阅读说明：本技术 一种多传感器融合动平衡分析方法、系统、设备、介质 (Multi-sensor fusion dynamic balance analysis method, system, equipment and medium ) 是由 李兵 张新翊 郎博 于 2021-08-24 设计创作，主要内容包括：本发明属于信号处理技术领域,公开了一种多传感器融合动平衡分析方法、系统、设备、介质,所述多传感器融合动平衡分析方法包括：从工业现场采集包含水平和垂直方向上的原始振动信号；通过幅值、相似性信息确定待分析信号源；借助目标经验模式分解对原始信号进行分解；根据分解后的结果与原信号间的相似性确定盲源分离待分析信号；采用目标盲源分离方法实现信号的分解；结合盲源分析结果识别动平衡成分；通过傅里叶变换计算动平衡分量所在盲源分离信号的一倍频幅值和相位信息；采用影响系数法进行动平衡配重计算,获得动平衡计算结果。本发明能有效去除掺杂在动平衡分析频率中的其他故障信息干扰,从本质上实现设备的动平衡,延长动平衡后设备的使用时间。(The invention belongs to the technical field of signal processing and discloses a multi-sensor fusion dynamic balance analysis method, a system, equipment and a medium, wherein the multi-sensor fusion dynamic balance analysis method comprises the following steps: acquiring an original vibration signal comprising horizontal and vertical directions from an industrial field; determining a signal source to be analyzed through the amplitude and the similarity information; decomposing the original signal by means of target empirical mode decomposition; determining a blind source to separate a signal to be analyzed according to the similarity between the decomposed result and the original signal; decomposing the signal by adopting a target blind source separation method; identifying dynamic balance components by combining blind source analysis results; calculating frequency multiplication amplitude and phase information of a blind source separation signal where the dynamic balance component is located through Fourier transform; and calculating the dynamic balance weight by adopting an influence coefficient method to obtain a dynamic balance calculation result. The invention can effectively remove the interference of other fault information doped in the dynamic balance analysis frequency, essentially realizes the dynamic balance of the equipment and prolongs the service time of the equipment after the dynamic balance.)

1. A multi-sensor fusion dynamic balance analysis method is characterized by comprising the following steps:

acquiring original vibration signals in horizontal and vertical directions from an industrial field or a laboratory bench;

determining a signal source to be analyzed through the amplitude and the similarity information;

decomposing the original signal by means of a set empirical mode decomposition method;

step four, determining a blind source to separate the signals to be analyzed according to the similarity between the decomposed result and the original signals;

step five, decomposing the signals by adopting a target blind source separation method;

step six, identifying dynamic balance components by combining blind source analysis results;

step seven, combining the key phase signal information, and calculating a frequency multiplication amplitude value and phase information of the blind source separation signal where the dynamic balance component is located through Fourier transform;

and step eight, calculating the dynamic balance weight by adopting an influence coefficient method, and finally obtaining a dynamic balance calculation result.

2. The multi-sensor fusion dynamic balance analysis method of claim 1, wherein in the second step, the determining of the signal source to be analyzed comprises:

determining the amplitude and the position of a plurality of groups of sensors which are vertically arranged; sorting according to the magnitude of the signal amplitude, and selecting the channel sensor information with the maximum amplitude to perform experimental analysis;

wherein, the selecting the channel sensor information with the maximum amplitude value to perform experimental analysis comprises the following steps: and selecting the first several mutually perpendicular sensor signals in the plurality of groups of channel sensors with the maximum amplitude value to perform dynamic balance analysis.

3. The multi-sensor fusion dynamic balance analysis method of claim 1, wherein in step three, the target empirical mode decomposition is a collective empirical mode decomposition.

4. The multi-sensor fusion dynamic balance analysis method according to claim 1, wherein in step five, the target blind source separation method is non-negative matrix factorization;

wherein, before the blind source separation, the number of the fault sources needs to be determined, which comprises:

and (3) decomposing the determined analysis signal source by using singular values, and selecting the analysis result with the maximum proportional relation as the number of fault sources according to the proportional relation between adjacent singular values.

5. The multi-sensor fusion dynamic balance analysis method according to claim 1, wherein in step six, the determination rule of the dynamic balance component comprises:

the dynamic balance is mainly concentrated on a frequency doubling, so that the magnitude of a frequency doubling amplitude, the frequency doubling ratio and whether other frequency doubling components exist in the blind source separation result are used as the determination standard of the dynamic balance component; the determination principle is determined according to the proportional relation between two frequency doubling, multiple frequency doubling components and the magnitude of a frequency doubling amplitude.

6. The multi-sensor fusion dynamic balance analysis method of claim 1, wherein in step seven, the calculating a frequency doubling amplitude and phase information of the blind source separation signal in combination with the key phase signal information comprises:

adding a balance weight, and repeating the steps from the first step to the seventh step according to the original vibration amplitude and phase information acquisition method to obtain signal first frequency multiplication amplitude and phase information under the trial weight working condition.

7. A multi-sensor fusion dynamic balance analysis system for implementing the multi-sensor fusion dynamic balance analysis method according to any one of claims 1 to 6, the multi-sensor fusion dynamic balance analysis system comprising:

the system comprises an original vibration signal acquisition module, a vibration signal acquisition module and a vibration signal acquisition module, wherein the original vibration signal acquisition module is used for acquiring an original vibration signal in the horizontal and vertical directions from an industrial field or a laboratory bench;

the signal source to be analyzed determining module is used for determining a signal source to be analyzed through the amplitude and the similarity information;

the original signal decomposition module is used for decomposing the original signal by means of ensemble empirical mode decomposition;

the signal to be analyzed determining module is used for determining a blind source to separate the signal to be analyzed according to the similarity between the decomposed result and the original signal;

the signal decomposition module is used for realizing the decomposition of the signal by adopting a target blind source separation method;

the dynamic balance component identification module is used for identifying dynamic balance components by combining blind source analysis results;

the blind source separation signal calculation module is used for calculating a frequency multiplication amplitude value and phase information of the blind source separation signal where the dynamic balance component is located through Fourier transform in combination with the key phase signal information;

and the dynamic balance weight calculation module is used for calculating the dynamic balance weight by adopting an influence coefficient method and finally obtaining a dynamic balance calculation result.

8. A computer device, characterized in that the computer device comprises a memory and a processor, the memory storing a computer program which, when executed by the processor, causes the processor to carry out the steps of:

acquiring original vibration signals in horizontal and vertical directions from an industrial field or a laboratory bench; determining a signal source to be analyzed through the amplitude and the similarity information; decomposing the original signal by means of target empirical mode decomposition; determining a blind source to separate a signal to be analyzed according to the similarity between the decomposed result and the original signal; decomposing the signal by adopting a target blind source separation method; identifying dynamic balance components by combining blind source analysis results; calculating a frequency doubling amplitude and phase information of the blind source separation signal where the dynamic balance component is located through Fourier transform in combination with the key phase signal information; and calculating the dynamic balance weight by adopting an influence coefficient method, and finally obtaining a dynamic balance calculation result.

9. A computer-readable storage medium storing a computer program which, when executed by a processor, causes the processor to perform the steps of:

acquiring original vibration signals in horizontal and vertical directions from an industrial field or a laboratory bench; determining a signal source to be analyzed through the amplitude and the similarity information; decomposing the original signal by means of target empirical mode decomposition; determining a blind source to separate a signal to be analyzed according to the similarity between the decomposed result and the original signal; decomposing the signal by adopting a target blind source separation method; identifying dynamic balance components by combining blind source analysis results; calculating a frequency doubling amplitude and phase information of the blind source separation signal where the dynamic balance component is located through Fourier transform in combination with the key phase signal information; and calculating the dynamic balance weight by adopting an influence coefficient method, and finally obtaining a dynamic balance calculation result.

10. An information data processing terminal, characterized in that the information data processing terminal is used for implementing the multi-sensor fusion dynamic balance analysis system according to claim 7.

Technical Field

The invention belongs to the technical field of signal processing, and particularly relates to a multi-sensor fusion dynamic balance analysis method, system, equipment and medium.

Background

In recent years, large-scale rotating machinery is rapidly developing in the directions of high efficiency, high power, high precision, automation and the like. As a key component of a rotating device, the operating state of the rotor is directly related to the system performance and the service life of the rotating device. Various faults are inevitably generated in the long-term running process of the rotor due to factors such as poor assembly, alternating load influence, thermal deformation and the like, and further vibration and noise of mechanical equipment are caused, and even mechanism damage is generated. Imbalance, misalignment, oil whirl, etc. are common failure modes of rotary machines. Wherein the inertial forces and moments generated by rotor imbalance are among the important influencing factors generated by the vibrations of the rotating machine. The data show that 70% of the vibration failures of rotating machines result from rotor system imbalance.

At present, the field dynamic balance technology is a common method for solving the original rotor unbalance, and an influence coefficient method and a modal analysis method are main means for field dynamic balance. Unbalanced rotors are usually present in the vibration signal as a fundamental component, and reducing the fundamental vibration becomes the ultimate goal of eliminating the imbalance. However, not only the imbalance fault is related to the fundamental frequency component, but also the mechanical looseness, the rubbing, the oil whirl, the pipeline excitation and the like all affect the fundamental frequency, and the blind dynamic balance analysis cannot essentially solve the problem of the vibration of the rotor system, and further cause the composite faults of the rubbing of the rotor, the foundation looseness and the like. Sometimes, although the vibration of the system is reduced by dynamic balance, the vibration is increased after a short time of operation because a vibration source is not found.

Therefore, under the condition that the starting and stopping times of the rotating equipment are not increased, the vibration of the rotor system is traced, a rapid, convenient and accurate dynamic balance analysis method is developed, the vibration caused by unbalance is effectively separated from other vibrations, the unbalance influence is fundamentally removed, and the method is very necessary for reducing the system vibration and improving the running quality of the rotor.

Through the above analysis, the problems and defects of the prior art are as follows:

(1) in practical application, not only imbalance faults are related to fundamental frequency components, but also mechanical loosening, rubbing, oil film vortex, pipeline excitation and the like all affect the fundamental frequency, and blind dynamic balance analysis cannot essentially solve the problem of vibration of a rotor system, and further cause composite faults such as rotor rubbing, foundation loosening and the like.

(2) In the field dynamic balance technology, although the dynamic balance reduces the vibration of the system, the vibration is increased after the transient operation because no vibration source is found.

The difficulty in solving the above problems and defects is:

(1) the coupling fault source is effectively extracted and separated out basically based on the existing signal analysis source, and is a necessary way for further fault elimination, however, different fault sources sometimes show similar fault characteristics.

(2) The response results of signal sources to rotor faults are different, a single signal source and an underdetermined analysis method cannot comprehensively and comprehensively cover all fault information, and the separation of coupling faults by the underdetermined signal source is always an important research direction.

The significance of solving the problems and the defects is as follows: the invention effectively combines the prior empirical mode decomposition method and the blind source separation method by means of multi-sensor fusion information, realizes the identification and separation of different types of coupling rotor faults, and prevents misjudgment and misdiagnosis caused by superposition faults. The extraction of the dynamic balance effective component is realized essentially, and the stripping and confirmation of the inherent unbalance component are realized. Meanwhile, after comprehensively evaluating the effect of various faults, engineering technicians reduce and eliminate various faults according to symptoms and medicament, effectively ensure long-time safe operation of the unit, and further provide powerful support for cost reduction and efficiency improvement of production enterprises.

Disclosure of Invention

Aiming at the problems in the prior art, the invention provides a multi-sensor fusion dynamic balance analysis method, a system, equipment and a medium, and particularly relates to a multi-sensor fusion dynamic balance analysis method, a system, equipment and a medium based on blind source separation for rotating equipment.

The invention is realized in such a way that a multi-sensor fusion dynamic balance analysis method comprises the following steps:

step one, acquiring original vibration signals in horizontal and vertical directions from an industrial field or a laboratory bench synchronously, and acquiring the information of the vibration fault source of the unit to the maximum extent.

Determining a signal source to be analyzed through the amplitude and the similarity information; the influence of analysis efficiency and analysis quality caused by information processing of multiple groups of sensors on algorithm effectiveness is reduced while the fault information amount is improved.

Decomposing the original signal by means of target empirical mode decomposition; useful information with strong correlation with the unbalanced fault is extracted, and interference of other fault components is abandoned.

Step four, determining a blind source to separate the signals to be analyzed according to the similarity between the decomposed result and the original signals; the influence of analysis efficiency and analysis quality caused by a plurality of groups of mode decomposition results on algorithm effectiveness is mainly reduced. Step five, decomposing the signals by adopting a target blind source separation method; different fault coupling components are decomposed into different signals for subsequent analysis.

Step six, identifying dynamic balance components by combining blind source analysis results; and separation and identification of different fault coupling information are realized.

Step seven, combining the key phase signal information, and calculating a frequency multiplication amplitude value and phase information of the blind source separation signal where the dynamic balance component is located through Fourier transform; effective input information is provided for dynamic balance analysis.

And step eight, calculating the dynamic balance weight by adopting an influence coefficient method, and finally obtaining a dynamic balance calculation result. Essentially achieving the final dynamic balance analysis

Further, in step two, the determining of the signal source to be analyzed includes:

determining the amplitude and the position of a plurality of groups of sensors which are vertically arranged; and sequencing according to the amplitude of the signal, and selecting the channel sensor information with the maximum amplitude to perform experimental analysis.

Wherein, the selecting the channel sensor information with the maximum amplitude value to perform experimental analysis comprises the following steps: and selecting the first several mutually perpendicular sensor signals in the plurality of groups of channel sensors with the maximum amplitude value to perform dynamic balance analysis.

Further, in step three, the target empirical mode decomposition is, but not limited to, a set empirical mode decomposition.

Further, in step five, the target blind source separation method is, but not limited to, non-negative matrix factorization;

wherein, before the blind source separation, the number of the fault sources needs to be determined, which comprises:

and (3) decomposing the determined analysis signal source by using singular values, and selecting the analysis result with the maximum proportional relation as the number of fault sources according to the proportional relation between adjacent singular values.

Further, in step six, the determination principle of the dynamic balance component includes:

the dynamic balance is mainly concentrated on a frequency doubling, so that the magnitude of a frequency doubling amplitude, the frequency doubling ratio and whether other frequency doubling components exist in the blind source separation result are used as the determination standard of the dynamic balance component; the determination principle is determined according to the proportional relation between two frequency doubling, multiple frequency doubling components and the magnitude of a frequency doubling amplitude.

Further, in step seven, the calculating a frequency multiplication amplitude and phase information of the blind source separation signal by combining the key phase signal information includes:

adding a balance weight, and repeating the steps from the first step to the seventh step according to the original vibration amplitude and phase information acquisition method to obtain signal first frequency multiplication amplitude and phase information under the trial weight working condition.

Another object of the present invention is to provide a multi-sensor fusion dynamic balance analysis system using the multi-sensor fusion dynamic balance analysis method, the multi-sensor fusion dynamic balance analysis system including:

the system comprises an original vibration signal acquisition module, a vibration signal acquisition module and a vibration signal acquisition module, wherein the original vibration signal acquisition module is used for acquiring an original vibration signal in the horizontal and vertical directions from an industrial field or a laboratory bench;

the signal source to be analyzed determining module is used for determining a signal source to be analyzed through the amplitude and the similarity information;

the original signal decomposition module is used for decomposing the original signal by means of target empirical mode decomposition;

the signal to be analyzed determining module is used for determining a blind source to separate the signal to be analyzed according to the similarity between the decomposed result and the original signal;

the signal decomposition module is used for realizing the decomposition of the signal by adopting a target blind source separation method;

the dynamic balance component identification module is used for identifying dynamic balance components by combining blind source analysis results;

the signal calculation module for the dynamic balance component is used for calculating a frequency multiplication amplitude value and phase information of the blind source separation signal of the dynamic balance component by combining the key phase signal information and through Fourier transform;

and the dynamic balance weight calculation module is used for calculating the dynamic balance weight by adopting an influence coefficient method and finally obtaining a dynamic balance calculation result.

It is a further object of the invention to provide a computer device comprising a memory and a processor, the memory storing a computer program which, when executed by the processor, causes the processor to perform the steps of:

acquiring original vibration signals in horizontal and vertical directions from an industrial field or a laboratory bench; determining a signal source to be analyzed through the amplitude and the similarity information; decomposing the original signal by means of target empirical mode decomposition; determining a blind source to separate a signal to be analyzed according to the similarity between the decomposed result and the original signal; decomposing the signal by adopting a target blind source separation method; identifying dynamic balance components by combining blind source analysis results; calculating a frequency doubling amplitude and phase information of the blind source separation signal where the dynamic balance component is located through Fourier transform in combination with the key phase signal information; and calculating the dynamic balance weight by adopting an influence coefficient method, and finally obtaining a dynamic balance calculation result.

It is another object of the present invention to provide a computer-readable storage medium storing a computer program which, when executed by a processor, causes the processor to perform the steps of:

acquiring original vibration signals in horizontal and vertical directions from an industrial field or a laboratory bench; determining a signal source to be analyzed through the amplitude and the similarity information; decomposing the original signal by means of target empirical mode decomposition; determining a blind source to separate a signal to be analyzed according to the similarity between the decomposed result and the original signal; decomposing the signal by adopting a target blind source separation method; identifying dynamic balance components by combining blind source analysis results; calculating a frequency doubling amplitude and phase information of the blind source separation signal where the dynamic balance component is located through Fourier transform in combination with the key phase signal information; and calculating the dynamic balance weight by adopting an influence coefficient method, and finally obtaining a dynamic balance calculation result.

Another object of the present invention is to provide an information data processing terminal for implementing the multi-sensor fusion dynamic balance analysis system.

By combining all the technical schemes, the invention has the advantages and positive effects that: the invention provides a multi-sensor fusion dynamic balance analysis method, in particular to a dynamic balance analysis method for rotating equipment based on blind source separation, which abandons the error zone that engineering technicians directly regard fundamental frequency components as dynamic balance components, and the method uses the fundamental frequency components as comprehensive frequency doubling components after different faults are superposed for demonstration and solution, can effectively remove other fault information interferences doped in dynamic balance analysis frequencies under the condition of not additionally increasing the starting and stopping times, essentially realizes the dynamic balance of the equipment, prolongs the service time of the equipment after the dynamic balance, effectively strips other fault components except the dynamic balance, and enables the engineering technicians to purposefully adopt different and targeted maintenance strategies aiming at different faults. Unnecessary starting caused by misdiagnosis and misjudgment is reduced, so that the diagnosis effect is improved, and the auxiliary equipment can work stably for a long time.

In order to effectively analyze the vibration source, accurately extract the rotor unbalance component in the vibration signal from a plurality of groups of sensors and further find the root cause of the fault, the invention provides a multi-sensor fusion dynamic balance analysis method based on blind source separation. The method comprises the steps of firstly collecting original vibration signals of a plurality of sensors to be detected under a stable rotating speed, analyzing the original signals based on ensemble empirical mode decomposition to obtain preprocessed signals, determining the number of fault sources by using a singular value decomposition method, then performing fault mode separation on the signals by using a blind source separation model, determining a dynamic balance component according to a frequency doubling amplitude ratio, finally obtaining vibration components caused by single dynamic balance influence, meanwhile, determining the generation reasons of other faults according to the obtained fault characteristics of other signals, tracing back to the source, eliminating fault interference and ensuring the long-period safe operation of production.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the embodiments of the present invention will be briefly described below, and it is obvious that the drawings described below are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.

Fig. 1 is a flowchart of a multi-sensor fusion dynamic balance analysis method according to an embodiment of the present invention.

Fig. 2 is a schematic diagram of a multi-sensor fusion dynamic balance analysis method provided by the embodiment of the invention.

FIG. 3 is a block diagram of a multi-sensor fusion dynamic balance analysis system according to an embodiment of the present invention;

in the figure: 1. an original vibration signal acquisition module; 2. a signal source determination module to be analyzed; 3. an original signal decomposition module; 4. a module for determining a signal to be analyzed; 5. a signal decomposition module; 6. a dynamic balance component identification module; 7. a blind source separation signal calculation module; 8. and a dynamic balance weight calculation module.

Fig. 4 is a schematic view of a rotor testing table used in a verification test provided by an embodiment of the present invention.

Fig. 5 is a schematic diagram of a blind source separation processing result according to an embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

Aiming at the problems in the prior art, the invention provides a multi-sensor fusion dynamic balance analysis method, a system, equipment and a medium, and the invention is described in detail with reference to the attached drawings.

As shown in fig. 1, the multi-sensor fusion dynamic balance analysis method provided by the embodiment of the present invention includes the following steps:

s101, acquiring original vibration signals in horizontal and vertical directions from an industrial field or a laboratory bench;

s102, determining a signal source to be analyzed through amplitude and similarity information;

s103, decomposing the original signal by means of target empirical mode decomposition;

s104, determining a blind source to separate the signals to be analyzed according to the similarity between the decomposed result and the original signals;

s105, decomposing the signals by adopting a target blind source separation method;

s106, identifying dynamic balance components by combining blind source analysis results;

s107, calculating a frequency multiplication amplitude value and phase information of the blind source separation signal where the dynamic balance component is located through Fourier transform by combining the key phase signal information;

and S108, calculating the dynamic balance weight by adopting an influence coefficient method, and finally obtaining a dynamic balance calculation result.

A schematic diagram of a multi-sensor fusion dynamic balance analysis method provided by the embodiment of the invention is shown in fig. 2.

As shown in fig. 3, the multi-sensor fusion dynamic balance analysis system provided by the embodiment of the present invention includes:

the system comprises an original vibration signal acquisition module 1, a vibration signal acquisition module and a vibration signal acquisition module, wherein the original vibration signal acquisition module is used for acquiring original vibration signals in the horizontal and vertical directions in a unit operation state from an industrial field or a laboratory bench;

the signal source to be analyzed determining module 2 is used for determining a signal source to be analyzed through the amplitude and the similarity information;

the original signal decomposition module 3 is used for decomposing the original signal by means of target empirical mode decomposition;

the signal to be analyzed determining module 4 is used for determining a blind source to separate the signal to be analyzed according to the similarity between the decomposed result and the original signal;

the signal decomposition module 5 is used for realizing the decomposition of the signal by adopting a target blind source separation method;

the dynamic balance component identification module 6 is used for identifying dynamic balance components by combining blind source analysis results;

the blind source separation signal calculation module 7 is used for calculating a frequency multiplication amplitude value and phase information of the blind source separation signal where the dynamic balance component is located through Fourier transform in combination with the key phase signal information;

and the dynamic balance weight calculating module 8 is used for calculating the dynamic balance weight by adopting an influence coefficient method and finally obtaining a dynamic balance calculating result.

The technical solution of the present invention is further described below with reference to specific examples.

In order to effectively analyze the vibration source, accurately extract the rotor imbalance component in the vibration signal from the multiple groups of sensors and further find the root cause of the fault, the embodiment provides the multi-sensor fusion dynamic balance analysis method based on blind source separation, the balance error caused by the incompleteness of single sensor information is reduced through the multi-sensor information fusion technology, the balance effect and the continuity of dynamic balance are improved, and meanwhile, the faults in other forms are judged by combining the separation result. During specific implementation, a plurality of original vibration signals of the sensor to be detected under a stable rotating speed are collected, the original signals are analyzed based on ensemble empirical mode decomposition to obtain preprocessed signals, the number of fault sources is determined by using a singular value decomposition method, then a blind source separation model is adopted to perform fault mode separation on the signals, dynamic balance components are determined according to a frequency multiplication amplitude ratio, vibration components caused by single dynamic balance influence are finally obtained, meanwhile, the generation reasons of other faults are determined according to the fault characteristics of the obtained other signals, root tracing is conducted, fault interference is eliminated, and long-period safe operation of production is ensured.

In order to verify the experimental effect, the dynamic balance of the rotor is completed on a specific experiment table, and the experiment table not only meets the requirements of dynamic balance experiments, but also provides implementation of constant radial force, so as to simulate the fault of 'pipeline excitation', and form the superposition of the unbalance fault and the fault of 'pipeline excitation'. The experiment was divided into two parts: the method has two working conditions of no radial force and radial force so as to simulate two faults of single unbalanced pipeline excitation and unbalanced pipeline excitation composite fault. Completing single unbalance fault according to a conventional dynamic balance method without blind source separation and other analysis, and recording the mass and the orientation of the added balance weight as m₀∠p₀It is used as reference alignment information. Then adding a certain radial force in a certain direction, respectively calculating the direct dynamic balance and the size and the direction of the counterweight proposed by the invention, and sequentially recording as m₁∠p₁And m₂∠p₂. The size and the direction m of the counterweight extracted by the invention₂∠p₂And m₀∠p₀And comparing to verify the analysis effect of the invention.

A rotor experiment table used in the verification experiment provided by the embodiment of the invention is shown in fig. 4.

Step (S1): and synchronously acquiring signals of the rotor experiment table, wherein the signals comprise four paths of eddy current vibration signals and a path of key phase signal. Without loss of generality, the signals in this example were acquired and implemented by a standard Bentley rotor bench.

Step (S2): and acquiring data information of two groups of sensors in mutually perpendicular directions. Table 1 shows the raw vibration pass frequency amplitude (under additional radial force) for each sensor.

TABLE 1 passband values extracted for four sets of original signals

Measuring point	Channel 1	Channel 2	Channel 3	Channel 4
					Value of pass frequency	15.3	22.1	53.4	36.4

Step (S3): and determining a signal to be analyzed according to a signal selection principle, and determining a sensor for trial analysis.

Step (S4): and processing the signal to be analyzed by using set modal decomposition, performing singular value SVD on the decomposition result, and calculating the number of fault sources according to a fault source number determination method.

Step (S5): and the blind source separation realizes the separation of the main components of the signals of the multiple sensors, and identifies and extracts the signals of the dynamic balance.

Step (S6): and intercepting vibration signals of not less than 3 complete cycles by combining the key phase information, performing order ratio analysis, and obtaining accurate amplitude and phase information by adopting Fourier transform.

Step (S7): adding the test weights, and repeating the steps (S1-S6) to calculate the accurate amplitude and phase information after the test weights are added after the rotating speed reaches the working rotating speed, wherein the part (S3) of sensor selection is directly consistent with the previous step.

Step (S8): and calculating the size and the position of the added balance weight by an influence coefficient method, comparing the dynamic balance weight results under the condition of no additional radial force, calculating errors and analyzing error sources.

The blind source separation processing result provided by the embodiment of the invention is shown in fig. 5.

TABLE 2 results of modal component singular value decomposition with number of eigenvalues

Feature(s)	Value of	Feature(s)	Value of
				λ1	28273.32	λ4	3459.35
λ2	28106.73	λ5	2187.63
				λ3	4004.33	λ6	2182.94

TABLE 3 results of conventional dynamic balance vibration analysis without additional radial force

Without additional force	Vibration amplitude/mum	Vibration phase/° c
			Original vibration	104.68	340
Test weight lift	50.82	344
			Balance vehicle	25.5	55

TABLE 4 results of conventional dynamic balance vibration analysis under additional radial force

With additional force	Vibration amplitude/mum	Vibration phase/° c
			Original vibration	46.04	24
Test weight lift	72.44	23
			Balance vehicle	8.91	127

TABLE 5 influence coefficient method equilibrium vector calculation without additional force

TABLE 6 influence coefficient method equilibrium vector calculation with additional force

TABLE 7 dynamic balance calculation results of the present invention with additional radial force

In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When used in whole or in part, can be implemented in a computer program product that includes one or more computer instructions. When loaded or executed on a computer, cause the flow or functions according to embodiments of the invention to occur, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, the computer instructions may be transmitted from one website site, computer, server, or data center to another website site, computer, server, or data center via wire (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL), or wireless (e.g., infrared, wireless, microwave, etc.)). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that includes one or more of the available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., Solid State Disk (SSD)), among others.

The above description is only for the purpose of illustrating the present invention and the appended claims are not to be construed as limiting the scope of the invention, which is intended to cover all modifications, equivalents and improvements that are within the spirit and scope of the invention as defined by the appended claims.