossing moments as a half-cycle and calculates its reliability for each half-cycle. The reliability reflects the degree to which each half-cycle belongs to a real vortex shear signal.

[0027] In this embodiment, the original analog signal is front-end processed to remove isolated spikes and subtract baseline drift, resulting in a processed analog signal. Each pair of adjacent sampling times in the processed analog signal is iterated. If the voltage values ​​of the pair of adjacent sampling times have opposite signs, a zero-crossing point is determined to exist between them. Using linear interpolation, the moment when the voltage value between the pair of adjacent sampling times is equal to zero is calculated based on their respective sampling times and voltage values, and this is taken as the zero-crossing moment. All detected zero-crossing moments are arranged in chronological order to obtain a zero-crossing moment sequence. Any two adjacent zero-crossing moments in the zero-crossing moment sequence are combined into a half-cycle pair. The earliest zero-crossing moment in the half-cycle pair is taken as the start point of the half-cycle, and the latest zero-crossing moment is taken as the end point of the half-cycle. A signal segment located between the end point and the start point of the half-cycle pair is extracted from the processed analog signal and taken as the corresponding half-cycle signal segment.

[0028] It should be noted that the isolated spike removal and baseline drift subtraction used in the front-end processing are conventional signal preprocessing techniques in this field, and can be directly adopted here without further elaboration on their specific implementation algorithms.

[0029] As we know, electromagnetic interference generated during the start-up and shutdown of the air supply pump enters the probe signal through the signal link coupling path, causing electrical spikes in the original analog signal. Fluid pressure pulsations, acting on the probe through the fluid medium conduction path, cause baseline drift in the original analog signal, resulting in an overall signal offset. Therefore, these path interference phenomena can introduce false zero crossings. Thus, it is necessary to first perform front-end processing on the original analog signal, removing isolated spikes and subtracting baseline drift to obtain a processed analog signal.

[0030] Opposite signs of voltage values ​​mean that the voltage values ​​at two adjacent sampling times are one positive and one negative. For example, if the voltage value at the previous sampling time is +0.3V and the voltage value at the next sampling time is -0.2V, it indicates that there is a zero-crossing point between these two sampling times.

[0031] It should be noted that using linear interpolation to determine the moment when the voltage is zero between two points is a conventional technique in this field and will not be elaborated upon here.

[0032] Two adjacent zero-crossing moments refer to two zero-crossing moments that are immediately adjacent in chronological order on the time axis.

[0033] The process of obtaining credibility is as follows: Figure 2 As shown, it includes: S201: For each half-cycle, obtain the integral value of the signal segment of that half-cycle as the first integral value; obtain the integral value of the signal segment of the next half-cycle adjacent to that half-cycle as the second integral value; calculate the absolute value of the difference between the first integral value and the second integral value as the absolute value of the difference; take the maximum value between the first integral value and the second integral value as the reference value; obtain the difference ratio based on the absolute value of the difference and the reference value; calculate the difference between the value 1 and the difference ratio as the first evaluation index of that half-cycle in the positive and negative symmetry dimension.

[0034] The integral value of a half-cycle signal segment refers to the integral of the absolute value of the signal within that segment with respect to time, i.e., the area enclosed by the signal waveform and the time axis. Specifically, it is obtained by multiplying the absolute voltage values ​​at each sampling time point within the segment by the sampling interval and then summing them up.

[0035] The integral value (whether it is the first integral value or the second integral value) reflects the intensity of the vortex shedding signal within the relevant half-cycle.

[0036] The absolute value of the difference reflects the degree of difference in signal intensity between adjacent half-cycles. Since a fundamental physical characteristic of a real vortex shear signal is the alternation of positive and negative half-cycles with roughly symmetrical amplitudes, a smaller absolute value of the difference indicates that the signal intensity of adjacent half-cycles (usually positive and negative half-cycles) is closer, the symmetry is better, and it better matches the physical characteristics of a real vortex shear. Conversely, a larger absolute value of the difference indicates that adjacent half-cycles are asymmetrical, and interference is more likely to exist.

[0037] Difference ratio = Absolute value of difference ÷ (Reference value + Preset minimum value).

[0038] To ensure that the difference ratio is unaffected by the overall signal strength and facilitates consistent comparisons under different flow rate conditions, a reference value is used to facilitate subsequent normalization of the absolute value of the difference.

[0039] The difference ratio is used to represent the relative difference in signal strength between adjacent half-cycles. A smaller difference ratio indicates better symmetry between adjacent half-cycles and a signal closer to the true vortex shedding signal; conversely, a larger difference ratio indicates a greater difference between adjacent half-cycles and a higher likelihood of false zero crossings caused by interference.

[0040] S202: Obtain the frequency value of the half-cycle; obtain the frequency values ​​of the first preset number of half-cycles before and the first preset number of half-cycles after the half-cycle, and combine them with the frequency value of the half-cycle itself to calculate the median value of all the frequency values ​​as the reference frequency; calculate the absolute value of the difference between the frequency value of the half-cycle and the reference frequency as the absolute value of the deviation; divide the absolute value of the deviation by the reference frequency to obtain the deviation ratio; normalize the difference between the value 1 and the deviation ratio to obtain the second evaluation index of the half-cycle in the frequency continuity dimension.

[0041] It is understood that the frequency of a real vortex shear signal changes continuously and slowly over a short period of time without sudden changes. If the frequency of a certain half-cycle deviates too much from the median frequency of several adjacent half-cycles, it indicates that the half-cycle may be caused by a false zero crossing.

[0042] The frequency value of a half-cycle is calculated based on the duration of the signal segment in that half-cycle (the end of the half-cycle - the beginning of the half-cycle). Specifically, frequency = 1 ÷ (2 × duration of the half-cycle).

[0043] It should be noted that the specific value of the preset first quantity can be set based on industry experience, and this embodiment does not impose a specific limitation. For example, the preset first quantity can be 2.

[0044] For example, assuming the first quantity is 2, and this half-cycle is the i-th half-cycle, then the first two half-cycles are the (i-2)-th and (i-1)-th, and the last two half-cycles are the (i+1)-th and (i+2)-th. The frequency values ​​of these four half-cycles, together with the frequency value of the half-cycle itself, are counted, and the median value is taken as the reference frequency.

[0045] The absolute value of the deviation reflects the degree to which the frequency of the half-cycle deviates from the overall frequency level of the adjacent half-cycle. The smaller the absolute value of the deviation, the closer the frequency of the half-cycle is to the frequency level of the adjacent half-cycle, the smoother the frequency change, and the more consistent it is with the physical characteristics of a real vortex street. Conversely, the larger the absolute value of the deviation, the more likely the frequency of the half-cycle has changed abruptly, which is more likely to be caused by false zero crossings or interference.

[0046] As an example, the initial difference is obtained by subtracting the deviation ratio from the value 1; if the initial difference is less than 0, it is set to 0; if the initial difference is greater than 1, it is set to 1. After the above normalization process, the value of the second evaluation index is limited to the range of 0 to 1.

[0047] The second evaluation index is used to measure the continuity between the frequency of the current half-cycle and the frequency level of the adjacent half-cycle, reflecting whether the frequency change of the vortex shear signal is stable. Specifically, the closer the second evaluation index is to 1, the closer the frequency of the current half-cycle is to the frequency level of the adjacent half-cycle, the more stable the frequency change, and the more it conforms to the characteristics of a true vortex shear. Conversely, the closer the index is to 0, the greater the deviation between the frequency of the current half-cycle and the frequency level of the adjacent half-cycle, and the more likely there is interference or false zero crossing.

[0048] S203: Obtain the integral value and frequency value of the signal segment of the previous half-cycle; calculate the rate of change of the integral value of the current half-cycle relative to the previous half-cycle as the area change rate; calculate the rate of change of the frequency of the current half-cycle relative to the previous half-cycle as the frequency change rate; calculate the absolute value of the difference between the area change rate and the frequency change rate as the absolute value of the rate of change difference; calculate the sum of the absolute values ​​of the area change rate and the absolute values ​​of the frequency change rate as the sum of the absolute values ​​of the rate of change; obtain the misscheduling based on the absolute value of the rate of change difference and the sum of the absolute values ​​of the rate of change; calculate the difference between the value 1 and the misscheduling as the third evaluation index of the current half-cycle in the amplitude-frequency coordination dimension.

[0049] The area change rate = (integral value of the current half-cycle - integral value of the previous half-cycle) ÷ integral value of the previous half-cycle, which reflects the relative change in signal strength between adjacent half-cycles. Similarly, the frequency change rate is calculated by subtracting the frequency value of the previous half-cycle from the frequency value of the current half-cycle, and then dividing by the frequency value of the previous half-cycle, reflecting the relative change in frequency between adjacent half-cycles.

[0050] Disruption = Absolute value of the difference in rate of change ÷ (Sum of absolute values ​​of rate of change + Preset minimum).

[0051] Discontinuity reflects the degree of asynchrony between amplitude and frequency changes, and is a quantitative indicator of whether the two are coordinated. The smaller the discontinuity, the closer the rate of change of area and the rate of change of frequency are, the more synchronized the changes of amplitude and frequency are, and the more consistent they are with the physical characteristics of a real vortex street. Conversely, the larger the discontinuity, the greater the difference between amplitude and frequency changes, and the more likely there are disturbances such as mechanical vibrations.

[0052] S204: Multiply the first evaluation index, the second evaluation index, and the third evaluation index to obtain the first product value; take the cube root of the first product value to obtain the reliability of this half-cycle.

[0053] It should be noted that when performing all division operations involved in the above credibility indicators (including the calculation of difference ratio, deviation ratio, area change rate and frequency change rate), if the value contained in the denominator is zero or close to zero, a preset minimum amount should be added to the denominator to prevent division overflow from affecting the calculation of credibility score.

[0054] It should be noted that the specific value of the preset minimum quantity can be determined based on the signal amplitude level and the analog-to-digital conversion accuracy to ensure that it does not interfere with normal numerical calculations. For example, a typical value for the preset minimum quantity is 10 to the power of negative six.

[0055] It's important to understand that the first, second, and third evaluation metrics each characterize the physical features of a real vortex street from different dimensions. Poor performance in any of these dimensions will significantly lower the credibility. Therefore, a geometric mean (e.g., the cube root) is used instead of an arithmetic mean, making the credibility more sensitive to the weakest link (i.e., the smallest evaluation metric). Credibility is only high when all three dimensions perform well.

[0056] S103: Based on the reliability of each half-cycle, a stable core segment is obtained. The median value of the frequency values ​​of each half-cycle within the stable core segment is taken as the reference frequency. Zero-crossing time matching is performed by extending from the stable core segment to both sides half-cycle at a time. The zero-crossing time that passes the matching is recorded as the protection zero-crossing point.

[0057] It's important to understand that while the reliability of a single half-cycle reflects its signal quality, it cannot guarantee consistent rhythm between adjacent half-cycles. Furthermore, there might be individual half-cycles with high reliability but surrounding half-cycles with lower reliability. Directly using these individual half-cycles as the benchmark for subsequent extended matching can easily lead to misjudging reliable signals, thus affecting the accuracy of the zero-crossing protection. Therefore, it's necessary to select a segment within a continuous time window—composed of multiple consecutive half-cycles—that has the highest overall reliability and best internal consistency, as the stable core segment.

[0058] In this embodiment, within a preset first processing window, a preset second number of consecutive half-cycles are extracted as candidate segments, starting from each half-cycle within the window, resulting in multiple candidate segments. The median and standard deviation of the confidence level of all half-cycles within each candidate segment are calculated to obtain the confidence median and confidence standard deviation. The highest confidence median is used as the priority screening condition, and the lowest confidence standard deviation is used as the secondary screening condition. One segment is selected from each candidate segment as a stable core segment. The median of the frequency values ​​of all half-cycles within the stable core segment is used as the reference frequency.

[0059] The default first processing window is a short time window that slides on the timeline to define the range of the half-cycle being processed.

[0060] It should be noted that the specific value of the preset first processing window can be adjusted by those skilled in the art based on the actual situation of the vortex shedding signal, and this embodiment does not impose a specific limitation. For example, the preset first processing window can be set to 5 complete vortex shedding cycles, that is, 10 half cycles.

[0061] To ensure that the frequency statistics within the selected segments are statistically significant, the selected segments typically need to contain at least two complete alternating positive and negative cycles, i.e., a complete alternating positive and negative cycle consisting of one positive half-cycle and one adjacent negative half-cycle. Based on this, the second preset number can be set to four half-cycles.

[0062] For example, assuming the preset second quantity is 4, the first to fourth half-cycles are extracted from the first half-cycle within the window as the first candidate segment, and the second to fifth half-cycles are extracted from the second half-cycle as the second candidate segment. The extraction direction is from the starting point in the direction of increasing time. In addition, if the number of consecutive half-cycles remaining at the edge of the window is less than the preset second quantity, then the starting point does not constitute a candidate segment and is skipped directly.

[0063] The median credibility value reflects the overall credibility level of a candidate segment.

[0064] The confidence standard deviation reflects the degree of dispersion of confidence among the half-cycles within a candidate segment.

[0065] For example, suppose there is candidate segment A with a median confidence level of 0.92 and a standard deviation of 0.05; and candidate segment B with a median confidence level of 0.90 and a standard deviation of 0.02. In this case, candidate segment A with the higher median confidence level is selected first. If the median confidence level of candidate segment A and candidate segment B is both 0.92, and the standard deviation of candidate segment A is 0.05 and the standard deviation of candidate segment B is 0.02, then candidate segment B is selected.

[0066] Statistically, the median is more robust than the arithmetic mean, being less affected by individual abnormal half-cycle frequency values. Therefore, the median is used to reliably represent the dominant frequency of the true vortex shedding signal within the window, providing a scientific data basis for subsequent prediction of zero-crossing positions and delineation of candidate search intervals.

[0067] In a preferred implementation, based on the reference frequency, the time length of a complete cycle is calculated, and then divided by two to obtain the reference half-cycle duration. Starting from the leftmost zero-crossing moment of the stable core segment, predictions are made by extrapolating half-cycles to the left according to the reference half-cycle duration to obtain the predicted zero-crossing moments on the left. Simultaneously, starting from the rightmost zero-crossing moment of the stable core segment, predictions are made by extrapolating half-cycles to the right according to the reference half-cycle duration to obtain the predicted zero-crossing moments on the right. For each predicted zero-crossing moment, the sequence of zero-crossing moments arranged chronologically is searched to see if there is an actual zero-crossing moment whose time interval with the predicted zero-crossing moment is within a preset sampling interval. If it exists, the match is successful, and the actual zero-crossing moment is included in the protection zero-crossing point set. The next zero-crossing moment is then predicted to both sides from the actual zero-crossing moment as a new starting point. If it does not exist, the match fails, and the extension operation stops.

[0068] The time length of a complete cycle is obtained by taking the reciprocal of the reference frequency, and then dividing by two to obtain the reference half-cycle length.

[0069] It should be noted that when there is no flow, the probe of the vortex flow meter outputs a noise signal and there is no effective vortex frequency. At this time, a stable core segment will not be formed, and it will not enter the subsequent extended matching process. Therefore, the boundary case with a reference frequency of zero has been eliminated in actual processing.

[0070] It should be noted that the preset sampling interval is determined by the preset sampling frequency, and this embodiment does not impose a specific limitation. For example, when the preset sampling frequency is 2000Hz, the preset sampling interval is 0.0005 seconds.

[0071] Specifically, after a successful match, the actual zero-crossing moment is taken as a new starting point, and the next zero-crossing moment is predicted on the outside of it according to the reference half-cycle duration. That is, the left side continues to predict to the left, and the right side continues to predict to the right, realizing a rolling extension of half-cycles until a match on one side fails.

[0072] The actual zero-crossing moment is the real zero-crossing point position extracted from the processed analog signal through zero-crossing detection. In extended matching, it represents the actual position of the real zero-crossing point in the signal and can be used as an immovable anchor point in subsequent interference suppression.

[0073] Furthermore, if the number of consecutive matching failures on a certain side reaches a preset failure threshold, the extension on that side will be stopped.

[0074] It should be noted that the specific value of the preset failure count threshold can be determined based on engineering experience, and this embodiment does not impose a specific limitation. For example, the preset failure count threshold can be 2.

[0075] S104: Define the candidate search interval based on the reference frequency, extract candidate frequency components within the candidate search interval, and calculate the mechanical similarity used to characterize the degree of interference based on the feature differences between the candidate frequency components and the reference frequency.

[0076] It is important to understand that when extracting candidate frequency components, it is impossible to search the entire frequency range generated by the processed analog signal. Furthermore, the dominant frequency of the actual vortex shedding signal is already given by the reference frequency. If the vibration frequency of interference sources such as the air supply pump happens to fall within the frequency measurement range of the vortex flowmeter and is close to the reference frequency, the interference will overlap with the actual vortex shedding signal in the frequency domain, making conventional frequency domain filtering indistinguishable. If the interference is far from the dominant frequency of the vortex shedding signal, it can be directly separated by frequency domain filtering and is not the focus of this invention. Therefore, this invention only defines a candidate search interval near the reference frequency to detect and process potential interference components that are close to the vortex shedding signal in the frequency domain. Moreover, this defined candidate search interval is not too large, preventing the introduction of irrelevant noise, and also avoids omitting interference components close to the reference frequency.

[0077] As a preferred implementation, a frequency range centered on a reference frequency and with a preset deviation as its radius is established as the candidate search range. Time-frequency analysis is performed on the processed analog signal to obtain its energy distribution in the time-frequency domain, which includes the energy values ​​corresponding to each sampling time and frequency. In the time-frequency domain, within the candidate search range, the continuous change trajectory of the frequency with the highest energy value at each sampling time is traced over time to obtain candidate ridges. Based on the frequencies corresponding to each sampling time in the candidate ridges, the candidate frequency range is determined. Signal components within the candidate frequency range are reconstructed from the processed analog signal as candidate frequency components. The candidate frequency components are time-series signals with time as the horizontal axis and voltage as the vertical axis.

[0078] It should be noted that the specific value of the preset deviation can be determined by comprehensively considering the median absolute deviation of the half-cycle frequency within the stable core segment, the frequency resolution of the time-frequency analysis, and the amount of sampling error propagating to the frequency domain at the zero-crossing moment. For example, the sum of these three factors can be used as the preset deviation, and this embodiment does not impose a specific limitation. For instance, the preset deviation can be set to 5Hz.

[0079] Median absolute deviation: First, the median frequency value is obtained, then the absolute deviation of each frequency from the median frequency value is calculated, and the median of these absolute deviations is the median absolute deviation. This is common statistical knowledge. The median absolute deviation of the half-cycle frequencies within the stable core segment reflects the natural dispersion of frequencies within the stable core segment.

[0080] Frequency resolution: The smallest resolvable interval on the frequency axis in time-frequency analysis, which can be calculated from the preset sampling frequency and can be applied directly here.

[0081] It should be noted that the amount of sampling error propagating to the frequency domain at the zero-crossing moment can be predetermined by those skilled in the art based on the sampling frequency and half-cycle duration, and will not be elaborated here.

[0082] It should be noted that time-frequency analysis (such as short-time Fourier transform or continuous wavelet transform) is a conventional technique in this field and can be used directly here without further elaboration on its specific implementation algorithm.

[0083] The time-frequency domain is a two-dimensional domain, with time on the horizontal axis and frequency on the vertical axis. Each point in the energy distribution represents the signal energy intensity at that sampling time and frequency.

[0084] The candidate ridge line reflects the trajectory of the frequency point with the highest energy at each sampling moment within the candidate search interval, which is the path of the change of the dominant frequency of the candidate frequency component over time.

[0085] Since the candidate ridges record the frequency values ​​corresponding to each sampling time, and these frequency values ​​may fluctuate within a range over time, the minimum and maximum values ​​among these frequency values ​​are taken, and the frequency range formed by the minimum and maximum values ​​is used as the candidate frequency range.

[0086] It is important to understand that since the candidate frequency components extracted within the candidate search interval may be either genuine vortex shedding signals or mechanical vibration interference adjacent to the vortex shedding frequency, their essential differences cannot be distinguished solely by frequency position. Therefore, it is necessary to comprehensively compare the characteristic differences between the candidate frequency components and the genuine vortex shedding signals from four dimensions: frequency stability, amplitude stability, amplitude-frequency correlation, and phase stability. A quantified mechanical similarity is then calculated to characterize the likelihood that the candidate frequency component belongs to mechanical vibration, thus providing a basis for determining whether to suppress the candidate frequency component subsequently.

[0087] In a preferred implementation, a Hilbert transform is performed on the candidate frequency components to obtain their instantaneous frequency information, amplitude envelope information, and instantaneous phase information. Taking each half-cycle corresponding to the protected zero-crossing point as the object, integration and frequency calculations are performed on each half-cycle to obtain the integral value and frequency value of each half-cycle, which serve as the half-cycle integral value and half-cycle frequency value corresponding to the protected zero-crossing point. The predicted zero-crossing time corresponding to each protected zero-crossing point is obtained by extrapolating the reference frequency from the stable core segment to both sides half-cycle by half-cycle, and the theoretical phase is calculated based on the time interval between each predicted zero-crossing time and the actual zero-crossing time. Based on the instantaneous frequency information and... The evaluation index for frequency stability is determined based on the differences between reference frequencies; the evaluation index for amplitude stability is determined based on the differences between amplitude envelope information and the integral values ​​of each half-cycle corresponding to the protection zero-crossing point; the evaluation index for amplitude-frequency correlation is determined based on the correlation between amplitude envelope information and the frequency values ​​of each half-cycle corresponding to the protection zero-crossing point; the evaluation index for phase stability is determined based on the differences between instantaneous phase information and theoretical phase; and the mechanical similarity of candidate frequency components is calculated by comprehensively considering the evaluation indexes of at least two of the following dimensions: frequency stability, amplitude stability, amplitude-frequency correlation, and phase stability.

[0088] It should be noted that the Hilbert transform is a conventional technique in the field for extracting instantaneous frequency, amplitude envelope, and instantaneous phase from a signal, and its specific implementation algorithm will not be elaborated here.

[0089] It should be noted that integrating a signal segment to obtain an integral value and calculating the frequency value based on the duration of the signal segment are both conventional signal processing methods in this field, and will not be elaborated here.

[0090] The theoretical phase is calculated by dividing the time deviation between the predicted and actual zero-crossing times by the reference half-cycle duration and then multiplying by... It was obtained through conversion.

[0091] The frequency stability dimension is used to evaluate whether the frequency variation of candidate frequency components is more stable than that of the actual vortex shedding signal.

[0092] It is known that mechanical vibration is constrained by motor speed and structural modes, and its frequency is usually more stable than that of the real vortex shedding signal (which is affected by fluid turbulence and has slight fluctuations). If the frequency dispersion of the candidate frequency component is less than that of the real vortex shedding signal, the frequency stability evaluation index is too high, indicating that the candidate frequency component is more likely to be mechanical vibration.

[0093] As an example, the instantaneous frequency information of candidate frequency components is obtained. The instantaneous frequency values ​​of each sampling time of the candidate frequency components are then discretized (e.g., median absolute deviation) to obtain the first degree of dispersion. The frequency value sequence corresponding to each original half-cycle within the stable core segment is then obtained, and its dispersion is also calculated to obtain the second degree of dispersion. The ratio of the second degree of dispersion to the first degree of dispersion is calculated to obtain the evaluation index for the frequency stability dimension. A larger ratio indicates that the frequency of the candidate frequency component is more stable than the actual vortex shedding signal, and is more likely to be mechanical vibration.

[0094] It should be noted that the median absolute deviation is a well-known concept in statistics, and its specific calculation method can be directly applied, so it will not be elaborated here. For example, for the instantaneous frequency values ​​at each sampling time within a candidate frequency component, the absolute deviation between the instantaneous frequency value at each sampling time and the median value of the instantaneous frequency values ​​at all times is calculated, and then the median value of all absolute deviations is taken to obtain the first dispersion.

[0095] Frequency value sequence corresponding to each original half-cycle within the stable core segment: When determining the stable core segment, the stable core segment contains several consecutive half-cycles. Each half-cycle has a frequency value calculated based on its own duration. These frequency values ​​are arranged in chronological order to form a frequency value sequence.

[0096] The amplitude stability dimension is used to evaluate whether the amplitude variation of candidate frequency components is more stable than that of the real vortex shedding signal.

[0097] It is known that the amplitude of mechanical vibration is relatively stable after being transmitted through the structure, while the amplitude of the real vortex shedding signal fluctuates naturally with the turbulence of the fluid. If the amplitude dispersion of the candidate frequency component is less than that of the real vortex shedding signal, the amplitude stability evaluation index is too high, indicating that the candidate frequency component is more likely to be mechanical vibration.

[0098] As an example, the amplitude envelope information of candidate frequency components is obtained, and the amplitude envelope values ​​of the candidate frequency components at each time point are discretized to obtain the third degree of dispersion; the integral values ​​of each half-cycle corresponding to the protected zero-crossing point are obtained, and the degree of dispersion is also obtained to obtain the fourth degree of dispersion; the ratio of the fourth degree of dispersion to the third degree of dispersion is calculated to obtain the evaluation index of amplitude stability. A larger ratio indicates that the amplitude of the candidate frequency component is more stable than the actual vortex shedding signal, and is more likely to be mechanical vibration.

[0099] The amplitude-frequency correlation dimension is used to evaluate whether the amplitude changes of candidate frequency components are not synchronized with the frequency changes of the real vortex shedding signal.

[0100] It is known that the amplitude and frequency of a real vortex shedding signal are usually positively correlated (i.e., the frequency increases as the flow rate increases, and the intensity of the vortex also increases). However, the amplitude change of mechanical vibration is not affected by the vortex shedding frequency. If the correlation between the amplitude change of a candidate frequency component and the vortex shedding frequency change is lower, the evaluation index of the amplitude-frequency correlation dimension is higher, indicating that the candidate frequency component is more likely to be mechanical vibration.

[0101] As an example, the amplitude envelope information of candidate frequency components is obtained, and the change in amplitude envelope between adjacent half-cycles is calculated to obtain an amplitude change sequence; the frequency values ​​of each half-cycle corresponding to the protection zero-crossing point are obtained, and the change between adjacent half-cycles is calculated to obtain a frequency change sequence; the Pearson correlation coefficient between the amplitude change sequence and the frequency change sequence is calculated; the absolute value of the correlation coefficient is subtracted from the numerical value of 1 to obtain the evaluation index of the amplitude-frequency correlation dimension. A higher evaluation index indicates that the amplitude change of the candidate frequency component is less synchronized with the vortex shedding frequency change, and is more likely to be mechanical vibration.

[0102] The phase stability dimension is used to evaluate whether the phase of a candidate frequency component is more stable than that of the actual vortex shedding signal.

[0103] It is known that real vortex shedding signals are affected by fluid turbulence, resulting in minute random disturbances in the zero-crossing phase; mechanical vibrations are constrained by the structure, leading to smaller and more regular zero-crossing phase diffusion. If the phase diffusion of a candidate frequency component is less than that of the real vortex shedding signal, the phase stability evaluation index is too high, indicating that the candidate frequency component is more likely to be mechanical vibration.

[0104] Since the amplitude envelope information is a continuous sequence arranged according to sampling times, its temporal resolution is higher than the half-cycle scale, and a single half-cycle contains multiple sampling points. Therefore, in order to facilitate the quantification of amplitude changes between adjacent half-cycles, the amplitude envelope sequence within each half-cycle needs to be reduced to a scalar feature value. Specifically, the integral value of the amplitude envelope sequence within each half-cycle is taken as the amplitude envelope feature value of that half-cycle. The integration method is to multiply the amplitude envelope values ​​at each sampling time within that half-cycle by the sampling interval and then sum them up. This integral value reflects the overall intensity level of the amplitude envelope within that half-cycle, which is consistent with the calculation method of the half-cycle integral value in step S201.

[0105] It should be noted that the difference in amplitude envelope eigenvalues ​​between adjacent half-cycles is used to represent the amount of change in amplitude envelope between adjacent half-cycles.

[0106] As an example, the predicted zero-crossing time corresponding to each protection zero-crossing point is obtained by extrapolating half-cycles outward from the stable core segment according to the reference frequency. The time deviation between the actual zero-crossing time and the corresponding predicted zero-crossing time of each protection zero-crossing point is converted into a phase deviation, resulting in the theoretical phase deviation for each protection zero-crossing point. The dispersion of each theoretical phase deviation is calculated to obtain the fifth dispersion. The instantaneous phase information of the candidate frequency components is obtained, and the phase residual is obtained after removing the linear trend. The dispersion of the phase residual is calculated to obtain the sixth dispersion. The ratio of the fifth dispersion to the sixth dispersion is calculated to obtain the evaluation index of the phase stability dimension. The larger the ratio, the more stable the phase of the candidate frequency component is compared with the actual vortex shedding signal, and the more likely it is to be mechanical vibration.

[0107] It should be noted that when calculating the evaluation indicators of the above dimensions, if the denominator is less than a preset minimum positive number, a preset minimum positive number is directly used to replace the denominator in the calculation to avoid division overflow. The specific value of the preset minimum positive number can be determined by those skilled in the art based on the actual application scenario. This embodiment does not impose a specific limitation. For example, the preset minimum positive number can be 10 to the power of negative six.

[0108] Furthermore, in the above calculation process, the evaluation indicators for each dimension are uniformly mapped to the interval between 0 and 1 using ratios or difference ratios. If the value exceeds 1 or is less than 0, it is also truncated to the interval between 0 and 1. Therefore, the evaluation indicators are numerically comparable and can be directly used for subsequent comprehensive calculations without additional normalization. For example, frequency stability and amplitude stability are both expressed as a ratio of dispersion, amplitude-frequency correlation is expressed as 1 minus the absolute value of the correlation coefficient, and phase stability is also expressed as a ratio of dispersion. The values ​​of each indicator all fall between 0 and 1.

[0109] Specifically, the selected effective evaluation indicators (at least two dimensions) are multiplied together, and then the square root is taken based on the number of effective indicators to obtain the mechanical similarity of the candidate frequency components. For example, if all four dimensions of evaluation indicators are effective, the four evaluation indicators are multiplied together and the fourth root is taken; if only three dimensions are effective, the three evaluation indicators are multiplied together and the third root is taken; if only two dimensions are effective, the two evaluation indicators are multiplied together and the square root is taken.

[0110] S105: Using the protection zero-crossing point as a boundary constraint, interference suppression is performed on candidate frequency components based on mechanical similarity to obtain the suppressed signal. The suppressed signal is then converted into flow rate, and the flow rate is output as the measurement result of the vortex flow meter.

[0111] As one possible implementation, a suppression weight for a candidate frequency component is determined based on its mechanical similarity. Specifically, if the mechanical similarity is greater than or equal to a first preset threshold, the suppression weight is set to a first weight value. If the mechanical similarity is less than the first preset threshold but greater than or equal to a second preset threshold, the mechanical similarity of the candidate frequency component is directly used as the suppression weight. If the mechanical similarity is less than the second preset threshold, the suppression weight is set to a second weight value, which is less than the first weight value. The suppression weight is multiplied by the voltage value of the candidate frequency component at each sampling time to obtain the suppressed component. Using the protected zero-crossing point as a boundary constraint and without changing the signal value corresponding to the protected zero-crossing point as a constraint condition, the suppressed component is removed from the processed analog signal to obtain the suppressed signal.

[0112] It should be noted that the specific values ​​of the first preset threshold and the second preset threshold can be preset by those skilled in the art according to the actual application scenario, and this embodiment does not impose specific limitations. For example, the first preset threshold can be 0.7, and the second preset threshold can be 0.3.

[0113] For example, the first weight value is set to 1, and the second weight value is set to 0.

[0114] The suppression weight is used to determine the degree of suppression of candidate frequency components; the larger the value, the stronger the suppression of candidate frequency components.

[0115] It is important to understand that since the protected zero-crossing is a highly reliable zero-crossing obtained through extended matching of a stable core segment, it represents reliable phase information of the actual vortex shedding signal. If the signal value corresponding to the protected zero-crossing is changed during interference suppression, it is equivalent to disrupting the phase boundary of the actual vortex shedding signal, which will lead to deviations in subsequent flow conversion. Therefore, it is necessary to ensure that the suppression operation does not change the signal value corresponding to the protected zero-crossing to protect the integrity of the actual vortex shedding signal.

[0116] As an example, the voltage values ​​corresponding to each protection zero-crossing point on the simplified analog signal are obtained; the voltage values ​​at the corresponding times of each protection zero-crossing point in the suppressed component are compared with the voltage values ​​at the corresponding times of each protection zero-crossing point on the simplified analog signal; if the absolute value of the voltage value of the suppressed component at the corresponding time of the protection zero-crossing point is greater than the absolute value of the voltage value of the protection zero-crossing point on the simplified analog signal, it is determined to be an abnormal protection zero-crossing point; otherwise, it is determined to be a normal protection zero-crossing point; for abnormal protection zero-crossing points, the absolute value of the voltage value of the abnormal protection zero-crossing point on the simplified analog signal is used as the limiting value, and the suppressed component in the preset neighborhood of the abnormal protection zero-crossing point is limited; for normal protection zero-crossing points, the suppressed component in the preset neighborhood of the normal protection zero-crossing point is retained; after traversing all protection zero-crossing points, the suppressed components in the neighborhood of each protection zero-crossing point that have been limited or retained are combined to obtain the limited suppressed component; the limited suppressed component is subtracted from the simplified analog signal to obtain the suppressed signal.

[0117] Additionally, it should be noted that since the zero-crossing time is calculated through linear interpolation between adjacent sampling times, the corresponding voltage value on the processed analog signal at the protected zero-crossing point is not strictly zero. Instead, it reflects the actual amplitude of the signal at that zero-crossing point. This amplitude is typically a minimum value, representing the physical offset of the real signal near the zero-crossing point, and can serve as an effective benchmark for amplitude limiting comparison rather than a theoretical zero value. If the amplitude is less than a preset minimum positive number, this preset minimum positive number is used instead in the comparison calculation to avoid abnormal judgments due to excessively small values.

[0118] The specific value of the preset minimum positive number can be determined by those skilled in the art based on the actual application scenario and the signal amplitude level; this embodiment does not impose specific limitations. For example, the preset minimum positive number can be taken as follows based on the signal amplitude level: to The values ​​between.

[0119] Abnormal protection zero-crossing reflects that the amplitude of the suppressed component at the protection zero-crossing point exceeds the amplitude of the original signal. That is, if the suppressed component is removed directly, the signal value at the protection zero-crossing point will change (such as zero-crossing point offset), thereby destroying the phase boundary of the real vortex shedding signal.

[0120] The suppressed components within the preset neighborhood refer to the suppressed component segments within a certain time range before and after the protected zero-crossing point. Suppressed components within this range are treated as a whole and subjected to amplitude limiting. For example, the preset neighborhood can be set to a time range of one sampling period before and after (i.e., the reciprocal of the preset sampling frequency).

[0121] It should be noted that for suppression components that do not fall within any of the preset neighborhoods of the protection zero-crossing points, their original values ​​are kept unchanged and they are directly included in the combination. During combination, the suppression components that have been clipped or retained as is within each protection zero-crossing neighborhood are sequentially spliced ​​with all the unchanged suppression components that do not fall within the above neighborhoods according to their original positions on the time axis, to obtain the clipped suppression components of the same length as the processed analog signal.

[0122] An example of the combination process is as follows: the suppressed component segments that have been clipped or retained in the neighborhood of each protected zero-crossing point are arranged sequentially according to their original positions on the time axis. The overlapping parts between the segments are averaged or spliced ​​together according to a linear transition to obtain the complete clipped suppressed component.

[0123] Mechanical vibration interference in the suppressed signal has been suppressed, and the signal value at the protected zero-crossing point has not changed, thus protecting the phase boundary of the real vortex shedding signal.

[0124] It should be noted that the method of converting the suppressed signal into flow rate is a well-known technique in the field of vortex flow meters, and will not be elaborated here. Specifically, the half-cycle duration between adjacent protection zero-crossing points is calculated based on the protection zero-crossing point in the suppressed signal, and the frequency of each half-cycle is obtained. The median value of all half-cycle frequencies is taken as the corrected vortex frequency, and then multiplied by the preset instrument coefficient to obtain the flow rate.

[0125] It should be noted that the preset instrument coefficient depends on the diameter and generator size of the vortex flowmeter, and is determined through calibration before the vortex flowmeter leaves the factory. This embodiment does not impose specific limitations on it. For example, the instrument coefficient of a certain vortex flowmeter is 320 pulses per liter, that is, 320 vortex signal pulses are generated for every 1 liter of fluid passing through.

[0126] A vortex flow meter includes: a signal acquisition module for acquiring the original analog signal of the vortex probe after pre-amplification; a signal processing module for executing a vortex flow meter signal processing method; and a flow output module for outputting the flow rate as the measurement result of the vortex flow meter.

Claims

1. A signal processing method for a vortex flowmeter, characterized in that, The method includes: Acquire the original analog signal from the vortex shear probe after pre-amplification; Extract the zero-crossing times of the original analog signal, define two adjacent zero-crossing times as a half-cycle, and calculate the reliability of each half-cycle based on the signal characteristics of each half-cycle. Stable core segments are obtained by screening based on the reliability of each half-cycle. The median value of the frequency values ​​of each half-cycle within the stable core segment is taken as the reference frequency. Zero-crossing time matching is performed by extending from the stable core segment to both sides half-cycle at a time. The zero-crossing time that passes the matching is recorded as the protection zero-crossing point. Candidate search intervals are defined based on the reference frequency. Candidate frequency components are extracted within the candidate search intervals. Based on the characteristic differences between the candidate frequency components and the reference frequency, the mechanical similarity used to characterize the degree of interference is calculated. Using the protection of the zero-crossing point as a boundary constraint, interference suppression is performed on the candidate frequency components based on mechanical similarity to obtain the suppressed signal. The suppressed signal is then converted into flow rate, and the flow rate is output as the measurement result of the vortex flow meter.

2. The signal processing method for a vortex flowmeter according to claim 1, characterized in that, The original analog signal is an analog voltage signal output by a vortex shear probe, which has been pre-amplified but has not yet undergone zero-shaping and flow conversion. The original analog signal is collected at a preset sampling frequency, and a voltage-time series is formed with the sampling time as the horizontal axis and the voltage value as the vertical axis. The extraction of the zero-crossing time of the original analog signal, defining two adjacent zero-crossing times as a half-cycle, includes: The original analog signal is front-end processed to remove isolated spikes and subtract baseline drift, resulting in a processed analog signal. For each pair of adjacent sampling times in the processed analog signal, if the voltage values ​​of the pair of adjacent sampling times have opposite signs, it is determined that there is a zero-crossing point between the pair of adjacent sampling times. Using linear interpolation, the time when the voltage value between the pair of adjacent sampling times is equal to zero is calculated based on the sampling time and voltage value of each pair of adjacent sampling times, and this time is taken as the zero-crossing time. Arrange all detected zero-crossing moments in chronological order to obtain a zero-crossing moment sequence; Form a half-cycle pair by any two adjacent zero-crossing moments in the zero-crossing moment sequence; take the earliest zero-crossing moment in the half-cycle pair as the start of the half-cycle and the latest zero-crossing moment as the end of the half-cycle; extract the signal segment between the end of the half-cycle and the start of the half-cycle pair from the processed analog signal, and take it as the half-cycle signal segment corresponding to the half-cycle.

3. The signal processing method for a vortex flowmeter according to claim 2, characterized in that, The process of obtaining the credibility includes: For each half-cycle, the integral value of the signal segment of that half-cycle is obtained as the first integral value; the integral value of the signal segment of the next half-cycle adjacent to that half-cycle is obtained as the second integral value; the absolute value of the difference between the first integral value and the second integral value is calculated as the absolute value of the difference; the maximum value between the first integral value and the second integral value is taken as the reference value; based on the absolute value of the difference and the reference value, the difference ratio is obtained; the difference between the value 1 and the difference ratio is calculated as the first evaluation index of that half-cycle in the positive and negative symmetry dimension. Obtain the frequency value of the half-cycle; obtain the frequency values ​​of the first preset number of half-cycles before and the first preset number of half-cycles after the half-cycle, and combine them with the frequency value of the half-cycle itself to calculate the median value of all frequency values ​​as the reference frequency; calculate the absolute value of the difference between the frequency value of the half-cycle and the reference frequency as the absolute value of the deviation; divide the absolute value of the deviation by the reference frequency to obtain the deviation ratio; normalize the difference between the value 1 and the deviation ratio to obtain the second evaluation index of the half-cycle in the frequency continuity dimension. Obtain the integral value and frequency value of the signal segment of the previous half-cycle; calculate the rate of change of the integral value of the current half-cycle relative to the previous half-cycle as the area change rate; calculate the rate of change of the frequency of the current half-cycle relative to the previous half-cycle as the frequency change rate; calculate the absolute value of the difference between the area change rate and the frequency change rate as the absolute value of the rate of change difference; calculate the sum of the absolute values ​​of the area change rate and the absolute values ​​of the frequency change rate as the sum of the absolute values ​​of the rate of change; based on the absolute values ​​of the rate of change difference and the sum of the absolute values ​​of the rate of change, obtain the misscheduling; calculate the difference between the value 1 and the misscheduling as the third evaluation index of the current half-cycle in the amplitude-frequency coordination dimension. Multiply the first evaluation index, the second evaluation index, and the third evaluation index to obtain the first product value; take the cube root of the first product value to obtain the credibility of this half-cycle.

4. The signal processing method for a vortex flowmeter according to claim 1, characterized in that, The process of obtaining the stable core fragment includes: Within a preset first processing window, a preset second number of consecutive half-cycles are extracted as candidate segments, starting from each half-cycle within the window, to obtain multiple candidate segments; the median and standard deviation of the confidence of all half-cycles within each candidate segment are calculated to obtain the confidence median and confidence standard deviation. The highest median confidence score is used as the primary screening criterion, and the lowest standard deviation of confidence score is used as the secondary screening criterion. One of the candidate segments is selected as the stable core segment. The median of the frequency values ​​of all half-cycles within the stable core segment is used as the reference frequency.

5. The signal processing method for a vortex flowmeter according to claim 4, characterized in that, The process of performing zero-crossing time matching by extending the process from the stable core segment to both sides in half-cycles, and recording the zero-crossing time that passes the matching as the protection zero-crossing point, includes: Based on the reference frequency, the time length of a complete cycle is calculated, and then divided by two to obtain the reference half-cycle length; Starting from the leftmost zero-crossing moment of the stable core segment, the prediction is extrapolated to the left half-cycle according to the baseline half-cycle duration to obtain the predicted zero-crossing moments on the left. At the same time, starting from the rightmost zero-crossing moment of the stable core segment, the prediction is extrapolated to the right half-cycle according to the baseline half-cycle duration to obtain the predicted zero-crossing moments on the right. For each predicted zero-crossing moment, search the sequence of zero-crossing moments obtained by arranging all zero-crossing moments in chronological order to find whether there is an actual zero-crossing moment whose time interval between the predicted zero-crossing moment and the actual zero-crossing moment is within the preset sampling interval; If it exists, the match is successful, and the actual zero-crossing time is included in the protected zero-crossing point set. The next zero-crossing time is then predicted to both sides from the actual zero-crossing time as the new starting point. If it does not exist, the match fails, and the extension operation stops.

6. The signal processing method for a vortex flowmeter according to claim 2, characterized in that, The extraction process of the candidate frequency components includes: Establish a frequency range centered on the reference frequency and with a preset deviation as the radius, as the candidate search range; Time-frequency analysis is performed on the processed analog signal to obtain the energy distribution of the processed analog signal in the time-frequency domain. The energy distribution includes the energy value corresponding to each sampling time and each frequency. In the time-frequency domain, within the candidate search interval, the continuous change trajectory of the frequency with the largest energy value at each sampling moment over time is tracked to obtain the candidate ridge line; The candidate frequency range is determined based on the frequency corresponding to each sampling time in the candidate ridge line; The signal components within the candidate frequency range are reconstructed from the processed analog signal and used as candidate frequency components; where the candidate frequency components are time series signals with time as the horizontal axis and voltage value as the vertical axis.

7. The signal processing method for a vortex flowmeter according to claim 1, characterized in that, The process of obtaining the mechanical similarity includes: Perform Hilbert transform on the candidate frequency components to obtain the instantaneous frequency information, amplitude envelope information, and instantaneous phase information of the candidate frequency components; Taking each half-cycle corresponding to the zero-crossing point of the protection as the object, the integral and frequency of each half-cycle are calculated separately to obtain the integral value and frequency value of each half-cycle, which are used as the half-cycle integral value and half-cycle frequency value corresponding to the zero-crossing point of the protection. The predicted zero-crossing time corresponding to each protection zero-crossing point is obtained by extrapolating the reference frequency from the stable core segment to both sides half-cycle, and the theoretical phase is calculated based on the time interval between each predicted zero-crossing time and the actual zero-crossing time. The evaluation index for frequency stability is determined based on the difference between instantaneous frequency information and reference frequency. The evaluation index for amplitude stability is determined based on the difference between the amplitude envelope information and the integral values ​​of each half-cycle corresponding to the protection zero crossing point. Based on the correlation between the amplitude envelope information and the frequency values ​​of each half-cycle corresponding to the zero-crossing point of the protection, the evaluation index of the amplitude-frequency correlation dimension is determined. The evaluation index for phase stability is determined based on the difference between instantaneous phase information and theoretical phase. The mechanical similarity of candidate frequency components is calculated by comprehensively considering evaluation metrics from at least two of the following dimensions: frequency stability, amplitude stability, amplitude-frequency correlation, and phase stability.

8. The signal processing method for a vortex flowmeter according to claim 2, characterized in that, The process of obtaining the suppressed signal includes: Based on the mechanical similarity of candidate frequency components, the suppression weight of the candidate frequency components is determined; wherein, if the mechanical similarity is greater than or equal to a first preset threshold, the suppression weight is set to a first weight value; if the mechanical similarity is less than the first preset threshold but greater than or equal to a second preset threshold, the mechanical similarity of the candidate frequency components is directly used as the suppression weight; if the mechanical similarity is less than the second preset threshold, the suppression weight is set to a second weight value, and the second weight value is less than the first weight value. The suppression weight is multiplied by the voltage value of the candidate frequency component at each sampling time to obtain the suppression component; Using the protected zero-crossing point as the boundary constraint and the signal value corresponding to the protected zero-crossing point as the constraint condition, the suppressed component is removed from the processed analog signal to obtain the suppressed signal.

9. The signal processing method for a vortex flowmeter according to claim 8, characterized in that, The process of removing suppressed components from the simplified analog signal, using the protected zero-crossing point as a boundary constraint and without changing the signal value corresponding to the protected zero-crossing point as a constraint condition, to obtain the suppressed signal includes: Obtain the voltage values ​​corresponding to each protection zero-crossing point on the processed analog signal; The voltage values ​​at the corresponding times of each protection zero-crossing point in the suppression component are compared with the voltage values ​​at the corresponding times of each protection zero-crossing point on the rectified analog signal. If the absolute value of the voltage value of the suppression component at the corresponding time of the protection zero-crossing point is greater than the absolute value of the voltage value of the protection zero-crossing point on the rectified analog signal, it is determined to be an abnormal protection zero-crossing point; otherwise, it is determined to be a normal protection zero-crossing point. For abnormal protection zero crossings, the absolute value of the voltage value on the rectified analog signal at the abnormal protection zero crossing is used as the limiting value, and the suppression component in the preset neighborhood of the abnormal protection zero crossing is limited. For normal protection zero crossings, the suppression component within the preset neighborhood of the normal protection zero crossing is retained; After traversing all protection zero-crossing points, the suppression components that have been limited or retained in the neighborhood of each protection zero-crossing point are combined to obtain the limited suppression components. The suppressed signal is obtained by subtracting the amplitude-limited suppressed component from the rectified analog signal.

10. A vortex flow meter, characterized in that, include: The signal acquisition module is used to acquire the original analog signal of the vortex shear probe after pre-amplification; The signal processing module is used to execute the vortex flowmeter signal processing method as described in any one of claims 1 to 9; The flow output module is used to output the flow rate as the measurement result of the vortex flow meter.

Images