Vibration sensor background noise test method based on vacuum cavity

Abstract

The invention discloses a vibration sensor background noise testing method based on a vacuum cavity, and the method comprises the steps: constructing an active-passive composite vibration isolation platform in a sealed vacuum cavity, rigidly fixing a to-be-tested sensor on the vibration isolation platform in a testing process, and guaranteeing that a sensitive axis of the to-be-tested sensor is orthogonal to a cavity reference; and an output signal is recorded for a long time at a sampling rate greater than or equal to 256Hz through a high-precision acquisition system. The data processing comprises polynomial detrending, sliding window de-averaging, multi-band elimination filtering and power spectral density (PSD) analysis so as to extract the equivalent amplitude and frequency domain characteristics of the background noise. In order to ensure the credibility of the result, the test needs to be repeated for three times, the consistency of the PSD curve is compared, and the deviation does not exceed 5%. The method combines the vacuum environment, composite vibration isolation, signal preprocessing and spectrum analysis technologies, can accurately quantify the background noise characteristics of the sensor, and is suitable for research and development verification and performance evaluation of the high-sensitivity sensor.

Description

Technical Field

[0001] The present invention belongs to the technical field of vibration sensor background noise testing, and more specifically, relates to a vibration sensor background noise testing method based on a vacuum chamber. Background Art

[0002] Vibration sensors are widely used in precision measurement, aerospace, earthquake monitoring, and other fields, playing a particularly important role in detecting minute vibrations and motion. The sensor's measurement precision and accuracy determine the performance of the error measurement system. Due to process imperfections, sensor measurement data is often accompanied by errors. These errors are primarily random and deterministic. These errors manifest themselves in the sensor's noise floor and sensitivity, or scale factor.

[0003] The output of a vibration sensor is primarily divided into two components: the effective output reflecting changes in the vibration excitation, and the randomly varying inherent current noise of the sensor, also known as the sensor's noise floor. Currently, existing testing methods mostly focus on sensitivity testing. Measuring the sensor's noise floor has stringent requirements for the measurement environment. Even slight external vibrations and even sounds can couple into the sensor's effective output, which is the excitation signal, affecting the proper measurement of the sensor's noise. Failure to perfectly eliminate the effects of external vibration interference will cause the sensor's noise floor to be overwhelmed by external interference, making it impossible to accurately measure the sensor's true noise floor signal.

[0004] Furthermore, most existing vibration sensor background noise tests simply use a vibration isolation base for basic isolation, but this method does not eliminate external environmental interference. To overcome the performance bottleneck of vibration sensors, a high-precision, highly anti-interference background noise detection system needs to be constructed. This system must integrate ultra-low noise acquisition and multi-dimensional environmental shielding technology to accurately separate the sensor's inherent noise from external coupling interference. The vacuum chamber-based testing method isolates air-conducted noise through a vacuum environment, and combines an active-passive composite vibration isolation mechanism to eliminate the influence of mechanical vibration, allowing for stable acquisition of the sensor's true background characteristics. Its core value lies in: through the synergistic effect of physical isolation and algorithmic noise reduction, microscale analysis of the noise spectral density (PSD) is achieved, providing a reliable data foundation for subsequent error compensation model establishment and sensitivity calibration, significantly improving the sensor's dynamic measurement accuracy and long-term stability. Summary of the Invention

[0005] The purpose of this invention is to provide a vacuum chamber-based vibration sensor background noise testing method that can effectively overcome the problems of environmental noise interference and insufficient background noise measurement accuracy, providing more accurate and stable testing and calibration support for vibration sensors. The method includes:

[0006] The S1 vacuum chamber and vibration isolation system are built, and an ultra-low frequency vibration isolation platform is installed at the bottom of the vacuum chamber to ensure the airtightness of the chamber. The vibration isolation platform adopts an active-passive composite vibration isolation design and is connected to the vacuum pump, signal acquisition system, and environmental monitoring module to provide a test environment for the sensor without external vibration interference.

[0007] S2 sensor installation and fixation: The vibration sensor to be measured is rigidly fixed on the vibration isolation platform in the vacuum chamber, ensuring that the sensor's sensitive axis is strictly orthogonal to the chamber reference plane. The sensor signal line is led out through the vacuum sealed interface and connected to the high-precision acquisition card outside the cavity;

[0008] S3 vacuum chamber sealing and vacuuming, seal the vacuum chamber and start the vacuum pump to reduce the pressure inside the chamber to 10 -3 Pa below, eliminating interference from air flow and sound wave conduction. Simultaneously monitor the temperature, humidity and residual vibration inside the cavity to ensure stable environmental parameters;

[0009] S4 background noise signal acquisition: Under vacuum and vibration isolation conditions, continuously collect sensor output signals at a sampling rate of ≥256Hz for ≥1 hour. During the acquisition process, turn off the extracavity power supply and electromagnetic equipment to avoid coupling of power frequency and electromagnetic noise.

[0010] S5 signal preprocessing and trend elimination: Perform polynomial least squares fitting on the original signal to eliminate the linear trend term caused by the gravity component or cavity tilt. Use the sliding window method to segmentally remove the mean to ensure the signal is zero-mean; use the polynomial least squares method to eliminate the trend term of the signal. According to the sampling frequency Fs set by the sensor, let the sampling time interval Δt = 1 / Fs, and set the polynomial function as:

[0011]

[0012] Pass-through We can get k+1 polynomial coefficients a i (i=0,1,…,k). Since the vibration isolation ground is not a plane that is perfectly perpendicular to the direction of gravity of the earth, the main factors that cause the signal to generate a trend term in the measurement experiment are not only the zero bias of the sensor itself, but also the inclination angle θ of the sensor's main axis to the standard horizontal direction, which causes the gravity component gsinθ to act on the inertial sensor and generate a linear trend term signal. Therefore, only the unknown coefficient a1 of the linear polynomial needs to be obtained, that is, the polynomial order k=1.

[0013] When k=1, the fitted polynomial function is:

[0014]

[0015] The signal after eliminating the gravity component trend is:

[0016]

[0017] To filter out S6 power frequency interference, a multi-band filter is designed to perform notch processing on 50 Hz and its harmonic frequency bands (100 Hz, 150 Hz, etc.). The filter amplitude-frequency response is optimized using the window function method to minimize power frequency interference; The design selects a multi-band filter to filter out the power frequency interference in the sensor's background noise.

[0018]

[0019] S7 power spectrum density analysis and noise quantization, the filtered signal is segmented and Hanning windowed, and the power spectrum density (PSD) is calculated by the average periodogram method. According to the PSD curve integration results, the equivalent magnitude and frequency domain distribution characteristics of the background noise are extracted; assuming that the sampling rate of the vibration sensor is Fs, according to the set resolution bandwidth is ωk, first use the formula Determine the signal length NFs of the fast Fourier transform, divide the noise signal x(n) into several segments of each length NFs, eliminate the trend term of each signal to obtain a preliminary purification of the signal, and select the hanning window, as shown in the formula As shown, the signal is windowed:

[0020]

[0021]

[0022] By formula The fast Fourier transform algorithm converts the time domain signal into the frequency domain signal X(k):

[0023]

[0024] x(n) is a time domain sequence of random noise; k=0,1,2,…,N-1; We can get:

[0025]

[0026] Finally, use the formula The square of the frequency domain signal amplitude is divided by the length of the signal to obtain an estimate of the power spectrum density function. The corresponding signal of each power spectrum density function is averaged to obtain an estimate of the power spectrum density function.

[0027]

[0028] S8 Result Verification and Repeatability Test: Repeat steps S4-S7 three times, comparing the consistency of the PSD curves. A deviation of ≤5% is considered valid. If out of tolerance, check the vacuum level, vibration isolation performance, and sensor installation status, and retest until the results are stable.

[0029] The present invention provides a method for testing the background noise of a vibration sensor based on a vacuum chamber, which has the following advantages:

[0030] (1) This system eliminates interference such as air flow and sound wave conduction through a 10^-3 Pa vacuum chamber, significantly reduces environmental noise coupling, and ensures the purity of the background noise signal.

[0031] (2) The active-passive vibration isolation platform of the present invention is combined with a vacuum cavity to suppress mechanical vibration and electromagnetic interference in multiple dimensions, and the test stability is improved by more than 60%.

[0032] (3) The present invention achieves accurate quantification of microscale noise levels based on power spectral density (PSD) analysis and average periodogram method.

[0033] (4) The present invention adopts polynomial fitting and sliding window method to adaptively eliminate the gravity component and the signal offset caused by cavity tilt, thereby improving data credibility.

[0034] (5) This device is compatible with multiple types of sensors and supports complex working conditions in laboratories and industrial sites. The calibration repeatability error is ≤5%, meeting the needs of high-precision measurement. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Attachment Figure 1 This is a schematic diagram of an installation device for a specific implementation example of the method for testing background noise of a vibration sensor based on a vacuum chamber according to the present invention;

[0036] Attachment Figure 2 This is a flow chart of a method for testing background noise of a vibration sensor based on a vacuum chamber;

[0037] Attachment Figure 3 Schematic diagram of the noise test effect under vacuum in the method of the present invention. DETAILED DESCRIPTION

[0038] To address the low precision and high environmental interference issues of existing calibration techniques, this paper proposes a sensitivity calibration method based on a vacuum chamber. This method isolates air flow and acoustic wave conduction noise through a vacuum chamber. Combined with an active-passive composite vibration isolation platform within the chamber, this method eliminates over 99% of mechanical vibration interference. Real-time synchronous data acquisition and a power spectral density algorithm are used to determine the noise floor, enabling micron-level vibration signal analysis. The calibration process supports dynamic calibration across the full frequency range (0.1-200 Hz).

[0039] refer to Figure 1 The present invention is a schematic diagram of an implementation example of the method of the present invention, comprising a vibration isolation base (1), a vacuum chamber base (2), a vacuum chamber (3), an active vibration isolation table (4), a vibration sensor (5), a data acquisition card (6), a host computer (7), and a vacuum pump (8). The vibration isolation base (1) fixes the vacuum chamber base (2), an active vibration isolation table (4) is installed in the vacuum chamber (3), and the vibration sensor (5) is rigidly fixed on the vibration isolation table, with the sensitive axis being orthogonal to the chamber reference. The vacuum pump (6) is started to reduce the chamber pressure to below 10-3 Pa, and the environmental stability is monitored. The external power supply is turned off, and the sensor output signal is continuously collected at a sampling rate of ≥256 Hz for a period of ≥1 hour through the data acquisition card (7). The host computer (8) performs polynomial fitting on the original signal to eliminate the trend term, and designs a multi-band stop filter to filter out 50 Hz power frequency interference. The power spectral density (PSD) is calculated by segmented Hanning windowing to extract the equivalent magnitude of the background noise. The test is repeated three times to verify the consistency of the PSD (deviation ≤ 5%). When the deviation is exceeded, the vacuum degree and vibration isolation performance are checked to ensure the reliability of the results. .

[0040] refer to Figure 2 The flowchart of the vibration sensor background noise test based on the vacuum chamber is shown in FIG. The present invention mainly includes the following steps:

[0041] S1: An ultra-low frequency vibration isolation platform is installed at the bottom of the vacuum chamber to ensure the chamber's tightness. The isolation platform uses an active-passive composite vibration isolation design and is connected to the vacuum pump, signal acquisition system, and environmental monitoring module to provide a test environment for the sensor free of external vibration interference.

[0042] S2: The vibration sensor to be measured is rigidly fixed to the vibration isolation platform inside the vacuum chamber, ensuring that the sensor's sensitive axis is strictly orthogonal to the chamber's reference plane. The sensor signal line is led out through a vacuum-sealed interface and connected to a high-precision data acquisition card outside the chamber.

[0043] S3: Seal the vacuum chamber and start the vacuum pump to reduce the pressure inside the chamber to below 10^-3 Pa to eliminate interference from air flow and sound wave conduction. Simultaneously monitor the temperature, humidity, and residual vibration inside the chamber to ensure stable environmental parameters.

[0044] S4: Continuously acquire sensor output signals at a sampling rate of ≥256 Hz for ≥1 hour under vacuum and vibration isolation conditions. Turn off the extracavity power supply and electromagnetic equipment during acquisition to avoid coupling of power frequency and electromagnetic noise.

[0045] S5: Perform polynomial least squares fitting on the original signal to eliminate the linear trend term caused by the gravity component or cavity tilt. Use the sliding window method to segmentally remove the mean to ensure zero mean of the signal.

[0046] S6: Design a multi-band rejection filter to notch the 50 Hz band and its harmonics (100 Hz, 150 Hz, etc.). Use a window function to optimize the filter's amplitude-frequency response to minimize power frequency interference.

[0047] S7: Divide the filtered signal into segments and apply a Hanning window. Calculate the power spectral density (PSD) using the average periodogram method. Extract the equivalent magnitude and frequency domain distribution of the background noise based on the integral results of the PSD curve.

[0048] S8: Repeat steps S4-S7 three times, comparing the consistency of the PSD curves. A deviation of ≤5% is considered valid. If the deviation is out of tolerance, check the vacuum level, vibration isolation performance, and sensor installation status, and retest until the results are stable.

[0049] The above description is a detailed description of an embodiment of the present invention and is not intended to limit the present invention in any form. Those skilled in the art may make a series of optimizations, improvements, and modifications based on the present invention. Therefore, the scope of protection of the present invention shall be defined by the appended claims.

Claims

1. A vibration sensor background noise testing method based on a vacuum chamber, characterized in that: The testing method comprises the following steps: S1: An ultra-low frequency vibration isolation platform is installed at the bottom of the vacuum chamber to ensure the chamber's tightness. The isolation platform adopts an active-passive composite vibration isolation design and is connected to the vacuum pump, signal acquisition system, and environmental monitoring module to provide a test environment for the vibration sensor free of external vibration interference. S2: The vibration sensor to be tested is rigidly fixed on the vibration isolation platform in the vacuum chamber, ensuring that the sensitive axis of the vibration sensor is strictly orthogonal to the reference plane of the chamber. The vibration sensor signal line is led out through the vacuum sealed interface and connected to the high-precision acquisition card outside the chamber. S3: Seal the vacuum chamber and start the vacuum pump to reduce the pressure inside the chamber to below 10^-3 Pa to eliminate interference from air flow and sound wave conduction; simultaneously monitor the temperature, humidity, and residual vibration inside the chamber to ensure stable environmental parameters; S4: Under vacuum and vibration isolation conditions, continuously collect sensor output signals at a sampling rate of ≥256 Hz for ≥1 hour. During the collection process, turn off the extracavity power supply and electromagnetic equipment to avoid coupling between power frequency and electromagnetic noise. S5: Perform polynomial least squares fitting on the original signal to eliminate the linear trend term caused by the gravity component or cavity tilt; use the sliding window method to segmentally remove the mean to ensure zero mean of the signal; S6: Design multi-band rejection filters to notch the 50 Hz frequency band and its harmonics. Use the window function method to optimize the filter amplitude-frequency response and suppress power frequency interference to the maximum extent; S7: The filtered signal is segmented and Hanning windowed, and the power spectral density (PSD) is calculated using the average periodogram method. According to the PSD curve integration results, the equivalent magnitude and frequency domain distribution characteristics of the background noise are extracted; S8: Repeat steps S4-S7 three times, compare the consistency of each PSD curve, and consider it valid if the deviation is ≤5%. If it exceeds the tolerance, check the vacuum degree, vibration isolation performance and sensor installation status, and retest until the results are stable.

2. The method for testing background noise of a vibration sensor based on a vacuum chamber according to claim 1, characterized in that: The specific method of S5 is: The trend term of the signal is eliminated by using the polynomial least squares method. According to the sampling frequency Fs set by the vibration sensor, the sampling time interval Δt=1 / Fs is set, and the polynomial function is set as: Pass-through Get k+1 polynomial coefficients a i , i=0,1,…,k; in addition, we need to obtain the unknown coefficient a1 of the first-order polynomial, that is, the polynomial order k=1; When k=1, the fitted polynomial function is: The signal after eliminating the gravity component trend is: 。 3. The method for testing background noise of a vibration sensor based on a vacuum chamber according to claim 2, characterized in that: The specific method of S6 is: Utilization The design selects a multi-band filter to filter out the power frequency interference in the background noise of the vibration sensor; 。 4. The method for testing background noise of a vibration sensor based on a vacuum chamber according to claim 1, characterized in that The specific method involved in S24 is: Assuming that the sampling rate of the vibration sensor is Fs, according to the set resolution bandwidth is ωk, first use the formula Determine the signal length NFs of the fast Fourier transform, divide the noise signal x(n) into several segments of each length NFs, eliminate the trend term of each signal to obtain a preliminary purification of the signal, and select the hanning window, as shown in the formula As shown, the signal is windowed: By formula The fast Fourier transform algorithm converts the time domain signal into the frequency domain signal X(k): x(n) is a time domain sequence of random noise; k=0,1,2,…,N-1; have to: Finally, use the formula Divide the square of the frequency domain signal amplitude by the length of the signal to obtain an estimate of the power spectrum density function, and average the corresponding signals of each power spectrum density function to obtain an estimate of the power spectrum density function; 。 5. The method for testing background noise of a vibration sensor based on a vacuum chamber according to claim 1, characterized in that: The device for implementing the test method comprises: a vibration isolation base (1), a vacuum chamber base (2), a vacuum chamber (3), an active vibration isolation table (4), a vibration sensor (5), a data acquisition card (6), a host computer (7), and a vacuum pump (8); the vibration isolation base (1) fixes the vacuum chamber base (2), the active vibration isolation table (4) is installed in the vacuum chamber (3), the vibration sensor (5) is rigidly fixed on the active vibration isolation table (4), and the sensitive axis is orthogonal to the chamber reference; the vacuum pump (8) is started to reduce the chamber pressure to 10 -3 Pa, monitor the environmental stability; turn off the external power supply, and continuously collect the sensor output signal at a sampling rate of ≥256 Hz through the data acquisition card (6) for ≥1 hour; the host computer (7) performs polynomial fitting on the original signal to eliminate the trend term, and designs a multi-band filter to filter out the 50 Hz power frequency interference; calculate the power spectral density PSD by segmented Hanning window, and extract the equivalent magnitude of the background noise; repeat the test three times to verify the consistency of PSD, and check the vacuum degree and vibration isolation performance when it exceeds the tolerance to ensure the reliability of the results.

Images