A method for monitoring steam turbine safety

Abstract

The present invention discloses a method for monitoring the safety of a steam turbine, comprising the following steps: collecting, storing, and preprocessing vibration data of the steam turbine to obtain preprocessed data; converting the preprocessed data into a system signal; and comparing the system signal with a fault signal in a database to obtain the operating status of the blades and determine the safety of the steam turbine. The method for monitoring the safety of a steam turbine of the present invention is based on various vibration state parameters during the operation of the steam turbine, obtains the changing characteristics of the blade vibration during the real-time operation according to the monitored values, compares the real-time changing characteristics with the fault characteristics, implements fault diagnosis, calculates the real-time life of the steam turbine, and implements real-time assessment of the life of the steam turbine; and the fault diagnosis and life assessment processes do not require interrupting the operation of the steam turbine.

Description

Technical Field

[0001] The present invention particularly relates to a method for monitoring the safety of a steam turbine and a method for evaluating its life span based on blade vibration data. Background Art

[0002] Steam turbines are essential mechanical equipment in the power generation industry, with their operation and power generation processes highly integrated. A typical steam turbine consists of a shaft, blades, bearings, sleeves, couplings, a casing, and a base. As critical components, the long blades of steam turbines are often subjected to high mechanical loads due to extreme operating conditions, making them a major source of failure.

[0003] Because steam turbines are rotating machines, and their long blades are the most severely affected, the vibrations they create often pose safety hazards, exposing power generation units to significant risks of failure and economic loss. Downtime can lead to declines in output, safety, environmental performance, and customer satisfaction. As the power generation industry continues to improve its O&M capabilities, efforts are underway to minimize the use of downtime to address safety hazards associated with long steam turbine blades.

[0004] In recent years, people have turned their attention to more effective fault diagnosis strategies and more proactive maintenance technologies. One of the more prominent strategies is condition-based maintenance technology. Compared with other more traditional maintenance strategies (such as unplanned downtime and planned preventive maintenance, etc.), this strategy has the advantage that new faults and anomalies can be captured and identified very early, thereby avoiding catastrophic events caused by faults.

[0005] To implement condition-based maintenance (CBM) for fault monitoring and diagnosis of long turbine blades, different monitoring and measurement methods are used to capture and track various operating parameters, such as vibration monitoring, lubricant and wear debris monitoring, infrared thermography, acoustic emission, process monitoring, and human senses. Since vibration of long turbine blades is one of the main causes of failure, understanding the vibration behavior of long turbine blades under healthy and suspected fault conditions over a certain time interval in order to analyze irregularities in the vibration characteristics of long turbine blades is a primary technical goal of CBM.

[0006] Current steam turbine safety monitoring and life assessment methods focus primarily on monitoring the shaft's lateral, torsional, and bearing vibrations. However, they lack real-time monitoring of the long blades, a critical component of the turbine. Relying solely on inspections during downtime for maintenance cannot accurately assess the blade strength of the long blades during operation, hindering accurate assessment of turbine safety and life. If a long blade fails during operation, existing analysis processes can only detect changes when blade vibrations are high enough to affect the shaft's vibration frequency. They are unable to detect frequency changes caused by microcracks in the long blades, hindering further improvement in turbine safety monitoring and accurate turbine life assessment.

[0007] The disclosure of the above background technology content is only used to assist in understanding the inventive concept and technical solution of the present invention. It does not necessarily belong to the prior art of this patent application. In the absence of clear evidence that the above content has been disclosed before the filing date of this patent application, the above background technology should not be used to evaluate the novelty and creativity of this application. Summary of the Invention

[0008] In view of this, in order to overcome the defects of the prior art, an object of the present invention is to provide a method for monitoring the safety of a steam turbine based on blade vibration parameters.

[0009] In order to achieve the above object, the present invention adopts the following technical solutions:

[0010] A method for monitoring the safety of a steam turbine comprises the following steps:

[0011] Collect, store and pre-process the vibration data of the steam turbine blades and obtain the pre-processed data;

[0012] converting the preprocessed data into a system signal;

[0013] The system signal is compared with the fault signal in the database to obtain the blade operation status and judge the safety of the steam turbine.

[0014] According to some preferred embodiments of the present invention, the steam turbine vibration data includes blade vibration data and vibration data of other steam turbine components reflecting blade vibration information; the blade vibration data is obtained using an optical sensor and / or an eddy current sensor. The steam turbine vibration data also includes blade amplitude, instantaneous angular velocity, and shaft torsional vibration data of bearing and casing vibration.

[0015] According to some preferred implementation aspects of the present invention, the blade amplitude is calculated using the following formula:

[0016] P=n·πD·ΔT / 60

[0017] Where P is the blade amplitude, n is the rotational speed, D is the diameter of the blade tip monitoring position, and ΔT is the time difference between the actual arrival time and the theoretical arrival time of the blade tip.

[0018] According to some preferred implementation aspects of the present invention, the preprocessing is performed according to the following steps: the vibration data is screened using the Raida law to eliminate data that is obviously out of limit due to abnormal sensor readings, and then filtering and normalizing are performed.

[0019] According to some preferred embodiments of the present invention, a plurality of groups of sensors are provided in the longitudinal extension direction of the steam turbine, one group of sensors corresponds to one group of blades, and each group of sensors is provided within the width direction range of one group of blades.

[0020] According to some preferred implementation aspects of the present invention, each group of sensors includes at least two optical sensors and one eddy current sensor, and the eddy current sensor is arranged between the two optical sensors.

[0021] According to some preferred embodiments of the present invention, the projections of adjacent sensors on the steam turbine cross section correspond to a central angle of 30 to 60 degrees; and / or the projections of multiple sensors on the steam turbine longitudinal section are located on different horizontal and vertical planes. The sensors are installed at an angle facing the long blades.

[0022] According to some preferred implementation aspects of the present invention, a channel signal conditioner is used to convert the pre-processed data into a system signal.

[0023] According to some preferred implementation aspects of the present invention, the system signal includes a time domain signal, a frequency domain signal, a short-time Fourier transform signal, an instantaneous angular velocity signal, and a time-synchronized average signal.

[0024] According to some preferred implementation aspects of the present invention, the database stores blade vibration information of the steam turbine blades in normal operating state or fault state.

[0025] According to some preferred implementation aspects of the present invention, the fault conditions include blade cracking, high cycle fatigue, blade friction, blade root loosening, erosion, creep and corrosion.

[0026] According to some preferred implementation aspects of the present invention, the comparison is to perform least squares calculation on the obtained system signal and the normal operating information and fault feature information in the database, respectively, to obtain multiple sets of regression equations, find the most matching regression equation, and obtain the operating status corresponding to the regression equation according to the database, thereby obtaining the current operating status of the long blade and realizing blade health monitoring and management.

[0027] The steps also include calculating the total life loss during the service period of the steam turbine by collecting the real-time temperature, stress and operating time of the blades under a certain working condition.

[0028] According to some preferred implementation aspects of the present invention, the total life loss is obtained using the following formula:

[0029]

[0030] Where, τ i Refers to the actual operation time of the turbine under the blade operation state i, τ b,j Refers to the life of the blade under operating state i, τ b Refers to the total life loss of the turbine during its service life.

[0031] Due to the adoption of the above technical solution, compared with the existing technology, the benefits of the present invention are: the turbine safety monitoring method of the present invention is based on the various vibration state parameters during the operation of the turbine, and obtains the changing characteristics of the blade vibration during the real-time operation according to the monitoring values, compares the real-time changing characteristics with the fault characteristics, realizes fault diagnosis, and calculates the real-time life of the turbine, realizing real-time evaluation of the turbine life; its fault diagnosis process and life evaluation process do not require interruption of turbine operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0033] Figure 1 2. It is a structural diagram of a steam turbine safety monitoring and life assessment system in a preferred embodiment of the present invention;

[0034] Figure 2 Schematic diagram of the installation of the sensor installed in the preferred embodiment of the present invention;

[0035] Figure 3 Schematic diagram of the process of acquiring blade vibration data in a preferred embodiment of the present invention. DETAILED DESCRIPTION

[0036] In order to enable those skilled in the art to better understand the technical solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0037] In response to the current problems of poor health monitoring of the operation process of long turbine blades and low economy, the present invention proposes a method for turbine safety monitoring and life assessment based on blade vibration information, which realizes real-time collection of turbine long blade operation information and analysis of various fault characteristics, and has good robustness.

[0038] The principle of the present invention is as follows: long blades are the components most likely to cause turbine failures. Long blades are subjected to various high dynamic loads during turbine operation, including thermal loads, inertial loads and bending loads. The failure modes of long blades are usually blade cracking, high cycle fatigue, blade friction, blade root loosening, erosion, creep and corrosion. Blade health monitoring can be determined by observing and comparing the vibration levels of the natural frequencies of the blades under normal conditions and faulty operating conditions. When each blade rotates past the sensor, the system records the actual arrival time of the blade. When the blade is not vibrating, the theoretical time for a given blade to arrive at the sensor for each revolution is the same. When the blade natural frequency matches the excitation frequency, blade resonance occurs. If the blade tip is deflected due to vibration, the blade tip will pass the sensor slightly earlier or later than expected, thereby generating a signal pulse that is inconsistent with the theoretical arrival time. By collecting the signal pulse, the time difference ΔT is obtained, and then the blade amplitude is obtained:

[0039] P=n·πD·ΔT / 60

[0040] Where P is the blade amplitude, n is the rotational speed, D is the diameter of the blade tip monitoring position, and ΔT is the time difference between the actual arrival time and the theoretical arrival time of the blade tip.

[0041] Example 1 Method for steam turbine safety monitoring and life assessment based on blade vibration parameters

[0042] Based on the above principles, the method for steam turbine safety monitoring and life assessment based on blade vibration parameters in this embodiment specifically includes the following steps:

[0043] Step 1: Collect vibration data of turbine blades based on the installed sensors, and collect vibration data of other components that can directly reflect blade vibration information, including but not limited to blade amplitude, instantaneous angular velocity, shaft torsional vibration of bearing vibration and housing vibration, and other vibration related information. Figure 3 As shown, the vibration data is pre-processed (data conditioning) and then stored.

[0044] The preprocessing method is as follows: the vibration data is screened using the Raida law to remove data that is obviously out of limit due to abnormal sensor readings, and then filtered and normalized.

[0045] The vibration data of the turbine blades described in step 1 and the vibration data of other components that can intuitively reflect the blade vibration information are obtained in the following manner: wherein, the vibration data of the turbine blades comes from the added sensors; the vibration data of other components that can intuitively reflect the blade vibration information comes from the monitoring system that comes with the turbine system when it leaves the factory.

[0046] The measurement principle, installation location, and selection of sensor parameters for the added sensor are described as follows:

[0047] Measurement Principle: During the rotation of the turbine's long blades, in addition to the turbine system's existing measurement points and monitoring methods, sensors are installed for non-contact measurement of the turbine's long blades. These sensors collect the signal pulses generated when the blades pass over the sensors. These pulses are subsequently compared with the theoretical arrival time of each blade at the sensor. If the blade tip deflects due to vibration, it may pass over the sensor slightly earlier or later than expected, generating a signal pulse that is inconsistent with the theoretical arrival time.

[0048] Installation position: The installation angle of the sensor is facing the long blades. The sensor can be selected as an optical sensor or an eddy current sensor according to the accuracy of the collected information. In this embodiment, multiple groups of sensors are provided in the length extension direction of the turbine. One group of sensors corresponds to one group of blades, and each group of sensors is provided within the width direction range of one group of blades. The number of sensors installed is three in a group: one eddy current sensor and two optical sensors. The three are spaced 45° apart in the circumferential direction and staggered. The sensor is fixed to the inner cylinder of the turbine by mechanical fixing, and its probe working surface faces the end face of the turbine moving blade. Figure 2 shown. Figure 2 The left side of the middle diagram is a left-view structure diagram. Its width is the width of a set of blades. Under this width and viewing angle, sensors are staggered in the horizontal and vertical directions ( Figure 2 ), and an eddy current sensor is set between the two optical sensors.

[0049] Sensor Parameters: The sensor required for non-contact measurement of long turbine blades has an operating temperature range of -25°C to 200°C. Measurements can be made on a full rotation of 8 to 120 blades, making it suitable for most commercial steam turbines in China. The eddy current sensor operates from a 5-24V DC power supply. The output signal is a rectangular pulse signal that can be modulated into a digital signal by a signal conditioner. The maximum measurement frequency is 1000Hz. The optical sensor has a maximum measurement frequency of 500Hz, a sensing distance range of 0.001-0.1m, and a DC power supply range of 10-30V.

[0050] Step 2: Convert the pre-processed data into system signals

[0051] A four-channel signal conditioner is used to amplify the long blade vibration signal obtained in step 1 and transmit the data to the data acquisition board. After conditioning by the signal conditioner, the blade vibration signal undergoes signal preprocessing and is converted into a time domain signal, a frequency domain signal, a short-time Fourier transform signal, an instantaneous angular velocity signal, and a time-synchronized average signal. These five types of signals serve as the data source for the next step of data analysis.

[0052] The calculation methods of these five types of signals are briefly described as follows: the time domain signal is a digital signal sampled and amplified by the signal conditioner; the frequency domain signal is a digital signal after the time domain signal undergoes Fourier transform; the short-time Fourier signal is to divide the time domain signal into smaller parts according to the time series, apply the Fourier transform to all parts separately, and then reassemble the data into a two-dimensional digital signal of time and frequency; the instantaneous angular velocity signal is extracted from the original signal of the signal conditioner, representing the torsional vibration of the blade and the change in the time interval between consecutive pulses of the signal conditioner; the time-synchronized average signal is the spectral average signal after FFT (fast Fourier transform) of the time domain signal, which removes any signal components unrelated to the blade speed and can reduce the noise in the complex signal spectrum.

[0053] Step 3: Compare the system signal with the fault signal in the database to obtain the blade operation status

[0054] The normal operating information and fault characteristic information in the database are obtained through finite element simulation. The purpose of finite element simulation is to obtain operational vibration information of blades in different states, such as defect-free, assembly defects, looseness, cracks, chatter, fretting wear, high-cycle fatigue, and low-cycle fatigue, to generate a long blade operation database. Among them, assembly defects are caused by differences in blade geometric characteristics and can be simulated by setting geometric characteristic differences; looseness is caused by excessive gaps between the blade root and the main shaft support structure and can be simulated by setting geometric characteristic differences; cracks are caused by foreign matter, manufacturing defects, high-cycle or low-cycle fatigue, resonance fatigue, and stress corrosion and can be simulated by setting different extreme operating conditions; chatter is caused by the interaction between aerodynamic forces and blade vibration displacement and can be simulated by calculating blade performance under different operating conditions; high-cycle fatigue and low-cycle fatigue can be simulated by setting a wide range of operating conditions and then using the corresponding calculation modules in the finite element software.

[0055] The time domain signal, frequency domain signal, short-time Fourier transform signal, instantaneous angular velocity signal, time-synchronized average signal obtained in step 2 and the normal operation information and fault feature information in the database are respectively calculated by least squares to obtain five sets of regression equations. The fault signal includes the time domain signal, frequency domain signal, short-time Fourier transform signal, instantaneous angular velocity signal, and time-synchronized average signal under different fault states.

[0056] Find the most matching regression equation, obtain the operating status corresponding to the regression equation according to the database, and thus obtain the current operating status of the long blade, realizing the health monitoring and management of the long blade.

[0057] Step 4: Obtain turbine safety through blade operating status

[0058] The blade operating status is a key evaluation indicator of steam turbine safety. The blade operating status information obtained in step 3 corresponds to the safety of the steam turbine. That is, the safety of the steam turbine is judged by the operating status of the long blades obtained in step 3.

[0059] Step 5: Evaluate the life of the turbine by considering the blade operating status or turbine safety.

[0060] When the blade is operating under a specific condition, the temperature, stress, and operating time at that moment are collected, and the corresponding lifespan is obtained by consulting the material's endurance strength. The total life loss during the turbine's service period is then obtained using the following formula, allowing for a turbine life assessment.

[0061]

[0062] Where, τ i Refers to the actual operating time of the turbine under the blade operating state i, τ b,jRefers to the life of the blade under operating state i, τ b Refers to the total life loss of the turbine during its service life.

[0063] Example 2 System for steam turbine safety monitoring and life assessment based on blade vibration parameters

[0064] To match the method in Example 1, this embodiment provides a steam turbine safety monitoring and life assessment system. Classified by implemented functions, it comprises four subsystems (a first subsystem, a second subsystem, a third subsystem, and a fourth subsystem). The signal acquisition and preprocessing layer corresponding to step 1 in Example 1 is integrated into the first subsystem, the signal conditioning layer in step 2 is integrated into the second subsystem, the fault identification and fault diagnosis layer in step 3 is integrated into the third subsystem, and the data display and human-computer interaction layer is integrated into the fourth subsystem. The core algorithm is a state-based steam turbine long blade fault monitoring and diagnosis algorithm.

[0065] The first subsystem integrates various sensors for monitoring the operating status of the long blades of the steam turbine, and pre-processes the operating data and monitoring information of the long blades of the steam turbine for subsequent analysis. The data pre-processed by the first subsystem is introduced into the second subsystem. The second subsystem expands the data signal after signal conditioning to generate a system signal and inputs it to the third subsystem and the fourth subsystem at the same time. The third subsystem uses the conditioned vibration information and operating condition information of the long blades of the steam turbine to monitor and diagnose faults, and outputs the corresponding conclusions to the fourth subsystem. The fourth subsystem integrates the output signals of the second and third subsystems, and outputs the health monitoring results of the long blades of the steam turbine in a concise and intuitive form through a preset human-machine display interface, which is convenient for analysis by power plant operation and maintenance personnel and equipment experts. Specifically,

[0066] 1) First subsystem

[0067] During the rotation of the long turbine blades, in addition to the existing measurement points and monitoring methods in the turbine system, sensors are installed for non-contact measurement of the long turbine blades. By measuring blade parameters (such as natural frequency, arrival time, and arrival angle), long blade faults, including fretting wear, cracks, bending, and loosening, can be detected. The first subsystem preprocesses the collected signals and introduces the preprocessed data into the second subsystem. For sensor settings and parameters, see Example 1.

[0068] 2) Second subsystem

[0069] A four-channel signal conditioner amplifies the long blade vibration signals from the first subsystem and transmits the data to the data acquisition board. The second subsystem, with a frequency response range of 0.05 Hz to 50,000 Hz, conditions the data from the first subsystem. The data output from the second subsystem is recorded using LabVIEW software and backed up on the fourth subsystem.

[0070] 3) The third subsystem

[0071] The conditioned vibration information and operating condition information of the long turbine blades are used for fault monitoring and diagnosis to identify the current operating status of the blades. Blade faults can also be identified, such as whether the blades have assembly defects, looseness, cracks, vibration, micro-wear, high-cycle fatigue, low-cycle fatigue and other defects, and the corresponding conclusions are output to the fourth subsystem.

[0072] 4) The fourth subsystem

[0073] At least a liquid crystal display is included, and the displayed content covers units such as the speed value of the long blade of the turbine, fault information, and operating status. The hardware part is equipped with a keyboard and a mouse. The user can easily change the program in the human-computer interaction interface to meet the operating requirements. The following operations can be performed as needed: (i) automatic calibration of sampling information, (ii) fixed or free setting of range, (iii) chart scaling, and (iv) display of operating status and fault information.

[0074] The present invention provides a technical solution for steam turbine fault diagnosis based on blade vibration, enabling safety monitoring and life assessment of steam turbines, and improving fault monitoring capabilities during the operation of long turbine blades. The first subsystem uses non-contact and non-destructive measurement methods to acquire long blade vibration information. The second subsystem performs signal conditioning on the acquired signals. The third subsystem, based on the blade vibration information, uses fault diagnosis methods to perform feature matching and fault identification on the status information. The fourth subsystem integrates various types of information into a human-computer interaction interface. Fault diagnosis conclusions are obtained by feature matching the system's real-time operating data, facilitating analysis by power plant operations and maintenance personnel and equipment experts, thereby improving the accuracy and portability of steam turbine safety monitoring and life assessment.

[0075] The method of the present invention is based on various vibration state parameters during the operation of the turbine (vibration-related information such as blade vibration amplitude, instantaneous angular velocity, shaft torsional vibration of bearing vibration and casing vibration), and obtains the change characteristics of the blade vibration during real-time operation according to the monitored values. The real-time change characteristics and fault characteristics are compared to realize fault diagnosis, calculate the real-time life of the turbine, and realize real-time evaluation of the turbine life; the fault diagnosis process and the life evaluation process do not need to interrupt the operation of the turbine.

[0076] The above embodiments are intended only to illustrate the technical concepts and features of the present invention. Their purpose is to enable those skilled in the art to understand the contents of the present invention and implement them accordingly. They are not intended to limit the scope of protection of the present invention. Any equivalent changes or modifications made in accordance with the spirit of the present invention are intended to be covered by the scope of protection of the present invention.

Claims

1. A method for monitoring the safety of a steam turbine, characterized in that: The steps include: Collect, store and pre-process the vibration data of the steam turbine and obtain the pre-processed data; converting the preprocessed data into a system signal; Comparing the system signal with the fault signal in the database to obtain the blade operation status and judge the safety of the steam turbine; A plurality of groups of sensors for collecting vibration data of the turbine blades are arranged in the longitudinal extension direction of the turbine, one group of sensors corresponding to one group of blades, and each group of sensors is arranged within the width direction range of one group of blades; Each sensor group includes at least two optical sensors and one eddy current sensor, wherein the eddy current sensor is arranged between the two optical sensors; The central angles corresponding to the projections of adjacent sensors on the cross section of the steam turbine are 30 to 60 degrees; the projections of the plurality of sensors on the longitudinal section of the steam turbine are located on different horizontal and vertical planes; The step also includes calculating the total life loss during the service period of the steam turbine by collecting the real-time temperature, stress and operating time of the blade under a certain working condition; The total life loss is calculated using the following formula: Where, τ i Refers to the actual operation time of the turbine under the blade operation state i, τ b，j Refers to the life of the blade under operating state i, τ b Refers to the total life loss of the steam turbine during its service life; The comparison is to perform least squares calculation on the obtained system signal and the normal operation information and fault feature information in the database, respectively, to obtain multiple sets of regression equations, find the most matching regression equation, and obtain the operating status corresponding to the regression equation according to the database, thereby obtaining the current operating status of the long blade and realizing blade health monitoring and management; The normal operation information and fault characteristic information in the database are obtained through finite element simulation. The purpose of finite element simulation is to obtain the operating vibration information of the blade in different states and generate a long blade operation database. The steam turbine safety monitoring method is implemented based on a steam turbine safety monitoring and life assessment system, which includes a first subsystem, a second subsystem, a third subsystem, and a fourth subsystem; the first subsystem preprocesses the collected vibration data and introduces the preprocessed data into the second subsystem; The second subsystem is used to amplify the blade vibration signal from the first subsystem and complete the data signal conditioning of the first subsystem. The signal output by the second subsystem is recorded by software and backed up and stored in the fourth subsystem. The third subsystem uses the conditioned turbine blade vibration signal and operating condition signal to perform fault monitoring and diagnosis, identify the current operating status of the blade, and output the corresponding conclusion to the fourth subsystem. The fourth subsystem performs data display and human-computer interaction.

2. The monitoring method according to claim 1, characterized in that: The vibration data of the steam turbine includes the vibration data of the blades and the vibration data of other components of the steam turbine that reflects the vibration information of the blades; the vibration data of the blades is obtained by using optical sensors and / or eddy current sensors; the vibration data of the steam turbine includes the shaft torsional vibration data of the blade amplitude, instantaneous angular velocity, bearing vibration and casing vibration.

3. The monitoring method according to claim 2, characterized in that: The blade amplitude is calculated by the following formula: P=n·πD·ΔT / 60 Where P is the blade amplitude, n is the rotational speed, D is the diameter of the blade tip monitoring position, and ΔT is the time difference between the actual arrival time and the theoretical arrival time of the blade tip.

4. The monitoring method according to claim 1, characterized in that: The pre-processing is performed according to the following steps: the vibration data is screened using the Raida law to remove data that is obviously out of range due to abnormal sensor readings, and then filtering and normalizing are performed.

5. The monitoring method according to claim 1, characterized in that: A channel signal conditioner is used to convert the pre-processed data into a system signal; the system signal includes a time domain signal, a frequency domain signal, a short-time Fourier transform signal, an instantaneous angular velocity signal, and a time-synchronized average signal.

6. The monitoring method according to claim 1, characterized in that: The database stores blade vibration information of the steam turbine blades in a normal operating state or a fault state.

7. The monitoring method according to claim 6, characterized in that: The failure conditions include blade cracking, high cycle fatigue, blade rubbing, blade root loosening, erosion, creep and corrosion.

Images