A full-enclosed compressor rotating speed measurement method based on multi-signal processing technology fusion

Abstract

This invention relates to a method for measuring the rotational speed of a fully enclosed compressor using a multi-signal processing technique, comprising the following steps: establishing a current signal speed measurement model; acquiring the current signal, which originates from the current signal of the fully enclosed compressor; preprocessing the current signal, which involves first removing the DC component of the signal, then passing it through a third-order low-pass Butterworth filter, followed by extracting the rotational speed signal using a HILBERT transform; setting a refined frequency band interval, then performing wavelet transform for spectral refinement, and then performing spectral correction using a ratio correction method; then using a Kalman filter to optimally estimate the rotational frequency of the fully enclosed compressor after correction, in order to smooth the rotational frequency of the fully enclosed compressor, and finally obtaining the rotational speed value and displaying it on a host computer; using the above method, the problems of low measurement accuracy and high computational resource consumption of fully enclosed compressor rotational speed can be solved, thereby improving measurement accuracy and computational speed.

Description

Technical Field

[0001] This invention relates to the field of measurement technology for the rotational speed of a fully enclosed compressor, and more specifically to a method for measuring the rotational speed of a fully enclosed compressor that integrates multiple signal processing technologies. Background Technology

[0002] In the field of measuring the speed of fully enclosed compressors, due to the inherent enclosed nature of these compressors, their speed can only be measured indirectly by installing external sensors. Common methods for measuring the speed of fully enclosed compressors include vibration measurement, exhaust pulsation measurement, magnetic leakage measurement, and current measurement. Although these methods use different principles, they are all based on the speed characteristics of the signal, obtaining a weak signal through an acquisition system and then processing it. The current measurement method is currently the most commonly used method for measuring the speed of fully enclosed compressors, as it offers high accuracy and strong anti-interference capabilities. While extracting the speed signal based on the current measurement method is relatively simple, and the speed signal is obtained through a spectrum analysis algorithm, different spectrum analysis algorithms have different performance characteristics depending on their application in different engineering fields. Choosing an unsuitable spectrum analysis algorithm can lead to problems. Existing methods, such as the ZFFT complex modulation spectrum refinement method, suffer from low accuracy and high computational cost. The ZFFT method, while requiring significant computational resources, also suffers from limitations in refinement accuracy and maximum refinement factor due to low-pass filter characteristics. The CZT spectrum refinement method cannot eliminate interference for signals with overlapping spectral peaks, and its computation speed is slower than ZFFT when achieving the same effect. Currently, most fully enclosed compressor speed measurements utilize embedded speed measurement devices. Existing methods often require storing intermediate data, which, if computationally expensive, negatively impacts real-time speed measurement performance. Therefore, a high-precision, fast-speed signal processing algorithm is needed to measure high-real-time speed.

[0003] Therefore, existing technologies require a method that improves the accuracy of speed measurement and has a faster calculation speed to solve the problem of achieving high-precision speed measurement of fully enclosed compressors. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention proposes a method for measuring the rotational speed of a fully enclosed compressor that integrates multiple signal processing techniques. This method can effectively solve the problem of low accuracy in measuring the rotational speed of fully enclosed compressors, and provides new ideas for subsequent research on spectrum refinement algorithms and engineering applications of fully enclosed compressor rotational speed measurement.

[0005] To address the aforementioned issues, this application proposes a method for measuring the rotational speed of a fully enclosed compressor that integrates multiple signal processing techniques, comprising the following steps:

[0006] 1) Establish a speed measurement model based on current signals;

[0007] 2) The current signal is acquired using a DSP data acquisition system;

[0008] 3) Preprocessing of current signals;

[0009] 4) Wavelet transform spectrum refinement of current signals;

[0010] 5) Spectral correction to refine the spectrum;

[0011] 6) Kalman filter estimates the rotational frequency of the fully enclosed compressor;

[0012] 7) Rotation speed calculation and display.

[0013] Prioritizing the establishment of a speed measurement model based on current signals, the specific steps are as follows: In the motor of a fully enclosed compressor, defective suspension ball bearings or improper installation can lead to rotor eccentricity. Under the influence of rotor eccentricity and rotor slotting, the air gap magnetic permeability changes. When the air gap magnetic permeability of the motor changes, it leads to uneven magnetic field distribution, thereby causing the generation of eccentric harmonics. During operation, the motor of the fully enclosed compressor also generates eccentric harmonics due to the unbalanced forces such as radial force, centrifugal force, and load exerted on the rotor by the moving mechanism. In this invention, the eccentric harmonics are formed by the superposition of the above two causes. In practical applications, most of the industrial frequency AC power used for equipment power supply is generated by inverter circuits. In addition to generating AC power, inverter circuits also generate a weak DC component; therefore, the actual current signal mainly includes the fundamental current, eccentric harmonics, current harmonics, and the DC component. Therefore, the speed measurement model based on current signals is established as follows:

[0014]

[0015] Prioritizing the acquisition of the fully enclosed compressor current signal using a DSP data acquisition system, the current signal of the fully enclosed compressor is input into the data acquisition system via a current sensor WBI411S97. The data acquisition system is a data acquisition circuit board with F28335 as the core and AD7610 chip as the sampling chip. F28335 sends the acquired raw signal to the host computer via 485 communication.

[0016] Pre-processing of the current signal is specifically as follows: The host computer uses the MATLAB API to call the m-scripts in the MATLAB programming environment for data processing. If measuring the speed of a fixed-frequency fully enclosed compressor, the original signal is first stripped of its DC component, then filtered through a third-order low-pass Butterworth filter to obtain the current signal. The current signal is then subjected to a HILBERT transform, and the square of the original current signal is added to the square of the HILBERT-processed current signal. The resulting signal is the speed signal. If measuring the speed of a variable-frequency fully enclosed compressor, the current signal does not require pre-processing after acquisition.

[0017] Prioritizes wavelet transform spectral refinement for spectral analysis of the rotational speed signal; its algorithm flowchart is as follows. Figure 2 As shown, if measuring the speed of a fixed-frequency fully enclosed compressor, taking a motor with 1 pole pair and a power supply fundamental frequency of 50Hz as an example, the refined frequency band can be determined to be 45-55Hz. A scaling factor of 0.15 is set, and the Morlet wavelet function is selected to perform wavelet transform spectral refinement on the speed signal. After spectral refinement, the refined spectrum Y(k) and the refined time-domain speed signal y(n) are obtained. If measuring the speed of a variable-frequency fully enclosed compressor, FFT is first used to roughly calculate the frequency band of the compressor's rotational frequency. Then, the refined frequency band, scaling factor, and wavelet function are set to perform wavelet transform spectral refinement on the speed signal.

[0018] Prior to this, the spectrum correction of the refined spectrum uses the ratio correction method of the Hanning window. After applying the ratio correction method with the Hanning window to the refined time-domain signal y(n) obtained from the refined spectrum, the accurate rotational frequency f of the fully enclosed compressor is obtained. m .

[0019] Prioritize estimating the rotational frequency of the fully enclosed compressor using the Kalman filter by setting the initial values ​​of the Kalman filter as follows: e MEAk =0.5e EST0 =0.2, and then the fully enclosed compressor rotational frequency f is obtained after processing multiple sets of data signals. m After passing through a Kalman filter, the rotational frequency of the fully enclosed compressor is estimated using the Kalman filter.

[0020] Prioritize the calculation and display of the rotational speed value as follows: from the rotational speed formula The rotational frequency of the fully enclosed compressor is estimated using a Kalman filter to calculate the rotational speed, which is then displayed on the host computer interface.

[0021] The beneficial effects of this invention are mainly reflected in:

[0022] 1) The present invention uses the removal of DC component in the preprocessing of current signal, which is beneficial for obtaining high signal-to-noise ratio speed signal by using HILBERT in the subsequent process.

[0023] 2) This invention uses wavelet transform spectrum to refine the rotation speed signal, utilizes wavelet function to design filter with good filtering characteristics, and the selectivity of wavelet function and scale parameter facilitates practical engineering application.

[0024] 3) This invention uses the Kalman filter optimal estimation algorithm to predict and smooth the rotational frequency of the fully enclosed compressor, reducing data fluctuations caused by measurement errors and process errors.

[0025] 4) This invention uses a multi-signal processing technology fusion method, with a rotational speed error within ±0.5 rpm and a measurement cycle of 1 s. Attached Figure Description

[0026] Figure 1 This is a flowchart of the measurement process for the fully enclosed compressor speed measurement method based on wavelet transform spectrum refinement according to the present invention.

[0027] Figure 2 This is a flowchart of the wavelet transform spectrum refinement algorithm;

[0028] Figure 3 This is a refined spectrum diagram after wavelet transform spectrum refinement in an embodiment of the present invention;

[0029] Figure 4 This is a measurement result diagram of the fully enclosed compressor speed measurement using wavelet transform spectrum refinement in an embodiment of the present invention. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0031] This invention presents a method for measuring the speed of a fully enclosed compressor that integrates multiple signal processing techniques, applicable to the measurement of the speed of a fully enclosed compressor. (Reference) Figure 1 As shown, this invention provides a method for measuring the rotational speed of a fully enclosed compressor that integrates multiple signal processing technologies, comprising the following steps:

[0032] Step 1: Establishing a speed measurement model based on current signals: The motor of the fully enclosed compressor is a single-phase squirrel-cage induction motor. In this motor, defective suspension ball bearings or improper installation can cause rotor eccentricity. Under the influence of rotor eccentricity and rotor slotting, the air gap magnetic permeability changes. When the air gap magnetic permeability of the motor changes, it leads to uneven magnetic field distribution, thereby causing the generation of eccentric harmonics. During operation, the motor of the fully enclosed compressor also generates eccentric harmonics due to the unbalanced forces such as radial force, centrifugal force, and load on the rotor. In this invention, the eccentric harmonics are formed by the superposition of the above two causes. Eccentric harmonics are typical speed-related components in the current of an induction motor. In practical applications, most of the AC power used for equipment power supply is generated by inverter circuits, which, in addition to generating AC power, also produce a weak DC component. Therefore, the actual current signal mainly includes the fundamental current, eccentric harmonics, current harmonics, and the DC component.

[0033] Because the magnetic field and electric potential generated by the operation of the motor change periodically, the current signal of a single-phase squirrel-cage induction motor is usually periodic. According to the Fourier expansion, any periodic function can be expanded into an infinite series. Therefore, expanding the original signal x1(t) into an infinite series yields:

[0034]

[0035] in

[0036] Since actual current signals mainly include the fundamental current wave, eccentric harmonics, current harmonics, and DC components, the velocity measurement model x1(t) for the current signal is specifically as follows:

[0037]

[0038] Step 2: Acquire signals using a DSP data acquisition system: The signal from the fully enclosed compressor is transmitted to the data acquisition circuit board using a WBI411S97 current sensor. The data acquisition circuit board is a data acquisition system based on the F28335 and the AD7610 chip as the sampling chip. The sampling frequency is 1000Hz. When the compressor is working stably, it continuously acquires data for 60 seconds, that is, continuously acquires 60,000 data points. The original signal x1(t) of the continuous sequence is sampled to obtain the original signal x1(n) of the discrete sequence, which is sent to the host computer via 485 communication. The host computer receives and saves the original signal x1(n).

[0039] Step 3: Preprocessing of the current signal: Divide the 60s data into 6 parts, each part being 10s of data. Using the same signal processing methods, a fully enclosed compressor rotational frequency value can be obtained. In fact, processing data in 10s increments can also achieve the requirement of obtaining a fully enclosed compressor rotational frequency per second. For example, if the data collected in the first 10s yields a rotational speed, the data collected in the previous 9s can be added to the data collected in the current second to obtain the current rotational speed value. Under stable motor operation, the rotational speed value will fluctuate within a certain range. Therefore, this example uses a 10s data point for analysis. The host computer uses MATLAB's external API to call m-scripts within the MATLAB programming environment for data processing.

[0040] If the rotational speed of a fixed-frequency, fully enclosed compressor is measured, the DC component of the original signal x1(n) is first removed. In the current signal, the amplitude of the current harmonics is relatively large. Since the sampling frequency is 1000Hz, according to Shannon's sampling theorem, the highest-order current harmonic analyzed is the 9th harmonic. Therefore, the current signal x2(n) is obtained after passing through a third-order low-pass Butterworth filter with a cutoff frequency of 150Hz. The Butterworth filter is characterized by having the flattest amplitude-frequency characteristic within the passband, which decreases smoothly and monotonically with increasing frequency; the higher the order, the closer the characteristic is to a rectangle, and the narrower the transition bandwidth; its transfer function is shown below:

[0041]

[0042] Next, the HILBERT transform is applied to the current signal x2(n) to extract the rotational speed signal. Specifically, the principle of using the HILBERT transform to extract the rotational speed signal from the current signal is as follows: The current signal mainly consists of two parts: the fundamental frequency of the power supply and the eccentric harmonics. The amplitudes of both parts are significantly greater than the noise. The fundamental frequency of the power supply is a high-frequency sine wave, while the eccentric harmonics are low-frequency sine waves. If spectral analysis is used directly, it is difficult to guarantee the accuracy of the spectral analysis. Therefore, the HILBERT transform is used to process the current signal. The specific steps are as follows:

[0043] Assume the specific expression for the current signal x2(n):

[0044] x2(n)=a1·sin(2πf1n)+a2·sin(2πf2n)

[0045] The current signal x2(n) is processed by HILBERT:

[0046]

[0047] Then the signal after HILBERT transform processing Add the square of x2(n) to the square of the signal before the HILBERT transform:

[0048]

[0049] At this time, the integrated signal x3(n) is the speed signal, and its frequency is the rotational frequency of the fully enclosed compressor.

[0050] If the speed of a variable frequency fully enclosed compressor is measured, the current signal does not need to be preprocessed after acquisition.

[0051] Step 4: Wavelet Transform Spectrum Refinement: In speed measurement, certain frequency components are often of interest. Therefore, the bandpass filtering characteristic of wavelet transform can be used to obtain a sufficiently narrow band signal, thereby measuring the rotational frequency of the fully enclosed compressor. However, obtaining the true spectrum of a certain frequency band is difficult with a single wavelet transform. Consider using a combination of multiple wavelets, which can achieve the bandpass filtering characteristic. If measuring the speed of a fixed-frequency fully enclosed compressor, given the number of motor pole pairs and the fundamental frequency of the power supply, the refinement band can be directly determined. Then, the scaling factor and wavelet function are set to refine the speed signal using wavelet transform. If measuring the speed of a variable-frequency fully enclosed compressor, first use FFT to roughly calculate the frequency band of the fully enclosed compressor to be analyzed. Then, set the refinement band, scaling factor, and wavelet function to refine the speed signal using wavelet transform.

[0052] This invention uses Morlet wavelets to construct combined wavelets for spectral refinement; the rotational speed signal obtained after step 2 is subjected to wavelet transform spectral refinement to obtain the refined spectrum of the rotational speed signal; spectral analysis of the rotational speed signal is performed through wavelet transform spectral refinement, and the algorithm flowchart is as follows. Figure 2 As shown.

[0053] The specific steps for wavelet transform spectrum refinement are as follows:

[0054] In step 4.1 of the wavelet transform spectrum refinement process, the frequency band f to be refined is first determined. L ~f H Then set the scale factor a;

[0055] Step 4.2 Select a suitable wavelet ψ a (n) Constructing a combined wavelet filter:

[0056] Step 4.3 After wavelet transform filtering, the frequency shift is calculated, and the frequency shift factor is...

[0057] Step 4.4 Resampling, where

[0058] Step 4.5 performs an FFT transform on the resampled signal to obtain the desired frequency band f. L ~f H The refined spectrum Y1(k) was refined by a factor of D;

[0059] By adjusting parameter a and different types of wavelets ψ a (n) can achieve the desired spectral refinement effect.

[0060] Preferred definition:

[0061] Where, ψ a (n) represents the wavelet, ψ c W(n) represents the combined wavelet, W(n) represents the filtered signal, W1(n) represents the frequency-shifted signal, y(n) represents the resampled signal, D represents the amplification factor, Δf represents the refined resolution, Y1(k) represents the refined spectrum, x3(n) represents the rotational speed signal, and n represents the discrete time unit.

[0062] Step 5: Spectrum Correction: The ratio correction method calculates the frequency correction amount by the ratio of the amplitudes of two adjacent spectral lines within the main lobe of the window function spectrum, thereby obtaining the frequency of the main lobe center, i.e., the true peak frequency of the signal. In the absence of noise, the ratio correction method with a Hanning window has very high accuracy, with a frequency error of less than 0.0001 frequency resolutions. To obtain a more accurate rotational frequency of the fully enclosed compressor, this invention uses the ratio formula correction method with a Hanning window. After ratio correction, the refined spectrum is obtained to arrive at the precise rotational frequency of the fully enclosed compressor.

[0063] The specific steps of the ratio correction method are as follows:

[0064] Hanning window main lobe function:

[0065] The spectrum of the signal to be analyzed, y(n) obtained by wavelet transform spectrum refinement, is multiplied in the time domain with h1(n) and then subjected to FFT to obtain the spectrum diagram: Y(k) = FFT(y(n)·h1(n))

[0066] Find two spectral lines near the frequency of maximum amplitude in Y(k). Using the Hanning window, calculate the barycenter coordinates using these two lines within the main lobe of the amplitude peak. Subtract k from the barycenter coordinates to obtain the correction value Δk.

[0067] After sorting, we can obtain:

[0068]

[0069] The corrected rotational frequency f of the fully enclosed compressor m The result is:

[0070]

[0071] Step 6: Kalman Filter Optimal Estimation: Due to the standard regulations for industrial speed measurement, the speed error cannot exceed 1%, 3000*0.01 / 60 = 0.5, which translates to a frequency error of no more than 0.5. In practice, using a Hall sensor to measure the actual speed of the fully enclosed compressor, the speed error is within ±5 rpm, 5 / 60 = 0.0833, which translates to a frequency error of no more than 0.0833. Therefore, the measurement error can be set to 0.08. The fully enclosed compressor motor has 1 pole pair, and the initial estimated value of the fully enclosed compressor rotational frequency is set to 50. Since the sampling frequency is 1000, there are 1000 data points per second. Therefore, there are 6 fully enclosed compressor rotational frequencies collected within 10 seconds, or one minute. Under stable motor operation, the speed value will have a fluctuation range. Due to process noise and measurement noise, there will be noise during the actual measurement process. This invention uses a Kalman filter to process the spectrally corrected rotational frequency of the fully enclosed compressor. After the rotational frequencies of the six fully enclosed compressors are processed by the Kalman filter, an estimated frequency value series is obtained. Whether it is a priori estimation or measurement estimation, it can make the posterior estimation closer to the true value. Finally, the rotational frequency of the fully enclosed compressor estimated by the Kalman filter is used.

[0072] The specific steps for optimal estimation of a Kalman filter are as follows:

[0073] Step 6.1 Initial settings for the Kalman filter: e MEAk =0.08e EST0 =0.2;

[0074] Step 6.2 Calculate the Kalman Gain:

[0075]

[0076] Step 6.3 Calculate the k-th estimated true data:

[0077]

[0078] Step 6.4 Update the k-th estimation error:

[0079] e ESTk =(1-k) k )e ESTk-1

[0080] Step 7: Speed ​​value display: Calculate the speed by using the speed calculation formula from the rotational frequency of the fully enclosed compressor estimated by the Kalman filter, record the data, and display the speed value on the host computer.

[0081] The results show that this method can effectively measure the speed value in the current signal of a fully enclosed compressor, with a speed error within ±0.5 rpm.

[0082] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for measuring the rotational speed of a fully enclosed compressor using a fusion of multiple signal processing techniques, characterized in that, Includes the following steps: 1) Establish a speed measurement model based on current signals: in Indicates the amplitude of the fundamental current wave. This indicates the amplitude of the eccentric harmonic. The frequency of eccentric harmonics, Indicates the frequency of the fundamental current wave. This represents the amplitude of the nth harmonic of the current. express The frequency of the secondary current harmonic Represents the DC component. Indicates a unit of continuous time. With sampling frequency The value is related to a finite value. 2) A DSP data acquisition system is used to acquire current signals; 3) Preprocessing of current signals: If the speed of a fixed-frequency fully enclosed compressor is measured, the DC component of the original signal is first removed, and then filtered through a third-order low-pass Butterworth filter to obtain the current signal. Then, the current signal is subjected to HILBERT transformation, and the square of the current signal before HILBERT transformation is added to the square of the current signal after HILBERT transformation. The signal obtained at this time is the speed signal. If the speed of a variable-frequency fully enclosed compressor is measured, the current signal does not need to be preprocessed after acquisition. 4) Wavelet transform spectrum refinement of current signals: To perform spectral analysis on the speed signal, if the speed of a fixed-frequency, fully enclosed compressor is measured, the refined frequency band is determined, a scaling factor and wavelet function are set, and wavelet transform is applied to refine the speed signal spectrum. The refined spectrum is then obtained. and the refined rotational speed time domain signal If measuring the rotational speed of a variable-frequency fully enclosed compressor, first use FFT to roughly calculate the frequency band of the rotational frequency of the fully enclosed compressor to be analyzed, and then set the refinement frequency band, set the scaling factor and wavelet function to refine the frequency spectrum of the rotational speed signal using wavelet transform. 5) Spectral correction to refine the spectrum: The ratio correction method using the Hanning window was employed to refine the time-domain signal of the refined spectrum. After applying the ratio correction method with a Hanning window, the precise rotational frequency of the fully enclosed compressor is obtained. ; 6) Kalman filter estimation of the rotational frequency of the fully enclosed compressor: Set the initial value of the Kalman filter, and then process multiple sets of data signals to obtain the rotational frequency of the fully enclosed compressor. After passing through the Kalman filter, the optimal estimate of the fully enclosed compressor rotational frequency is obtained. 7) Speed ​​value calculation and display: From the speed formula The rotational frequency of the fully enclosed compressor is estimated using a Kalman filter to calculate the rotational speed, and the rotational speed is then displayed on the host computer interface. Indicates the power supply frequency. This indicates the number of magnetic pole pairs.

2. The method for measuring the rotational speed of a fully enclosed compressor using multi-signal processing technology fusion as described in claim 1, characterized in that, The specific steps for wavelet transform spectrum refinement are as follows: 1) In the wavelet transform spectrum refinement process, the frequency band to be refined is first determined. Then set the scale factor. ; 2) Select an appropriate wavelet Constructing a combined wavelet filter: ; Indicates combined wavelets, Represents wavelets, This represents the superposition of the spectrum of each wavelet between the upper and lower limits of the passband of the combined wavelet; 3) After wavelet transform filtering, the frequency shift is, and the frequency shift factor is... ; 4) Resampling, including periodic impulse trains ; Discrete-time unit impulse function Represents a discrete time unit. Indicates the first The starting point of each cycle, Indicates the refinement factor; 5) Perform an FFT transform on the resampled signal to obtain the desired frequency band. The refined spectrum was reduced by a factor of D. ; By adjusting the parameters and different types of wavelets This achieves the desired spectral refinement effect.

3. The method for measuring the rotational speed of a fully enclosed compressor using multi-signal processing technology fusion as described in claim 1, characterized in that, The specific steps of the ratio correction method are as follows: Hanning window main lobe function: ; The signal to be analyzed is obtained by refining the wavelet transform spectrum. and The spectrum is obtained by multiplying in the time domain and then performing an FFT: ; exist Find two spectral lines near the frequency of maximum amplitude. Using the Hanning window, calculate the barycenter coordinates using the two spectral lines within the main lobe of the amplitude peak. Subtract k from the barycenter coordinates to obtain the correction value. ; After sorting, we can obtain: ； Corrected rotational frequency of the fully enclosed compressor The result is: 。 4. The method for measuring the rotational speed of a fully enclosed compressor using multi-signal processing technology fusion as described in claim 1, characterized in that, The specific steps for optimal estimation of the Kalman filter are as follows: 1) Initial settings for the Kalman filter; 2) Calculate Kalman Gain: ； in This represents the estimation error for the (k-1)th iteration. This represents the measurement error of the kth measurement. 3) Calculate the k-th estimated true data: ； 4) Update the error of the k-th estimation: 。

Images