Automatic acquisition system and early warning method for crack initiation and expansion information of rotating wheel of large-scale water-turbine generator set

Abstract

The invention provides an automatic acquisition system and an early warning method for crack initiation and propagation information of a runner of a large-scale water-turbine generator set, and relates to the field of monitoring of health conditions of runners of water turbines, on-line monitoring of cracks of runner bodies is carried out by adopting an acoustic emission technology, and acoustic emission source signals are generated at material deformation or fracture positions; the material is used as a medium to be spread to the surface of the material, mechanical vibration is generated on the surface of the material, the acoustic emission sensor converts surface displacement generated by the mechanical vibration into an electric signal, the signal is amplified and filtered to obtain a characteristic signal, and crack defect initiation and expansion signal monitoring of the water turbine runner in the in-service state can be achieved. And data support is provided for water turbine runner state maintenance.

Description

Technical Field

[0001] The invention relates to the field of water turbine runner health status monitoring, and in particular to an automatic collection system and early warning method for crack initiation and expansion information of a runner of a large water turbine generator set. Background Art

[0002] The turbine runner is one of the most critical components of the mechanical system of a hydropower station, and its operating health is related to the safe and stable operation of the entire unit. During the operation of the unit, the violent and frequent interactions between the turbine components will cause potential damage to the safe operation of the unit. Especially for turbine units that have been in operation for decades, most of the hydraulic operating conditions have changed significantly compared to the beginning of production. These changes in operating conditions may cause excessive vibration of the unit and lead to cracks or even damage to turbine components. Therefore, the automatic collection and early warning of the initiation and extension of cracks in the turbine runner is of great significance to promoting the condition maintenance and safe and stable operation of hydropower units. Summary of the invention

[0003] The purpose of the present invention is to provide an automatic collection system and early warning method for crack initiation and expansion information of a large-scale hydro-turbine generator set. The acoustic emission technology is used to perform online monitoring of cracks in the turbine body. The acoustic emission source signal is generated at the deformation or fracture of the material, and is transmitted to the surface of the material with the material itself as the medium, causing mechanical vibration on the surface of the material. The acoustic emission sensor converts the surface displacement generated by the mechanical vibration into an electrical signal. The signal is amplified and filtered to obtain a characteristic signal. The initiation and expansion signals of crack defects in the turbine runner under in-service conditions can be monitored, and data support is provided for the condition inspection and maintenance of the turbine runner.

[0004] In order to realize the above technical features, the object of the present invention is achieved as follows: a large-scale hydro-generator unit runner crack initiation and extension information automatic collection system, comprising: An acoustic emission sensor, which is used to collect sound wave signals generated by a sound source inside the material during the operation of the runner of the hydro-generator set; A vibration sensor, wherein the vibration sensor is used to collect vibration signals during the operation of the runner of the hydro-generator set; A data collector, which is connected to the acoustic emission sensor and the vibration sensor at the same time, thereby forming a data acquisition system together; The data acquisition system is connected to the collector ring via a signal line; The slip ring is connected to a remote receiving computer via a signal line; The remote receiving computer is used to store data in a database, and to call the data in the database through the algorithm computer to perform algorithm calculation and data display.

[0005] The acoustic emission sensor and the vibration sensor are arranged and installed on the main shaft and the flange of the runner of the hydro-generator set, or are installed by drilling holes.

[0006] The acoustic emission sensor, the vibration sensor and the data collector are integrally installed inside the waterproof housing and are filled with waterproof material to form a data collection system.

[0007] The data acquisition system is connected to a specially installed collector ring via a 485 communication line.

[0008] The data acquisition system is configured with at least 8 channels for data acquisition, each channel has a sampling frequency of at least 5 MHz, can store current data samples, selects a core board STM32H743 single-chip microcomputer, and uses an AD620 chip amplifier for the sampling circuit.

[0009] The output signal voltage of the vibration sensor is 5-15V. After five-frequency division, the signal is compressed to 1-3V. The divided signal is differentially input into the AD620 chip and then input into the microcontroller pin after passing through a voltage follower.

[0010] It also includes a sensor power supply module, which is used to supply power to the data acquisition system; It also includes a filtering module, which is used to filter and pre-process the sound wave signal and vibration signal collected by the data acquisition system; The filtering module is connected to the signal amplifying module and is used to amplify the filtered signal; The signal amplification module is connected to the data acquisition system; Also includes a gain switching module and a filter bandwidth switching module; The data acquisition system is connected to the gain switching module and the filter bandwidth switching module at the same time, the gain switching module is connected to the signal amplification module, and the filter bandwidth switching module is connected to the filter module.

[0011] A method for automatically collecting and warning information on the initiation and extension of cracks in a large-scale hydro-turbine generator set, the method is implemented by using the automatic collection system for the initiation and extension of cracks in a large-scale hydro-turbine generator set, and comprises the following steps: Step 1, installation and layout of data acquisition system: According to the specific signal source to be monitored, the data acquisition system is arranged and installed at the appropriate position of the turbine generator runner; Step 2, data collection and transmission of data acquisition system: During the operation of the turbine generator set rotor, the data acquisition system uses 8-channel ADC signal acquisition and 16-bit acquisition to achieve multi-channel signal acquisition; Step 3, conditioning of the collected signal: Wavelet packet decomposition is used to pre-process the collected signals for denoising, and it is applied before machine learning fault identification to quickly process the signals and improve the signal-to-noise ratio. Step 4: Extraction of crack fault signal of turbine generator set runner blade: The variational mode decomposition algorithm is used to extract crack fault features from the preprocessed signal, thereby improving the applicability of crack acoustic emission signal extraction mixed in strong background noise.

[0012] Also includes system testing steps: Through the above-mentioned automatic data collection system, the defect forms on the impeller blades of various hydro-turbine generator sets are tested at no-load and full-load to obtain monitoring data and verify the monitoring mechanism. Finally, the developed automatic data collection system is used for on-site application in a running turbine to obtain raw data, and the status of the turbine runner is analyzed based on the raw data. The monitoring results are corrected in combination with the status of the turbine runner during maintenance to provide support for long-term monitoring.

[0013] Beneficial effects of the present invention: 1. The method for automatically collecting and warning information on crack initiation and expansion of a large-scale hydro-turbine generator set described in the present invention adopts structural health monitoring technology to realize online monitoring of the working status of the runner, diagnose and evaluate damage conditions such as crack defects, without affecting the normal operation of the hydro-turbine unit, which is of great significance for reducing unnecessary maintenance work and ensuring the safety and reliability of the unit.

[0014] 2. The acquisition and early warning method of the present invention adopts the core board STM32H743 single-chip microcomputer and AD620 chip operational amplifier circuit, which has the characteristics of small size, low power consumption, low noise, low offset, high open-loop gain, etc., and is suitable for measuring and amplifying weak signals.

[0015] 3. The hardware circuit design of the present invention has 8-channel ADC signal acquisition and 16-bit acquisition, which can realize multi-channel signal acquisition.

[0016] 4. The present invention introduces wavelet packet analysis to perform signal denoising preprocessing on the measured signal, and applies it before fault identification by machine learning to quickly process the signal and improve the signal-to-noise ratio of the signal.

[0017] 5. The present invention introduces an optimized VMD algorithm to extract fault features of signals to improve the applicability of crack AE signal extraction mixed in strong background noise. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0019] Figure 1 This is a schematic diagram of the system of the present invention.

[0020] Figure 2 This is a diagram showing the arrangement of the acoustic emission sensors of the present invention.

[0021] Figure 3 It is the signal transmission route map of the acoustic emission sensor of the present invention.

[0022] Figure 4 This is the AD620 chip operational amplifier circuit diagram of the present invention.

[0023] Figure 5 This is the overall block diagram of the system of the present invention. DETAILED DESCRIPTION

[0024] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0025] Embodiment 1: When cracks on the impeller blades are generated and developed, a large number of acoustic emission signals are generated. The present invention realizes the automatic collection and early warning functions of crack initiation and expansion information of the impeller of a large hydropower unit based on acoustic emission technology, thereby realizing the health status monitoring and evaluation of the impeller of the hydropower unit, and providing data support for the status maintenance of the core components of the hydropower unit.

[0026] 1. Layout method of acoustic emission signal sensor: The operation of the turbine is accompanied by high-intensity background noise, which mainly includes water flow impact noise, vibration noise, unit rotation noise, etc. The background noise intensity is relatively high, while the amplitude of the impeller crack acoustic emission signal is relatively small. The crack acoustic emission signal is generally submerged in these strong background noises, so the selection of the layout location of the acoustic emission sensor is crucial.

[0027] The present invention installs the acoustic emission sensor, vibration sensor and data collector on the main shaft and the runner flange, or installs them by drilling holes. The acoustic emission sensor, vibration sensor and data collector are integrated into one and waterproof filling is adopted. The data acquisition system is connected to the specially installed collector ring through 485 connection, the digital signal is brushed out, connected to the remote receiving computer through 485, and the data is stored in the database. The algorithm computer is responsible for algorithm calculation and data display. The system implementation principle is as follows Figure 1 shown.

[0028] 2. Data acquisition system data acquisition and transmission: The data acquisition system is equipped with 8 data acquisition channels, each channel has a sampling frequency of up to 5MHz, and has a large memory that can store current data samples. In order to meet the above requirements, the present invention selects the core board STM32H743 single-chip microcomputer. The sampling circuit uses the AD620 chip amplifier, which uses a high-precision amplifier. The AD620 chip circuit is as follows Figure 4 .

[0029] The AD620 chip has the characteristics of small size, low power consumption, low noise, low offset, high open-loop gain, etc., which is suitable for measuring and amplifying weak signals. The output signal voltage of the vibration sensor is 5-15V. After five-frequency division, the signal is compressed to 1-3V, which is within the reference voltage range. The divided signal is differentially input into the AD620 chip and then input into the microcontroller pin after the voltage follower.

[0030] In order to meet the needs of data acquisition, the hardware circuit is designed with 8-channel ADC signal acquisition and 16-bit acquisition, which can realize multi-channel signal acquisition.

[0031] 3. Signal conditioning: The key point of the acoustic emission signal processing technology for impeller crack monitoring lies in signal extraction and identification, and the focus is on signal preprocessing. First, signal denoising is performed. Noise signal is an inevitable problem in vibration monitoring. The on-site environment when the turbine is running is noisy, and operation and environmental noise are inevitable. If the noise signal is not processed, the useful fault signal will be masked by the noise signal. In the subsequent fault feature extraction process, information cannot be accurately extracted, affecting the accuracy of subsequent machine learning fault pattern recognition. Therefore, signal denoising is one of the problems of crack monitoring.

[0032] When measuring a signal, it is necessary to denoise the signal. Wavelet denoising is a method of processing the coefficients of a signal after decomposition through wavelet transform. It amplifies useful signal coefficients and eliminates useless noise coefficients by setting thresholds.

[0033] Wavelet packet decomposition is based on wavelet transform, and improves the shortcoming of low resolution and accuracy of wavelet denoising in the high-frequency band of the signal. The use of wavelet packet transform can decompose the harmonic signal more finely in time and frequency, and the detailed processing of the high-frequency part of the signal is better than wavelet analysis. This method has the advantages of higher time-frequency resolution and high denoising accuracy for fine processing of signal components in the high-frequency band. This method overcomes the shortcomings of Fourier transform in non-steady-state signal analysis, improves the ability of local analysis of signals, and can better retain useful signals during denoising. In summary, the present invention introduces wavelet packet decomposition to perform signal denoising preprocessing on the measured signal, and applies it before machine learning fault identification to quickly process the signal and improve the signal-to-noise ratio of the signal.

[0034] The fault signal of the crack of the turbine runner blade is often non-stationary, and contains many non-stationary impact parts. The fault vibration signal is actually a non-stationary multi-component amplitude modulation-frequency modulation signal. At this time, traditional methods such as Fourier transform and wavelet analysis will have certain limitations. These transforms lack adaptability and affect the accuracy of fault diagnosis. The present invention will introduce an optimized variational mode decomposition (VMD) algorithm to extract the fault features of the signal to improve the applicability of crack acoustic emission (AE) signal extraction mixed in strong background noise.

[0035] 4. Data acquisition system test: Using the above data acquisition system, the sensor arrangement method and signal processing method of the turbine runner damage acoustic emission signal are tested by no-load, full-load, and various defect forms on the turbine runner to obtain monitoring data, and the aforementioned monitoring mechanism is verified. Finally, the developed data acquisition system is used on-site in the running turbine to obtain the original data, and the status of the turbine runner is analyzed based on the data. The monitoring results are corrected in combination with the status of the turbine runner during maintenance to provide support for long-term monitoring. The overall block diagram of the acquisition system is as follows Figure 5 shown.

[0036] Embodiment 2: A method for automatically collecting and warning information on the initiation and extension of cracks in a large-scale hydro-turbine generator set, the method is implemented by using the automatic collection system for the initiation and extension of cracks in a large-scale hydro-turbine generator set, and comprises the following steps: Step 1, installation and layout of data acquisition system: According to the specific signal source to be monitored, the data acquisition system is arranged and installed at the appropriate position of the turbine generator runner; Step 2, data collection and transmission of data acquisition system: During the operation of the turbine generator set rotor, the data acquisition system uses 8-channel ADC signal acquisition and 16-bit acquisition to achieve multi-channel signal acquisition; Step 3, conditioning of the collected signal: Wavelet packet decomposition is used to pre-process the collected signals for denoising, and it is applied before machine learning fault identification to quickly process the signals and improve the signal-to-noise ratio. Step 4: Extraction of crack fault signal of turbine generator set runner blade: The variational mode decomposition algorithm is used to extract crack fault features from the preprocessed signal, thereby improving the applicability of crack acoustic emission signal extraction mixed in strong background noise.

[0037] The above description is only a specific example of the present invention and does not constitute any limitation to the present invention. It is obvious that for professionals in this field, after understanding the content and principle of the present invention, various modifications and changes in form and details may be made without departing from the principle and structure of the present invention, but these modifications and changes based on the idea of ​​the present invention are still within the scope of protection of the claims of the present invention.

Claims

1. A large-scale hydro-generator unit runner crack initiation and expansion information automatic collection system, characterized in that: include: An acoustic emission sensor, which is used to collect sound wave signals generated by a sound source inside the material during the operation of the runner of the hydro-generator set; A vibration sensor, wherein the vibration sensor is used to collect vibration signals during the operation of the runner of the hydro-generator set; A data collector, which is connected to the acoustic emission sensor and the vibration sensor at the same time, thereby forming a data acquisition system together; The data acquisition system is connected to the collector ring via a signal line; The slip ring is connected to a remote receiving computer via a signal line; The remote receiving computer is used to store data in a database, and to call the data in the database through the algorithm computer to perform algorithm calculation and data display.

2. According to claim 1, a large-scale hydro-generator set runner crack initiation and expansion information automatic acquisition system is characterized in that: The acoustic emission sensor and the vibration sensor are arranged and installed on the main shaft and the flange of the runner of the hydro-generator set, or are installed by drilling holes.

3. According to claim 2, a large-scale hydro-generator unit runner crack initiation and expansion information automatic acquisition system is characterized in that: The acoustic emission sensor, the vibration sensor and the data collector are integrally installed inside the waterproof housing and are filled with waterproof material to form a data collection system.

4. According to claim 3, a large-scale hydro-generator unit runner crack initiation and expansion information automatic acquisition system is characterized in that: The data acquisition system is connected to a specially installed collector ring via a 485 communication line.

5. According to claim 3, a large-scale hydro-generator unit runner crack initiation and expansion information automatic acquisition system is characterized in that: The data acquisition system is configured with at least 8 channels for data acquisition, each channel has a sampling frequency of at least 5 MHz, can store current data samples, selects a core board STM32H743 single-chip microcomputer, and uses an AD620 chip amplifier for the sampling circuit.

6. According to claim 5, a large-scale hydro-generator unit runner crack initiation and expansion information automatic acquisition system is characterized in that: The output signal voltage of the vibration sensor is 5-15V. After five-frequency division, the signal is compressed to 1-3V. The divided signal is differentially input into the AD620 chip and then input into the microcontroller pin after passing through a voltage follower.

7. According to claim 5, a large-scale hydro-generator set runner crack initiation and expansion information automatic acquisition system is characterized in that: It also includes a sensor power supply module, which is used to supply power to the data acquisition system; It also includes a filtering module, which is used to filter and pre-process the sound wave signal and vibration signal collected by the data acquisition system; The filtering module is connected to the signal amplifying module and is used to amplify the filtered signal; The signal amplification module is connected to the data acquisition system.

8. According to claim 7, a large-scale hydro-generator set runner crack initiation and expansion information automatic acquisition system is characterized in that: Also includes a gain switching module and a filter bandwidth switching module; The data acquisition system is connected to the gain switching module and the filter bandwidth switching module at the same time, the gain switching module is connected to the signal amplification module, and the filter bandwidth switching module is connected to the filter module.

9. A method for automatically collecting and warning information on crack initiation and expansion of a large hydro-turbine generator, characterized in that: The method is implemented by using the large-scale hydro-generator set runner crack initiation and extension information automatic acquisition system as described in claim 8, and comprises the following steps: Step 1, installation and layout of data acquisition system: According to the specific signal source to be monitored, the data acquisition system is arranged and installed at the appropriate position of the turbine generator runner; Step 2, data collection and transmission of data acquisition system: During the operation of the turbine generator set rotor, the data acquisition system uses 8-channel ADC signal acquisition and 16-bit acquisition to achieve multi-channel signal acquisition; Step 3, conditioning of the collected signal: Wavelet packet decomposition is used to pre-process the collected signals for denoising, and it is applied before machine learning fault identification to quickly process the signals and improve the signal-to-noise ratio. Step 4: Extraction of crack fault signal of turbine generator set runner blade: The variational mode decomposition algorithm is used to extract crack fault features from the preprocessed signal, thereby improving the applicability of crack acoustic emission signal extraction mixed in strong background noise.

10. A method for automatically collecting and warning information on crack initiation and expansion of a large-scale hydro-generator unit according to claim 9, characterized in that: Also includes system testing steps: Through the above-mentioned automatic data collection system, the defect forms on the impeller blades of various hydro-turbine generator sets are tested at no-load and full-load to obtain monitoring data and verify the monitoring mechanism. Finally, the developed automatic data collection system is used for on-site application in a running turbine to obtain raw data, and the status of the turbine runner is analyzed based on the raw data. The monitoring results are corrected in combination with the status of the turbine runner during maintenance to provide support for long-term monitoring.

Images