Air compressor rotor testing method and testing device

Abstract

The invention discloses an air compressor rotor testing method and device, and relates to the technical field of air compressor rotor testing. Vibration signal data of a rotor at multiple rotating speeds are collected, a fast Fourier transform algorithm is adopted for analysis, frequency and phase distribution is extracted, weighted analysis is carried out on the phase distribution, and the frequency and the phase distribution are extracted; and extracting the coordinates of the initial position with uneven mass distribution to realize accurate positioning of the position with unbalanced mass of the rotor. Fitting the offset by using a least square method to obtain a mass deviation coordinate, generating an initial counterweight adjustment scheme, optimizing initial counterweight parameters by combining a gradient descent algorithm to obtain an optimized counterweight adjustment scheme, and re-iterating the counterweight parameters of the optimized counterweight adjustment scheme by using a particle swarm optimization algorithm to obtain a target counterweight scheme. And the vibration signal of the rotor meets the vibration balance standard. According to the invention, the dynamic balance test efficiency and precision of the air compressor rotor are improved, and the operation stability of rotary mechanical equipment is ensured.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of air compressor rotor testing, and in particular to an air compressor rotor testing method and testing device. BACKGROUND

[0002] As a core component of the compressor, the dynamic balance of the air compressor rotor directly affects the stability and service life of the equipment, and is an important research direction in the field of industrial manufacturing. If the rotor has uneven mass distribution during high-speed rotation, it will cause eccentric vibration, leading to accelerated equipment wear and even failure, which seriously affects production efficiency and safety. Therefore, developing a high-precision rotor dynamic balance testing method and device is of key significance to improving the performance of air compressors.

[0003] Current testing methods mainly rely on traditional vibration measurement and manual weight adjustment, and generally have the problems of low efficiency and insufficient precision. The traditional method often needs to be adjusted multiple times, the test cycle is long, and the adaptability to complex rotor structures is poor. Especially in the face of diversified rotational speed and load conditions, it is difficult to quickly capture small vibration signals and accurately determine the position of mass deviation. In addition, the existing technology lacks real-time and automation capabilities when dealing with nonlinear vibration responses, resulting in a reliance on manual experience during the adjustment process, which makes it difficult to meet the high-precision industrial demand.

[0004] The core challenge of dynamic balance testing is how to accurately measure and analyze the vibration characteristics of the rotor at different speeds, and to achieve rapid and automatic adjustment of the weight. Vibration characteristic measurement requires capturing small signals through sensors, but the vibration signal is affected by factors such as rotational speed, geometric shape, and support stiffness when the rotor is rotating at high speed, showing complex nonlinear characteristics, making it difficult to directly determine the specific position of uneven mass distribution. For example, in a certain air compressor rotor test, the sensor may detect significant vibrations, but due to inaccurate frequency and phase analysis, repeated trials are often required when adjusting the weight, which is time-consuming and ineffective. More complex is that adjusting the weight position and mass requires comprehensive consideration of the dynamic response of the rotor and the actual working conditions, and the traditional method is difficult to calculate accurate weight values in real time, resulting in low testing efficiency.

[0005] Therefore, how to develop a self-adaptive correction algorithm based on real-time vibration frequency and phase analysis to automatically calculate and adjust the weight parameters has become a key problem to improve the efficiency and precision of air compressor rotor dynamic balance testing. SUMMARY

[0006] The present application provides an air compressor rotor testing method and testing device to realize real-time vibration frequency and phase analysis-based automatic calculation and adjustment of weight parameters, and to improve the efficiency and precision of air compressor rotor dynamic balance testing.

[0007] The present application provides an air compressor rotor testing method executed by a computer, comprising: Obtaining vibration signal data of the rotor at multiple rotation speeds, analyzing the vibration signal data by using a fast Fourier transform algorithm to obtain vibration frequency and phase distribution; If the vibration frequency exceeds a preset frequency threshold, performing weighted analysis on the phase distribution to determine an initial position coordinate of the uneven mass distribution; Based on the initial position coordinate, fitting the offset of the rotor by using a least square method to obtain a mass deviation coordinate; Based on the mass deviation coordinate, generating an initial counterweight adjustment scheme, and optimizing an initial counterweight parameter of the initial counterweight adjustment scheme by using a gradient descent algorithm to obtain an optimized counterweight adjustment scheme; Based on the counterweight parameter of the optimized counterweight adjustment scheme, performing dynamic response simulation in a simulation environment to obtain a simulation vibration signal; If the simulation vibration signal exceeds a balance threshold, re-iterating and updating the counterweight parameter of the optimized counterweight adjustment scheme by using a particle swarm optimization algorithm to obtain a target counterweight scheme; The target counterweight scheme is used to be applied to the rotor for testing, so that the balance state of the rotor reaches a set standard.

[0008] The air compressor rotor testing method provided by the application further comprises: Applying the target counterweight scheme to the rotor by using an automatic execution mechanism to obtain real-time vibration feedback signals; If the vibration feedback signals contain residual vibration, performing secondary analysis on the vibration feedback signals to update the mass deviation coordinate and determine a feedback counterweight parameter; Based on the feedback counterweight parameter, continuously iteratively optimizing and real-time feedback adjusting to generate a dynamic balance counterweight scheme, wherein the working state of the rotor reaches a set standard when the dynamic balance counterweight scheme is applied to the rotor.

[0009] The air compressor rotor testing method provided by the application, if the vibration frequency exceeds a preset frequency threshold, performing weighted analysis on the phase distribution to determine an initial position coordinate of the uneven mass distribution, comprises: If the vibration frequency exceeds a preset frequency threshold, obtaining vibration signal data of the rotor, and calculating a vibration frequency value of the vibration signal data by using a fast Fourier transform algorithm to obtain frequency data; Based on the frequency value of the frequency data exceeding the preset frequency threshold, generating a phase distribution matrix based on the phase distribution; Based on the phase distribution matrix, calculating a weighted phase value of the phase distribution matrix by using a weighted analysis algorithm to determine a phase distribution feature; The phase distribution feature is compared with a preset mass distribution model, a mass distribution curve of the unbalanced mass of the rotor on a balance plane is fitted by using a least square method algorithm, and an uneven degree parameter is determined, wherein the uneven degree parameter is an evaluation index of the unbalanced state of the rotor. Based on the uneven degree parameter, a spatial deviation of the mass distribution is calculated, and the initial position coordinate is determined.

[0010] According to the air compressor rotor test method provided by the application, based on the initial position coordinate, the offset of the rotor is fitted by using the least square method, and the mass deviation coordinate is obtained, which includes: Based on the initial position coordinate and a preset coordinate system, the offset of the rotor relative to the central axis is calculated by using the Euclidean distance calculation method, and an offset data set is obtained; If the offset in the offset data set exceeds a preset offset threshold, the offset data set is denoised by using the median filtering method to obtain a smooth offset data set; Based on the smooth offset data set, the offset is fitted by using the least square method to obtain an eccentric vector parameter; Based on the direction and amplitude information in the eccentric vector parameter, the vector decomposition method is used to determine the mass deviation coordinate.

[0011] According to the air compressor rotor test method provided by the application, based on the mass deviation coordinate, an initial counterweight adjustment scheme is generated, the initial counterweight parameter of the initial counterweight adjustment scheme is optimized by using the gradient descent algorithm, and an optimized counterweight adjustment scheme is obtained, which includes: Based on the mass deviation coordinate, a deviation vector is calculated to obtain an initial mass deviation distribution; Based on the initial mass deviation distribution, the counterweight mass and the counterweight position are initialized, the initial counterweight adjustment scheme and the initial counterweight parameter corresponding to the initial counterweight adjustment scheme are determined; If the error analysis result of the initial counterweight parameter is greater than a preset error threshold, the counterweight mass and the counterweight position of the initial counterweight parameter are iteratively updated by using the gradient descent algorithm to obtain an optimized parameter; Based on the optimized parameter, the mass deviation distribution is recalculated to obtain the optimized counterweight adjustment scheme.

[0012] According to the air compressor rotor test method provided by the application, based on the counterweight parameter of the optimized counterweight adjustment scheme, dynamic response simulation is performed in a simulation environment to obtain a simulation vibration signal, which includes: The initial counterweight parameter is loaded in the simulation environment, and dynamic response simulation is performed to obtain vibration signal data; Based on the vibration signal data, a signal extraction technique is used to extract the vibration characteristics thereof, and vibration characteristic parameters are obtained; If the vibration characteristic parameters exceed a preset balance threshold, a gradient descent algorithm is used to adjust the counterweight parameters of the optimized counterweight adjustment scheme to generate new counterweight parameters; Based on the new counterweight parameters, dynamic response simulation is performed again in a simulation environment to obtain updated vibration signals; Based on the updated vibration signals, a signal processing technique is used to extract new vibration characteristics thereof, and the simulation vibration signals are determined.

[0013] According to the air compressor rotor test method provided by the application, if the simulation vibration signal exceeds the balance threshold, the particle swarm optimization algorithm is used to iteratively update the counterweight parameters of the optimized counterweight adjustment scheme to obtain a target counterweight scheme, which includes: Based on the simulation vibration signal, a Fourier transform is used to perform frequency domain decomposition thereon to obtain frequency domain vibration characteristics; If the amplitude of the frequency domain vibration characteristics exceeds the balance threshold, a signal processing algorithm is used to denoise the frequency domain vibration characteristics to obtain smoothed vibration characteristics; Based on the smoothed vibration characteristics, the particle swarm optimization algorithm is used to iteratively update the counterweight parameters of the optimized counterweight adjustment scheme to generate the target counterweight scheme.

[0014] The application also provides an air compressor rotor testing device, which includes: The acquisition module is configured to acquire vibration signal data of the rotor at multiple rotation speeds, and analyze the vibration signal data using a fast Fourier transform algorithm to obtain vibration frequency and phase distribution; The analysis module is configured to perform weighted analysis on the phase distribution to determine the initial position coordinates of the uneven mass distribution if the vibration frequency exceeds a preset frequency threshold; The fitting module is configured to fit the offset of the rotor based on the initial position coordinates using a least squares method to obtain mass deviation coordinates; The optimization module is configured to generate an initial counterweight adjustment scheme based on the mass deviation coordinates, optimize the initial counterweight parameters of the initial counterweight adjustment scheme using a gradient descent algorithm, and obtain an optimized counterweight adjustment scheme; The simulation module is configured to perform dynamic response simulation on the optimized counterweight adjustment scheme in a simulation environment based on the counterweight parameters thereof to obtain simulation vibration signals; The update module is configured to iteratively update the counterweight parameters of the optimized counterweight adjustment scheme using a particle swarm optimization algorithm to obtain a target counterweight scheme if the simulation vibration signals exceed a balance threshold. The target weight balancing scheme is applied to the rotor for testing so that the balance state of the rotor reaches a set standard.

[0015] The present invention provides an air compressor rotor testing method and device to address the problem of uneven vibration distribution caused by uneven mass distribution during high-speed rotor rotation. By collecting vibration signal data from the rotor at multiple speeds, the device uses a fast Fourier transform (FFT) algorithm to analyze the data, extracting the frequency and phase distributions. The device then performs a weighted analysis of the phase distribution to extract the initial position coordinates of the uneven mass distribution, thereby accurately locating the position of the rotor's mass imbalance. The device then uses the least squares method to fit the offsets, obtaining mass deviation coordinates, generating an initial weight adjustment scheme, and optimizing the initial weight adjustment parameters using a gradient descent algorithm to obtain an optimized weight adjustment scheme. The optimized weight adjustment scheme is then simulated in a simulation environment, and the particle swarm optimization (PSO) algorithm is further used to iterate the weight parameters of the optimized weight adjustment scheme to obtain a target weight adjustment scheme, ensuring that the rotor's vibration signal meets vibration balance standards. The present invention automatically calculates and adjusts the weight parameters based on real-time vibration frequency and phase analysis, ensuring that the rotor meets stable dynamic balance standards. This improves the efficiency and accuracy of air compressor rotor dynamic balance testing, reduces residual vibration, ensures the stability of rotating machinery operation, extends equipment service life, and extends the use of dynamic balance. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is one of the flow charts of the air compressor rotor testing method provided by an embodiment of the present invention; Figure 2 This is the second flow chart of the air compressor rotor testing method provided by an embodiment of the present invention; Figure 3 This is the third flow chart of the air compressor rotor testing method provided by an embodiment of the present invention; Figure 4 This is a fourth flow chart of the air compressor rotor testing method provided by an embodiment of the present invention; Figure 5 This is the fifth flow chart of the air compressor rotor testing method provided by an embodiment of the present invention; Figure 6 This is the sixth flow chart of the air compressor rotor testing method provided by an embodiment of the present invention; Figure 7 This is the seventh flow chart of the air compressor rotor testing method provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0017] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments of the present application, all the other embodiments obtained by a person of ordinary skill in the art without creative work fall within the scope of the present application.

[0018] With reference to Figure 1 The embodiment of the present application provides a test method for a compressor rotor, comprising the following steps. In step 100, vibration signal data of the rotor at multiple rotating speeds is acquired, and a fast Fourier transform algorithm is used to analyze the vibration signal data to obtain vibration frequency and phase distribution. The vibration signal data of the rotor at multiple rotating speeds can be collected by a high-precision sensor array, and then the fast Fourier transform algorithm is used to perform frequency domain conversion on the vibration signal data to obtain frequency spectrum data. According to the frequency spectrum data, the vibration frequency is extracted, and the amplitude and phase distribution of each frequency component are calculated. The core of this step is to locate the mass distribution defect of the rotor by dynamic vibration analysis. During the test process, the time domain vibration signal of the rotor is first collected at multiple preset rotating speed conditions. This multi-rotating speed coverage strategy can comprehensively capture the response characteristics of the rotor under different dynamic states. Then, the fast Fourier transform is used to perform frequency spectrum decomposition on the vibration signal data, and the time domain vibration waveform is converted into frequency domain representation, so that the main vibration frequency component of the rotor and the corresponding phase angle information are accurately extracted, and the vibration frequency and phase distribution are obtained. The phase distribution reflects the space-time characteristics of each vibration source of the rotor in the circumferential direction.

[0019] For example, when acquiring rotor vibration signal data through a high-precision sensor array, MEMS acceleration sensors can be arranged in the X, Y, and Z directions of the rotor bearing seat, and vibration signals at rotation speeds of 1000 rpm, 2000 rpm, and 3000 rpm are collected, with a sampling time of 10 seconds, to generate time series data (10000 points per second, a total of 100000 points). The data is transmitted to an embedded processor through a 24-bit ADC converter in I2C protocol. Then, the signal is processed using the fast Fourier transform algorithm, which is implemented using the Cooley-Tukey algorithm, with the input data point number N = 8192 (2 raised to the power of 13), and the time domain signal is decomposed into the frequency domain through butterfly operation, and the frequency resolution Δf = 10000 / 8192 ≈ 1.22 Hz is calculated. When analyzing the frequency spectrum, the main vibration frequency is extracted, for example, at 2000 rpm (33.33 Hz), a significant peak is observed at 33 Hz, 66 Hz (2 times frequency), and 100 Hz (3 times frequency), with amplitudes of 0.5g, 0.2g, and 0.1g, respectively. The phase distribution is calculated from the complex results output by the fast Fourier transform algorithm to obtain the phase angle of each frequency component, for example, the phase at 33 Hz is 45°, and the phase at 66 Hz is 90°. Combined with the rotor dynamics model, the linear relationship between frequency and rotation speed is verified to determine whether there is imbalance.

[0020] In step 200, if the vibration frequency exceeds the preset frequency threshold, the phase distribution is analyzed by weighting to determine the initial position coordinates of the uneven mass distribution. When a certain vibration frequency component is detected to exceed the preset safety threshold, it indicates that the rotor has an abnormal excitation source, usually caused by uneven mass distribution, at which time the weighting analysis mechanism is started: based on the reliability of the phase data at different rotation speeds and the vibration energy size, the phase distribution is dynamically weighted. The dynamic weight distribution can be a phase data weight coefficient, which strengthens the high-energy vibration section and weakens the noise interference area, so that environmental noise and harmonic interference can be effectively stripped, and the initial angular position of the mass imbalance in the rotor circumferential coordinate system can be accurately positioned. Therefore, the high-energy vibration area and the noise interference area can be determined by the vibration frequency, and the high-energy vibration area and the noise interference area can be respectively analyzed by the phase data weight coefficient to obtain more accurate initial position coordinates of the uneven mass distribution. The initial position coordinates are used to guide the subsequent dynamic balance correction of the counterweight operation, and the initial position coordinates include the radius and angle of the mass distribution uneven site.

[0021] In step 300, based on the initial position coordinates, the least squares method is used to fit the offset of the rotor to obtain the mass deviation coordinates. After determining the initial position coordinates of the rotor mass distribution imbalance, a precise counterweight optimization process is further implemented. First, a mathematical model of the rotor spatial displacement is established using the least squares method: by aggregating the detected vibration phase shift data under multiple rotational speed conditions, a mass deviation distribution function in the rotor circumferential direction is constructed to fit the displacement of the rotor through the mass deviation distribution function. The least squares method can effectively eliminate single-point measurement error interference, and through the mass deviation distribution function, a continuous displacement trajectory of the actual mass center relative to the geometric center is fitted in a three-dimensional coordinate system, and finally the mass deviation coordinates representing the maximum imbalance point are output. Among them, the mass deviation coordinates include angle position information and the spatial vector amplitude of the mass defect.

[0022] Step 400, generating an initial counterweight adjustment scheme based on the mass deviation coordinates, using a gradient descent algorithm to optimize the initial counterweight parameters of the initial counterweight adjustment scheme to obtain an optimized counterweight adjustment scheme; Based on the mass deviation coordinates, an initial counterweight adjustment scheme is generated, which includes the initial mass parameters of the counterweight blocks and the angle positions of the counterweight block installation. In short, the initial counterweight adjustment scheme includes the counterweight mass and the counterweight position. To overcome the precision limitations caused by the nonlinearity of rotor dynamics in traditional counterweight schemes, a gradient descent algorithm is introduced for parameter optimization: the algorithm starts with the initial counterweight parameters, calculates the residual vibration energy values under different counterweight combinations, constructs a multi-dimensional response surface of the counterweight effect, and iteratively adjusts the counterweight mass and the counterweight position along the negative gradient direction of the surface, so that the parameters automatically converge to the optimal solution with the minimum vibration energy. Finally, the optimized counterweight adjustment scheme is obtained. This dynamic optimization process based on the gradient descent algorithm can adaptively compensate for the influence of complex factors such as aerodynamic effects and bearing stiffness changes.

[0023] Step 500, based on the counterweight parameters of the optimized counterweight adjustment scheme, performing dynamic response simulation in a simulated environment to obtain simulation vibration signals; After obtaining the optimized counterweight adjustment scheme, the multi-physical field coupling simulation platform can be used to virtually verify the optimized counterweight adjustment scheme. Based on the dynamics equation of the rotor-bearing system and the aerodynamic load model, the counterweight parameters in the optimized counterweight adjustment scheme are imported for full-speed domain transient simulation, accurately reproducing the dynamic response characteristics of the rotor after the counterweight blocks are applied. The simulation outputs simulation vibration signals, which are used to evaluate the simulation effect of the optimized counterweight adjustment scheme. Therefore, the frequency spectrum characteristics of the simulation vibration signals have engineering comparability with the actual test data, and can pre-act the vibration behavior of the rotor under real working conditions, without the need to apply it to the air compressor rotor for testing, which can reduce the loss of the air compressor rotor and improve the simulation accuracy and effect.

[0024] Step 600: If the simulated vibration signal exceeds the balance threshold, a particle swarm optimization algorithm is used to iteratively update the weight parameters of the optimized weight adjustment scheme to obtain a target weight adjustment scheme. The target weight balancing scheme is applied to the rotor for testing so that the balance state of the rotor reaches a set standard.

[0025] If the equivalent amplitude of the simulated vibration signal exceeds the preset balance threshold, it indicates that the optimized solution still has residual imbalance and the counterweight solution needs to be further optimized. The balance threshold takes into account the safety margin and process requirements, and represents the state threshold when the rotor reaches a dynamic balance state. At this time, the particle swarm optimization algorithm is started to re-optimize the parameters. The particle swarm optimization algorithm maps the counterweight mass and counterweight position in the optimized counterweight adjustment solution to the position vector of the particle in multidimensional space, and uses the residual vibration energy value as the fitness function. The particle swarm searches in parallel in the parameter space through a group collaboration mechanism, and uses the dual guidance of the historical optimal solution and the global optimal solution to dynamically adjust the particle flight direction and step size. This distributed intelligent optimization can effectively escape the local optimal trap, gradually approach the global optimal solution during the iterative process, and finally output the target counterweight solution that meets the balance standard.

[0026] The present invention provides an air compressor rotor testing method and device to address the problem of uneven vibration distribution caused by uneven mass distribution during high-speed rotor rotation. By collecting vibration signal data from the rotor at multiple speeds, the device uses a fast Fourier transform (FFT) algorithm to analyze the data, extracting the frequency and phase distributions. The device then performs a weighted analysis of the phase distribution to extract the initial position coordinates of the uneven mass distribution, thereby accurately locating the position of the rotor's mass imbalance. The device then uses the least squares method to fit the offsets, obtaining mass deviation coordinates, generating an initial weight adjustment scheme, and optimizing the initial weight adjustment parameters using a gradient descent algorithm to obtain an optimized weight adjustment scheme. The optimized weight adjustment scheme is then simulated in a simulation environment, and the particle swarm optimization (PSO) algorithm is further used to iterate the weight parameters of the optimized weight adjustment scheme to obtain a target weight adjustment scheme, ensuring that the rotor's vibration signal meets vibration balance standards. The present invention automatically calculates and adjusts the weight parameters based on real-time vibration frequency and phase analysis, ensuring that the rotor meets stable dynamic balance standards. This improves the efficiency and accuracy of air compressor rotor dynamic balance testing, reduces residual vibration, ensures the stability of rotating machinery operation, extends equipment service life, and extends the use of dynamic balance.

[0027] In one embodiment, see Figure 2 , also includes: Step 700: applying the target weight balancing scheme to the rotor through an automated actuator to obtain a real-time vibration feedback signal; Step 800, if the vibration feedback signal contains residual vibration, performing secondary analysis on the vibration feedback signal, updating the mass deviation coordinates, and determining the feedback counterweight parameter; Step 900, based on the feedback counterweight parameter, continuously iterating optimization and real-time feedback adjustment to generate a dynamic balance counterweight scheme, wherein the working state of the rotor reaches the set standard when the dynamic balance counterweight scheme is applied to the rotor.

[0028] The rotor is adjusted according to the target counterweight scheme by an automatic actuator, and vibration data feedback from the rotor is collected to obtain real-time vibration feedback signal. Then it is detected whether the vibration feedback signal contains residual vibration. The vibration feedback signal can be analyzed by fast Fourier transform algorithm to obtain vibration frequency characteristics, and then it is determined whether the vibration feedback signal contains residual vibration by whether the vibration frequency characteristics exceed the preset range. It should be noted that if the vibration frequency characteristics exceed the preset range, it means that the vibration feedback signal contains residual vibration. Further, the counterweight scheme adjustment mechanism can be that if it is detected that the vibration frequency characteristics exceed the preset range, the counterweight correction amount is calculated; the counterweight parameter of the target counterweight scheme is updated according to the counterweight correction amount, the updated vibration data is collected, and it can also be judged whether the updated vibration data meets the dynamic balance state by support vector machine algorithm.

[0029] If it is detected that the vibration feedback signal contains residual vibration, the vibration signal is analyzed in frequency domain by Fourier transform to obtain vibration frequency components. According to the vibration frequency components, the main vibration mode is extracted by principal component analysis to determine the mass deviation characteristics. If the mass deviation characteristics exceed the preset threshold, the deviation data is fitted by least square method to calculate the deviation coordinate value. According to the deviation coordinate value, the mass deviation coordinates are updated by linear interpolation method to obtain the updated mass deviation coordinates. The feedback counterweight parameter is generated by the updated mass deviation coordinates. If the feedback counterweight parameter and the historical parameter differ by more than the preset range, the feedback counterweight parameter is optimized and adjusted by Kalman filter to obtain the final adjustment value, and the dynamic balance counterweight scheme is generated.

[0030] In another embodiment, the vibration feedback signal of the rotor in operation is obtained when the target counterweight scheme is applied to the rotor by the automatic actuator. Then, the vibration feedback signal is analyzed by using fast Fourier transform to obtain the unbalance frequency component. According to the unbalance frequency component, the initial counterweight position and mass are calculated to determine the first intermediate counterweight scheme. The first intermediate counterweight scheme is verified by simulation to obtain intermediate simulation vibration data. If the intermediate simulation vibration data deviates from the preset standard deviation by more than a preset threshold, the first intermediate counterweight scheme is iterated by using a gradient descent algorithm to obtain a second intermediate counterweight scheme. The real-time vibration data is obtained by adjusting the counterweight position and configuration mass of the rotor according to the second intermediate counterweight scheme. If the real-time vibration data meets the preset standard, it is determined that the balance state is stable, and the final counterweight scheme is output as the second intermediate counterweight scheme.

[0031] In this embodiment, the full-process automatic closed loop of virtual optimization, entity execution, and online feedback is realized, the limitations of traditional dynamic balancing relying on manual trial and error are eliminated, and the balancing correction period is shortened and the optimization efficiency is improved through real-time data feedback decision-making. And the secondary analysis module realizes dynamic compensation of uncertain interference, effectively separating the steady-state unbalance and transient interference.

[0032] In one embodiment, referring to Figure 3 , if the vibration frequency exceeds the preset frequency threshold, the phase distribution is analyzed by weighting to determine the initial position coordinates of the uneven mass distribution, including: Step 201, if the vibration frequency exceeds the preset frequency threshold, the vibration signal data of the rotor is obtained, and the vibration frequency value of the vibration signal data is calculated by using fast Fourier transform algorithm to obtain frequency data; Step 202, based on the frequency value of the frequency data exceeding the preset frequency threshold, a phase distribution matrix is generated based on the phase distribution; Step 203, based on the phase distribution matrix, the weighted phase value of the phase distribution matrix is calculated by using a weighted analysis algorithm to determine the phase distribution feature; Step 204, the phase distribution feature is compared with a preset mass distribution model, and the mass distribution curve of the unbalance mass of the rotor on the balance plane is fitted by using a least squares method algorithm to determine an uneven degree parameter, wherein the uneven degree parameter is an evaluation index of the unbalance state of the rotor; Step 205, based on the uneven degree parameter, the spatial deviation of the mass distribution is calculated to determine the initial position coordinates.

[0033] When the vibration frequency exceeds the preset safety threshold, a multi-level diagnostic protocol is automatically triggered. First, the vibration signal data of the rotor in the full speed range is obtained, and real-time spectral decomposition is performed using fast Fourier transform to extract frequency data containing the fundamental frequency, multiple frequencies and high-order harmonics. After filtering by a threshold filter, only the over-standard frequency bands are retained for further analysis. For the phase information corresponding to the over-standard frequency, a spatio-temporal coupled phase distribution matrix is constructed. The matrix has the rotor speed gradient as the longitudinal dimension and the sensor array spatial coordinates as the transverse dimension, forming a dynamic topological mapping of the rotor surface vibration phase. Based on the signal-to-noise ratio and vibration energy entropy value of each node in the phase distribution matrix, a weighted phase value is calculated using a weighted analysis algorithm: higher weight is given to high-confidence measurement points, and phase distortion caused by electromagnetic interference or bearing clearance is automatically attenuated, obtaining phase distribution characteristics representing the nature of the quality defect. The weighted analysis algorithm can be an adaptive fuzzy weighted algorithm.

[0034] The phase distribution characteristics are input into the preset mass distribution model of the rotor, which embeds the rigid body dynamics equation and the material density distribution rule. The residual field of the measured phase and the theoretical phase is fitted by the least squares method, and a continuous mass distribution probability curve is reconstructed in the rotor balance plane. The gradient extreme point and the curvature radius of the mass distribution probability curve jointly generate the unevenness parameter, wherein the unevenness parameter is a standardized measure of the mass eccentricity. Finally, based on the unevenness parameter, vector synthesis operation is performed to solve the offset angle and radial deviation of the mass core in the polar coordinate system, and output the initial position coordinates available for engineering.

[0035] For example, assuming that the vibration frequency is collected in real time through a sensor, and the preset threshold is set to 50 Hz, the system detects that the current vibration frequency is 60 Hz, which exceeds the threshold, triggering the phase distribution weighted analysis. First, the phase data of the vibration signal is collected, and the time domain signal is converted into the frequency domain using Fourier transform to obtain a phase distribution matrix, assuming that the matrix is [30°, 45°, 60°, 90°], corresponding to four key points of the device. The weighted analysis uses a weighted average algorithm, and the weights are determined according to the historical vibration contribution rate of each point, for example, the weight vector is [0.4, 0.3, 0.2, 0.1]. The weighted phase is calculated as follows: (30° x 0.4 + 45° x 0.3 + 60° x 0.2 + 90° x 0.1) = 46.5°. To determine the initial position of the uneven mass distribution, the system compares the weighted phase with a standard phase library, which stores the phase distribution of the device under normal operation, for example, the standard value is 20°. The deviation is calculated as follows: 46.5° - 20° = 26.5°, which is greater than the threshold of 5°, indicating that there is uneven mass. Combined with the geometric model of the device, assuming that the device is a disc with a radius of 0.5 m, the position of the uneven mass distribution is mapped through the phase deviation, and the eccentric angle of 26.5° corresponds to an arc of 0.462 rad, and the coordinates of the position are (0.5 x cos(0.462), 0.5 x sin(0.462)) = (0.448 m, 0.223 m). The system records the coordinates and generates a mass adjustment instruction, which is output to the control module to automatically adjust the counterweight to correct the unevenness and ensure that the vibration frequency returns to the threshold range.

[0036] In this embodiment, a dynamic threshold triggering mechanism is realized, and real-time threshold comparison of frequency data replaces the traditional fixed threshold to adapt to the safety margin of different rotating speed conditions; a phase topology denoising technology is realized, and the phase distribution matrix is analyzed through the correlation of time and space dimensions to effectively separate the real mass signal from the environmental noise; and through the mass distribution probability curve, the traditional discrete point fitting limitation is broken through to realize the continuous field quantitative description of the rotor imbalance state.

[0037] In one embodiment, referring to Figure 4 , the initial position coordinates are used to fit the offset of the rotor by the least square method to obtain mass deviation coordinates, including: Step 301, based on the initial position coordinates and a preset coordinate system, the offset of the rotor relative to the center axis is calculated by the Euclidean distance calculation method to obtain an offset data set; Step 302, if there is an offset in the offset data set that exceeds a preset offset threshold, the offset data set is denoised by a median filter method to obtain a smoothed offset data set; Step 303, based on the smoothed offset data set, the offset is fitted by the least square method to obtain an eccentric vector parameter; Step 304, based on the direction and amplitude information in the eccentricity vector parameter, the vector decomposition method is used to determine the mass deviation coordinate.

[0038] Based on the initial position coordinates, the calculation of spatial offset is carried out, and in the preset rotor three-dimensional coordinate system, the radial offset of each measuring point relative to the theoretical central axis is calculated by using the Euclidean distance calculation model, and the offset data set containing the spatial offset vector under multiple rotating speed conditions is generated. The offset data set essentially represents the dynamic fluctuation characteristics of the rotor geometric center trajectory. When it is detected that any one of the offset data set exceeds the preset dynamic offset threshold, the anti-impulse interference mechanism is automatically activated, and the offset data set is denoised by the median filtering algorithm: the spatial offset is sorted and statistically filtered with the rotating speed as the time sequence axis, so that the wild point generated by the instantaneous collision of the bearing or electromagnetic interference is effectively suppressed, and the smooth offset data set conforming to the rigid body motion law is output. This process can eliminate a large amount of transient noise interference while retaining the true offset trend.

[0039] Then, based on the smooth data set, the offset trajectory fitting is carried out by using the weighted least squares method: according to the vibration energy of different rotating speed points, the differential weight is given, and the continuous function of the offset with the angle change is constructed. The least squares method breaks through the limitation of traditional average fitting, generates the eccentricity vector parameter containing the direction angle and the amplitude, and the eccentricity vector parameter represents the spatial deviation vector of the mass center relative to the rotating axis. Finally, the mass imbalance component in each balance plane is decoupled out by the vector decomposition method of the polar coordinate, and the mass deviation coordinate is output.

[0040] In this embodiment, the dynamic threshold adaptive system is realized, the characteristics of the offset threshold automatically adjusting with the rotating speed are solved, the misjudgment problem of the constant threshold under variable working conditions is solved. The kinematics guided denoising technology is realized, the median filtering is combined with the rigid body motion constraint, the real offset characteristics are retained while the non-mechanism noise is eliminated. The energy weighted fitting is realized, the weight distribution strategy based on the vibration energy is used to strengthen the contribution degree of the high rotating speed working condition data. The dynamic accuracy of the eccentricity vector parameter is improved.

[0041] In one embodiment, please refer to Figure 5 , the initial counterweight adjustment scheme is generated based on the mass deviation coordinate, the initial counterweight parameter of the initial counterweight adjustment scheme is optimized by using the gradient descent algorithm, and the optimized counterweight adjustment scheme is obtained, which includes: Step 401, based on the mass deviation coordinate, the deviation vector is calculated to obtain the initial mass deviation distribution; Step 402, based on the initial mass deviation distribution, the counterweight mass and the counterweight position are initialized, the initial counterweight adjustment scheme and the initial counterweight parameter corresponding to the initial counterweight adjustment scheme are determined; Step 403, if the error analysis result of the initial counterweight parameter is greater than the preset error threshold, iteratively update the counterweight mass and counterweight position of the initial counterweight parameter through the gradient descent algorithm to obtain an optimized parameter; Step 404, based on the optimized parameter, recalculate the mass deviation distribution to obtain the optimized counterweight adjustment scheme.

[0042] Based on the accurate positioning of the mass deviation coordinates, first calculate the deviation vector in three-dimensional space, which integrates the radial deviation amplitude, the circumferential angle, and the axial position, to generate an initial mass deviation distribution reflecting the unbalance state of the rotor. The initial mass deviation distribution is presented in polar coordinate form and contains the core area of mass defects and gradient diffusion characteristics. Based on the initial mass deviation distribution, automatically trigger intelligent initialization of the counterweight parameter: according to the rotor dynamics characteristics and material density constraints, derive the recommended mass range and installation angle range of the counterweight block through reverse vector operation, generate an initial counterweight adjustment scheme containing the initial counterweight position and the initial counterweight mass, the initial counterweight position is the installation angle range of the counterweight block, and the initial counterweight mass is the recommended mass range of the counterweight block. The scheme synchronously outputs the initial counterweight parameter of the key performance indicator, which can be the mass tolerance, the phase tolerance, and the estimated residual vibration energy value. The initial counterweight parameter can include the initial counterweight position and the initial counterweight mass, or it can be an effectiveness parameter evaluating the counterweight mass and the counterweight position.

[0043] Then, based on the initial counterweight parameter, perform error analysis to obtain an error analysis result. When the error analysis result shows that the residual vibration prediction value exceeds the dynamic error threshold, which is related to the rotor safety level, the system activates the gradient-driven optimization engine. Starting from the initial parameter, the algorithm iteratively searches along the negative gradient direction of the vibration energy function: by virtually perturbing the counterweight mass and angle parameters, calculate the sensitivity matrix of the target function of the residual vibration amplitude, and dynamically adjust the parameter update step. After multiple rounds of intelligent iteration, output the optimized parameter that makes the vibration energy converge to the safety domain. Finally, based on the optimized parameter, reconstruct the mass deviation distribution model to generate an engineering executable optimized counterweight adjustment scheme.

[0044] For example, generate a counterweight adjustment scheme based on mass deviation coordinates. First, obtain the mass deviation coordinates from the input data. Suppose the initial mass deviation coordinates are (x=0.5m, y=0.3m, z=0.2m), which represent the offset of the object's center of mass from the ideal position. The goal is to adjust the center of mass to the origin (0,0,0). Based on the gradient descent algorithm, optimize the counterweight mass and position, initialize the counterweight parameters, set the initial counterweight mass m=1kg, and the position coordinates are (x=0.1m, y=0.1m, z=0.1m). The initial counterweight adjustment scheme contains the initial counterweight position and the initial counterweight mass. c =0.1m, y c =0.1m, z c= 0.1m). Define the objective function as the sum of the square of the centroid offset, the calculation formula of the objective function E is as follows: E = x 2 + y 2 + z 2 Where (x, y, z) is the initial mass deviation coordinate, the centroid coordinate (x, y, z) is updated by calculating the influence of the counterweight on the centroid. The calculation formula is as follows: And Similar, where m is the initial counterweight mass, (x, y, z) is the coordinate value of the initial counterweight position, m c is the mass deviation value of the initial mass deviation distribution, (x c , y c , z c ) is the position deviation value of the initial mass deviation distribution, and m is assumed to be 10kg. The gradient descent algorithm iteratively updates the counterweight parameters, and the learning rate α = 0.01, the partial derivative of the objective function with respect to m c , x c , y c , z c . Assuming that the initial E = 0.38, after 10 iterations, E decreases to 0.05, and the counterweight parameters converge to m c = 1.2kg, x c = 0.08m, y c = 0.06m, z c = 0.04m. The analysis process verifies that the counterweight position is close to the deviation direction, and the mass is slightly increased to balance the deviation. The final output counterweight scheme is to add 1.2kg counterweight at (0.08m, 0.06m, 0.04m), and the centroid deviation is reduced to 0.05, meeting the accuracy requirement. If the deviation is still above the threshold, the number of iterations can be increased or the learning rate can be adjusted.

[0045] In this embodiment, the vectorization algorithm is realized, the three-dimensional deviation vector breaks through the limitation of traditional plane projection, and the axial asymmetric imbalance is accurately captured; the dynamic guidance initialization is realized, the counterweight parameter initialization is fused with the rotor modal shape data, and the resonance risk is avoided; the gradient descent algorithm based on vibration energy sensitivity iteratively updates the algorithm, which makes the optimization efficiency several times higher than that of traditional methods.

[0046] In one embodiment, referring to Figure 6 , the counterweight parameters based on the optimized counterweight adjustment scheme are simulated in a dynamic response simulation environment to obtain simulation vibration signals, including: Step 501, load the initial counterweight parameters in the simulation environment, perform dynamic response simulation, and obtain vibration signal data; Step 502, based on the vibration signal data, extract its vibration characteristics using signal extraction technology to obtain vibration characteristic parameters; Step 503, if the vibration characteristic parameters exceed the preset balance threshold, adjust the counterweight parameters of the optimization counterweight adjustment scheme through the gradient descent algorithm to generate new counterweight parameters; Step 504, based on the new counterweight parameters, re-perform dynamic response simulation in the simulation environment to obtain updated vibration signals; Step 505, based on the updated vibration signals, extract new vibration characteristics using signal processing technology to determine the simulation vibration signals.

[0047] After obtaining the initial counterweight parameters, they are imported into a multi-physical field coupled simulation environment, which can be a finite element model based on a rotor, bearing and shell system. This model can accurately simulate the dynamic behavior of the rotor after the counterweight block is applied. Through full-speed domain transient simulation, high-fidelity vibration signal data is output. Based on the vibration signal data, signal extraction techniques are used to extract key vibration characteristics to obtain vibration characteristic parameters. For example, background noise can be stripped through wavelet packet decomposition, and Hilbert transform demodulation can be used to demodulate the modulation sideband to condense vibration characteristic parameters including fundamental frequency amplitude, frequency energy ratio, phase coherence, etc. The signal extraction technique can be an adaptive signal deconstruction technique.

[0048] When the vibration characteristic parameters exceed the dynamic balance threshold, gradient-driven re-optimization is started. With the current counterweight parameters as the initial point, the gradient descent algorithm is used to calculate the partial derivative matrix of the vibration energy function with respect to the counterweight mass and angle, and the parameter step size is adjusted adaptively along the negative gradient direction. After generating new counterweight parameters each time, the simulation environment is triggered to run again, forming a closed loop of parameter updating and simulation verification, and finally obtaining simulation vibration signals that meet the balance standard.

[0049] For example, in the simulation environment, the dynamic response simulation of the preliminary counterweight parameters is first constructed, assuming that the rotating mechanical rotor system is 10 kg, the rotating speed is 1500 rpm, the initial eccentric counterweight is 0.1 kg, and the radius is 0.05 m. The rotor model is established using multi-body dynamics software, the material property is set to steel (density 7850 kg / m³), and the rotating constraint is applied. The dynamic response simulation uses the Runge-Kutta algorithm to solve the motion equation, the time step is 0.001 s, the simulation time is 2 s, and the vibration displacement of the rotor under different counterweights is calculated. The vibration signal is obtained through a virtual acceleration sensor, the sensor sampling frequency is 1000 Hz, the output vibration acceleration time sequence is obtained, and the peak acceleration is assumed to be 0.5 m / s². The signal processing uses fast Fourier transform to convert the time domain signal to frequency domain, extracts the main frequency component (25 Hz, corresponding to the rotating speed), and analyzes the vibration amplitude. If the main frequency amplitude exceeds the threshold value of 0.3 m / s², the counterweight is adjusted. The optimization algorithm uses genetic algorithm, the initial population size is 50, the iteration is 20 times, the objective function is to minimize the main frequency amplitude, the counterweight mass is adjusted to 0.12 kg, and the position angle is rotated by 15°. The new vibration signal is obtained by simulation again, and the peak acceleration is reduced to 0.25 m / s². When the balance threshold is judged, the main frequency amplitude is compared with the threshold value of 0.3 m / s². If it is less than or equal to the threshold value, it is determined to meet the balance requirement. The final result shows that the amplitude after adjustment is 0.25 m / s², which meets the requirement.

[0050] In this embodiment, the intelligent signal decoupling technology is realized, the adaptive signal decomposition breaks through the limitation of traditional algorithm, and the weak fault characteristics in strong noise background are accurately extracted. The closed-loop gradient optimization is realized, the real-time linkage mechanism is realized by combining the simulation and optimization process, the time consumption of single iteration is reduced, and the efficiency is improved several times compared with the traditional offline optimization.

[0051] In one embodiment, referring to Figure 7 , if the simulation vibration signal exceeds the balance threshold, a particle swarm optimization algorithm is used to re-iterate and update the counterweight parameters of the optimization counterweight adjustment scheme to obtain a target counterweight scheme, including: Step 601, based on the simulation vibration signal, a Fourier transform is used to perform frequency domain decomposition to obtain frequency domain vibration characteristics; Step 602, if the amplitude of the frequency domain vibration characteristics exceeds the balance threshold, a signal processing algorithm is used to denoise the frequency domain vibration characteristics to obtain smooth vibration characteristics; Step 603, based on the smooth vibration characteristics, a particle swarm optimization algorithm is used to re-iterate and update the counterweight parameters of the optimization counterweight adjustment scheme to generate the target counterweight scheme.

[0052] The vibration signal acquired in the simulation environment is first subjected to high-resolution Fourier transform for frequency domain decomposition, which can suppress spectral leakage, accurately separate the fundamental frequency, harmonics and subsynchronous components, and generate a frequency domain vibration feature containing amplitude, frequency and phase. The frequency domain vibration feature quantifies the residual vibration energy distribution of the counterweight rotor, meeting the analysis requirements of the super-speed rotor. When the amplitude of the frequency domain vibration feature exceeds the dynamic balance threshold, the intelligent frequency domain noise reduction channel is activated, and a signal processing algorithm is used to denoise the frequency domain vibration feature, eliminating non-mechanical components such as electromagnetic interference and background noise while retaining the main vibration energy, and outputting a smooth vibration feature that meets the rigid body dynamics law. The signal processing algorithm can be adaptive wavelet threshold filtering and coherent averaging technology.

[0053] Based on the smooth vibration feature, the particle swarm optimization engine is called, the counterweight mass and angle parameters are mapped to the position vector of the particle swarm, the residual vibration energy is taken as the fitness function, the counterweight parameters of the optimized counterweight adjustment scheme are iteratively updated, and the particle swarm is distributed in the parameter space through the group cooperation mechanism: individual particles remember the historical optimal solution, global particles share optimal position information, and combined with the adaptive adjustment strategy of inertia weight, efficient global optimization is realized, and the target counterweight scheme that minimizes the vibration energy is output.

[0054] For example, to simulate the vibration signal and optimize the counterweight parameters, first, the vibration signal is collected by the sensor, assuming that the collected signal frequency is 50 Hz and the amplitude is 0.02 mm, and the threshold is set to 0.015 mm. The simulation system uses finite element analysis software to establish a rotor model, which contains 4 counterweight points, each with an initial counterweight mass of 0.5 kg and an initial phase angle of 0°. After running the simulation, the vibration amplitude is calculated to be 0.018 mm, which exceeds the threshold of 0.015 mm, triggering the optimization process. The system automatically calls the particle swarm optimization algorithm, sets the particle swarm size to 30, the maximum number of iterations to 100, the inertia weight w from 0.9 to 0.4 linearly, and sets the learning factor c1=c2=2. The position of each particle represents a set of counterweight parameters, including mass and phase angle, and the search space limits the mass to 0.1-1.0 kg and the phase angle to 0-360°. The particle swarm optimization algorithm iteratively updates the particle velocity and position, and calculates the objective function, which is the square of the difference between the vibration amplitude and the threshold. For example, at the 68th iteration, an optimized solution is found: counterweight point 1: mass 0.62 kg, phase angle 45°, counterweight point 2: mass 0.55 kg, phase angle 120°, counterweight point 3: mass 0.48 kg, phase angle 200°, counterweight point 4: mass 0.51 kg, phase angle 300°. Input this scheme into the simulation model, recalculate the vibration amplitude, and the result is 0.013 mm, which is lower than the threshold of 0.015 mm, meeting the requirements.

[0055] The embodiment realizes a dynamic guidance noise reduction mechanism, adaptive filtering based on a rotor motion equation, effective reservation of a frequency band strongly related to a quality defect, realization of global optimization of swarm intelligence, a particle swarm algorithm breaking through local convergence limitation of gradient optimization, residual vibration reduction compared with a gradient method, realization of cross-dimension parameter mapping, an intelligent mapping mechanism of vibration characteristics to counterweight parameters, and reduction of optimization iteration times.

[0056] The air compressor rotor testing device provided by the application is described below, and the air compressor rotor testing device described below can be correspondingly referred to the air compressor rotor testing method described above.

[0057] The application also provides an air compressor rotor testing device, comprising: An acquisition module is configured to acquire vibration signal data of a rotor at multiple rotation speeds, and analyze the vibration signal data by using a fast Fourier transform algorithm to obtain vibration frequency and phase distribution. An analysis module is configured to perform weighted analysis on the phase distribution to determine initial position coordinates of uneven mass distribution if the vibration frequency exceeds a preset frequency threshold. A fitting module is configured to fit the offset of the rotor by using a least square method based on the initial position coordinates to obtain mass deviation coordinates. An optimization module is configured to generate an initial counterweight adjustment scheme based on the mass deviation coordinates, optimize initial counterweight parameters of the initial counterweight adjustment scheme by using a gradient descent algorithm, and obtain an optimized counterweight adjustment scheme. A simulation module is configured to perform dynamic response simulation on the optimized counterweight adjustment scheme in a simulation environment based on counterweight parameters of the optimized counterweight adjustment scheme to obtain simulation vibration signal. An update module is configured to re-iteratively update the counterweight parameters of the optimized counterweight adjustment scheme by using a particle swarm optimization algorithm to obtain a target counterweight scheme if the simulation vibration signal exceeds a balance threshold. The target counterweight scheme is used to be applied to the rotor for testing, so that the balance state of the rotor reaches a set standard.

[0058] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the application, and not to limit them; although the application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the application.

Claims

1. An air compressor rotor testing method, characterized by, Computer-executed, comprising: Obtaining vibration signal data of the rotor at multiple rotational speeds, analyzing the vibration signal data using a fast Fourier transform algorithm to obtain vibration frequency and phase distribution; If the vibration frequency exceeds a preset frequency threshold, performing weighted analysis on the phase distribution to determine the initial position coordinates of the uneven mass distribution; Based on the initial position coordinates, fitting the rotor's offset using the least squares method to obtain the mass deviation coordinates; Based on the mass deviation coordinates, generating an initial counterweight adjustment scheme, optimizing the initial counterweight parameters of the initial counterweight adjustment scheme using the gradient descent algorithm to obtain an optimized counterweight adjustment scheme; Based on the counterweight parameters of the optimized counterweight adjustment scheme, performing dynamic response simulation in a simulated environment to obtain simulation vibration signals; If the simulation vibration signals exceed the balance threshold, using the particle swarm optimization algorithm to re-iterate and update the counterweight parameters of the optimized counterweight adjustment scheme to obtain a target counterweight scheme; Wherein, the target counterweight scheme is used to apply to the rotor for testing, so that the balance state of the rotor reaches the set standard.

2. The air compressor rotor testing method of claim 1, wherein, Also includes: Applying the target counterweight scheme to the rotor through an automatic actuator to obtain real-time vibration feedback signals; If the vibration feedback signals contain residual vibrations, performing secondary analysis on the vibration feedback signals to update the mass deviation coordinates and determine the feedback counterweight parameters; Based on the feedback counterweight parameters, continuously iteratively optimizing and real-time feedback adjusting to generate a dynamic balance counterweight scheme, wherein the working state of the rotor reaches the set standard when the dynamic balance counterweight scheme is applied to the rotor.

3. The air compressor rotor testing method of claim 1, wherein, If the vibration frequency exceeds a preset frequency threshold, obtaining the vibration signal data of the rotor and calculating the vibration frequency value of the vibration signal data using a fast Fourier transform algorithm to obtain frequency data; Based on the frequency value of the frequency data exceeding the preset frequency threshold, generating a phase distribution matrix based on the phase distribution; Based on the phase distribution matrix, calculating the weighted phase value of the phase distribution matrix using a weighted analysis algorithm to determine the phase distribution characteristics; Comparing the phase distribution characteristics with a preset mass distribution model, fitting the mass distribution curve of the unbalanced mass of the rotor on the balance plane using the least squares method algorithm to determine the uneven degree parameter, wherein the uneven degree parameter is an evaluation index of the unbalanced state of the rotor; Based on the uneven degree parameter, calculating the spatial deviation of the mass distribution to determine the initial position coordinates. Based on the initial position coordinates and a preset coordinate system, calculating the rotor's offset relative to the central axis using the Euclidean distance calculation method to obtain the offset data set; 4. The air compressor rotor testing method of claim 1, wherein, ​ ​ If the offset data set has an offset exceeding a preset offset threshold, a median filtering method is used to denoise the offset data set to obtain a smoothed offset data set; Based on the smoothed offset data set, a least square method is used to fit the offset to obtain an eccentric vector parameter; Based on the direction and amplitude information in the eccentric vector parameter, a vector decomposition method is used to determine the mass deviation coordinates.

5. The air compressor rotor testing method of claim 1, wherein, The initial counterweight adjustment scheme is generated based on the mass deviation coordinates, and the initial counterweight parameters of the initial counterweight adjustment scheme are optimized using a gradient descent algorithm to obtain an optimized counterweight adjustment scheme, including: Based on the mass deviation coordinates, a deviation vector is calculated to obtain an initial mass deviation distribution; Based on the initial mass deviation distribution, the counterweight mass and the counterweight position are initialized to determine the initial counterweight adjustment scheme and the initial counterweight parameters corresponding to the initial counterweight adjustment scheme; If the error analysis result of the initial counterweight parameters is greater than a preset error threshold, the counterweight mass and the counterweight position of the initial counterweight parameters are iteratively updated by a gradient descent algorithm to obtain optimized parameters; Based on the optimized parameters, the mass deviation distribution is recalculated to obtain the optimized counterweight adjustment scheme.

6. The air compressor rotor testing method of claim 1, wherein, The counterweight parameters based on the optimized counterweight adjustment scheme are simulated in a simulated environment to obtain a simulation vibration signal, including: The initial counterweight parameters are loaded in the simulated environment, and dynamic response simulation is performed to obtain vibration signal data; Based on the vibration signal data, signal extraction technology is used to extract its vibration characteristics to obtain vibration characteristic parameters; If the vibration characteristic parameters exceed a preset balance threshold, the counterweight parameters of the optimized counterweight adjustment scheme are adjusted by a gradient descent algorithm to generate new counterweight parameters; Based on the new counterweight parameters, dynamic response simulation is performed again in the simulation environment to obtain updated vibration signals; Based on the updated vibration signals, signal processing technology is used to extract new vibration characteristics to determine the simulation vibration signals.

7. The air compressor rotor testing method of claim 1, wherein, If the simulation vibration signals exceed the balance threshold, a particle swarm optimization algorithm is used to iteratively update the counterweight parameters of the optimized counterweight adjustment scheme to obtain a target counterweight scheme, including: Based on the simulation vibration signals, a Fourier transform is used to perform frequency domain decomposition to obtain frequency domain vibration features; If the amplitude of the frequency domain vibration features exceeds the balance threshold, a signal processing algorithm is used to denoise the frequency domain vibration features to obtain smoothed vibration features; Based on the smoothed vibration features, a particle swarm optimization algorithm is used to iteratively update the counterweight parameters of the optimized counterweight adjustment scheme to generate the target counterweight scheme.

8. An air compressor rotor testing apparatus, characterized by, including: An acquisition module is configured to acquire vibration signal data of a rotor at multiple rotation speeds, and a fast Fourier transform algorithm is used to analyze the vibration signal data to obtain vibration frequency and phase distribution; An analysis module is configured to analyze the phase distribution by weighting if the vibration frequency exceeds a preset frequency threshold to determine the initial position coordinates of the mass distribution imbalance; a fitting module configured to fit the offset of the rotor based on the initial position coordinates to obtain mass deviation coordinates by using a least square method; an optimization module configured to generate an initial counterweight adjustment scheme based on the mass deviation coordinates, and to optimize initial counterweight parameters of the initial counterweight adjustment scheme by using a gradient descent algorithm to obtain an optimized counterweight adjustment scheme; a simulation module configured to perform dynamic response simulation on the optimized counterweight adjustment scheme in a simulation environment based on counterweight parameters of the optimized counterweight adjustment scheme to obtain a simulation vibration signal; an updating module configured to, if the simulation vibration signal exceeds a balance threshold, update the counterweight parameters of the optimized counterweight adjustment scheme by using a particle swarm optimization algorithm to obtain a target counterweight scheme; wherein the target counterweight scheme is used to be applied to the rotor for testing, so that the balance state of the rotor reaches a set standard.

Images