WO2023273378A1

WO2023273378A1 - Bolt fault diagnosis method and apparatus

Info

Publication number: WO2023273378A1
Application number: PCT/CN2022/078811
Authority: WO
Inventors: 张猛; 周军长; 陈旭; 郭靖; 铎林; 刘云平; 宋敏; 唐磊; 赵政雷
Original assignee: 东方电气集团东方电机有限公司
Priority date: 2021-06-29
Filing date: 2022-03-02
Publication date: 2023-01-05
Also published as: CN113656989A

Abstract

Disclosed in the present invention are a bolt fault diagnosis method and apparatus. The method comprises the following steps: S1, performing a bolt fault simulation test, and recording monitoring data of a bolt fault monitoring point; S2, simulating a bolt fault by means of a finite element method, and establishing a virtual fault case library; and S3, extracting a fault feature from the virtual fault case library, and performing artificial intelligence modeling on a real machine bolt fault. By means of the present invention, the problem in the prior art of it being impossible to perform effective diagnosis on all bolts by means of a limited number of test points is solved, thereby solving the problem of performing monitoring and fault diagnosis of a bolt group of existing devices.

Description

Method and device for bolt fault diagnosis

technical field

The invention relates to the technical field of bolt fault diagnosis, in particular to a bolt fault diagnosis method and device.

Background technique

The top cover is one of the key components of the turbine, and its main function is to fix the water guide mechanism and bear the axial load of the unit. The top cover and the seat ring are connected by bolts. The vibration of the unit caused by uncertain factors will cause the bolts to bear variable amplitude loads, which will have a huge impact on the mechanical strength and fatigue life of the bolts. The Russian Sayan Hydropower Station abroad caused a major accident due to the fatigue failure of the top cover-seat ring bolts. The top cover-seat ring bolts of a pumped storage power station in China were fatigued and fractured when the turbine of a certain unit was shut down in an emergency, and the top cover was lifted. Cause major safety accidents (such as flooding of factory buildings, etc.).

The bolt is a typical multi-notch part, and its fatigue performance is affected by various factors such as the structure, material, manufacturing process, dynamic stress level and environmental factors of the bolt. The strength and fatigue checks are carried out under the preset working conditions, but the influence of the actual installation value of the bolts of the real machine, the actual operating state of the real machine and the real operation history on the stress state of the actual bolts cannot be accurately considered. In the industry, the failure probability of important bolts of hydraulic turbines is extremely low, but once it happens, the damage will be huge, ranging from equipment damage to flooding of workshops and major personnel safety accidents.

Restricted by sensor technology (the number of independent channels of high-performance ultrasonic load cells in the industry is up to 24), site layout conditions, and economical implementation costs, most of the existing monitoring solutions monitor a part of the bolts of the integral flange (such as the turbine roof The number of cover-seat ring joint bolts is relatively large, generally about 64 to 120, and only a part of the bolts are monitored), and it is difficult to arrange measuring points on each bolt during field implementation. Therefore, it is an important and urgent problem to carry out effective monitoring, fault warning and fault diagnosis on all bolts of the entire flange through limited measuring points.

The safety of hydropower equipment is very high, and there are very few fault samples of loose and broken bolts, and due to confidentiality reasons, it is difficult to obtain public real technical information; for the safety of equipment and personnel, artificially manufacture and obtain fault samples on the real machine, The same is not feasible.

Contents of the invention

In order to overcome the deficiencies of the prior art, the present invention provides a method and device for bolt fault diagnosis, which solves the problem in the prior art that all bolts cannot be effectively diagnosed through limited measuring points, thereby solving the problems of existing equipment bolt group monitoring and Troubleshooting problems.

The technical solution adopted by the present invention to solve the above problems is:

A bolt fault diagnosis method, comprising the following steps:

S1, carry out the bolt failure simulation test, and record the monitoring data of the bolt failure monitoring point;

S2, simulating bolt faults through the finite element method, and establishing a virtual fault case library;

S3, extract the fault features in the virtual fault case library, and perform artificial intelligence modeling on the bolt fault of the real machine.

The present invention solves the problems of limited number of bolt fault monitoring points, limited arrangement of bolt fault monitoring points, and rapid diagnosis of fault points through bolt fault simulation tests, fault simulation, artificial intelligence modeling, and the like. Through the finite element simulation, the virtual fault case library is established, and the automatic diagnosis of the bolt faults of the real machine can be realized, which is convenient for the diagnosis of all bolts. In actual use, the following settings can be adopted: if a bolt without a measuring point is loosened or broken, according to the finite element simulation analysis, the axial force of the bolts near the measuring point will change to varying degrees, and the combination of the changing values can be As the corresponding fault characteristics, the looseness and fracture of the bolts without measuring points can be reversed, so as to realize the location of the fault bolts and quantitative fault diagnosis analysis, and complete the intelligent diagnosis of this fault through artificial neural network technology.

As a preferred technical solution, the bolt failure simulation test in step S1 includes a bolt loosening failure simulation test and/or a bolt breaking failure simulation test.

Bolt loosening and bolt breakage are common bolt faults. Through the above simulation tests, the scope of application of the present invention is wider.

As a preferred technical solution, the monitoring data of the bolt fault monitoring point in step S1 includes the bolt axial force value and/or the bolt elongation value of the bolt fault monitoring point.

The bolt axial force value and bolt elongation value are used as numerical indicators of bolt failure, which is convenient for statistical analysis and modeling.

As a preferred technical solution, step S2 also includes the following steps: analyzing simulation errors, and verifying the bolt axial force monitoring instrument.

Through the above steps, the present invention can verify the performance of the bolt axial force monitoring instrument before construction, verify the fault diagnosis method, and improve the self-adaptive ability, fault tolerance and robustness of diagnosing the flange bolt fault of the real machine .

As a preferred technical solution, in step S2, the real machine working condition data, the real machine operating state data, and the initial preload installation value of the unit are used as boundary conditions for simulation input.

This is convenient to improve the accuracy and authenticity of the simulation and diagnosis of bolt faults, and is conducive to improving the quality of diagnosis.

As a preferred technical solution, step S3 further includes the following step: enabling the bolt fault model of the real machine to have an online self-learning function.

The online self-learning function further improves the error-correcting ability of diagnosis and makes it more intelligent.

A bolt fault diagnosis device applied to the bolt fault diagnosis method includes a bolt, a bolt axial force monitoring instrument connected to the bolt, and a load simulation device connected to the bolt.

The load simulation device provides simulated load for bolt failure, and the bolt axial force monitoring instrument monitors the bolt axial force in the process of bolt failure in real time, so as to facilitate the diagnosis of bolt failure.

As a preferred technical solution, the load simulation device is a jack.

The jack applies loads well and is easy to install.

As a preferred technical solution, the bolt axial force monitoring instrument is an ultrasonic load cell.

Ultrasonic force sensor has high force measurement accuracy, non-destructive monitoring, relatively sensitive and easy to install.

As a preferred technical solution, a bolt length monitoring instrument is also included.

This facilitates monitoring of length changes during bolt failure, further improving simulation and diagnostics. As a preference, the bolt length monitoring instrument can choose a dial indicator, which is convenient for installation.

Compared with the prior art, the present invention has the following beneficial effects:

(1) The present invention solves the problems of limited number of bolt fault monitoring points, limited location of bolt fault monitoring points, and rapid diagnosis of fault points through bolt fault simulation tests, fault simulation, artificial intelligence modeling, etc.; through finite element simulation, Establish a virtual fault case library, and realize automatic diagnosis of bolt faults on real machines, which facilitates the diagnosis of all bolts; the invention can reverse the loosening and fracture of bolts without measuring points, so as to realize the positioning and quantitative fault diagnosis of faulty bolts Analysis, complete the intelligent diagnosis of this fault through artificial neural network technology;

(2) The bolt fault simulation test described in step S1 of the present invention comprises a bolt loosening fault simulation test and/or a bolt fracture fault simulation test; bolt loosening and bolt fracture are common bolt faults, and by the above simulation tests, the scope of application of the present invention is wider wide;

(3) The monitoring data of the bolt fault monitoring point described in step S1 of the present invention includes the bolt axial force value and/or the bolt elongation value of the bolt fault monitoring point; the bolt axial force value and the bolt elongation value are used as the numerical index of the bolt fault, which is convenient Convenient statistical analysis and modeling;

(4) Step S2 of the present invention also includes the following steps: analyze the simulation error, and verify the bolt axial force monitoring instrument; through the above steps, the present invention can verify the performance of the bolt axial force monitoring instrument before construction, and diagnose the fault The method is verified, which improves the adaptive ability, fault tolerance and robustness of the diagnosis of the flange bolt fault of the real machine;

(5) In the step S2 of the present invention, the boundary condition of simulation input is taken as the boundary condition of simulation input with real machine working condition data, real machine operating state data, unit initial preload installation value; It is beneficial to improve the quality of diagnosis;

(6) step S3 of the present invention also includes the following steps: make the bolt failure model of the real machine possess an online self-learning function, and the online self-learning function further improves the error correction ability of diagnosis and is more intelligent;

(7) The load simulation device of the present invention provides simulated loads for bolt failures, and the bolt axial force monitoring instrument monitors the bolt axial force in the process of bolt failure in real time, thereby facilitating the diagnosis of bolt failures;

(8) The load simulation device of the present invention is a jack, which has a good effect of applying load and is easy to install;

(9) The bolt axial force monitoring instrument of the present invention is an ultrasonic load cell, which has high force measurement accuracy, non-destructive monitoring, is more sensitive, and is easy to install;

(10) The present invention also includes a bolt length monitoring instrument; this is convenient for monitoring the length change in the process of bolt failure, thereby further improving the effect of simulation and diagnosis.

Description of drawings

Fig. 1 is the fault diagnosis modeling flow chart of the bolt fault simulation test;

Fig. 2 is the fault diagnosis modeling flow chart of the real machine bolt fault;

Fig. 3 is the structural representation of the bolt fault diagnosis device of the present invention;

Fig. 4 is the P direction view of Fig. 3;

Fig. 5 is a cross-sectional view along A-A plane of Fig. 3 .

The marks in the drawings and the names of corresponding parts: 1. Bolts, 2. Bolt axial force monitoring instrument, 3. Load simulation device, 4. Bolt length monitoring instrument, 5. Base, 6. Base fixing bolts, 7. Flange, 8, test nut, 9, jack support plate, 10, force measuring washer.

detailed description

The present invention will be further described in detail below in conjunction with the embodiments and the accompanying drawings, but the embodiments of the present invention are not limited thereto.

Example 1

As shown in Figures 1 to 5, a bolt fault diagnosis method includes the following steps:

Example 2

As shown in Figures 1 to 5, as a further optimization of Embodiment 1, this embodiment includes all the technical features of Embodiment 1. In addition, this embodiment also includes the following technical features:

A bolt fault diagnosis device applied to the bolt fault diagnosis method described above includes a bolt 1 , a bolt axial force monitoring instrument 2 connected to the bolt 1 , and a load simulation device 3 connected to the bolt 1 .

The load simulation device 3 provides simulated load for the failure of the bolt 1, and the bolt axial force monitoring instrument 2 monitors the bolt axial force in the process of the failure of the bolt 1 in real time, so as to facilitate the diagnosis of the failure of the bolt 1.

As a preferred technical solution, the load simulation device 3 is a jack.

The jack applies loads well and is easy to install.

As a preferred technical solution, the bolt axial force monitoring instrument 2 is an ultrasonic load cell.

As a preferred technical solution, a bolt length monitoring instrument 4 is also included.

This is convenient for monitoring the length change during the failure process of the bolt 1, thereby further improving the effect of simulation and diagnosis. As a preference, the bolt length monitoring instrument 4 can choose a dial indicator, which is convenient for installation.

Example 3

As shown in FIGS. 1 to 5 , on the basis of Embodiment 1 and Embodiment 2, this embodiment provides a more detailed implementation manner.

The fault simulation device consists of base 5, base fixing bolt 6, flange 7, test bolt (i.e. bolt 1), test nut 8, jack (i.e. load simulation device 3), jack support plate 9, dial indicator (i.e. Bolt length monitoring instrument 4), bolt force sensor (ie bolt axial force monitoring instrument 2), force measuring washer 10, see Figure 1 for details.

When in use, the base 5 is fixed on the test bench by the base fixing bolt 6; the flange 7 is connected to the base 5 according to the set preload by the test bolt and the test nut; the external load of the work is accurately simulated by the jack . The bolt force sensor is consistent with the real machine, and the ultrasonic force sensor is preferred. The elongation value of the test bolt is monitored by a dial gauge.

The test bolts of this device are 8 M30-8.8-O bolts, and the bolt axial force measuring points are arranged on 2#, 4#, 6#, 8# bolts; the jack range is 100 tons, and the simulated axial working load is set to 80 tons .

When simulating bolt loosening faults, pre-tighten 8 bolts to the specified load, such as 180kN; loosen the specified bolts to a certain or several proportions of the pre-tightening value, such as loosening to 50% of the pre-tightening force, while loosening the bolts, record each The reading of the bolt measuring point changes.

When simulating bolt fracture failure, pre-tighten 8 bolts to the specified load, such as 180kN; completely loosen and remove the specified bolts, and record the reading changes of each bolt measuring point.

Artificial neural network is a mature and widely used intelligent algorithm in the field of fault diagnosis. This technology can fit nonlinear models finely enough to realize complex nonlinear mapping (pattern recognition) between input features and faults, and has self-adaptive capability, strong fault tolerance and robustness.

In the bolt loosening fault simulation and bolt fracture fault simulation, the measured data of the fault sample can be obtained; based on the finite element simulation, the corresponding virtual simulation calculation value can be obtained, and compared and corrected with the measured data of the simulation device. Then, a complete fault sample library is further established through artificial intelligence technology, and the intelligent diagnosis of all bolt faults of the flange 7 based on the limited bolt axial force monitoring data is realized. The fault diagnosis modeling process is shown in Figure 1 for details. Table 1 is a schematic diagram of fault feature extraction based on finite element calculation. It can be seen from Table 1 that after the bolts (1#, 3#, 5#, 7#) of the non-axial force monitoring point are loosened and broken, there is axial force The axial force values of the monitoring point bolts (2#, 4#, 6#, 8#) change accordingly, and the combination of the changing values depends on the location and degree of loosening/breaking of the specific faulty bolts.

Table 1 Schematic diagram of fault feature extraction based on finite element calculation

Due to the diversity and complexity of real machine fault samples, virtual fault samples are established through finite element simulation.

Taking the real machine working condition data (output, water head, tail water level, flow, etc.), real machine operating state data (pressure, pressure pulsation, unit vibration, etc.), and the initial preload installation value of the unit as the simulation input boundary conditions, through finite element simulation , perform virtual simulation on bolt loosening and fracture faults, and extract the variation value of bolt axial force at the measuring point. Afterwards, the fault feature information similar to Table 1 is established; finally, the artificial neural network technology is used to model the fault diagnosis through the artificial intelligence algorithm, so as to realize the automatic evaluation of the bolt safety state and the automatic fault diagnosis.

The fault modeling flow chart of the real machine bolt is shown in Figure 2 for details.

The invention provides a fault simulation device and an intelligent diagnosis method for the loosening and fracture of top cover-seat ring bolts. The method is based on the operating condition data of the unit, the operating state data of the unit, the bolt pre-tightening installation data, and the measuring points of some bolts. On-line monitoring data, finite element calculation and artificial neural network technology can automatically judge and automatically on-line early warning the possible loosening and fracture of all bolts in the flange 7, which can reduce the implementation cost and maintenance cost of bolt monitoring equipment. The performance verification of the force sensor in the factory (before the construction of the power station) provides a reliable guarantee for the safe operation and condition maintenance of the power station.

The present invention can be simultaneously applied to bolt health status evaluation and intelligent fault diagnosis of flange 7 connection structures such as draft pipe entry doors, volute casing entry doors, and ball valves. It has the following advantages and functions:

1) Introduce the real-time working condition data of the unit, the state monitoring data and the bolt initial pre-tightening installation data, simulate the bolt fault through the finite element method, and extract the fault characteristics;

2) Simulate the bolt fault by simulating tooling, verify the performance of the bolt force sensor, and verify the accuracy and reliability of the bolt fault diagnosis method;

3) Through the artificial intelligence algorithm and artificial neural network technology, model the loosening and fracture fault diagnosis of real machine bolts, based on the change characteristics of bolt measurement point data, realize the health status evaluation of real machine bolts and intelligent diagnosis of bolt faults, and With online self-learning function of fault diagnosis model.

The present invention can verify the performance of the bolt force sensor in the factory (before the construction of the power station), and perform health status evaluation and fault diagnosis on all the bolts of the flange of the real machine, which can reduce the implementation cost and maintenance cost of the bolt monitoring equipment, and provide a better solution for the power station. Safe operation and state-of-the-art maintenance provide a reliable basis.

If all the bolts of the real machine are equipped with measuring points, the fault can be directly judged according to the data of the bolt axial force measuring points. If only some bolts of all flange bolts are equipped with bolt sensor measuring points, the health status assessment and fault diagnosis of all flange bolts can be carried out through limited measuring points.

As described above, the present invention can be preferably carried out.

All features disclosed in all embodiments in this specification, or steps in all implicitly disclosed methods or processes, except for mutually exclusive features and/or steps, can be combined and/or extended and replaced in any way.

The above are only preferred embodiments of the present invention, and are not intended to limit the present invention in any form. According to the technical essence of the present invention, within the spirit and principles of the present invention, any simple changes made to the above embodiments The modification, equivalent replacement and improvement, etc., all still belong to the protection scope of the technical solution of the present invention.

Claims

A bolt fault diagnosis method is characterized in that, comprising the following steps:

S1, carry out the bolt failure simulation test, and record the monitoring data of the bolt failure monitoring point;

S2, simulating bolt faults through the finite element method, and establishing a virtual fault case library;

S3, extract the fault features in the virtual fault case library, and perform artificial intelligence modeling on the bolt fault of the real machine.
The bolt fault diagnosis method according to claim 1, wherein the bolt fault simulation test in step S1 includes a bolt loosening fault simulation test and/or a bolt breaking fault simulation test.
The bolt fault diagnosis method according to claim 1, wherein the monitoring data of the bolt fault monitoring point in step S1 includes the bolt axial force value and/or the bolt elongation value of the bolt fault monitoring point.
A bolt fault diagnosis method according to claim 1, characterized in that step S2 further comprises the following steps: analyzing simulation errors and verifying the bolt axial force monitoring instrument.
A bolt fault diagnosis method according to claim 1, characterized in that in step S2, the real machine working condition data, real machine operating state data, and unit initial preload installation value are used as boundary conditions for simulation input.
A bolt fault diagnosis method according to claim 1, characterized in that step S3 further comprises the following step: enabling the bolt fault model of the real machine to have an online self-learning function.
A bolt fault diagnosis device applied to a bolt fault diagnosis method according to any one of claims 1 to 6, characterized in that it comprises a bolt (1), a bolt axial force monitoring instrument connected to the bolt (1) ( 2), a load simulation device (3) connected with the bolt (1).
A bolt fault diagnosis device according to claim 7, characterized in that the load simulation device (3) is a jack.
A bolt fault diagnosis device according to claim 7, characterized in that the bolt axial force monitoring instrument (2) is an ultrasonic load cell.
A bolt fault diagnosis device according to claim 7, further comprising a bolt length monitoring instrument (4).