Pumped storage power station unit runner crack diagnosis and service life evaluation method

Abstract

The invention relates to the technical field of pumped storage power stations, in particular to a pumped storage power station unit runner crack diagnosis and life evaluation method, which comprises the following steps: S1, accurately positioning a runner crack detection position; s2, accurately analyzing the shape and size of the runner crack; and S3, analyzing the performance microstructure of the runner material. According to the method, the crack condition, the material characteristics and the stress distribution under the actual working condition of the runner can be comprehensively and accurately evaluated by detecting the runner crack, analyzing the shape and the size of the crack, analyzing the microstructure of the material performance and finally simulating and analyzing the stress distribution of the working load of the runner; the method provides a solid data basis and theoretical basis for a crack propagation mechanism, life cycle prediction and runner life evaluation, can effectively realize multi-dimensional and multi-technical-means fusion, ensures the precision of crack detection and analysis, and can provide omnibearing data support for subsequent crack propagation prediction and life evaluation.

Description

Technical Field

[0001] The invention relates to the technical field of pumped storage power stations, and in particular to a method for diagnosing cracks and assessing the life of a runner of a pumped storage power station unit. Background Art

[0002] As an important energy regulation and storage system, pumped storage power stations are widely used to balance grid loads and regulate power supply. In pumped storage power stations, the unit runner is one of the core components. It is responsible for driving the water flow and realizing the function of pumping or generating electricity through rotation. Since the unit runner operates under high load and complex working conditions for a long time, it faces huge mechanical stress and fatigue load, and is prone to cracks or other damage. Therefore, ensuring the structural safety and operating life of the runner is the key to ensuring the efficient and safe operation of the pumped storage power station.

[0003] At present, traditional methods for crack detection and life assessment of turbine runners mainly rely on manual inspection, conventional ultrasonic testing, visual inspection and other means. Although these methods can detect the existence of cracks, they have great limitations in crack morphology, size analysis and prediction of crack expansion trends. Traditional crack detection methods are often difficult to comprehensively and accurately evaluate the damage condition of the turbine runner, especially for the evaluation of complex factors such as crack micromorphology and stress distribution. Traditional technologies often cannot provide sufficient data support. In addition, the crack expansion process is usually complex and changeable. Relying solely on manual experience and conventional technical means cannot effectively predict the crack expansion trend and the remaining life of the unit.

[0004] With the advancement of technology, some crack assessment technologies based on finite element analysis, fatigue testing and other methods have been gradually applied to runner crack diagnosis, but these technologies still have shortcomings such as lack of real-time performance, limited accuracy, and inability to fully combine factors such as material properties and load stress in crack propagation prediction. Therefore, we provide a method for crack diagnosis and life assessment of pumped storage power station units. Summary of the invention

[0005] In view of the above-mentioned shortcomings of the prior art, the first object of the present invention is to provide a method for crack diagnosis and life assessment of a runner of a pumped storage power station unit, so as to solve the problems in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A method for crack diagnosis and life assessment of a runner of a pumped storage power station unit comprises the following steps:

[0008] S1. Accurate positioning of the wheel crack detection position;

[0009] S2. Accurate analysis of the shape and size of the runner crack;

[0010] S3, performance and microstructure analysis of runner materials;

[0011] S4. Simulation analysis of stress distribution of runner working load;

[0012] S5. Crack propagation mechanism and life cycle prediction;

[0013] S6. Prediction of remaining service life of runner in life assessment;

[0014] S7. Maintenance preventive decision-making based on crack diagnosis;

[0015] S8. Data feedback mechanism for crack diagnosis optimization.

[0016] The present invention is further configured as follows: in the step S1, in the precise positioning of the wheel crack detection position:

[0017] S1.1. In the operation state of the power station, the vibration characteristics of the runner are monitored in real time through a vibration sensor. A highly sensitive PCB 353B18 acceleration sensor is used for data acquisition. The sampling frequency should be set to above 10 kHz to ensure that the vibration fluctuations caused by high-frequency cracks can be captured. Through time domain analysis and frequency domain analysis, combined with the natural frequency and vibration mode of the runner, fast Fourier transform is used for spectrum analysis to identify possible crack locations and crack source areas. By comparing the vibration data under different working conditions, the evolution trend and location information of the crack can be effectively determined;

[0018] S1.2. Use an Olympus 38DL PLUS ultrasonic flaw detector with a broadband sensor to scan the rotor using a handheld probe or an automated robot platform. The test is carried out in a stationary rotor environment, and the ultrasonic sensor is placed on the outer edge of the rotor. Array probe technology is used to accurately determine the shape and depth of the crack through angle scanning and depth scanning, combined with sound wave attenuation and reflection time. The return signal is processed using waveform analysis software, and the location and size of the crack are determined using reflection intensity and echo time.

[0019] S1.3. Under heating conditions, the heat conduction characteristics of the cracks will be different from those of the surrounding areas, resulting in thermal imaging differences. In the experimental environment, the wheel needs to be loaded to a certain working condition, and the FLIR T640 infrared thermal imager is used for data acquisition. The resolution of the thermal imager should be 640x480, and the measurement accuracy should be ±0.1℃. The surface of the wheel is evenly heated using an electric heating film, and the temperature change is monitored in real time to capture the thermal anomalies of surface cracks and damage. The thermal imaging images are processed by MATLAB software, using the image difference analysis method to determine the precise location and extension characteristics of the cracks;

[0020] S1.4. FBG sensors are placed on the surface of the runner to monitor the strain distribution of various parts of the runner. During the experiment, the installation position of the sensor needs to be optimized to capture the strain changes in key stress areas. The experiment should be carried out under full load operation. A high-precision spectrometer is used for data collection with a sampling frequency of 1 kHz. By analyzing different strain distributions and combining finite element simulation data, the possible areas where cracks may occur are determined, and the detection strategy is adjusted through real-time data feedback.

[0021] The present invention is further configured as follows: in the step S2, in the precise analysis of the wheel crack shape and size:

[0022] S2.1. In the experiment, a high-resolution GE phoenix v|tome|x industrial CT scanner is used to perform three-dimensional imaging of the sample. The scanning resolution can reach the micron level. The experimental environment should ensure that the rotor sample is in a stationary state, and the rotor needs to be cut non-destructively for scanning. The CT data processing software of VGStudio MAX can reconstruct the obtained two-dimensional tomography data into three dimensions, accurately presenting the three-dimensional morphology, depth and position of the crack. Through three-dimensional modeling and virtual cutting technology, the morphological characteristics of the crack can be accurately determined, and then its expansion trend can be analyzed;

[0023] S2.2. During the experiment, a Zeiss Axio Imager high-power optical microscope was used to observe cracks on the sample surface. In the experimental environment, surface pretreatment was required in the area where the cracks were clearer. The sample surface was treated with metallographic polishing and etching agents to facilitate observation. The resolution of the optical microscope should reach 0.1 microns, which can effectively display the microstructural changes of the cracks and the starting position of the microcracks. Combined with Image-Pro Plus image processing technology, the width and depth of the cracks can be accurately measured, and the crack extension direction can be analyzed.

[0024] S2.3. In the experiment, the Keyence LJ-V7080 high-precision laser scanner was used to scan the surface of the rotor. The scanning accuracy can reach 1 micron. The experimental environment requires the rotor to be stationary and ensure the suitability of the scanning angle and surface finish. The laser scanning data is processed by a computer to form a three-dimensional image of the rotor surface. Combined with the image analysis software, the specific size of the crack can be extracted. Through quantitative analysis, the width, length and depth of the crack can be obtained, providing accurate geometric data for subsequent crack extension analysis.

[0025] S2.4. During the experiment, the crack area was photographed with a Canon EOS 5D Mark IV high-resolution camera. The image was processed for uniform illumination to eliminate the influence of uneven surface light. Then, the image processing toolbox in MATLAB was used to extract the crack boundary using the Canny algorithm. The length, width, and morphological characteristics of the crack were calculated using image segmentation and feature extraction algorithms. The crack propagation speed and possible affected area were determined using statistical analysis.

[0026] The present invention is further configured as follows: in the step S3, the microstructure analysis of the performance of the runner material:

[0027] S3.1. In the experiment, a Rockwell HRB hardness tester is used to test the hardness of the wheel surface. The hardness range should be between 60HRA and 80HRA. By sampling at different positions, the hardness distribution of the material is evaluated and possible crack initiation areas are identified. The crack propagation experiment can be carried out by an Instron 8800 fatigue testing machine to simulate the stress changes of the wheel during long-term operation. Combined with the material hardness data, its crack resistance performance is evaluated;

[0028] S3.2. In the experiment, the fatigue life test of the runner is carried out using the MTS 810 high-load fatigue testing machine. Alternating loads of different amplitudes and frequencies are applied. Material samples should be sampled from the runner for standardized tests. The test sample size is 20mm×20mm×10mm. By gradually loading and monitoring the development of cracks, the fatigue life and crack growth rate of the runner are recorded, and a runner fatigue life model is established to provide a basis for crack growth prediction.

[0029] S3.3, FEI Quanta 450 scanning electron microscope was used to perform surface and fracture analysis on the wheel material. The scanning acceleration voltage was set to 15 kV and the magnification was 500 to 5000 times. The composition distribution of the material was analyzed by electron probe to identify the weak area where the crack occurred. The microscopic mechanism of crack initiation was determined in combination with fracture analysis.

[0030] S3.4. Oxford Instruments X-Strata 980 spectrometer was used to analyze the chemical composition of the runner material in detail, especially the content of elements such as carbon, chromium, and nickel, to ensure that the alloy composition of the material meets the design requirements. By comparing the actual composition with the standard requirements, the crack resistance and long-term operation reliability of the runner were evaluated.

[0031] The present invention is further configured as follows: in the step S4, in the simulation analysis of the stress distribution of the runner working load:

[0032] S4.1. By defining the material properties of the runner, the yield strength (about 350 MPa), elastic modulus (210 GPa) and Poisson's ratio (0.3) of the steel, the simulation results are ensured to be consistent with the actual performance. The actual operating data, including the running speed, operating temperature (200°C), pressure and load acting on the runner, are input. The stress distribution under different load conditions is simulated, especially in the runner blades, connecting parts and supporting areas, with a focus on the high stress areas in the stress concentration area, which may become the starting point of the crack. This analysis provides basic data for subsequent crack propagation prediction and life assessment.

[0033] S4.2. Deploy a sensor network in the actual working environment to collect the load data (torque, vibration, temperature, pressure, etc.) of the runner in real time. Sensors such as strain gauges, accelerometers and thermocouples can be used to monitor the stress state and temperature changes of the runner under different working conditions. The data is transmitted to the data analysis platform through a wireless communication system to calculate the dynamic load and stress distribution of the runner in real time. According to the changes in real-time data, the working performance of the runner under different working conditions is analyzed, potential load fluctuations are identified, and compared with the results of finite element analysis to further accurately evaluate the working state of the runner and provide data support for real-time monitoring of stress concentration areas and crack initiation points;

[0034] S4.3. By combining finite element analysis with real-time monitoring data, focus on identifying areas on the runner where stress concentration may occur, usually including the root of the runner blade, bolt holes and runner support parts. Use stress intensity factor (K) to evaluate the crack growth potential in these areas. The fracture toughness data of the material needs to be introduced during the analysis. Based on the linear fracture mechanics properties of the material, the critical stress intensity factor (Kc) is used to compare the stress in these areas with the material's resistance to crack growth. If the stress in the stress concentration area exceeds the critical value of the material, the risk of crack growth will increase significantly. This analysis can provide key clues for the prediction of subsequent crack growth paths.

[0035] S4.4. Based on the historical operating data of the runner, the mining fatigue theory is applied to analyze the impact of cumulative load on the runner life. By collecting load data under different working conditions, the stress amplitude of each load cycle is calculated, and the load amplitude-cycle number relationship is used to evaluate the fatigue damage of the runner material under different stress levels. When analyzing historical data, the changes in factors such as temperature, rotation speed, and pressure should be considered, and the cumulative effect should be calculated in combination with the SN curve of the material. Through periodic fatigue tests, the damage accumulation characteristics of the runner under different load histories are verified. The research results provide a theoretical basis for the possible fatigue damage points of the runner, especially the crack initiation that may occur in the stress concentration area.

[0036] The present invention is further configured as follows: in the step S5, crack propagation mechanism and life cycle prediction:

[0037] S5.1. Using the Paris rule ( ) is used to predict the crack growth rate, where is the range of stress intensity factor, and m are material constants. First, the cracks in the runner are preliminarily diagnosed to determine their initial size and location. Then, the changes in stress intensity factors under different loading conditions are calculated. The experimental data (C and m values ​​of steel) are used to calibrate the Paris law, and then the crack growth rate under different loads during the use of the runner is predicted.

[0038] S5.2. Based on the theory of fracture mechanics, a crack propagation path model of the runner under different working conditions is established to analyze the direction and path of crack propagation. Through the stress field and the relative position of the crack, finite element analysis is used to simulate the crack propagation process under different loads. During the simulation process, the starting position of the crack, the geometric shape of the runner and the fracture toughness of the material should be considered. The crack propagation path near the root of the runner blade and the bolt hole is analyzed in detail. In the experiment, the crack propagation path can be verified by CT scanning or X-ray imaging technology to ensure the accuracy of the model. Through the simulation analysis of the crack propagation path, it can be determined whether there is crack merging or damage to the supporting structure;

[0039] S5.3. Simulate the crack growth process under dynamic loads, simulate the influence of periodic or random loads (speed changes, load fluctuations) on crack growth that the runner bears in actual work, use finite element software with time domain or frequency domain analysis functions to analyze the dynamic response of the runner under complex loads, and calculate the influence of dynamic loads on crack growth rate. The crack growth rate will not only depend on the static stress intensity factor, but also on factors such as load frequency and amplitude. In order to verify the accuracy of the simulation model, the crack growth rate under dynamic loads can be calibrated by comparing laboratory fatigue tests and field monitoring data;

[0040] S5.4. Combined with the crack monitoring data collected during the operation of the runner, the crack propagation model is updated regularly. Crack propagation monitoring technology (ultrasonic, X-ray, magnetic particle testing) is used to track the size changes of the cracks in real time, and the actual detection results are fed back to the crack propagation prediction model. Through the continuously updated crack data, the prediction model of the crack propagation rate and path is optimized to ensure that the model can reflect the actual performance of the runner under different working conditions. Dynamic monitoring equipment can be used in the experiment to perform non-destructive testing on the runner regularly, and the prediction model can be adjusted according to the test results to ensure a more accurate assessment of the remaining life of the runner.

[0041] The present invention is further configured as follows: in the step S6, in the prediction of the remaining service life of the runner life assessment:

[0042] S6.1. Based on the maintenance records, failure modes and operating conditions of the runner in previous years, an empirical model is constructed using multivariate regression analysis or machine learning algorithms (support vector machine, random forest) to predict the remaining service life of the runner. By analyzing historical data, the key factors affecting the life of the runner (speed, load, vibration) are identified, and based on the changing trends of these factors, the fatigue damage of the runner in the future is predicted. In the experimental environment, the accuracy of the model can be verified by comparing it with field data through long-term laboratory tests and necessary corrections can be made;

[0043] S6.2. Combine the fatigue data of the material (SN curve, fatigue limit) and the stress analysis results of the runner, use Miner's law or multi-axial fatigue damage model to predict the fatigue life, and predict the fatigue life of the runner by accumulating fatigue damage under different load conditions, especially at the critical fatigue points of key components (blades, support rings). In the experiment, a fatigue testing machine is used to simulate the actual working conditions, and the fatigue limits under different conditions are measured to verify the accuracy of the model prediction. Through the accumulation analysis of fatigue damage, the critical points where the runner may fail are identified and corresponding maintenance strategies are formulated;

[0044] S6.3. Use reliability analysis methods (Weibull distribution, Monte Carlo simulation) to evaluate the overall life distribution of the runner. According to the use environment, loading history and material properties of the runner, establish a probability model of the runner life. Through a large number of simulations, evaluate the failure probability of the runner under different conditions. This process needs to combine the failure mode of the runner, the mechanical properties of the material and the actual workload to evaluate the possible failure path of the runner during use. In the experiment, the reliability of the model can be verified through statistical analysis and test data to provide an accurate quantitative basis for the life prediction of the runner.

[0045] S6.4. Considering that the runner is affected by multiple physical fields such as temperature, stress, and vibration in actual operation, a multi-physical field coupling analysis method (thermal-mechanical coupling, vibration-stress coupling) is used to comprehensively evaluate the life of the runner. In the simulation, the interaction between the temperature field, stress field, and vibration field is combined to establish a multi-physical field coupling model to comprehensively predict the fatigue life of the runner. These coupling effects are verified through experiments, especially the fatigue behavior and crack propagation characteristics of the runner material under high temperature and high pressure environments. Environmental simulation tests (high temperature and high pressure fatigue tests) can be used in the experiment to verify the effectiveness of the model and correct the prediction results.

[0046] The present invention is further configured as follows: in the step S7, in the decision-making of repair and preventive maintenance based on crack diagnosis:

[0047] S7.1. Generate a detailed crack assessment report based on crack monitoring data and life prediction results to assess the current crack hazard level, expansion speed and possible failure consequences. Risk assessment is combined with the operating environment of the runner (load, speed, temperature), and the impact of cracks on the safe operation of the runner is assessed using fault tree analysis or failure mode and effect analysis methods. In the experiment, the accuracy of the risk assessment model is verified by combining field measured data with simulation results to provide a quantitative basis for maintenance decisions.

[0048] S7.2. Combine the crack growth model and real-time monitoring data to analyze the crack growth trend in the future and determine the maintenance timing of the runner. Use the relationship between crack growth rate and service life to evaluate the speed and impact of crack growth under different working conditions. Obtain crack growth data under different working conditions through experiments to improve the model and ensure accurate maintenance decisions. If the crack growth rate is close to the critical value, maintenance or component replacement should be carried out in advance to prevent failures.

[0049] S7.3. Based on the crack growth analysis results, make recommendations for component replacement and strengthening measures for the runner, replace high-risk areas (components with rapid crack growth) in a timely manner, and take strengthening measures (spraying wear-resistant layers, using high-strength materials) for stress-concentrated components. In the experiment, verify the effects of different strengthening measures through material tests and crack growth tests to ensure the effectiveness of the maintenance plan;

[0050] S7.4. Develop a scientific regular maintenance plan based on the crack expansion and remaining life of the runner. Combined with the monitoring data of key components, set key monitoring nodes and conduct crack detection after a certain number of hours of regular operation. During the experiment, the runner is regularly inspected at different time points to verify the accuracy and timeliness of the monitoring strategy, so as to ensure the safe and efficient operation of the runner within its remaining life.

[0051] The present invention is further configured as follows: in the step S8, in the crack diagnosis optimization of the data feedback mechanism:

[0052] S8.1. Real-time data collection of the rotor operation data is carried out through a high-precision sensor system (strain sensor, optical fiber sensor, accelerometer), and data processing and analysis is performed through a cloud platform. During the data analysis process, methods such as time series analysis and spectrum analysis are applied to detect the crack expansion state in real time and determine the risk level of the crack. In the experiment, real-time data is used to compare the simulation results to verify the accuracy of the data collection and analysis methods, providing continuous data support for crack diagnosis;

[0053] S8.2. Use machine learning technologies such as deep learning and convolutional neural networks to optimize crack detection algorithms. By training models with big data and combining a large amount of crack sample data, optimize the algorithm model and improve the algorithm's ability to identify tiny cracks and complex crack morphologies. By testing crack samples under different working conditions, verify the performance of the algorithm optimized by machine learning, and continuously adjust model parameters to improve diagnostic accuracy.

[0054] S8.3. Establish a fault case database, store the historical fault data, crack growth history and maintenance records of the runner in the database, analyze common problems through data mining, combine new technologies and methods, and continuously update the fault case database to ensure that the database can reflect the latest technological progress in runner crack diagnosis and optimize crack diagnosis models and maintenance plans;

[0055] S8.4. Combine crack monitoring results with equipment operation feedback to form a closed-loop optimization mechanism. Regularly update crack diagnosis methods based on detection results and optimize related algorithms. Continuously improve crack diagnosis systems and life prediction models through technology upgrades to ensure continued safe operation of equipment. During the experiment, new diagnostic technologies are continuously tested and applied to actual monitoring.

[0056] Beneficial Effects

[0057] Compared with the known public technology, the technical solution provided by the present invention has the following beneficial effects:

[0058] 1. The present invention can comprehensively and accurately evaluate the crack condition, material properties and stress distribution of the runner under actual working conditions, from the detection of the runner crack to the crack shape and size analysis and the microstructure analysis of the material properties, and finally includes the simulation analysis of the stress distribution of the runner working load, so as to provide a solid data foundation and theoretical basis for the crack propagation mechanism, life cycle prediction and runner life assessment, effectively realize the integration of multi-dimensional and multi-technical means, ensure the accuracy of crack detection and analysis, and provide all-round data support for subsequent crack propagation prediction and life assessment.

[0059] 2. The present invention predicts crack propagation by using the Paris law and finite element analysis, and then combines fatigue data, reliability analysis and multi-physical field coupling analysis to accurately evaluate the remaining service life and possible failure paths of the runner. Based on the crack propagation trend and maintenance history data, scientific maintenance strategies and optimization plans are formulated. By combining the data feedback mechanism with machine learning technology, the crack diagnosis and life prediction models are updated in real time, and the crack diagnosis system and maintenance decisions are optimized. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1The present invention is a flow chart of a method for diagnosing cracks and assessing the life of a runner of a pumped storage power station unit. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are 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 creative work are within the scope of protection of the present invention.

[0062] The present invention will be further described below in conjunction with the embodiments.

[0063] Example 1

[0064] like Figure 1 As shown, the present invention provides a technical solution: a method for diagnosing cracks and assessing the life of a runner of a pumped storage power station unit, wherein in step S1, the runner crack detection position is accurately located:

[0065] S1.1. In the operation state of the power station, the vibration characteristics of the runner are monitored in real time through a vibration sensor. A highly sensitive PCB 353B18 acceleration sensor is used for data acquisition. The sampling frequency should be set to above 10 kHz to ensure that the vibration fluctuations caused by high-frequency cracks can be captured. Through time domain analysis and frequency domain analysis, combined with the natural frequency and vibration mode of the runner, fast Fourier transform is used for spectrum analysis to identify possible crack locations and crack source areas. By comparing the vibration data under different working conditions, the evolution trend and location information of the crack can be effectively determined;

[0066] S1.2. Use an Olympus 38DL PLUS ultrasonic flaw detector with a broadband sensor to scan the rotor using a handheld probe or an automated robot platform. The test is carried out in a stationary rotor environment, and the ultrasonic sensor is placed on the outer edge of the rotor. Array probe technology is used to accurately determine the shape and depth of the crack through angle scanning and depth scanning, combined with sound wave attenuation and reflection time. The return signal is processed using waveform analysis software, and the location and size of the crack are determined using reflection intensity and echo time.

[0067] S1.3. Under heating conditions, the heat conduction characteristics of the cracks will be different from those of the surrounding areas, resulting in thermal imaging differences. In the experimental environment, the wheel needs to be loaded to a certain working condition, and the FLIR T640 infrared thermal imager is used for data acquisition. The resolution of the thermal imager should be 640x480, and the measurement accuracy should be ±0.1℃. The surface of the wheel is evenly heated using an electric heating film, and the temperature change is monitored in real time to capture the thermal anomalies of surface cracks and damage. The thermal imaging images are processed by MATLAB software, using the image difference analysis method to determine the precise location and extension characteristics of the cracks;

[0068] S1.4. FBG sensors are placed on the surface of the runner to monitor the strain distribution of various parts of the runner. During the experiment, the installation position of the sensor needs to be optimized to capture the strain changes in the key stress area. The experiment should be carried out under full load operation. A high-precision spectrum analyzer is used for data collection. The sampling frequency should reach 1 kHz. By analyzing different strain distributions and combining finite element simulation data, the possible areas where cracks may appear are determined, and the detection strategy is adjusted through real-time data feedback.

[0069] In the step S2, accurate analysis of the wheel crack shape and size:

[0070] S2.1. In the experiment, a high-resolution GE phoenix v|tome|x industrial CT scanner is used to perform three-dimensional imaging of the sample. The scanning resolution can reach the micron level. The experimental environment should ensure that the rotor sample is in a stationary state, and the rotor needs to be cut non-destructively for scanning. The CT data processing software of VGStudio MAX can reconstruct the obtained two-dimensional tomography data into three dimensions, accurately presenting the three-dimensional morphology, depth and position of the crack. Through three-dimensional modeling and virtual cutting technology, the morphological characteristics of the crack can be accurately determined, and then its expansion trend can be analyzed;

[0071] S2.2. During the experiment, a Zeiss Axio Imager high-power optical microscope was used to observe cracks on the sample surface. In the experimental environment, surface pretreatment was required in the area where the cracks were clearer. The sample surface was treated with metallographic polishing and etching agents to facilitate observation. The resolution of the optical microscope should reach 0.1 microns, which can effectively display the microstructural changes of the cracks and the starting position of the microcracks. Combined with Image-Pro Plus image processing technology, the width and depth of the cracks can be accurately measured, and the crack extension direction can be analyzed.

[0072] S2.3. In the experiment, the Keyence LJ-V7080 high-precision laser scanner was used to scan the surface of the rotor. The scanning accuracy can reach 1 micron. The experimental environment requires the rotor to be stationary and ensure the suitability of the scanning angle and surface finish. The laser scanning data is processed by a computer to form a three-dimensional image of the rotor surface. Combined with the image analysis software, the specific size of the crack can be extracted. Through quantitative analysis, the width, length and depth of the crack can be obtained, providing accurate geometric data for subsequent crack extension analysis.

[0073] S2.4. During the experiment, the crack area was photographed by a Canon EOS 5D Mark IV high-resolution camera. The image was processed for uniform illumination to eliminate the influence of uneven surface light. Then, the image processing toolbox in MATLAB was used to extract the crack boundary using the Canny algorithm. The length, width and morphological characteristics of the crack were calculated by image segmentation and feature extraction algorithms. The crack propagation speed and possible affected area were determined by combining statistical analysis.

[0074] In the step S3, the microstructure analysis of the performance of the runner material:

[0075] S3.1. In the experiment, a Rockwell HRB hardness tester is used to test the hardness of the wheel surface. The hardness range should be between 60HRA and 80HRA. By sampling at different positions, the hardness distribution of the material is evaluated and possible crack initiation areas are identified. The crack propagation experiment can be carried out by an Instron 8800 fatigue testing machine to simulate the stress changes of the wheel during long-term operation. Combined with the material hardness data, its crack resistance performance is evaluated;

[0076] S3.2. In the experiment, the fatigue life test of the runner is carried out using the MTS 810 high-load fatigue testing machine. Alternating loads of different amplitudes and frequencies are applied. Material samples should be sampled from the runner for standardized tests. The test sample size is 20mm×20mm×10mm. By gradually loading and monitoring the development of cracks, the fatigue life and crack growth rate of the runner are recorded, and a runner fatigue life model is established to provide a basis for crack growth prediction.

[0077] S3.3, FEI Quanta 450 scanning electron microscope was used to perform surface and fracture analysis on the wheel material. The scanning acceleration voltage was set to 15 kV and the magnification was 500 to 5000 times. The composition distribution of the material was analyzed by electron probe to identify the weak area where the crack occurred. The microscopic mechanism of crack initiation was determined in combination with fracture analysis.

[0078] S3.4. Oxford Instruments X-Strata 980 spectrometer was used to analyze the chemical composition of the runner material in detail, especially the content of carbon, chromium, nickel and other elements, to ensure that the alloy composition of the material meets the design requirements. By comparing the actual composition with the standard, the crack resistance and long-term operation reliability of the runner were evaluated.

[0079] In the step S4, simulation analysis of the stress distribution of the runner working load:

[0080] S4.1. By defining the material properties of the runner, the yield strength (about 350 MPa), elastic modulus (210 GPa) and Poisson's ratio (0.3) of the steel, the simulation results are ensured to be consistent with the actual performance. The actual operating data, including the running speed, operating temperature (200°C), pressure and load acting on the runner, are input. The stress distribution under different load conditions is simulated, especially in the runner blades, connecting parts and supporting areas, with a focus on the high stress areas in the stress concentration area, which may become the starting point of the crack. This analysis provides basic data for subsequent crack propagation prediction and life assessment.

[0081] S4.2. Deploy a sensor network in the actual working environment to collect the load data (torque, vibration, temperature, pressure, etc.) of the runner in real time. Sensors such as strain gauges, accelerometers and thermocouples can be used to monitor the stress state and temperature changes of the runner under different working conditions. The data is transmitted to the data analysis platform through a wireless communication system to calculate the dynamic load and stress distribution of the runner in real time. According to the changes in real-time data, the working performance of the runner under different working conditions is analyzed, potential load fluctuations are identified, and compared with the results of finite element analysis to further accurately evaluate the working state of the runner and provide data support for real-time monitoring of stress concentration areas and crack initiation points;

[0082] S4.3. By combining finite element analysis with real-time monitoring data, focus on identifying areas on the runner where stress concentration may occur, usually including the root of the runner blade, bolt holes and runner support parts. Use stress intensity factor (K) to evaluate the crack growth potential in these areas. The fracture toughness data of the material needs to be introduced during the analysis. Based on the linear fracture mechanics properties of the material, the critical stress intensity factor (Kc) is used to compare the stress in these areas with the material's resistance to crack growth. If the stress in the stress concentration area exceeds the critical value of the material, the risk of crack growth will increase significantly. This analysis can provide key clues for the prediction of subsequent crack growth paths.

[0083] S4.4. Based on the historical operating data of the runner, the mining fatigue theory is applied to analyze the impact of cumulative load on the runner life. By collecting load data under different working conditions, the stress amplitude of each load cycle is calculated, and the load amplitude-cycle number relationship is used to evaluate the fatigue damage of the runner material under different stress levels. When analyzing historical data, the changes in factors such as temperature, rotation speed, and pressure should be considered, and the cumulative effect should be calculated in combination with the SN curve of the material. Through periodic fatigue tests, the damage accumulation characteristics of the runner under different load histories are verified. The research results provide a theoretical basis for the possible fatigue damage points of the runner, especially the crack initiation that may occur in the stress concentration area.

[0084] In this embodiment, from the detection of the impeller crack (vibration monitoring, ultrasonic scanning, infrared thermal imaging monitoring and other technologies) to the crack shape and size analysis (using CT scanning, optical microscopy, laser scanning and other technologies) and the microstructure analysis of material properties (hardness test, fatigue test, fracture analysis, etc.), the simulation analysis of the stress distribution of the impeller working load is also included. Through these steps, the crack situation, material properties and stress distribution of the impeller under actual working conditions can be comprehensively and accurately evaluated, providing a solid data foundation and theoretical basis for the crack propagation mechanism, life cycle prediction and impeller life assessment, and effectively realizing the integration of multi-dimensional and multi-technical means to ensure the accuracy of crack detection and analysis, and can provide all-round data support for subsequent crack propagation prediction and life assessment.

[0085] Example 2

[0086] like Figure 1 As shown, the present invention provides a technical solution: a method for diagnosing cracks and assessing the life of a runner of a pumped storage power station unit, wherein in step S5, crack propagation mechanism and life cycle prediction:

[0087] S5.1. Using the Paris rule ( ) is used to predict the crack growth rate, where is the range of stress intensity factor, and m are material constants. First, the cracks in the runner are preliminarily diagnosed to determine their initial size and location. Then, the changes in stress intensity factors under different loading conditions are calculated. The experimental data (C and m values ​​of steel) are used to calibrate the Paris law, and then the crack growth rate under different loads during the use of the runner is predicted.

[0088] S5.2. Based on the theory of fracture mechanics, a crack propagation path model of the runner under different working conditions is established to analyze the direction and path of crack propagation. Through the stress field and the relative position of the crack, finite element analysis is used to simulate the crack propagation process under different loads. During the simulation process, the starting position of the crack, the geometric shape of the runner and the fracture toughness of the material should be considered. The crack propagation path near the root of the runner blade and the bolt hole is analyzed in detail. In the experiment, the crack propagation path can be verified by CT scanning or X-ray imaging technology to ensure the accuracy of the model. Through the simulation analysis of the crack propagation path, it can be determined whether there is crack merging or damage to the supporting structure;

[0089] S5.3. Simulate the crack growth process under dynamic loads, simulate the influence of periodic or random loads (speed changes, load fluctuations) on crack growth that the runner bears in actual work, use finite element software with time domain or frequency domain analysis functions to analyze the dynamic response of the runner under complex loads, and calculate the influence of dynamic loads on crack growth rate. The crack growth rate will not only depend on the static stress intensity factor, but also on factors such as load frequency and amplitude. In order to verify the accuracy of the simulation model, the crack growth rate under dynamic loads can be calibrated by comparing laboratory fatigue tests and field monitoring data;

[0090] S5.4. Combined with the crack monitoring data collected during the operation of the runner, the crack growth model is updated regularly. The crack growth monitoring technology (ultrasonic, X-ray, magnetic particle testing) is used to track the size changes of the cracks in real time, and the actual test results are fed back to the crack growth prediction model. Through the continuously updated crack data, the prediction model of the crack growth rate and path is optimized to ensure that the model can reflect the actual performance of the runner under different working conditions. In the experiment, dynamic monitoring equipment can be used to conduct non-destructive testing on the runner regularly, and the prediction model can be adjusted according to the test results to ensure a more accurate assessment of the remaining life of the runner;

[0091] In the step S6, the remaining service life prediction of the runner life assessment:

[0092] S6.1. Based on the maintenance records, failure modes and operating conditions of the runner in previous years, an empirical model is constructed using multivariate regression analysis or machine learning algorithms (support vector machine, random forest) to predict the remaining service life of the runner. By analyzing historical data, the key factors affecting the life of the runner (speed, load, vibration) are identified, and based on the changing trends of these factors, the fatigue damage of the runner in the future is predicted. In the experimental environment, the accuracy of the model can be verified by comparing it with field data through long-term laboratory tests and necessary corrections can be made;

[0093] S6.2. Combine the fatigue data of the material (SN curve, fatigue limit) and the stress analysis results of the runner, use Miner's law or multi-axial fatigue damage model to predict the fatigue life, and predict the fatigue life of the runner by accumulating fatigue damage under different load conditions, especially at the critical fatigue points of key components (blades, support rings). In the experiment, a fatigue testing machine is used to simulate the actual working conditions, and the fatigue limits under different conditions are measured to verify the accuracy of the model prediction. Through the accumulation analysis of fatigue damage, the critical points where the runner may fail are identified and corresponding maintenance strategies are formulated;

[0094] S6.3. Use reliability analysis methods (Weibull distribution, Monte Carlo simulation) to evaluate the overall life distribution of the runner. According to the use environment, loading history and material properties of the runner, establish a probability model of the runner life. Through a large number of simulations, evaluate the failure probability of the runner under different conditions. This process needs to combine the failure mode of the runner, the mechanical properties of the material and the actual workload to evaluate the possible failure path of the runner during use. In the experiment, the reliability of the model can be verified through statistical analysis and test data to provide an accurate quantitative basis for the life prediction of the runner.

[0095] S6.4. Considering that the runner is affected by multiple physical fields such as temperature, stress, and vibration in actual operation, a multi-physical field coupling analysis method (thermal-mechanical coupling, vibration-stress coupling) is used to comprehensively evaluate the runner life. In the simulation, the interaction between the temperature field, stress field, and vibration field is combined to establish a multi-physical field coupling model to comprehensively predict the fatigue life of the runner. These coupling effects are verified experimentally, especially the fatigue behavior and crack growth characteristics of the runner material under high temperature and high pressure environments. Environmental simulation tests (high temperature and high pressure fatigue tests) can be used in the experiment to verify the effectiveness of the model and calibrate the prediction results.

[0096] In the step S7, decision-making on maintenance prevention based on crack diagnosis:

[0097] S7.1. Generate a detailed crack assessment report based on crack monitoring data and life prediction results to assess the current crack hazard level, expansion speed and possible failure consequences. Risk assessment is combined with the operating environment of the runner (load, speed, temperature), and the impact of cracks on the safe operation of the runner is assessed using fault tree analysis or failure mode and effect analysis methods. In the experiment, the accuracy of the risk assessment model is verified by combining field measured data with simulation results to provide a quantitative basis for maintenance decisions.

[0098] S7.2. Combine the crack growth model and real-time monitoring data to analyze the crack growth trend in the future and determine the maintenance timing of the runner. Use the relationship between crack growth rate and service life to evaluate the speed and impact of crack growth under different working conditions. Obtain crack growth data under different working conditions through experiments to improve the model and ensure accurate maintenance decisions. If the crack growth rate is close to the critical value, maintenance or component replacement should be carried out in advance to prevent failures.

[0099] S7.3. Based on the crack growth analysis results, make recommendations for component replacement and strengthening measures for the runner, replace high-risk areas (components with rapid crack growth) in a timely manner, and take strengthening measures (spraying wear-resistant layers, using high-strength materials) for stress-concentrated components. In the experiment, verify the effects of different strengthening measures through material tests and crack growth tests to ensure the effectiveness of the maintenance plan;

[0100] S7.4. According to the crack extension and remaining life of the runner, a scientific regular maintenance plan is formulated. Combined with the monitoring data of key components, key monitoring nodes are set. Crack detection is carried out after a certain number of hours of regular operation. During the experiment, the runner is regularly inspected at different time nodes to verify the accuracy and timeliness of the monitoring strategy, so as to ensure the safe and efficient operation of the runner within its remaining life.

[0101] In the step S8, data feedback mechanism crack diagnosis optimization:

[0102] S8.1. Real-time data collection of the rotor operation data is carried out through a high-precision sensor system (strain sensor, optical fiber sensor, accelerometer), and data processing and analysis is performed through a cloud platform. During the data analysis process, methods such as time series analysis and spectrum analysis are applied to detect the crack expansion state in real time and determine the risk level of the crack. In the experiment, real-time data is used to compare the simulation results to verify the accuracy of the data collection and analysis methods, providing continuous data support for crack diagnosis;

[0103] S8.2. Use machine learning technologies such as deep learning and convolutional neural networks to optimize crack detection algorithms. By training models with big data and combining a large amount of crack sample data, optimize the algorithm model and improve the algorithm's ability to identify tiny cracks and complex crack morphologies. By testing crack samples under different working conditions, verify the performance of the algorithm optimized by machine learning, and continuously adjust model parameters to improve diagnostic accuracy.

[0104] S8.3. Establish a fault case database, store the historical fault data, crack growth history and maintenance records of the runner in the database, analyze common problems through data mining, combine new technologies and methods, and continuously update the fault case database to ensure that the database can reflect the latest technological progress in runner crack diagnosis and optimize crack diagnosis models and maintenance plans;

[0105] S8.4. Combine crack monitoring results with equipment operation feedback to form a closed-loop optimization mechanism. Regularly update crack diagnosis methods based on detection results and optimize related algorithms. Continuously improve crack diagnosis systems and life prediction models through technology upgrades to ensure continued safe operation of equipment. During the experiment, new diagnostic technologies are continuously tested and applied to actual monitoring.

[0106] In this embodiment, the Paris law and finite element analysis are used to predict crack extension, and then fatigue data, reliability analysis and multi-physics field coupling analysis are combined to accurately evaluate the remaining service life and possible failure paths of the runner. Based on the crack extension trend and maintenance history data, scientific maintenance strategies and optimization plans are formulated. Through the combination of data feedback mechanism and machine learning technology, crack diagnosis and life prediction models are updated in real time, and crack diagnosis systems and maintenance decisions are optimized. Through these steps, continuous monitoring, accurate early warning, and flexible maintenance decision-making plans can be provided to ensure the safety and efficiency of the runner during long-term operation, thereby improving the reliability of the equipment, extending its service life, and reducing the risk of failure.

[0107] The present invention,

[0108] In actual use, the impeller cracks are fully monitored through a variety of detection technologies. In operation, vibration sensors, ultrasonic flaw detectors, infrared thermal imagers and other equipment are used to accurately detect and locate the initial position, shape, depth and expansion trend of the impeller cracks. Then, high-precision methods such as CT scanning, optical microscopy, and laser scanning are used to analyze the crack shape and size, combined with strain monitoring and material performance testing to further evaluate the microstructure and crack resistance of the impeller material.

[0109] Through these detailed detection and analysis data, combined with finite element simulation analysis methods, it is possible to accurately simulate the stress distribution of the runner under different working conditions, identify stress concentration areas, and provide basic data for crack growth mechanism and life cycle prediction. By using the Paris law and dynamic load simulation, combined with the historical operation data and crack monitoring results of the runner, the crack growth model is regularly updated, the changes of cracks are tracked in real time, and a scientific basis is provided for the remaining life assessment.

[0110] The present invention ensures the accuracy of crack diagnosis and the reliability of life prediction through multi-dimensional data fusion and real-time feedback mechanism, and optimizes the maintenance decision-making process. Combined with the application of machine learning algorithms, this method continuously optimizes crack diagnosis technology, improves detection accuracy and efficiency, and ultimately achieves efficient and safe operation of equipment, prolongs the service life of equipment, and reduces the risk of failure.

[0111] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for crack diagnosis and life assessment of a runner of a pumped storage power station unit, characterized in that: The following steps are involved: S1. Accurate positioning of the wheel crack detection position; S2. Accurate analysis of the shape and size of the runner crack; S3, performance and microstructure analysis of runner materials; S4. Simulation analysis of stress distribution of runner working load; S5. Crack propagation mechanism and life cycle prediction; S6. Prediction of remaining service life of runner in life assessment; S7. Maintenance preventive decision-making based on crack diagnosis; S8. Data feedback mechanism for crack diagnosis optimization.

2. The method for crack diagnosis and life assessment of a pumped storage power station unit according to claim 1, characterized in that: In the step S1, the wheel crack detection position is accurately positioned: S1.

1. Use vibration sensors to monitor the vibration data generated during the operation of the runner and find abnormal vibration signals; S1.

2. Scan the wheel using an ultrasonic sensor to determine whether there is a crack and its exact location; S1.

3. Use infrared thermal imaging technology to monitor the temperature distribution on the surface of the runner and determine whether there are cracks based on the temperature difference; S1.

4. By arranging optical fiber sensors, the rotor is monitored in real time and the occurrence of cracks is identified through strain changes.

3. The method for crack diagnosis and life assessment of a pumped storage power station unit according to claim 1, characterized in that: In the step S2, accurate analysis of the wheel crack shape and size: S2.

1. Obtain three-dimensional morphological data of cracks by X-ray CT scanning; S2.

2. Observe the cracks under a microscope and analyze the shape and direction of the cracks; S2.3, using a laser scanner to perform three-dimensional reconstruction of the crack surface to obtain the size and depth of the crack; S2.

4. Extract the geometric features of the cracks through high-resolution image processing technology.

4. The method for crack diagnosis and life assessment of a pumped storage power station unit according to claim 1, characterized in that: In the step S3, the microstructure analysis of the performance of the runner material: S3.

1. Use a Rockwell hardness tester to test the hardness of the wheel material and evaluate the material's resistance to crack growth; S3.

2. Conduct fatigue tests on materials to evaluate their fatigue resistance under cyclic stress; S3.

3. Use scanning electron microscopy and transmission electron microscopy to analyze the microstructure of the material and determine the sensitive areas where cracks occur; S3.

4. Analyze the chemical composition of the material through energy spectrum analysis technology to determine the possible crack-prone areas.

5. The method for crack diagnosis and life assessment of a pumped storage power station unit according to claim 1, characterized in that: In the step S4, simulation analysis of the stress distribution of the runner working load: S4.

1. Use finite element analysis software to simulate and analyze the stress distribution of the runner under different working conditions; S4.2, real-time monitoring of environmental factors such as load, speed and temperature of the runner during operation as input parameters; S4.

3. Identify stress concentration areas in the runner where cracks may occur through stress analysis; S4.

4. Analyze the historical load data of the runner during operation and infer the critical load section where cracks may occur.

6. A method for crack diagnosis and life assessment of a pumped storage power station unit runner according to claim 1, characterized in that: In the step S5, crack propagation mechanism and life cycle prediction: S5.

1. Use Paris's law to predict crack growth rates at different stress levels. S5.

2. Use fracture mechanics theory, combined with crack size and shape, to calculate the crack propagation process; S5.

3. Dynamically simulate the crack propagation process of the runner under complex loads and predict its propagation path; S5.

4. Regularly monitor the crack growth of the runner and update the prediction model by tracking the changes in the cracks.

7. A method for crack diagnosis and life assessment of a pumped storage power station unit according to claim 1, characterized in that: In the step S6, the remaining service life prediction of the runner life assessment: S6.

1. Establish an empirical assessment model for the unit runner life based on historical data and literature; S6.

2. Construct a fatigue life assessment model based on the stress and fatigue test data of the runner; S6.

3. Use reliability theory, combined with the use and failure statistics of the runner, to establish a life assessment model; S6.

4. Combine multiple physical fields such as temperature, stress, and electromagnetic field to construct a comprehensive life prediction model.

8. A method for crack diagnosis and life assessment of a pumped storage power station unit runner according to claim 1, characterized in that: In the step S7, decision-making on maintenance prevention based on crack diagnosis: S7.

1. Generate a detailed crack report based on the above crack diagnosis and evaluation results to provide a basis for maintenance; S7.

2. Determine the best time for repair based on crack growth prediction and life assessment; S7.

3. Based on the evaluation results, propose necessary component replacement or strengthening measures; S7.

4. Develop a long-term regular inspection and maintenance plan.

9. A method for crack diagnosis and life assessment of a pumped storage power station unit runner according to claim 1, characterized in that: In the step S8, data feedback mechanism crack diagnosis optimization: S8.

1. Collect real-time monitoring data during the operation of the unit and conduct comprehensive analysis; S8.

2. Continuously optimize the crack diagnosis and life assessment model based on new data feedback; S8.

3. Analyze the failure cases in detail and summarize the rules; S8.

4. Establish a knowledge base for unit runner crack diagnosis and life assessment and continuously update technical achievements.

Images