An electromagnetic flowmeter

Abstract

This application relates to the field of flow measurement technology and provides an electromagnetic flowmeter, comprising: an excitation device for outputting an excitation signal to alternate between high-frequency and low-frequency excitation, wherein a complete excitation cycle of the excitation signal includes a first high-frequency band, a first low-frequency band, a second high-frequency band, and a second low-frequency band, wherein the high-frequency band signal can be used to effectively reduce the influence of flow noise, and the low-frequency band signal can be used to ensure the zero-point stability of the flowmeter; electrodes for detecting induced electromotive force and outputting electrode signals; and a signal conditioning and acquisition device for amplifying and analog filtering the electrode signals and converting them into digital signals, wherein the digital signals include high-frequency electrode signals and low-frequency electrode signals; and eliminating zero-point drift through a signal processing algorithm to obtain a stable and effective high-frequency electrode signal.

Description

Technical Field

[0001] This invention relates to the field of flow measurement technology, specifically to an electromagnetic flowmeter based on alternating high and low frequency excitation, suitable for high-precision measurement of general slurries or low conductivity fluids.

[0002] Background Technology / Prior Application Content

[0003] An electromagnetic flowmeter is a flow measurement instrument based on Faraday's law of electromagnetic induction. Its main components include a pipe cavity, a pair of signal electrodes, an excitation coil, an excitation drive unit, and a signal processing unit. The excitation drive unit drives the excitation coil to generate a magnetic field B perpendicular to the fluid flow direction. When the fluid flows through the pipe cavity at an average velocity V, the signal electrodes output an induced electromotive force signal. ,in, This is a coefficient related to factors such as pipe diameter. The signal processing unit calculates the average flow velocity V of the fluid based on the induced electromotive force signal E, and calculates the flow rate by combining this with the cross-sectional area of ​​the pipe.

[0004] Currently, most mainstream electromagnetic flowmeters use low-frequency square wave excitation, suitable for general flow measurement applications. However, in some special cases, such as measuring fluids containing a certain amount of slurry or low conductivity, flowmeters with low-frequency square wave excitation may experience significant output fluctuations. This is because in these situations, the noise spectrum exhibits a 1 / f characteristic, and low-frequency interference close to the flowmeter's operating frequency has a significant impact on the flowmeter's output.

[0005] One approach is to use electromagnetic flowmeters with high-frequency excitation (such as mains frequency excitation), which can significantly reduce output fluctuations, but is prone to zero-point instability. This zero-point stability issue is the main reason why, in general applications, the more recently developed low-frequency square wave excitation method has replaced the mains frequency excitation method.

[0006] One improved method is to simultaneously apply high-frequency and low-frequency excitation, as exemplified by the dual-frequency excitation electromagnetic flowmeter from Japan, application number CN87101677.X (Kobayashi Tamotsu, Kuromori Kenichi, Goto Shigeru, Matsunaga Yoshinori, Torimaru Takashi, Sukuya Norihiro, Takanashi Takanagami, Electromagnetic Flowmeter). With simultaneous application of both frequencies, the amplitudes of the low-frequency and high-frequency signals from the sensor signal are extracted separately using signal sampling and amplitude demodulation. These are then filtered separately using a low-pass filter with a large time constant and a high-pass filter with an equal time constant. The two filtering results are summed to obtain the response speed of the high-frequency signal and the stability of the low-frequency signal.

[0007] However, this typical dual-frequency excitation control and corresponding signal processing system is quite complex and has a high implementation cost. To date, it has been produced primarily by Yokogawa Corporation of Japan. Some domestic companies, after weighing the complexities of dual-frequency and high-frequency technologies, are still developing in the direction of high-frequency excitation, such as CN201010215831 (Xu Kejun, Yang Shuanglong, Wang Gang, Liang Liping, Zhang Ran, Shi Lei, Yang Yibing, A DSP-based electromagnetic flowmeter signal processing system); some companies are using new signal extraction methods on the basis of dual-frequency excitation, such as CN201910055218 (Ge Liang, Li Longhai, Li Junlan, Lai Xin, Wei Guohui, Yang Qing, Jia Hu, Shi Mingjiang, Deng Kui, Huang Long, Huang Qi, An electromagnetic flowmeter system and signal extraction method based on dual-frequency excitation); and some companies are adding ternary excitation on the basis of dual-frequency excitation to obtain a stable zero point, such as CN202410387058 (Cheng Zhuming, Xu Liusheng, He Rongbo, Hu Xuefeng, An electromagnetic flowmeter system and signal processing method based on dual-frequency ternary rectangular wave).

[0008] For the measurement of slurries or fluids with low conductivity, the same applicant has disclosed an electromagnetic flowmeter using a novel dual-frequency excitation method in its priority document (application number 202510782467.4), in which the absolute value sequence of the trapezoidal weighted coefficient sequence has 0.5 at the beginning and end and 1 for the rest.

[0009] Compared with the prior application, the main improvements of this invention are: the original high- and low-frequency shared excitation section has been redefined as a high-frequency excitation section for simplification; a longer, multi-step weighted coefficient sequence is adopted, thereby significantly improving the ability to resist low-frequency noise, while maintaining the original low cost and ease of implementation, demonstrating the advantages of implementation methods and cost-effectiveness. Summary of the Invention

[0010] This invention provides an electromagnetic flowmeter for measuring slurries or low-conductivity fluids, offering both zero-point stability and resistance to flow noise. The flowmeter inherits the advantages of low-frequency square-wave electromagnetic flowmeters in its hardware design, such as simplicity and low cost, while simultaneously addressing problems in existing technologies through alternating high and low frequency excitation.

[0011] This application provides an electromagnetic flowmeter, see [link]. Figure 1 It includes: an excitation device; electrodes for detecting induced electromotive force and outputting electrode signals; a signal conditioning and acquisition device for amplifying and filtering the electrode signals and converting them into digital signals, including high-frequency electrode signals and low-frequency electrode signals; and an MCU that integrates a computing unit and an output unit.

[0012] The excitation device is configured to output alternating high-frequency and low-frequency excitation, and a complete excitation cycle of the excitation signal includes a first high-frequency segment, a first low-frequency segment, a second high-frequency segment, and a second low-frequency segment arranged sequentially (e.g., ...). Figure 2 , Figure 3 (As shown). The excitation device specifically includes an excitation control circuit, a constant current drive circuit, and an excitation coil connected in sequence. The excitation control circuit is controlled by the MCU to output an excitation current with alternating high-frequency and low-frequency bands. The excitation device is also equipped with a current detection circuit to reflect whether the excitation current is normal.

[0013] The MCU's output unit is configured to output control signals to the excitation device according to a preset excitation timing sequence; the MCU's arithmetic unit is configured to execute the algorithm specified in this invention during excitation to perform calculations on the digital signals output by the signal conditioning and acquisition device, so as to realize the signal calculation, weighted accumulation and zero-point drift elimination steps.

[0014] In this embodiment, the excitation configuration is such that the excitation directions of the first low-frequency band and the second low-frequency band are opposite, and the electrode signals of these two low-frequency bands are exactly the same as the electrode signals obtained under ordinary low-frequency square waves; the starting excitation direction of the second high-frequency band is opposite to the starting excitation direction of the first high-frequency band, so that the high-frequency signal of the high-frequency band is also the same as the low-frequency signal, and will have positive and negative signs depending on the starting excitation direction; the excitation level directions of the beginning and end of the high-frequency band are the same, and the corresponding duration is an odd multiple of the half cycle of the high-frequency excitation, so as to use a weighted coefficient sequence of odd length.

[0015] Using the weighted coefficient sequence selector of claim 5, different pre-stored weighted coefficient sequences can be selected as needed. As an example, the pre-stored weighted coefficient sequences available for selection by the weighted coefficient sequence selector in this patent include, but are not limited to:

[0016] w1 = [-1 / 6, 1 / 2, -5 / 6, 1, -1, 1, -5 / 6, 1 / 2, -1 / 6]

[0017] w2 = [-1 / 4, 1 / 2, -3 / 4, 1, -1, 1, -3 / 4, 1 / 2, -1 / 4]

[0018] w3 = [-1 / 6, 1 / 2, -5 / 6, 1, -1, 1, -1, 1, -1, 1, -1, 1, -5 / 6, 1 / 2, -1 / 6]

[0019] w4 = [-1 / 4, 1 / 2, -3 / 4, 1, -1, 1, -1, 1, -1, 1, -1, 1, -3 / 4, 1 / 2, -1 / 4]

[0020] In this embodiment, the high-frequency signal is 75Hz. When the selector is w1 or w2, a total of 9 high-frequency half-cycles are involved in the calculation, which meets the requirements of claim 2 (4): the excitation length is an odd multiple of the high-frequency excitation half-cycle. The weighting coefficient sequence has 4 values ​​with different absolute values; the sum of the absolute values ​​of the weighting coefficient sequence is an integer N=6; correspondingly, the total duration of the N high-frequency excitation half-cycles is 40ms, which is exactly twice the power frequency cycle of 20ms, which helps to eliminate power frequency interference components.

[0021] When the selector is selected as w3 or w4, a total of 15 high-frequency half-cycles participate in the calculation, which also meets the requirements of claim 2 (4). The weighted coefficient sequence has 4 values ​​with different absolute values; the sum of the absolute values ​​of the weighted coefficient sequence is 12, and the length of the 12 75Hz high-frequency half-cycles is 80ms, which is an integer multiple of the power frequency.

[0022] In this embodiment, the arithmetic unit is equipped with a preset signal processing algorithm, which includes the following steps:

[0023] Electrode signal amplitude calculation;

[0024] High-frequency zero-point drift calculation;

[0025] Calculation of effective high-frequency electrode signals.

[0026] In this embodiment, the specific steps for calculating the electrode signal amplitude include:

[0027] Calculate the amplitude of the high-frequency electrode signal and the amplitude of the low-frequency electrode signal using high-frequency electrode signals and low-frequency electrode signals respectively;

[0028] The amplitudes of the high-frequency electrode signal and the low-frequency electrode signal are respectively subjected to first-order low-pass filtering with the same parameters and a large time constant. The difference between the two filtering results is then used as the high-frequency zero-point drift.

[0029] In this embodiment, the specific steps for calculating the effective high-frequency electrode signal include: removing the high-frequency zero-point drift from the high-frequency amplitude signal and using it as the effective high-frequency electrode signal of the flow meter.

[0030] The technical solution of this invention has the following technical effects: It employs alternating high and low frequency excitation, with high-frequency and low-frequency excitation occurring alternately to form a complete excitation cycle. The high-frequency portion has a higher frequency, effectively reducing the influence of flow noise; the low-frequency portion ensures the zero-point stability of the flowmeter. Furthermore, the new sequence formed by taking the absolute value of the weighted coefficient sequence used in this invention contains at least three different values, which can better suppress low-frequency noise. It not only has good suppression capability within the 0~10Hz range but is also applicable to the measurement of fast-flowing slurries or slurries containing bubbles with a noise spectrum extending to 10~30Hz. Attached Figure Description

[0031] Figure 1 This is a system topology diagram of an electromagnetic flowmeter in one embodiment of this application;

[0032] Figure 2 This is a schematic diagram of the excitation waveform when w1 or w2 is selected in one embodiment of this application;

[0033] Figure 3 This is a schematic diagram of the excitation waveform when w3 or w4 is selected in one embodiment of this application;

[0034] Figure 4(a) is the theoretical spectrum diagram of the weighted sequences w1 and w2 that satisfy features (A) to (E) of claim 4;

[0035] Figure 4(b) shows the theoretical spectrum of the weighted sequences w3 and w4 that satisfy features (A) to (E) of claims 4;

[0036] Figure 5 (a) shows the experimental suppression characteristics verification data of the w1 sequence;

[0037] Figure 5 (b) shows the experimental suppression characteristics verification data of the w2 sequence. Detailed Implementation

[0038] Symbol explanation:

[0039] - w1, w2, w3, w4: Implementation examples of weighted coefficient sequences pre-stored in this application, wherein w1 and w2 are 9-point sequences, and w3 and w4 are 15-point sequences;

[0040] - N: The sum of the absolute values ​​of the weighted coefficient sequence, which also represents the equivalent number of high-frequency excitation half-cycles used for integration with the power frequency.

[0041] - L: The length of the weighted coefficient sequence.

[0042] The technical solution of this application will now be described in detail with reference to the accompanying drawings. Figure 1 As shown, the electromagnetic flowmeter includes an excitation device, electrodes, a signal conditioning and acquisition device, and an MCU. The excitation device, controlled by the MCU, outputs alternating high-frequency and low-frequency excitation. The electrodes detect the induced electromotive force and output electrode signals. The signal conditioning and acquisition device amplifies, filters, and converts the electrode signals from analog to digital, outputting high-frequency and low-frequency electrode signals. The MCU integrates a processing unit to calculate the high-frequency and low-frequency electrode signals to eliminate zero points and obtain a valid high-frequency electrode signal.

[0043] In the embodiment, when w1 or w2 is selected, the excitation signal waveform of a complete excitation cycle is as follows: Figure 2As shown. Both the first and second high-frequency bands contain nine half-cycles of 75Hz square wave signals, with a total duration of 60ms. The initial excitation direction of the first high-frequency band is negative, and the initial excitation direction of the second high-frequency band is positive, to obtain the high-frequency signals of the corresponding high-frequency bands. The duration of both the first and second low-frequency bands is 20ms. The excitation directions of the first and second low-frequency bands are opposite, consistent with the excitation method of the low-frequency square wave flowmeter, thus obtaining the low-frequency electrode signals related to low-frequency excitation. The excitation direction of the first low-frequency band is the same as the initial excitation direction of the first high-frequency band, and the excitation direction of the second low-frequency band is the same as the initial excitation direction of the second high-frequency band. Since the tail-end level direction is the same as its respective initial level direction, the tail-end level direction of the high-frequency band is consistent with the level direction of the adjacent low-frequency band. Therefore, the actual effective low-frequency band duration is the low-frequency band duration plus the duration of the high-frequency half-cycle, which is beneficial for the stability of the low-frequency electrode signals. The entire excitation cycle length is 160ms.

[0044] The weighted coefficient sequences corresponding to the nine high-frequency excitation half-cycles are -1 / 6, 1 / 2, -5 / 6, 1, -1, 1, -5 / 6, 1 / 2, -1 / 6 or -1 / 4, 1 / 2, -3 / 4, 1, -1, 1, -3 / 4, 1 / 2, -1 / 4. The sum of the absolute values ​​of the weighted coefficient sequences is N=6. The duration of these six high-frequency half-cycles (40ms) and the duration of half excitation cycle (80ms) are both set to integer multiples of the power frequency interference duration of 20ms (power frequency noise frequency is 50Hz), which helps to reduce the impact of power frequency interference.

[0045] The high-frequency signal in the high-frequency band is calculated using the weighted coefficient sequence conforming to claim 4. Under positive flow conditions, the signal calculated from this weighted coefficient sequence is negative in the first high-frequency band and positive in the second high-frequency band. Therefore, the positive / negative sign of the signal in the high-frequency band is the same as that under ordinary square wave excitation.

[0046] In the embodiment, when w3 or w4 is selected, the excitation signal waveform of a complete excitation cycle is as follows: Figure 3 As shown. Both the first and second high-frequency bands contain 15 half-cycles of square wave signals at a frequency of 75Hz, with a total duration of 100ms. The initial excitation direction of the first high-frequency band is negative, and the initial excitation direction of the second high-frequency band is positive. The duration of both the first and second low-frequency bands is 20ms. The excitation directions of the first and second low-frequency bands are opposite, but the excitation direction of the first low-frequency band is the same as the initial excitation direction of the first high-frequency band; the entire excitation cycle length is 240ms.

[0047] The weighted coefficient sequences for the 15 half-cycles are -1 / 6, 1 / 2, -5 / 6, 1, -1, 1, -1, 1, -1, 1, -5 / 6, 1 / 2, -1 / 6 or -1 / 4, 1 / 2, -3 / 4, 1, -1, 1, -1, 1, -1, 1, -1, 1, -3 / 4, 1 / 2, -1 / 4. The sum of the absolute values ​​of the weighted coefficient sequences is N=12. The duration of these 12 high-frequency half-cycles (80ms) and the duration of half excitation cycle (120ms) are both set to integer multiples of the power frequency interference duration of 20ms (power frequency noise frequency is 50Hz), which helps to reduce the impact of power frequency interference.

[0048] In the patent application filed by the same applicant (application number 202510782467.4), an electromagnetic flowmeter with alternating high and low frequency excitation has been disclosed, which uses a single-point trapezoidal weighted coefficient sequence: in one embodiment, the number of half cycles of the high frequency signal is 13, the absolute value of the first and last two weighted coefficient sequences is 0.5, and the rest are 1.

[0049] This patent uses different time-domain window functions: due to power frequency interference, the Hanning or Hamming window cannot be directly used to reduce the amplitude of the sidelobes, i.e., the low-frequency part of this patent. Instead, a window function with a 3-step weighted coefficient sequence is used: there are three additional steps for both the rising and falling frequencies. Different sequence lengths of the window function can be selected. The longer the sequence, the smaller the window truncation effect, but the update speed becomes slower, and the sampling interval for low frequencies also becomes longer; too long a sequence will affect accuracy. When the sequence length is selected, different sequences correspond to different bands of suppression capability.

[0050] Both the weighted coefficient sequences in this patent and the prior application belong to the first type of linear-phase FIR filter. The results of the theoretical analysis are shown in Figure 4(a). The comparison objects are the prior patents with w1, w2, and 7 points.

[0051] As can be seen, the amplitude-frequency characteristics of this patent are:

[0052] Both W1 and W2 have a higher suppression capability than prior patents throughout the entire process, and are at least no less than prior patents.

[0053] w1: At 30Hz, the suppression capability is improved by at least 10 dB (3 times) compared to the prior patent.

[0054] w2: At 25~50Hz, the suppression capability is improved by at least 10 dB (3 times) compared to the prior patent.

[0055] New patent features:

[0056] w1: At 30Hz, the attenuation is better than -39 dB (1 / 89).

[0057] w2: At 50Hz, the attenuation is better than -28 dB (1 / 25).

[0058] This patent reduces the settling time of low-frequency excitation by 13.333 ms by adding a 75Hz full cycle. However, due to the arrangement of the excitation direction in claim 2 (4), the end level direction of the high-frequency segment immediately preceding the low-frequency segment is consistent with the level direction of the low-frequency segment. Under this arrangement, the actual duration of the low-frequency segment will be four times the duration of the high-frequency half-cycle, thus still ensuring that the electrode signal fully enters the steady state.

[0059] The theoretical analysis results of w3 and w4 are shown in Figure 4(b). The longer sequence has better low-frequency suppression capability, the entire excitation cycle is 240ms, and the operating frequency is reduced from 6.25Hz to 4.167Hz.

[0060] The experimental verifications for w1 and w2 are shown in Figures 5(a) and (b), where a sine wave is directly applied to the ADC input. It can be seen that the experimental data match well below 56Hz, but there is some error at 60Hz and above. This is because these frequencies are close to the high-frequency cycle, making them prone to error. Even so, it is still clear that below 50Hz, the amplitude decreases significantly, especially in the low-frequency portion below 30Hz in Figure 5(a), where the amplitude is indeed very low.

[0061] Since this invention focuses on the ability to suppress low-frequency flow noise, as shown in Figures 5(a) and 5(b), the two figures clearly show that within 56Hz, the measured attenuation is in high agreement with the theoretical curve that is very easy to obtain, which fully verifies the core purpose of the invention; the high-frequency band has certain errors, but this does not affect the technical effect of the invention in the target application scenario.

[0062] Since the actual output signal is the result of the superposition of responses at different frequencies, the above sine wave frequency sweep method can be used to verify that the present invention does indeed have a strong suppression effect on 1 / f.

[0063] [Technical Principles of Weighted Coefficient Sequences]

[0064] To reduce the amplitude of low frequencies, conventional window functions such as the Hanning window cannot be used. Although the Hanning window can significantly reduce sidelobes and thus reduce the amplitude of low frequencies, it generally cannot simultaneously suppress power frequency frequencies.

[0065] The weighted coefficient sequence used in this patent can simultaneously achieve low-frequency noise suppression and power frequency interference elimination. Its core mechanism is jointly guaranteed by claims 4(A) to (E), as analyzed below:

[0066] (A) Alternating positive and negative polarity: It is strictly synchronized with the high-frequency excitation polarity to form a synchronous demodulation base, effectively extracting the fundamental wave of the flow signal and suppressing DC bias and even harmonics.

[0067] (B) (C) Multi-step symmetrical window: The absolute value sequence gradually decreases at both ends, is 1 in the middle, and is symmetrical about the center (containing at least 3 different values), which creates an effect of increasing the main lobe and reducing the side lobes in the frequency domain. Depending on the selection of the weighting coefficient sequence selector, it can significantly suppress flow noise in a certain range. For example, compared with trapezoidal weighting of only 0.5 at the beginning and end (as in the earlier application), this patent improves the attenuation capability of the frequency band below 30Hz by more than 10dB when using w1 (see Figure 4(a)).

[0068] (D)(E) Power Frequency Phase Folding: (D) ensures that the duration of N high-frequency half-cycles is an integer multiple of the power frequency cycle, so that the sampling points at index n and n+N are in the same power frequency phase; combined with the constraint w[n]+w[n+N]=±1 in (E), the paired sampling points can be algebraically merged into a unit weight during weighted accumulation. The entire sequence degenerates into N equivalent units with weights of ±1, and the equivalent accumulation window exactly covers an integer number of power frequency cycles, forming a transfer function depth zero at 50Hz (see Figures 4(a)(b) and 5(a)(b)).

[0069] The aforementioned low-frequency suppression and power frequency suppression are achieved through a single weighted calculation, requiring no additional filtering circuitry. Any weighted coefficient sequence conforming to claims (A) to (E), due to its symmetry, is a first-type linear-phase FIR filter when the length L of the weighted coefficient sequence is odd, and a fourth-type linear-phase FIR filter when the length L of the weighted coefficient sequence is even. Its frequency response can be conveniently calculated using DTFT to obtain the graphs in Figures 4(a) and (b).

[0070] An example of a weighted coefficient sequence with an even length L is [-0.25, 0.75, -1, 1, -1, 1, -0.75, 0.25]. However, it is preferable to use an odd length L, as this ensures that the weighted coefficient sequence is strictly symmetric about time, rather than just symmetric in its absolute value. This allows the output to remain essentially unchanged when the flowmeter signal fluctuates slowly, exhibiting a rapid rise or fall over a short period. The antisymmetric nature of the weighted coefficient sequence with an even length L makes it easier to capture the slope of the change, and therefore it is generally used for differential operations.

[0071] To verify the necessity of feature (E) of claim 4, this application provides a weighted coefficient sequence that satisfies only features (A) to (D) but not feature (E) for comparison:

[0072] wE = [0.25, -0.6, 0.65, -1, 1, -1, 0.65, -0.6, 0.25] (L=9)

[0073] Technical effect comparison:

[0074] Substituting wE into the discrete-time Fourier transform (DTFT) to calculate its amplitude-frequency response curve, the results show:

[0075] - In the low-frequency range of 0~30Hz, the attenuation capability of wE is similar to that of the w2 sequence that satisfies feature (E) (both are better than -28dB), verifying the effectiveness of features (A)~(D) in suppressing low-frequency noise;

[0076] - However, at the 50Hz power frequency, the attenuation of wE is only -18.3dB, while the w2 sequence forms a deep zero of >60dB at 50Hz due to the power frequency phase folding mechanism of characteristic (E) (see Figure 4(a)).

[0077] Therefore, the numerical complementary constraint w[n]+w[n+N]=±1 of feature (E) is an algebraic necessary condition for realizing the "power frequency synchronous integration window" to strictly generate the amplitude zero point in the frequency domain.

[0078] The signal calculation steps are as follows:

[0079] Step 1: Extract electrode signals.

[0080] The output frequency of the synchronous ADC is set to an integer multiple of the 75Hz excitation frequency, such as 30kHz or 15kHz, and the flowmeter performs equal-interval sampling. The average value of the data from the start of stabilization after an excitation change until the end before the next excitation change is taken as the electrode signal for that period. High-frequency electrode signals are extracted from the time-division high-frequency signals, and low-frequency electrode signals are extracted from the time-division low-frequency signals. For example, the range of high-frequency signal data is 1 / 2, 1 / 3, etc., of the 75Hz half-cycle, and the range of low-frequency signal data is 10ms, etc.

[0081] The high-frequency electrode signal and the low-frequency electrode signal are processed separately to obtain the corresponding electrode signal amplitude. Specifically, this includes obtaining the low-frequency electrode signal amplitude and obtaining the high-frequency electrode signal amplitude.

[0082] Step 2: Acquisition of low-frequency excitation electrode amplitude signal.

[0083] This part is the traditional calculation method for low-frequency square wave excitation electromagnetic flowmeters, which integrates electrode signals under positive and negative excitation.

[0084] The signals during the nth negative low-frequency excitation of the first low-frequency band and the positive low-frequency excitation of the second low-frequency band are respectively denoted as... , The synthesized electrode signal is denoted as .

[0085] when After obtaining it, calculate using the following formula. :

[0086]

[0087] when After obtaining it, use the following formula to calculate. :

[0088]

[0089] Step 3: Acquisition of high-frequency electrode signal amplitude.

[0090] Further, it can be divided into the following steps:

[0091] Step 3A: Calculate the high-frequency signal in the high-frequency band using the weighted coefficient sequence:

[0092] ,

[0093] in:

[0094] - The weighted coefficient sequence is w1, w2, w3, w4. When w1 and w2, i = 0, 1, …, 8. When w3 and w4, i = 0, 1, …, 14. The sampling point index is within the high-frequency half-cycle.

[0095] - These are the sampled values ​​of the high-frequency electrode signals acquired by the flow meter under high-frequency excitation.

[0096] - H represents the high-frequency electrode signal amplitude in the high-frequency band calculated according to the above weighted coefficient sequence, which is further distinguished according to the different excitation directions:

[0097] · The electrode signal amplitude calculated in the first high-frequency band, as explained earlier, is generally negative in the case of positive flow.

[0098] · The electrode signal amplitude calculated in the second high-frequency band, as explained earlier, is generally positive in the case of positive flow.

[0099] Step 3B:

[0100] When the amplitude of the high-frequency signal (usually negative) in the first high-frequency band... Amplitude of high-frequency signals (generally positive) in the second high-frequency band After obtaining it, it is synthesized using the same method as the low-frequency excitation signal.

[0101] When sampling the current n periods, obtain Recorded as The result of the previous period is recorded as When sampling the current n periods, obtain Recorded as The result of the previous period is recorded as .

[0102] when or After obtaining the signal, calculate the amplitude of the combined high-frequency electrode signal, and record it as follows: .

[0103] when After obtaining it, calculate according to the following formula. :

[0104]

[0105] when After obtaining it, calculate according to the following formula. :

[0106]

[0107] Step 4: Calculate the high-frequency zero-point drift.

[0108] For current high frequency signals With low frequency signals The same low-pass filter with a large time constant (e.g., 2-20 minutes) is used for filtering, and the filtered result is denoted as... , And calculate the difference. .

[0109] Step 5: Calculation of effective high-frequency electrode signals.

[0110] The effective electrode signal is calculated using the following formula. Eliminate high-frequency zero-point drift:

[0111]

[0112] The difference in the right half of the above equation The flow-related components were eliminated, leaving mainly the high-frequency zero-point drift. Since the low-frequency signal itself is highly stable and further noise is suppressed after filtering with the same parameters, and the high-frequency component is inherently insensitive to flow noise, the difference between the two mainly reflects the high-frequency zero-point drift component.

[0113] Final difference calculation This eliminates high-frequency zero-point drift components.

[0114] The embodiments described above are only for illustrating the meaning of this application and should not be construed as limiting the scope of the patent application. For example, the frequency of the high-frequency signal and / or the number of high-frequency half-cycles may differ from the example, or the high-frequency starting excitation may be positive.

[0115] In particular, for Figure 3 The waveform shown illustrates that the computational unit can employ a segmented calculation strategy: the 15 high-frequency half-cycles are divided into two 9-half-cycle windows (overlapping by 3 half-cycles), and weight sequences 1 / 4, -1 / 2, 3 / 4, -1, 1, -1, 3 / 4, -1 / 2, and 1 / 4 are applied for calculation respectively, followed by a synthesis of the results. This strategy can maintain noise immunity while reusing functions such as... Figure 2 The computing module of the illustrated embodiment improves the data update rate. This implementation, due to its additional three steps, still falls within the protection scope of claim 4.

[0116] Therefore, the scope of protection of this application shall be determined by the contents covered by the claims.

Claims

1. An electromagnetic flowmeter, comprising: Excitation device, used to output excitation signal; Electrodes are used to detect induced electromotive force and output electrode signals. The signal conditioning and acquisition device is used to amplify and analog filter the electrode signal and convert it into a digital signal, wherein the digital signal includes a high-frequency electrode signal and a low-frequency electrode signal. The MCU integrates an arithmetic unit and an output unit. It is configured to first output control signals to the excitation device according to the preset excitation timing to drive the excitation device to generate a magnetic field; and during the excitation period, it performs calculations on the digital signals output by the signal conditioning and acquisition device to realize signal calculation, weighted accumulation and zero drift elimination steps. The excitation device is characterized in that it is controlled by an MCU and configured to output alternating high-frequency excitation and low-frequency excitation, and a complete excitation cycle of the excitation signal includes a first high-frequency band, a first low-frequency band, a second high-frequency band, and a second low-frequency band arranged in sequence.

2. The electromagnetic flowmeter according to claim 1, characterized in that, The excitation device is further configured as follows: Within a complete excitation cycle, the excitation polarity relationship of each frequency band shall satisfy the following conditions: (1) The starting excitation direction of the second high-frequency band is opposite to the starting excitation direction of the first high-frequency band; (2) The excitation direction of the first low-frequency band is opposite to the excitation direction of the second low-frequency band; (3) The excitation direction of the first low-frequency band is the same as the initial excitation direction of the first high-frequency band; (4) The tail level direction of the first high frequency band and the second high frequency band is the same as the starting level direction of each band, and the duration of each high frequency band is an odd multiple of the high frequency excitation half cycle, wherein the high frequency excitation half cycle refers to the duration of the single polarity of the high frequency excitation signal.

3. The electromagnetic flowmeter according to claim 1, characterized in that, The arithmetic unit is configured as follows: The amplitude of the low-frequency electrode signal is calculated using the positive and negative excitation period difference method. For the high-frequency electrode signal, a weighted accumulation algorithm is first used to calculate the high-frequency signal of each high-frequency band, and then the amplitude of the high-frequency electrode signal is obtained by comprehensive calculation based on the positive and negative excitation period difference method.

4. The electromagnetic flowmeter according to claim 3, characterized in that, The computation unit pre-stores a weighted coefficient sequence, and the computation unit is further configured as follows: The sampled values ​​of the high-frequency electrode signal are weighted and accumulated using the weighted coefficient sequence, and the weighted coefficient sequence satisfies: (A) The coefficients alternate between positive and negative and are synchronized with the high-frequency excitation polarity to extract the flow signal; (B) The sequence after taking the absolute value contains at least 3 different values, forming a multi-step window function; (C) Symmetrical distribution: The coefficient with an absolute value of 1 is located in the middle of the sequence, and the coefficients of the other values ​​are symmetrical about the center of the sequence; (D) Power frequency synchronization: The sum of the absolute values ​​of the coefficients is an integer N, and the total duration of N high-frequency excitation half cycles is exactly an integer multiple of the power frequency signal period; (E) Numerical complementarity: For all sampling point indices n that satisfy n+N < L, we have w[n] + w[n+N] = 1 or -1, where L is the length of the weighted coefficient sequence, and n and n+N correspond to two sampling points that are N high-frequency excitation half-cycles apart.

5. The electromagnetic flowmeter according to claim 4, characterized in that, The arithmetic unit further includes a weighted coefficient sequence selector, which is configured as follows: Depending on the operating conditions, one of the pre-stored weighted coefficient sequences is selected to calculate the amplitude of the high-frequency electrode signal.

6. The electromagnetic flowmeter according to any one of claims 3 to 5, characterized in that, The arithmetic unit is further configured as follows: The amplitudes of the high-frequency electrode signal and the low-frequency electrode signal are respectively subjected to first-order low-pass filtering with the same parameters and a large time constant. The difference between the two filtering results is then used as the high-frequency zero-point drift.

7. The electromagnetic flowmeter according to claim 6, characterized in that, The arithmetic unit is further configured as follows: The effective high-frequency electrode signal is obtained by removing the high-frequency zero-point drift from the amplitude of the high-frequency electrode signal.

Images