CN109323853B - Fan blade fan vibration detection device and fault detection prediction method - Google Patents
Fan blade fan vibration detection device and fault detection prediction method Download PDFInfo
- Publication number
- CN109323853B CN109323853B CN201811148445.9A CN201811148445A CN109323853B CN 109323853 B CN109323853 B CN 109323853B CN 201811148445 A CN201811148445 A CN 201811148445A CN 109323853 B CN109323853 B CN 109323853B
- Authority
- CN
- China
- Prior art keywords
- fan
- vibration
- model
- connecting shaft
- blade
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M13/00—Testing of machine parts
Abstract
The invention provides a fan blade vibration detection and fault prediction system, wherein a fan rotates around a connecting shaft at a certain rotating speed in a circular motion manner, and the system comprises a fan blade vibration detection device serving as a hardware part and a fan fault prediction method serving as a software part.
Description
Technical Field
The invention relates to the field of fan mechanical fault diagnosis, in particular to a fan blade vibration detection and fault prediction system.
Background
The fan is a power device for providing air flow, the fan forms the air flow through the fan blade, the fan blade can generate mechanical vibration in the rotating process, if the mechanical vibration is too large, parts such as an internal connecting shaft and the like can generate cracks, and the cracks are enlarged to cause irreversible damage, so that the whole fan is failed or even scrapped. Therefore, how to carry out preventive detection and judge the crack damage condition of mechanical parts in the fan so as to carry out maintenance and replacement in advance is a key and difficult problem for researching the operation stability of the fan. There are many ways to do this, with reference to other similar information in the country. For example: a vibration method research [ J ] of fan blade fault prediction of Beijing Union university (Zhang BaoQin, Lei Baozhen, etc.; electronic measurement and instrumental report, 2014,28(3):285 plus 291) provides a fan blade fault prediction method based on fan acceleration network model data fusion, reflects the integral operation mode of a fan by a three-dimensional vector decomposition, conversion and extraction technology, designs a vibration monitoring system of a digital signal processing technology, performs time domain and frequency domain analysis on low-frequency or ultra-low-frequency signals with amplitude exceeding limit, identifies that the fault rate of the fan caused by vibration in a low-frequency region of 4-30 Hz is the highest, and successfully realizes the advance prediction of 75% of fan blade mechanical fault rate. Wuhan university (Zhouquan. direct air-cooling fan bridge vibration research [ D ]. doctor's university thesis of Wuhan university, 2013) provides design suggestions and damping suggestion schemes of a fan disturbance force calculation method on the basis of comprehensively analyzing actual measurement data of a fan bridge vibration field and finite element numerical simulation analysis results. The application of Wuhan university (Shenluo, Hakkeming, combined prediction in fan vibration prediction [ J ]. Chinese Equipment engineering, 2006,9(3):43-45) proposes that a combined prediction method is introduced into the state prediction of fan measuring point vibration intensity, and the prediction results of a combined prediction model and other single prediction models are compared, which shows that the vibration prediction by adopting the combined prediction method has higher precision. The university of Wuhan's rational engineering (courage of couples, research on fan vibration noise analysis and numerical simulation research [ D ]. Master's graduate thesis of Wuhan's rational engineering, 2007) proposes that the limited volume method is used for simulating the internal flow field of the fan, the finite element method is used for analyzing and calculating the vibration of the fan and the shell, and the finite element method and the boundary element method are combined for predicting and numerically simulating the structural noise of the fan shell. The failure prediction method of the main components of the wind turbine based on regression analysis of the university of north china power [ D ] a university of north china power graduate paper, 2013) proposes another new failure prediction method based on the houtt control graph algorithm. According to the method, an autoregressive moving average (ARMA) model is established under the normal operation state of a monitored component, and the ARMA model temperature prediction residual is used as input information for main components of a transmission system of a wind turbine generator. An improved negative selection algorithm is provided by the northeast electric power university (Wuchuming, silver Hairzel. blower vibration fault diagnosis method for improving negative selection algorithm [ J ]. Chinese mechanical engineering, 2016,4: 479-. The Mahalanobis distance is introduced into a traditional negative selection algorithm to carry out primary screening on vibration data, and the algorithm is applied to prediction of vibration faults of the wind turbine generator. China university of science and technology (Lelinling, Huangqibai. Fan blade noise model research [ J ]. Mechanical engineering journal, 2004,40(7): 114-. And simulating the change condition of the blade noise under the condition that the noise model parameters change, and obtaining main factors influencing the fan noise. New method for predicting fan surge boundary [ J ] fan technology, 2005(4):52-53) of northwest university of industry (Lexijiayu, Liangff et al) proposes to predict fan surge boundary using shaft vibration, shaft displacement and fan noise signal and verifies its feasibility in actual engineering. Chongqing university (Xuguanji, Berlin. research on wind turbine noise prediction based on multiple linear regression analysis [ J ]. Chinese test, 2010,36(5):21-23) proposes that the sound pressure level of the noise of a wind turbine generator set is predicted by adopting a multiple linear regression analysis method, a regression equation is established through measurement of parameters which are easy to realize, a regression coefficient is solved by using a least square method, and finally a model for predicting the sound pressure level of the noise of the wind turbine generator set is established. Yanshan university (Leili. residual life prediction research of fan mechanical transmission components based on extreme learning machine [ D ]. Yanshan university Master graduate 2015) provides that a principal component analysis method and an extreme learning machine algorithm are used for performing prediction analysis on the operation trend and residual life of components, and a bearing component in a mechanical transmission component is used as a research object for experimental verification, so that a certain basis is provided for trend prediction and residual life prediction of the mechanical transmission component of a wind driven generator. Ningxia university (Wangchao, Wangchunxiu, etc.. Fan Low speed input shaft bearing fatigue life analysis [ J ].2017,6(12):43-48) proposes that the influence of bearing radial play and lubrication on the fan low speed input shaft bearing fatigue life is analyzed based on Romax Designer simulation, and discusses a fan low speed input shaft bearing fatigue life prediction formula considering the radial play. Shanghai institute of electrical machinery (the long-nosed. fan blade crack damage state identification technology research [ D ]. shanghai institute of master graduate, 2016) proposed a dual-spectrum analysis method of characteristic modal functions to extract crack signals. A fan blade finite element analysis [ J ] based on ANSYS is proposed by Shandong science and technology university (Liu, Zheng Wei and so on, 2017,3(18):61-64) to carry out vibration mode and static force analysis on the blade by using ANSYS finite element calculation software, calculate the natural frequency, the vibration pattern diagram of each order, the stress and the strain cloud diagram of the blade, and provide an analysis method for avoiding the resonance of the blade and prolonging the service life of the blade. The university of Anhui engineering (Xinjianzhi, desulfurization and dust removal controllable fan structural design and performance prediction research [ D ]. Anhui engineering university Master thesis, 2014) proposes a method combining numerical flow field simulation and test optimization, analyzes the influence of the structural parameters of a fan impeller for desulfurization and dust removal on the performance based on the flow field, designs orthogonal test structure test samples, optimizes and researches the structural parameters of the fan for desulfurization and dust removal by using a complex shape method and a genetic algorithm on the basis of the rule that the influence of the change of the structural parameters of the impeller on the performance of the fan is researched by using a range analysis method, and finds out an optimal structural parameter combination.
In the operation process of the fan, the moving parts such as the connecting shaft of the fan can inevitably vibrate due to installation gaps and mechanical friction, the vibration can generate energy, and along with the increase of vibration amplitude and frequency, the energy can be accumulated to a certain degree and break through a critical point, so that the mechanical damage of the moving parts, such as cracks of the connecting shaft, can be caused. Because the mechanical damage of the moving part can not be visually identified in the earlier stage, a device for detecting mechanical vibration in real time and a corresponding mechanical damage prediction method which are suitable for a fan need to be adopted, and the fan failure risk is reduced.
The fan rotates the blades connected with the shaft around the connecting shaft in a circular motion at a certain rotating speed under the action of electromagnetism. The fan blade fan can generate mechanical vibration when rotating, if the vibration amplitude is within a certain range, the vibration energy can be consumed by the metal resistance of the fan blade fan, but the vibration amplitude is increased along with the aggravation of the abrasion of each mechanical connecting piece in the fan blade fan, once the vibration energy exceeds the critical point of the metal resistance, a part in the fan blade fan, such as a connecting shaft, can generate cracks, until the moving part of the fan is completely broken along with the aggravation of the cracks, and finally the fan blade fan is irreversibly damaged, so that the whole fan is failed or even scrapped.
The fan directly influences the circular flow of air in the fan once the fan breaks down, and equipment is stopped. Therefore, real-time sampling of fan vibration signals is required to detect fan blade fan vibration and pre-treat fan inner connecting shaft fission and anomaly so as to take preventive maintenance. In the real-time sampling process of the fan vibration signals, the fan vibration signals are found to be in a nonlinear relation, the fan working process has the characteristic of complex motion, and the diversity of judgment results is determined under different working conditions, so that analysis and prediction cannot be performed, and an applicable fault detection and prediction system must be found.
Disclosure of Invention
In order to solve the above problems, the present invention provides a fan blade vibration detection and failure prediction system, wherein a fan rotates in a circular motion around a connecting shaft at a certain rotation speed, and the system is characterized in that: the system comprises a fan blade vibration detection device as a hardware part and a fan fault prediction method as a software part, wherein the fan blade vibration detection device comprises:
the vibration sensor is arranged on the connecting shaft, takes the vibration mechanical quantity of the connecting shaft as input quantity, and receives the input quantity by a receiving part of the vibration mechanical quantity and converts the input quantity into electric quantity pulse output;
the signal preprocessing module is used for storing and amplifying the electric quantity pulse signal transmitted by the vibration sensor through the built-in vector processor;
the A/D conversion module converts the analog signals acquired by the signal preprocessing module into digital signals so as to carry out the analysis and processing of the next step;
and the data transmission module transmits the digital signal through a transmission medium.
The fan fault prediction method comprises the following steps:
step 1) assuming that a connecting shaft of the fan blade fan has cracks, the cracks are continuously opened and closed in the rotating process of the connecting shaft, and the rigidity of the shaft surface is caused to generate nonlinear change along with time, and then a crack opening and closing function f is defined as:
the further calculation is as follows:
wherein, in the formula II, sz1Is the size of the inner diameter of a fan connecting shaft, sz2Is the outer diameter size m of a fan connecting shaft1For the crack opening quality of the connecting shaft, m2Connecting shaft crack closing mass, t is fan running time, c is an external damping coefficient, k is bending rigidity of a rotating shaft when no crack exists, delta k is rigidity variable quantity caused when the crack is completely opened, β is an initial phase angle, omega is fan rotating speed, and f is a crack opening and closing function;
step 2), judging the crack opening and closing function f by adopting a square wave function, wherein the function is as follows:
in formula IV, tau is the dynamic displacement of the fan connecting shaft;
step 3), after dimensionless parameters are introduced into the square wave function, a leaf fan vibration model f' can be obtained, namely:
with the increase of the running time t of the fan, the external damping coefficient c is in a descending trend, and the dynamic displacement tau of the connecting shaft is in an ascending trend, so that the hidden trouble that the connecting shaft cracks is also in an ascending trend.
Preferably, when the vibration frequency generated by the fan connecting shaft is inconsistent, the blade fan vibration model f 'is in a bending mode, the target function of the model in the bending mode is discontinuous, and the blade fan vibration model f' is described by a topology optimization method, so that the original model parameters are subjected to finite element dispersion, namely:
s.t.:
β-ω0 2≤0
0＜Δk≤1
preferably, the detection sensitivity of the sensor is attenuated according to the use period, and when the detection of the sensor is in error, the attenuation of the model is automatically compensated.
Preferably, when attenuation compensation is carried out on the sensitivity of the fan sensor in the fusion mode, an adaptive complex frequency incremental model is constructed to solve the compensation problem,
the structure of the complex frequency increment model in the fusion mode of the leaf fan vibration model f' is as follows:
s.t.:
[β-(ω0 2+Δωj 2)]≤0,j＝1,2,...t
0＜ρ≤(ρ0+Δρ)≤1
β-ω0 2≤0
0＜Δk≤1
Δωt 2＝fΔρ
in formula VII, the parameter ρ represents a frequency vector, and Δ ρ represents a frequency change amount.
Preferably, the complex frequency increment model of the leaf fan vibration model f' in the bending mode can be optimized by adopting an iteration method based on characteristic frequency optimization, the optimization target can select maximization of a characteristic frequency value of a specified iteration order and maximization of a characteristic frequency value of a specified adjacent order, namely, a model structure is optimized more ideally through a convergence condition, the characteristic frequency value of complex frequency sensitivity is effectively inhibited after iterative calculation, and failure of a calculation result caused by frequency distortion is avoided.
The invention provides a software main body. The method adopts a mathematical modeling mode to establish sub-models of each layer, the models of each layer are mutually linked, each layer of model is the derivation and supplement of the model of the previous layer, and the sequence cannot be reversed.
The crack opening and closing function f provided by the invention is designed based on the physical characteristics that the cracks are opened and closed in the rotation process of the connecting shaft and the rigidity of the shaft surface is changed, and related parameters are defined in the crack opening and closing function f and serve as the basis for the derivation of other submodels of the invention.
The invention provides a blade fan vibration model f', which is obtained after dimensionless parameters are introduced, and the model is deduced to be in two modes, namely a fusion mode and a bending mode according to the consistency of the vibration frequency of a fan connecting shaft. The model changes in the two modes are different, as are the applicability.
The invention provides a leaf fan vibration model f', which is described by adopting a topological optimization method, so that the parameters of an original model are subjected to finite element dispersion, and the method restrains the parameters in the model to ensure that an objective function curve becomes continuous and smooth, thereby screening the vibration frequency data of a fan connecting shaft to a certain extent, ensuring that the vibration frequency of the fan connecting shaft is kept consistent (namely single frequency) and increasing the stability of data output.
The invention provides a leaf fan vibration model f', which adopts a topological method of incremental complex frequency to compensate the sensitivity attenuation of a vibration sensor under two modes. When the fusion model is in a fusion mode, the complex frequency incremental model can be uniformly used for solving whether the compensation of single-frequency sensitivity or the compensation of complex frequency sensitivity is carried out; when the model is in a bending mode, an iteration method based on characteristic frequency optimization is adopted, the model structure is optimized more ideally through a convergence condition, the characteristic frequency value of the complex frequency sensitivity is effectively inhibited after iterative computation, and the failure of a computation result caused by frequency distortion (also called linear distortion) is avoided.
The invention is different from the disclosed failure prediction method in that:
firstly, the fan is disassembled, and the possibility of the fan blade generating faults is indirectly predicted by detecting the vibration of a connecting shaft connected with the fan blade and predicting the possibility and the size of the cracks possibly generated by the connecting shaft, so that the fan faults are prevented; in the existing method, the fan is used as a whole to detect and judge the occurrence of the fault or directly judge the damage condition of the fan blade to judge, so that the judgment result may be one-sidedness.
Secondly, the invention adopts a reverse derivation mode, namely, after dimensionless parameters are introduced on the premise of assuming that the connecting shaft has cracks, a leaf fan vibration model is established, and the model is continuously corrected by setting and increasing partial parameters of the model, so that the method can be suitable for different working conditions and operating conditions. The existing methods disclosed in the prior art are derived from the positive direction, that is, under the condition that the fan is in good condition, the trend prediction and the residual life prediction of mechanical transmission parts, the fan vibration (resonance and amplitude) trend, the fan noise trend, the mechanical fatigue trend and other analysis methods are used for detection and prediction, and the methods are only suitable for the fan in good condition, and are not suitable for the fan which has or tends to have faults, so that the judgment result is possibly one-sidedness.
Thirdly, the applicability of the fan vibration model of the fan under different vibration frequencies is also considered, and different modes of the model are introduced because the fan has the characteristic of complex motion in the working process and the diversity of the judgment result is determined by different working conditions. By adopting a topological optimization method, parameters in the model are constrained, the stability of data output is improved, and the model prediction result is more accurate. However, the existing methods do not consider the complex motion characteristics of the fan, and do not adopt proper convergence for data output of different vibration frequencies, so that the judgment result may be one-sidedness.
Finally, the invention also takes into account the problem of attenuation of the sensitivity of the detection device, so that a leaf fan vibration model with complex frequency sensitivity attenuation in different modes (fusion mode and bending mode) is introduced, and the model can be used as a sub-model (correction model) or used as a model independently. The model is divided into different segments, the content of the segment can be modified or added under different design requirements, and a certain segment can be independently used as a sub-model (the content of the segment head definition needs to be additionally added) without damaging the whole model structure. The existing methods are designed as an integral model, and cannot be divided or a certain section of the model is modified, so that the judgment result is one-sidedness.
The invention discloses a fan blade fan vibration detection device and a fault prediction method. As a fan blade vibration detection device, the device has the characteristics of high sensitivity, accurate measurement, stable data transmission, convenient operation and the like. As a fan fault prediction method, the method has the characteristics of wide use coverage, strong data output capacity, high popularization and the like.
Drawings
Fig. 1 is a schematic view of a fan blade fan vibration detection device provided by the present invention;
FIG. 2 is a flow chart of the fan blade vibration detection device;
FIG. 3 is a fan blade vibration model;
fig. 4 is a coordinate diagram of a section of the connecting shaft.
Detailed Description
A fan blade vibration detection and fault prediction system comprises a fan blade vibration detection device and a fan fault prediction method.
1.1 Fan blade fan vibration detection device
The structure of the device is shown in figure 1, and the fan blade fan vibration detection device comprises a vibration sensor, a signal preprocessing module, an A/D conversion module and a data transmission module.
The vibration sensor is arranged on the connecting shaft, and takes the vibration mechanical quantity of the connecting shaft as input quantity, and the input quantity is received by the receiving part and converted into electric quantity pulse output. The vibration generated when the fan runs mainly comes from the connecting shaft, the vibration amplitudes generated by the connecting shafts with different outer diameter sizes (sz2) and inner diameter sizes (sz1) in a running period (t) are different (namely, dynamic displacement tau), the vibration amplitudes can affect two crack parameters (m1 and m2), and when the vibration amplitudes exist and are increased all the time, the connecting shaft can crack and expand infinitely, and finally, the crack maximum opening (delta k maximum limit value) is generated, so that the connecting shaft is completely broken. Therefore, vibration amplitude data (namely dynamic displacement tau) are collected in real time through a vibration sensor arranged on the connecting shaft, and vibration trend prediction is carried out on the fan connecting shaft through a blade fan vibration model f', so that the service cycle of the component is determined.
The signal preprocessing module carries out simple processing on the electric quantity pulse signal transmitted by the vibration sensor through a built-in vector processor, such as signal storage and signal amplification, so that signal distortion and loss are avoided, and the signal at the moment is an analog quantity.
The A/D conversion module converts the analog signal into a digital signal so as to carry out the next analysis and processing.
And the data sending module sends the digital signal through a transmission medium.
The working flow chart of the fan blade vibration detection device is shown in fig. 2.
1.2 a fan failure prediction method, the method steps are as follows:
1.2.1, step 1) assuming that a connecting shaft of the fan blade fan has cracks, the cracks are continuously opened and closed in the rotating process of the connecting shaft, and the rigidity of the shaft surface changes nonlinearly along with time, the crack opening and closing function f can be defined as:
the further calculation is as follows:
wherein sz1 is the inner diameter of the fan connecting shaft, sz2 is the outer diameter of the fan connecting shaft, m1 is the crack opening mass of the connecting shaft, m2 is the crack closing mass of the connecting shaft, t is the fan running time, c is the external damping coefficient, k is the bending rigidity of the rotating shaft without cracks, delta k is the rigidity variation caused when the cracks are completely opened, β is the initial phase angle, omega is the fan rotating speed, and f is the crack opening and closing function;
1.2.2 step 2) the crack opening and closing function f as described above can be determined using a square wave function, which is:
wherein tau is the dynamic displacement of the fan connecting shaft.
1.2.3 step 3) the square wave function as described above can obtain a leaf fan vibration model f' after introducing dimensionless parameters, namely:
the sectional coordinate diagrams of the fan vibration model and the connecting shaft are shown in fig. 3 and 4.
1.2.4 the blade fan vibration model adopts a numerical analysis method to predict the cracks of the blade fan connecting shaft, as the running time t of the fan increases, the external damping coefficient c is in a descending trend, and the dynamic displacement tau of the connecting shaft is in an ascending trend, so that the hidden danger of cracks generated on the connecting shaft is also in an ascending trend.
1.2.5 the blade fan vibration model f ' as described above, in terms of applicability, two problems should be considered, one is that when the vibration frequency generated by the fan connecting shaft is always consistent, the blade fan vibration model f ' is in a fusion mode, and the model is applicable, but when the vibration frequency is inconsistent, each parameter in the algorithm is in a nonlinear change, at this time, the blade fan vibration model f ' is in a bending mode, and the objective function of the model in the bending mode is discontinuous, which brings difficulty to the numerical solution of the algorithm, and is reflected in that the iterative curve of the connecting shaft vibration frequency oscillates repeatedly and is difficult to converge, so the model must be optimized; secondly, the detection sensitivity of the vibration sensor on the connecting shaft can be attenuated along with the use period, and the detection of the sensor has errors, so that the attenuation of the model must be automatically compensated.
1.2.6 the fan vibration model f' is optimized as described previously. Describing by adopting a topological optimization method, and carrying out finite element dispersion on the original model parameters, namely:
s.t.:
β-ω0 2≤0
0＜Δk≤1
according to the method, parameters in the model are constrained, so that a target function curve becomes continuous and smooth, vibration frequency data of the fan connecting shaft can be screened to a certain extent, the vibration frequency of the fan connecting shaft is kept consistent (namely single frequency), and the stability of data output is improved.
1.2.7 attenuation compensation of the sensitivity of the leaf fan vibration model f' as described above. When the sensitivity of the fan sensor is subjected to attenuation compensation, the vibration frequency characteristic of a connecting shaft of the fan sensor is required to be utilized to carry out topology design on model parameters, so that a vibration model for solving the compensation problem is constructed.
1.2.7.1 approximation submodel to compensate for the problem as previously described. When the sensitivity attenuation of the sensor occurs, the traditional single-frequency sensitivity compensation can not realize real-time compensation and error-free along with the periodic change, so that a complex-frequency sensitivity compensation method is required.
1.2.7.2 the sensitivity compensation method for complex frequency as described above, which is a topological model in incremental form. The relationship between the characteristic value corresponding to the model and the incremental lifting parameter of the method after topological optimization is a single-frequency incremental model which cannot be directly used for solving a leaf fan vibration model with characteristic frequency topological optimization expressed by full quantity. For this purpose, an adaptive complex frequency incremental model is constructed.
1.2.7.3 the complex frequency incremental model as described above, which is solved in two modes, i.e. the leaf fan vibration model f' is in a fusion mode and a bending mode.
1.2.7.3.1 the fan vibration model f' is a complex frequency increment model in the fusion mode, and the generalized eigenvalues and corresponding orthogonal normalization condition of modal vector with respect to the model to be satisfied are omitted for the sake of brevity.
The complex frequency increment model in the fusion mode is constructed as follows:
where the parameter ρ represents a frequency vector and Δ ρ represents a frequency change amount.
According to the complex frequency increment model, when the dynamic displacement tau of the fan connecting shaft is small enough, max { min (tau) & min (β) } is equal to max { min (tau) }, and the complex frequency increment model automatically steps back to be a single frequency increment model (same as 1.2.6).
1.2.7.3.2 the fan vibration model f' is a complex frequency incremental model in bending mode. If an analysis method for switching from single-frequency sensitivity to complex-frequency sensitivity is simply adopted in the bending mode, the calculation result may be invalid, and adjacent characteristic values may not coincide. Therefore, a convergence condition needs to be set, and as long as the adjacent feature values satisfy the range of the convergence condition, the analysis method of the complex frequency sensitivity can be adopted.
(1) As for the complex frequency increment model with the convergence condition in the bending mode of the leaf fan vibration model f', an iteration method based on characteristic frequency optimization can be adopted to optimize the complex frequency increment model, the optimization target can select the maximization of the characteristic frequency value of a specified iteration order and also can select the maximization of the characteristic frequency value of a specified adjacent order, namely, the model structure is optimized more ideally through the convergence condition, the characteristic frequency value of complex frequency sensitivity is effectively inhibited after iterative computation, and the failure of the computation result caused by the generation of frequency distortion (also called linear distortion) is avoided.
(2) As mentioned above, the iterative method based on characteristic frequency optimization optimizes the complex frequency incremental model, taking the problem of maximizing characteristic frequency values of nth order and n-1 order as an example, and considering the complex frequency sensitivity compensation method based on the topological optimization model of the incremental form (technical scheme 1.2.7.2), the structure is as follows:
in the above formula, in the iteration process, convergence conditions are set to be j and q, and n times and n-1 times are iterated respectively.
The invention discloses a fan blade fan vibration detection device and a fault prediction method. As a fan blade vibration detection device, the device has the characteristics of high sensitivity, accurate measurement, stable data transmission, convenient operation and the like. As a fan fault prediction method, the method has the characteristics of wide use coverage, strong data output capacity, high popularization and the like.
Claims (5)
1. The utility model provides a fan blade fan vibration detection and failure prediction system, the fan uses certain rotational speed to be circular motion rotation around the connecting axle as the center, its characterized in that: the system comprises a fan blade vibration detection device as a hardware part and a fan failure prediction method as a software part,
fan blade fan vibration detection device includes:
the vibration sensor is arranged on the connecting shaft, takes the vibration mechanical quantity of the connecting shaft as input quantity, and receives the input quantity by a receiving part of the vibration mechanical quantity and converts the input quantity into electric quantity pulse output;
the signal preprocessing module is used for storing and amplifying the electric quantity pulse signal transmitted by the vibration sensor through the built-in vector processor;
the A/D conversion module converts the analog signals acquired by the signal preprocessing module into digital signals so as to carry out the analysis and processing of the next step;
a data transmission module for transmitting the analyzed and processed digital signal through a transmission medium;
the fan fault prediction method comprises the following steps:
step 1) assuming that a connecting shaft of the fan blade fan has cracks, the cracks are continuously opened and closed in the rotating process of the connecting shaft, and the rigidity of the shaft surface is caused to generate nonlinear change along with time, and then a crack opening and closing function f is defined as:
the further calculation is as follows:
wherein, in the formula I-formula II, sz1Is the size of the inner diameter of a fan connecting shaft, sz2Is the outer diameter size m of a fan connecting shaft1For the crack opening quality of the connecting shaft, m2Connecting shaft crack closing mass, t is fan running time, c is an external damping coefficient, k is bending rigidity of a rotating shaft when no crack exists, delta k is rigidity variable quantity caused when the crack is completely opened, β is an initial phase angle, omega is fan rotating speed, and f is a crack opening and closing function;
step 2), judging the crack opening and closing function f by adopting a square wave function, wherein the function is as follows:
in formula IV, tau is the dynamic displacement of the fan connecting shaft;
step 3), after dimensionless parameters are introduced into the square wave function, obtaining a leaf fan vibration model f':
with the increase of the running time t of the fan, the external damping coefficient c is in a descending trend, and the dynamic displacement tau of the connecting shaft is in an ascending trend, so that the hidden trouble that the connecting shaft cracks is also in an ascending trend.
2. The fan blade vibration detection and failure prediction system of claim 1, wherein: when the vibration frequency generated by the fan connecting shaft is inconsistent, the blade fan vibration model f 'is in a bending mode, the target function of the blade fan vibration model f' in the bending mode is discontinuous, and the blade fan vibration model f 'is described by a topological optimization method, so that the parameters of the original blade fan vibration model f' are subjected to finite element dispersion:
s.t.:
β-ω0 2≤0
0＜Δk≤1
3. the fan blade vibration detection and failure prediction system of claim 1, wherein: the detection sensitivity of the vibration sensor can be attenuated along with the use period, and the detection of the vibration sensor has errors at the moment, so that the attenuation of the leaf fan vibration model f' is automatically compensated.
4. The fan blade vibration detection and failure prediction system of claim 3, wherein: when the fusion model is in a fusion mode, a corresponding complex frequency increment model is constructed to solve the compensation problem when the attenuation compensation is carried out on the sensitivity of the vibration sensor,
the structure of the complex frequency increment model in the fusion mode of the leaf fan vibration model f' is as follows:
s.t.:
[β-(ω0 2+Δωj 2)]≤0,j＝1,2,...t
0＜ρ≤(ρ0+Δρ)≤1
β-ω0 2≤0
0＜Δk≤1
Δωt 2＝fΔρ
in formula VII, the parameter ρ represents a frequency vector, and Δ ρ represents a frequency change amount.
5. The fan blade vibration detection and failure prediction system of claim 4, wherein: the leaf fan vibration model f 'is in a bending mode, and an iteration method based on characteristic frequency optimization is adopted to optimize the leaf fan vibration model f' to be a complex frequency increment model in a fusion mode.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811148445.9A CN109323853B (en) | 2018-09-29 | 2018-09-29 | Fan blade fan vibration detection device and fault detection prediction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811148445.9A CN109323853B (en) | 2018-09-29 | 2018-09-29 | Fan blade fan vibration detection device and fault detection prediction method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109323853A CN109323853A (en) | 2019-02-12 |
| CN109323853B true CN109323853B (en) | 2020-06-30 |
Family
ID=65266194
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811148445.9A Active CN109323853B (en) | 2018-09-29 | 2018-09-29 | Fan blade fan vibration detection device and fault detection prediction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109323853B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109946029B (en) * | 2019-02-22 | 2020-07-10 | 北京理工大学 | Multi-frequency fatigue test method for turbocharger |
| US11334414B2 (en) * | 2019-03-04 | 2022-05-17 | Mitsubishi Heavy Industries, Ltd. | Abnormality detecting apparatus, rotating machine, abnormality detection method, and non- transitory computer readable medium |
| CN111259737B (en) * | 2020-01-08 | 2023-07-25 | 科大讯飞股份有限公司 | Method and device for predicting failure of steering wheel of vehicle, electronic equipment and storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US3816712A (en) * | 1971-11-29 | 1974-06-11 | Zeiss Stiftung | Method and apparatus for detecting relative deviations in position of two parts moved in a predetermined desired relationship |
| CN103063428A (en) * | 2013-01-25 | 2013-04-24 | 国电联合动力技术有限公司 | Wireless monitoring system and method for modal parameter of blade of fan |
| CN106153824A (en) * | 2016-06-22 | 2016-11-23 | 北京工业大学 | A kind of Prediction method for fatigue life based on crack closure effect |
| CN206017210U (en) * | 2016-08-25 | 2017-03-15 | 观为监测技术无锡股份有限公司 | A kind of blower fan apparatus with vibration monitoring |
| CN206177542U (en) * | 2016-09-20 | 2017-05-17 | 天津航天瑞莱科技有限公司 | Fatigue test device of engine blade |
| CN107314893A (en) * | 2017-05-08 | 2017-11-03 | 上海交通大学 | A kind of modular multi-function rotor testbed |
| CN108469377A (en) * | 2018-03-22 | 2018-08-31 | 九江学院 | A kind of measurement method of crack opening stress |
-
2018
- 2018-09-29 CN CN201811148445.9A patent/CN109323853B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US3816712A (en) * | 1971-11-29 | 1974-06-11 | Zeiss Stiftung | Method and apparatus for detecting relative deviations in position of two parts moved in a predetermined desired relationship |
| CN103063428A (en) * | 2013-01-25 | 2013-04-24 | 国电联合动力技术有限公司 | Wireless monitoring system and method for modal parameter of blade of fan |
| CN106153824A (en) * | 2016-06-22 | 2016-11-23 | 北京工业大学 | A kind of Prediction method for fatigue life based on crack closure effect |
| CN206017210U (en) * | 2016-08-25 | 2017-03-15 | 观为监测技术无锡股份有限公司 | A kind of blower fan apparatus with vibration monitoring |
| CN206177542U (en) * | 2016-09-20 | 2017-05-17 | 天津航天瑞莱科技有限公司 | Fatigue test device of engine blade |
| CN107314893A (en) * | 2017-05-08 | 2017-11-03 | 上海交通大学 | A kind of modular multi-function rotor testbed |
| CN108469377A (en) * | 2018-03-22 | 2018-08-31 | 九江学院 | A kind of measurement method of crack opening stress |
Non-Patent Citations (1)
| Title |
|---|
| 转轴上裂纹开闭模型的研究;高建民 等;《应用力学学报》;19920331;第9卷(第1期);第108-112页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109323853A (en) | 2019-02-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109323853B (en) | Fan blade fan vibration detection device and fault detection prediction method | |
| CN102980651B (en) | Monitoring method and monitoring device and monitoring system of wind turbine generator condition | |
| CN107223206B (en) | Fatigue test | |
| EP2026160B1 (en) | Event monitoring via combination of signals | |
| CN107192446B (en) | Method for monitoring natural frequency of tower of wind generating set | |
| EP2520794A1 (en) | Method of checking a wind turbine in a wind farm for a yaw misalignment, method of monitoring a wind turbine in a wind farm and monitoring apparatus | |
| CN107239621B (en) | A kind of critical rotor speed analysis method based on probability box framework | |
| CN112084583B (en) | Rotor blade life prediction method and system based on digital twin | |
| CN111400961B (en) | Wind generating set blade fault judging method and device | |
| CN107762730B (en) | A kind of large-scale variable pitch turbine control system and control method with trailing edge flaps | |
| CN103925155A (en) | Self-adaptive detection method for abnormal wind turbine output power | |
| CN105352586A (en) | Method used for realizing blade fatigue crack detection through resonance frequency detection | |
| CN109630449B (en) | Three-proofing ventilation equipment fault prediction system based on RBF algorithm | |
| Noppe et al. | Modeling of quasi-static thrust load of wind turbines based on 1 s SCADA data | |
| CN117113200B (en) | Rotor fault diagnosis method, device, electronic equipment and medium | |
| CN117077064B (en) | Fault detection method, device and equipment for wind power equipment and storage medium | |
| Mehlan et al. | Rotor imbalance detection and diagnosis in floating wind turbines by means of drivetrain condition monitoring | |
| CN111550429B (en) | Centrifugal compressor rotor system stability comprehensive control method and device | |
| Wait et al. | Wind-induced instabilities and monitoring of wind turbine | |
| Rusinski et al. | Monitoring and testing of high power industrial fans vibration | |
| CN113027699B (en) | Monitoring method, device and system of wind generating set | |
| Kuseyri | Condition monitoring of wind turbines: Challenges and opportunities | |
| CN112632726A (en) | Flow field reconstruction method for aeroelasticity simulation of impeller mechanical blade | |
| CN113435395B (en) | Wind wheel fault self-adaptive diagnosis method based on self-adaptive Kalman filter | |
| Lei et al. | Research on the Models of Coupling Dynamics and Damage Classification for Vehicle-Engine Vibration |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |