Hydroelectric generating set online monitoring data analysis method

Abstract

The invention relates to the technical field of data modeling, in particular to a hydroelectric generating set online monitoring data analysis method. The method comprises the following steps: obtaining online monitoring data of the hydroelectric generating set, and carrying out data preprocessing on the online monitoring data of the hydroelectric generating set so as to obtain online monitoring data of the hydroelectric generating set to be processed; performing generator vibration sensing monitoring data extraction according to the to-be-processed hydroelectric generating set online monitoring data so as to obtain generator vibration sensing monitoring data; performing vibration anomaly analysis according to the vibration sensing monitoring data of the generator to obtain vibration anomaly data; and acquiring generator sensing data, and performing environment correction on the abnormal vibration data according to the generator sensing data so as to obtain abnormal vibration correction data of the generator. The monitoring accuracy of the hydroelectric generating set monitoring model is improved based on the data modeling technology.

Description

Technical Field

[0001] The present invention relates to the technical field of online monitoring of hydropower units, and in particular to a method for analyzing online monitoring data of hydropower units. Background Art

[0002] With the continuous growth of global energy demand and the widespread application of renewable energy, hydropower, as an important form of clean energy, has received more and more attention. The operating efficiency and stability of hydropower units directly affect the overall performance of the power system. Therefore, real-time monitoring and analysis of hydropower units to ensure their efficient and safe operation has become a hot topic in current research. Traditional monitoring methods for hydropower units often rely on offline data analysis, which is difficult to reflect the status of equipment in real time and cannot detect potential problems in time. With the rapid development of information technology, especially the maturity of big data and artificial intelligence technology, online monitoring and data analysis of hydropower units have become a new research direction. Traditional monitoring methods for hydropower units face massive and complex real-time data, and these methods have certain limitations in applicability and accuracy. Summary of the invention

[0003] The present invention aims to provide a method for analyzing online monitoring data of a hydropower unit to solve at least one of the above technical problems.

[0004] Based on this, it is necessary for the present invention to provide a method for analyzing online monitoring data of a hydropower unit to solve at least one of the above technical problems.

[0005] To achieve the above object, a method for analyzing online monitoring data of a hydropower unit comprises the following steps:

[0006] Step S1: acquiring online monitoring data of a hydropower unit, and performing data preprocessing on the online monitoring data of the hydropower unit, thereby obtaining online monitoring data of the hydropower unit to be processed;

[0007] Step S2: extracting generator vibration sensor monitoring data according to the online monitoring data of the hydropower unit to be processed, thereby obtaining generator vibration sensor monitoring data; performing vibration anomaly analysis according to the generator vibration sensor monitoring data, thereby obtaining vibration anomaly data;

[0008] Step S3: acquiring the generator sensor data, and performing environmental correction on the abnormal vibration data according to the generator sensor data, thereby obtaining the generator abnormal vibration correction data;

[0009] Step S4: constructing a generator vibration monitoring model according to the generator abnormal vibration correction data, thereby obtaining a generator vibration monitoring model; performing generator online abnormal vibration detection on the hydropower unit online monitoring data according to the generator vibration monitoring model, thereby obtaining generator online abnormal vibration data;

[0010] Step S5: Perform vibration risk assessment based on the online abnormal vibration data of the generator to obtain the generator vibration risk data, and upload it to the hydropower unit management platform to execute the generator risk warning task.

[0011] The present invention can clean up noise, fill missing values, and standardize data by acquiring and preprocessing the online monitoring data of the hydropower unit, thereby improving the data quality. This provides an accurate basis for subsequent analysis. The real-time acquisition and preprocessing of the online monitoring data ensure the timeliness of the data, facilitate rapid response and processing of potential problems, and the preprocessing process helps to integrate data from different sensors and sources into a unified format, which is convenient for subsequent analysis and modeling. By extracting the vibration sensor data of the generator and performing vibration anomaly analysis, vibration anomalies can be discovered early to prevent equipment failure or damage. Vibration anomaly analysis can help identify specific abnormal patterns and trends, locate the source of the fault, and improve the accuracy of diagnosis. The abnormal analysis results provide data support for maintenance and overhaul decisions, reducing the risk of relying on experience judgment. By using the generator sensor data for environmental correction, the interference of environmental factors on the vibration data can be eliminated, making anomaly detection more accurate. The abnormal vibration data after environmental correction can better reflect the state of the equipment itself, reducing false alarms and missed reports. Correction is performed according to different environmental conditions, so that the analysis results are more in line with the actual operating environment, and the practical application value of the data is improved. The construction of a generator vibration monitoring model can make predictions based on historical data and corrected data, and detect potential abnormalities in advance. The use of the constructed model for online abnormal vibration detection realizes automated monitoring, reduces the burden of manual monitoring, and improves efficiency. By continuously updating and optimizing the model, the accuracy and robustness of the model can be improved, so that the system can cope with different operating conditions and abnormal situations. By conducting risk assessment on the online abnormal vibration data of the generator, potential risks can be predicted, so that measures can be taken in advance to avoid failures. The risk data can be uploaded to the hydropower unit management platform and risk warning tasks can be performed, which realizes real-time monitoring and timely warning of risks, reduces downtime and losses, and the risk assessment results can provide a basis for the formulation of maintenance strategies, making maintenance activities more targeted and effective, and improving the reliability and operating efficiency of equipment.

[0012] Optionally, step S1 specifically includes:

[0013] Step S11: obtaining online monitoring data of the hydropower unit;

[0014] Step S12: performing data cleaning on the online monitoring data of the hydropower unit, thereby obtaining online monitoring cleaning data;

[0015] Step S13: Fill missing values ​​in the online monitoring cleaning data to obtain online monitoring filling data;

[0016] Step S14: normalizing the online monitoring filling data to obtain the online monitoring data of the hydropower unit to be processed.

[0017] Through real-time monitoring, the present invention can collect the operating data of the hydropower unit in time to ensure that the status and performance of the equipment can be tracked. This is the basis for subsequent data processing and analysis. Cleaning data can remove noise and erroneous information to ensure the accuracy and consistency of the data. This helps to improve the reliability of the analysis results and provide a more accurate basis for subsequent decision-making. Processing missing values ​​can fill in the gaps in the data and avoid distortion of the analysis results due to missing data. The filled data can more comprehensively reflect the actual condition of the equipment and improve the integrity of data analysis. Normalization can convert data to a unified scale, eliminate the influence of different data dimensions, and make the influence of each feature fairly compared. This helps to improve the accuracy of subsequent analysis and modeling and ensure the consistency and comparability of data processing.

[0018] Optionally, step S2 specifically includes:

[0019] Step S21: extracting generator vibration sensor monitoring data according to the online monitoring data of the hydropower unit to be processed, thereby obtaining generator vibration sensor monitoring data;

[0020] Step S22: performing vibration displacement anomaly analysis on the generator vibration sensor monitoring data, thereby obtaining vibration displacement anomaly data;

[0021] Step S23: performing vibration mode abnormality analysis on the generator vibration sensor monitoring data, thereby obtaining vibration mode abnormality data;

[0022] Step S24: integrating the vibration abnormality data according to the vibration displacement abnormality data and the vibration mode abnormality data, so as to obtain the vibration abnormality data.

[0023] The present invention can obtain real-time vibration status information by extracting monitoring data from the generator vibration sensor, which provides an accurate raw data basis for subsequent analysis and ensures the effectiveness and accuracy of the detection. Abnormal analysis of vibration displacement can identify mechanical failures or abnormal vibration patterns. This step helps to detect potential equipment problems early, thereby avoiding more serious damage and downtime. Abnormal analysis of vibration patterns can reveal abnormal vibration patterns in equipment operation, which is important for identifying specific fault types or determining the causes of faults. It helps to formulate targeted maintenance measures. Integrating abnormal vibration displacement data and abnormal vibration pattern data can comprehensively evaluate the vibration condition of the equipment and form a comprehensive abnormal vibration data set. This integrated analysis can provide more accurate fault diagnosis information, help optimize maintenance strategies and improve equipment operation reliability.

[0024] Optionally, step S22 is specifically:

[0025] Step S221: performing vibration displacement time domain analysis on the generator vibration sensor monitoring data, thereby obtaining vibration displacement time domain data;

[0026] Step S222: performing frequency domain conversion according to the vibration displacement time domain data, thereby obtaining a vibration displacement spectrum diagram;

[0027] Step S223: performing vibration displacement threshold statistics according to the vibration displacement spectrum diagram, thereby obtaining vibration displacement threshold data;

[0028] Step S224: extracting vibration displacement monitoring data from the generator vibration sensor monitoring data, thereby obtaining vibration displacement monitoring data;

[0029] Step S225: performing vibration displacement detection on the vibration displacement monitoring data according to the vibration displacement threshold data, thereby obtaining vibration displacement abnormality data.

[0030] The present invention provides detailed information about the change of vibration displacement over time through analysis. This analysis helps to understand the dynamic characteristics of vibration, such as the amplitude, frequency change and periodicity of vibration. Through time domain data, engineers can identify whether there is obvious vibration anomaly or trend change. Time domain analysis helps to capture early signals of equipment failure, such as drastic changes in vibration or periodic abnormal fluctuations. Converting time domain data into frequency domain data (spectrogram) can more clearly identify the frequency components of the vibration signal. The spectrum diagram reveals the frequency distribution of vibration and helps to identify specific vibration modes and potential resonant frequencies. Frequency domain analysis can reveal abnormal features in frequency, such as harmonics, fundamental frequencies or subharmonic components, which can indicate specific failure modes of equipment, such as imbalance, alignment problems or gear failures. By statistically analyzing the vibration displacement data in the spectrum diagram, the vibration thresholds under normal operating conditions can be determined. These thresholds are used for subsequent abnormal detection to help identify vibration conditions beyond the normal range, provide a standardized basis for consistent abnormal detection and judgment, reduce human judgment errors, and improve the reliability of detection. Extracting vibration displacement information from the original monitoring data makes the data more focused on the key parameters required for analysis. This helps reduce data noise, focus on information related to vibration displacement, improve the efficiency of data processing, facilitate subsequent detailed analysis and processing, and ensure the relevance and accuracy of monitoring data. The extracted data is tested using the previously determined threshold to identify whether there is any vibration outside the normal range. This can help detect anomalies in a timely manner and take appropriate maintenance measures. By comparing with the set threshold, the vibration anomaly in the equipment can be quickly located, reducing the risk of equipment damage and improving the reliability and safety of the system.

[0031] Optionally, step S23 is specifically:

[0032] Step S231: Performing Fourier transform on the generator vibration sensor monitoring data to obtain a generator vibration spectrum diagram;

[0033] Step S232: performing a reference spectrum waveform analysis on the generator vibration spectrum diagram, thereby obtaining generator vibration reference spectrum waveform data;

[0034] Step S333: extracting peak spectrum waveform features from the generator vibration spectrum diagram to obtain peak spectrum waveform data, and performing reference spectrum waveform detection on the peak spectrum waveform data according to the generator vibration reference spectrum waveform data to obtain generator vibration abnormal waveform data;

[0035] Step S234: performing frequency statistics according to the generator vibration abnormal waveform data, thereby obtaining the generator vibration high-frequency abnormal waveform data;

[0036] Step S235: performing abnormal vibration mode recognition on the generator vibration spectrum diagram according to the generator vibration high-frequency abnormal waveform data, thereby obtaining vibration mode abnormal data.

[0037] By converting time domain data into frequency domain data, the present invention can clearly see different frequency components and help identify potential vibration problems. It provides spectrum benchmark data under normal operating conditions, so that subsequent vibration anomaly detection has a standard reference. By extracting the peak features in the spectrum, vibration anomalies can be more accurately detected and compared with benchmark data, thereby improving the sensitivity of anomaly detection. By counting the frequency of abnormal waveforms, the severity of the anomaly can be further confirmed and specific high-frequency vibration problems can be identified. Identifying abnormalities in vibration patterns helps to locate the source of the problem, so that measures can be taken to repair or prevent equipment damage.

[0038] Optionally, step S3 specifically includes:

[0039] Step S31: acquiring generator sensor data, and extracting temperature sensor data and inductive air gap sensor data according to the generator sensor data, thereby obtaining temperature sensor data and inductive air gap sensor data;

[0040] Step S32: performing generator temperature anomaly detection on the online monitoring data of the hydropower unit to be processed according to the temperature sensing data, thereby obtaining generator online temperature anomaly data;

[0041] Step S33: performing generator air gap anomaly detection on the online monitoring data of the hydropower unit to be processed according to the inductive air gap sensor data, thereby obtaining generator online air gap anomaly data;

[0042] Step S34: integrating abnormal sensor data according to the generator online temperature abnormality data and the generator online air gap abnormality data, thereby obtaining the generator abnormality sensor data;

[0043] Step S35: Perform environmental correction on the abnormal vibration data according to the abnormal generator sensing data, so as to obtain abnormal vibration correction data of the generator.

[0044] The present invention helps to fully understand the operating status of the generator by acquiring and extracting temperature and inductive air gap sensing data. These data are the basis for subsequent detection and analysis, ensuring the accuracy and integrity of the data. By performing temperature anomaly detection through temperature sensing data, the overheating problem of the generator can be discovered in time. This helps to prevent potential damage, ensure the safe operation of the equipment and extend the service life. By performing air gap anomaly detection through inductive air gap sensing data, the change problem of the generator air gap, such as uneven air gap, can be discovered. This helps to identify mechanical problems at an early stage, thereby preventing further equipment failures. Integrating temperature and air gap anomaly data provides a comprehensive anomaly detection result. This comprehensive analysis helps to more accurately diagnose the health status of the generator and provide comprehensive maintenance recommendations. By performing environmental correction on abnormal vibration data, interference caused by environmental factors can be eliminated. This makes the abnormal vibration data more reliable, facilitates further analysis and decision-making, and improves the accuracy of detection.

[0045] Optionally, step S32 is specifically:

[0046] Step S321: Acquire ambient temperature monitoring data, and perform temperature statistics on the ambient temperature monitoring data, thereby obtaining high temperature environment data;

[0047] Step S322: extracting the generator stator temperature monitoring data and the generator rotor temperature monitoring data from the temperature sensing data, thereby obtaining the generator stator temperature monitoring data and the generator rotor temperature monitoring data;

[0048] Step S323: extracting the core temperature feature of the generator stator temperature monitoring data, thereby obtaining the core temperature data;

[0049] Step S324: extracting rotor winding temperature characteristics from the generator rotor temperature monitoring data, thereby obtaining rotor winding temperature data;

[0050] Step S325: constructing a generator thermal coupling model according to the core temperature data and the rotor winding temperature data, thereby obtaining a generator thermal coupling model;

[0051] Step S326: performing a high temperature simulation of the generator according to the generator thermal coupling model, thereby obtaining high temperature simulation data of the generator;

[0052] Step S327: performing time distribution statistics on the generator high temperature simulation data and the high temperature environment data, so as to obtain the generator high temperature simulation time distribution data and the high temperature environment time distribution data;

[0053] Step S328: Calculate the time distribution overlap according to the generator high temperature simulation time distribution data and the high temperature environment time distribution data, so as to obtain the generator temperature abnormality data;

[0054] Step S329: comparing the generator temperature anomaly data with the online monitoring data of the hydropower unit to be processed, thereby obtaining the generator online temperature anomaly data.

[0055] The present invention can identify high temperature environmental conditions by acquiring and counting ambient temperature data, which helps to understand the potential impact of the external environment on the temperature of the generator. The temperature data of the generator stator and rotor are extracted, so that different parts of the generator can be monitored separately, providing a detailed data basis for subsequent analysis. The core temperature characteristics are extracted to accurately evaluate the temperature state of the core, thereby discovering potential overheating problems or performance degradation. The rotor winding temperature characteristics are extracted to monitor the key parts of the rotor and provide data support for identifying temperature anomalies related to the winding. A thermal coupling model of the generator is constructed to integrate the core and rotor temperature data to help understand the propagation and mutual influence mechanism of temperature inside the generator. Using the thermal coupling model for high temperature simulation can predict the performance of the generator under different temperature conditions and assist in predicting temperature anomalies. Time distribution statistics of high temperature simulation data and high temperature environment data are performed to provide time characteristics of temperature anomalies, which helps to identify the law of temperature changes. The time distribution overlap is calculated to compare the simulation data with the actual environment data to evaluate the actual situation of the generator temperature anomaly and the consistency of the simulation prediction. By comparing the generator temperature anomaly data with the online monitoring data, the temperature state of the generator can be monitored and verified in real time to ensure timely response to any abnormal situation.

[0056] Optionally, step S33 is specifically:

[0057] Step S331: Acquire environmental humidity monitoring data, and perform humidity statistics on the environmental humidity monitoring data, thereby obtaining high humidity environmental data;

[0058] Step S332: extracting generator air gap monitoring data from the online monitoring data of the hydropower unit to be processed, thereby obtaining the generator air gap monitoring data;

[0059] Step S333: performing humidity statistics on the generator air gap monitoring data, thereby obtaining the generator air gap monitoring high humidity data;

[0060] Step S334: performing material dielectric loss factor analysis according to the generator air gap monitoring high humidity data, thereby obtaining material dielectric loss factor data;

[0061] Step S335: performing air gap anomaly correlation on the generator air gap monitoring high humidity data according to the material dielectric loss factor data, thereby obtaining the generator high humidity air gap anomaly data;

[0062] Step S336: performing time distribution statistics according to the high humidity environment data and the generator high humidity air gap abnormality data, so as to obtain the high humidity environment time distribution data and the generator high humidity air gap abnormality time distribution data;

[0063] Step S337: Calculate the time distribution overlap according to the high humidity environment time distribution data and the generator high humidity air gap abnormality time distribution data, so as to obtain the generator air gap abnormality data.

[0064] The present invention can understand the change of environmental humidity in real time by acquiring environmental humidity monitoring data. This helps to identify the potential impact of humidity on generators and other equipment. Statistical analysis of environmental humidity data can provide long-term humidity trends and fluctuations, thereby providing data support for equipment maintenance and operation decisions, timely discovering the occurrence of high humidity environments, and facilitating taking measures to reduce humidity or preventive maintenance of equipment, thereby reducing the risk of equipment failure. Extracting the generator air gap monitoring data can obtain the real-time state of the air gap, thereby understanding the important operating parameters inside the generator. Monitoring the generator air gap data helps to timely discover potential failure risks, such as uneven air gaps, and then conduct fault warnings and timely processing. By performing humidity statistics on the air gap monitoring data, the impact of humidity on the air gap state can be analyzed. Excessive humidity can affect the electrical insulation performance of the air gap, leading to equipment failure. Understanding the impact of humidity on the air gap can help optimize the operating conditions of the generator and reduce the negative impact of humidity on the air gap. The analysis of the material dielectric loss factor can evaluate the performance changes of insulating materials under high humidity conditions, understand the degree of reduction in their insulation capacity, and by understanding the loss of materials, corresponding maintenance and replacement strategies can be formulated to ensure the long-term stable operation of the equipment. Correlating material dielectric loss factor with air gap monitoring data can help identify air gap anomalies due to high humidity. This helps to detect potential problems early and take corrective measures. By identifying and resolving air gap anomalies, the operating reliability and safety of the generator can be improved. Time distribution statistics can help identify the occurrence patterns of high humidity environments and air gap anomalies. Understanding these patterns can help predict and prevent equipment failures. Statistical data can be used to optimize equipment operation and maintenance scheduling, avoid high-load operations during high humidity periods, and reduce the risk of failures. Time distribution overlap calculation can analyze the correlation between ambient humidity and air gap anomalies. This helps to understand the specific impact of humidity on air gap anomalies. By understanding the overlap of the time distribution of humidity and air gap anomalies, the early warning system can be improved, and the ability to foresee future anomalies caused by high humidity can be improved, thereby optimizing maintenance strategies and operating procedures.

[0065] Optionally, step S4 is specifically:

[0066] Step S41: extracting the generator amplitude data and the generator vibration frequency data from the generator abnormal vibration correction data, thereby obtaining the generator amplitude data and the generator vibration frequency data;

[0067] Step S42: performing amplitude range statistics on the generator amplitude data, thereby obtaining the generator amplitude range data;

[0068] Step S43: performing period statistics on the generator vibration frequency data, thereby obtaining the generator vibration frequency period data;

[0069] Step S44: constructing a generator vibration monitoring model according to the generator amplitude range data and the generator vibration frequency period data, thereby obtaining a generator vibration monitoring model;

[0070] Step S45: performing generator online abnormal vibration detection on the online monitoring data of the hydropower unit according to the generator vibration monitoring model, thereby obtaining generator online abnormal vibration data.

[0071] The present invention provides basic data for further analysis and monitoring by extracting generator amplitude and vibration frequency data from abnormal vibration correction data. These data are the foundation for all subsequent analysis and model building. Directly extracting these data can help quickly locate abnormal behavior of the generator, thereby discovering potential problems at an early stage. By counting the range of amplitude data, the normal range of generator vibration and its abnormal fluctuations can be identified. This helps to understand the working state of the generator and whether there are abnormal conditions beyond the normal working range. Amplitude range statistics can provide fault warning signals to help identify abnormal conditions that cause generator damage in advance. Through period statistics, the periodic pattern of generator vibration and its abnormal changes can be identified. This is crucial for understanding the operating characteristics of the generator and diagnosing potential problems. Period statistics help identify periodic anomalies in frequency data, which helps to improve the diagnostic accuracy of generator vibration problems. By incorporating amplitude range and frequency cycle data into the model, a monitoring model for generator vibration can be established. This makes the monitoring process more systematic and scientific. The monitoring model can realize automated analysis and abnormality detection, reduce manual intervention, and improve detection efficiency and accuracy. Abnormal vibration detection through online monitoring data can achieve real-time monitoring of the working state of the generator. This is crucial for rapid response and troubleshooting. Online detection can detect abnormal vibrations in a timely manner. Preventive maintenance can reduce equipment downtime and improve the reliability and life of generators. Real-time data analysis and detection can optimize the operating status of generators and improve the efficiency and performance of the overall power generation system.

[0072] Optionally, step S45 is specifically:

[0073] Step S451: performing generator online abnormal vibration detection on the online monitoring data of the hydropower unit according to the generator vibration monitoring model, thereby obtaining generator abnormal vibration data and generator normal vibration data;

[0074] Step S452: performing a normal vibration parameter comparison based on the abnormal vibration data of the generator and the normal vibration data of the generator, thereby obtaining a vibration error parameter set;

[0075] Step S453: extracting abnormal vibration speed features from the vibration error parameter set, thereby obtaining abnormal vibration speed data;

[0076] Step S454: Compare the online vibration speed of the generator with the online monitoring data of the hydropower unit according to the abnormal vibration speed data, so as to obtain the online abnormal vibration data of the generator.

[0077] The present invention can obtain the vibration data of the generator in real time through the online monitoring system, timely discover the vibration anomaly, and reduce the impact of potential mechanical failures on the generator set. The use of the vibration monitoring model can improve the accuracy of abnormal detection and avoid false alarms caused by human errors or other non-fault factors. By distinguishing normal and abnormal vibration data, maintenance personnel can take preventive maintenance measures before the problem develops into a serious failure. By comparing normal and abnormal vibration parameters, the specific characteristics and causes of abnormal vibration can be more accurately identified, providing a basis for fault diagnosis. The obtained vibration error parameter set can help formulate more effective maintenance and repair strategies, reduce resource waste and unnecessary downtime. By analyzing the vibration error, problems in equipment design or installation can be found, thereby optimizing equipment performance and design. Extracting the vibration abnormal speed feature can help identify the nature and mode of abnormal vibration, so as to understand its specific impact on the generator set. Accurate speed feature extraction can discover potential mechanical problems in advance and implement more timely intervention measures, thereby avoiding serious equipment damage. By accurately analyzing the vibration speed, the cause of the abnormality can be more accurately determined, and the accuracy and reliability of diagnosis can be improved. By comparing the vibration velocity data, it is possible to confirm whether there is actual abnormal vibration, thereby avoiding false alarms and improving the reliability of detection. Based on the comparative analysis of real-time data, the monitoring and maintenance strategies can be dynamically adjusted to adapt to changes in the operating status of the equipment. Continuous comparison and analysis can help optimize the operating parameters and maintenance plans of the generator and improve the overall efficiency and reliability of the generator set. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments thereof made with reference to the following drawings:

[0079] Figure 1 A schematic diagram of the steps of the on-line monitoring data analysis method for a hydropower unit according to the present invention;

[0080] Figure 2 Detailed step flow diagram of step S1 in the present invention;

[0081] Figure 3 Detailed step flow diagram of step S2 in the present invention;

[0082] Figure 4 Detailed step flow diagram of step S3 in the present invention;

[0083] Figure 5 Detailed step flow diagram of step S4 in the present invention;

[0084] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0085] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0086] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. 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.

[0087] Example

[0088] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. 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 technicians in this field without creative work are within the scope of protection of the present invention.

[0089] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0090] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0091] To achieve this, please refer to Figures 1 to 5 The present invention provides a method for analyzing online monitoring data of a hydropower unit, the method comprising the following steps:

[0092] Step S1: acquiring online monitoring data of a hydropower unit, and performing data preprocessing on the online monitoring data of the hydropower unit, thereby obtaining online monitoring data of the hydropower unit to be processed;

[0093] In this embodiment, when obtaining the online monitoring data of the hydropower unit, it is necessary to first obtain data from each sensor (such as temperature, pressure, flow sensor) through a real-time data acquisition system. These data need to be pre-processed, including denoising, data smoothing, standardization, etc., to ensure data quality. After pre-processing, it is stored in a database for further analysis and processing. For example, a data cleaning tool such as the Pandas library in Python is used to remove outliers and fill in missing data, and then normalization is performed.

[0094] Step S2: extracting generator vibration sensor monitoring data according to the online monitoring data of the hydropower unit to be processed, thereby obtaining generator vibration sensor monitoring data; performing vibration anomaly analysis according to the generator vibration sensor monitoring data, thereby obtaining vibration anomaly data;

[0095] In this embodiment, the generator vibration sensor monitoring data is extracted from the processed data, and the vibration signal features are extracted using a frequency domain analysis method such as Fourier transform. Then, an abnormal analysis is performed using a vibration anomaly detection algorithm (such as a threshold-based method or a machine learning algorithm) to identify abnormal vibration data. For example, a support vector machine (SVM) is used to classify normal and abnormal vibration data, and output an abnormal detection result.

[0096] Step S3: acquiring the generator sensor data, and performing environmental correction on the abnormal vibration data according to the generator sensor data, thereby obtaining the generator abnormal vibration correction data;

[0097] In this embodiment, the generator sensor data (such as temperature, load, etc.) is obtained, and the abnormal vibration data is corrected for the environment using these data. The environmental correction method includes performing multivariate regression analysis on the vibration data and the environmental data (such as temperature and load changes) to eliminate the influence of environmental factors on the vibration. For example, a linear regression model is used to adjust the abnormal vibration data, and the corrected data is the abnormal vibration correction data of the generator.

[0098] Step S4: constructing a generator vibration monitoring model according to the generator abnormal vibration correction data, thereby obtaining a generator vibration monitoring model; performing generator online abnormal vibration detection on the hydropower unit online monitoring data according to the generator vibration monitoring model, thereby obtaining generator online abnormal vibration data;

[0099] In this embodiment, a generator vibration monitoring model is constructed based on the correction data. A neural network training model can be used to predict and monitor the vibration state of the generator. The model is used to perform real-time abnormality detection on the online monitoring data of the hydropower unit. For example, a deep learning model is constructed using TensorFlow, correction data is input for training, and abnormal states of generator vibration are detected in real time to generate online abnormal vibration data of the generator.

[0100] Step S5: Perform vibration risk assessment based on the online abnormal vibration data of the generator to obtain the generator vibration risk data, and upload it to the hydropower unit management platform to execute the generator risk warning task.

[0101] In this embodiment, the vibration risk assessment is performed on the online abnormal vibration data of the generator. Risk assessment methods such as Fault Tree Analysis (FTA) or Bayesian network are used to assess the risk level and generate risk data. Finally, the risk data is uploaded to the hydropower unit management platform to perform risk warning tasks. For example: the risk data is transmitted to the hydropower unit management platform through the API (Application Programming Interface), and the hydropower unit management platform triggers the warning mechanism according to the risk data and notifies the maintenance personnel to conduct inspection and processing.

[0102] Optionally, step S1 specifically includes:

[0103] Step S11: obtaining online monitoring data of the hydropower unit;

[0104] In this embodiment, sensors and instruments installed on the hydropower units, including temperature sensors, vibration sensors, pressure sensors and flow meters, transmit real-time data to the data acquisition system through industrial protocols (such as Modbus, OPC). The collected data is sent to the central data server using Ethernet or wireless network. The data server is configured with a real-time data receiving and storage module that can process data streams from multiple hydropower units. The central data server stores the received data in a database (such as MySQL, PostgreSQL or NoSQL database) to ensure the integrity and traceability of the data. The data includes timestamps, sensor identification, sensor readings, etc.

[0105] Step S12: performing data cleaning on the online monitoring data of the hydropower unit, thereby obtaining online monitoring cleaning data;

[0106] In this embodiment, statistical methods (such as standard deviation method) and machine learning methods (such as isolation forest) are used to identify and mark outliers in the data. Outliers are caused by sensor failure or data transmission errors. By comparing timestamps and sensor IDs, duplicate records are identified and deleted. Duplicate records are caused by multiple data collections or system errors. The data is converted to a unified format, such as unifying the timestamp to UTC time and converting the numerical units to a unified standard (such as unifying the pressure unit to Pascal) to ensure data consistency.

[0107] Step S13: Fill missing values ​​in the online monitoring cleaning data to obtain online monitoring filling data;

[0108] In this embodiment, the missing values ​​in the data set are detected, and the missing patterns are discovered by checking the data integrity report or using data visualization tools. For numerical data, the missing values ​​are filled with the mean of the data. For example, the average value of a sensor over the past period of time is calculated and the missing data is filled with this average value. Check whether the filled data is reasonable to ensure that the filling method does not introduce additional errors. The effectiveness of the filling can be verified by comparing the statistical characteristics of the data before and after filling.

[0109] Step S14: normalizing the online monitoring filling data to obtain the online monitoring data of the hydropower unit to be processed.

[0110] In this embodiment, the data is scaled to the range [0,1]. For example, for a sensor with a reading range between 50 and 200, its data is normalized to between 0 and 1 using the formula (x-min) / (max-min). The parameters used for normalization (such as minimum, maximum, mean, and standard deviation) are saved for subsequent processing and verification. Check the distribution of the normalized data to ensure that the data is within the target range and verify whether the normalization effect meets expectations.

[0111] Optionally, step S2 specifically includes:

[0112] Step S21: extracting generator vibration sensor monitoring data according to the online monitoring data of the hydropower unit to be processed, thereby obtaining generator vibration sensor monitoring data;

[0113] In this embodiment, the raw data of the vibration sensor is extracted from the real-time monitoring system. The raw data is usually in the unit of accelerometer (m / s 2) records. Store the extracted data in a database. The database should include fields such as timestamp, sensor ID, vibration acceleration value, etc. Convert the raw data into a format suitable for further analysis, such as converting acceleration data into vibration displacement data, usually using Fourier transform or integration method for conversion. For example: The system extracts vibration data from the acceleration sensor installed on the generator bearing every second, and stores this vibration data in a local database for subsequent analysis.

[0114] Step S22: performing vibration displacement anomaly analysis on the generator vibration sensor monitoring data, thereby obtaining vibration displacement anomaly data;

[0115] In this embodiment, the extracted data is subjected to denoising, for example, by using a filter to eliminate high-frequency noise. The vibration displacement feature is extracted from the denoised data, and the displacement data is usually obtained by integrating the acceleration data. A displacement threshold is set, and a statistical method or a machine learning model is used to detect whether the displacement exceeds the normal range. Anomaly detection can be accomplished by comparing with historical data or setting a standard for absolute values. For example: if the vibration displacement exceeds the normal range (for example, the displacement exceeds ±2mm), the system will mark it as abnormal and generate an alarm.

[0116] Step S23: performing vibration mode abnormality analysis on the generator vibration sensor monitoring data, thereby obtaining vibration mode abnormality data;

[0117] In this embodiment, spectrum analysis technology (such as Fast Fourier Transform, Fast Fourier Transform, FFT) is used to identify the spectrum pattern of the vibration signal. Frequency features, such as main frequency and harmonic components, are extracted from the spectrum diagram. The current vibration pattern is compared with the normal operating mode to determine whether there is an abnormality. A pattern recognition algorithm or a machine learning model can be used for comparison. For example: If a new frequency component (such as an additional harmonic) or a significant frequency shift is detected, the system will mark it as an abnormal vibration pattern and provide a detailed spectrum analysis report.

[0118] Step S24: integrating the vibration abnormality data according to the vibration displacement abnormality data and the vibration mode abnormality data, so as to obtain the vibration abnormality data.

[0119] In this embodiment, the displacement anomaly data and the mode anomaly data are aggregated into a unified data table, ensuring that each data record includes information such as the anomaly type, timestamp, and severity. Data fusion techniques (such as multidimensional data analysis or a comprehensive scoring system) are used to evaluate the overall impact of the anomaly. A comprehensive anomaly report is generated, including the abnormal conditions of the vibration displacement and vibration mode, and recommended measures are provided, such as components that need to be inspected and recommended maintenance operations. For example: the system integrates the data of the displacement anomaly and the mode anomaly to form a report that points out the potential failure caused by the vibration anomaly of the generator and recommends further inspection and maintenance operations.

[0120] Optionally, step S22 is specifically:

[0121] Step S221: performing vibration displacement time domain analysis on the generator vibration sensor monitoring data, thereby obtaining vibration displacement time domain data;

[0122] In this embodiment, the original vibration signal is collected from the vibration sensor of the generator. These sensors are usually installed in the key parts of the generator, such as the bearing seat and the stator core. The received original vibration signal is recorded at a high sampling rate to ensure the accuracy and details of the data. In the data processing stage, the original vibration signal is processed by applying time domain analysis technology, for example, the statistical characteristics such as the mean and variance of the data are calculated using the sliding window method. Algorithms such as Short-Time Fourier Transform (STFT) can be used to analyze the time variation of the vibration signal in segments to obtain the vibration displacement time domain data. For example: suppose the vibration sensor of the generator collects data at a sampling frequency of 10,000 times per second, and the data collection lasts for 10 minutes. The collected data is processed in segments, each segment contains 1 second of data, and then the mean and variance of each segment of data are calculated to determine the time domain characteristics of the vibration displacement.

[0123] Step S222: performing frequency domain conversion according to the vibration displacement time domain data, thereby obtaining a vibration displacement spectrum diagram;

[0124] In this embodiment, the vibration displacement time domain data is converted into frequency domain data using Fourier transform technology. This can be achieved through the fast Fourier transform (FFT) algorithm, which converts the time domain signal into a frequency domain signal. The vibration displacement spectrum diagram will show the vibration intensity of different frequency components, helping to identify periodic vibrations and potential fault modes. Frequency domain conversion requires the selection of an appropriate window function (such as a Hanning window or a Hamming window) to reduce spectral leakage and ensure that the resolution of the FFT is high enough to capture subtle frequency changes. For example: Apply FFT to the vibration data collected every second to generate a spectrum diagram showing the vibration intensity in the frequency range of 0-1000Hz. For example, the spectrum diagram reveals significant vibration peaks at 50Hz and 150Hz, indicating that the generator is resonant or faulty at these frequencies.

[0125] Step S223: performing vibration displacement threshold statistics according to the vibration displacement spectrum diagram, thereby obtaining vibration displacement threshold data;

[0126] In this embodiment, after obtaining the vibration displacement spectrum diagram, statistical analysis is used to set the threshold. By analyzing historical data, the spectrum characteristics and peak range under normal operation are determined. For example, a threshold model can be established by calculating the mean and standard deviation of each frequency band in the spectrum, and the frequency components exceeding the threshold are considered abnormal. The setting of the threshold can be based on statistical methods, such as a 95% confidence interval, or a fixed standard value can be set based on empirical data. For example: based on historical normal operation data, a vibration displacement threshold of 50Hz is set to 0.05mm. If the actual measured vibration displacement exceeds 0.05mm at this frequency, it indicates that an abnormality exists.

[0127] Step S224: extracting vibration displacement monitoring data from the generator vibration sensor monitoring data, thereby obtaining vibration displacement monitoring data;

[0128] In this embodiment, data related to vibration displacement is extracted from the collected vibration monitoring data. This includes extracting vibration displacement data for a specific time period from the original monitoring signal, and preprocessing it (such as denoising and standardization). The data extraction process can use data mining techniques, such as feature selection algorithms, to extract vibration features related to faults. For example: extract vibration displacement data every minute in the past 24 hours from the monitoring data of the generator, and perform standardization to remove background noise in the data. The extracted data will be used for subsequent anomaly detection.

[0129] Step S225: performing vibration displacement detection on the vibration displacement monitoring data according to the vibration displacement threshold data, thereby obtaining vibration displacement abnormality data.

[0130] In this embodiment, the monitoring data is detected according to the vibration displacement threshold. The actual vibration displacement value is compared with the set threshold, and if some data is found to exceed the threshold, it is marked as abnormal. This can be achieved by writing an automated script or applying an anomaly detection algorithm, such as rule-based detection or machine learning methods to identify abnormal patterns. For example: the extracted vibration displacement data is compared with the set threshold. If vibration data exceeding the set threshold is found within the detection period (for example, every hour) (such as the vibration displacement at 50Hz exceeds 0.05mm), it is recorded as abnormal vibration displacement data, and an alarm is generated to notify maintenance personnel to conduct further inspection.

[0131] Optionally, step S23 is specifically:

[0132] Step S231: Performing Fourier transform on the generator vibration sensor monitoring data to obtain a generator vibration spectrum diagram;

[0133] In this embodiment, high-precision vibration sensors (such as accelerometers) are installed at different parts of the generator to collect vibration signals in real time. These sensors need to have a high sampling rate (such as 5000 samples per second) to ensure that sufficient vibration data is obtained. The collected time domain vibration data is input into the fast Fourier transform (FFT) algorithm. The FFT algorithm converts this data from the time domain to the frequency domain. The FFT algorithm used should have high resolution so that subtle changes in the spectrum can be identified. The FFT calculation results are plotted as a spectrum diagram. The spectrum diagram will show the amplitude of the generator vibration signal at different frequencies. This requires ensuring that the resolution of the spectrum diagram is high enough to clearly distinguish the vibration components of different frequencies. This process can be completed using Matlab or Python's scientific computing libraries (such as SciPy).

[0134] Step S232: performing a reference spectrum waveform analysis on the generator vibration spectrum diagram, thereby obtaining generator vibration reference spectrum waveform data;

[0135] In this embodiment, when the generator is operating normally, a vibration spectrum diagram for a period of time is recorded as baseline data. These data should cover the vibration characteristics of the generator under various operating conditions. The baseline spectrum diagram is analyzed in detail to determine the characteristic frequencies (such as the fundamental frequency and its harmonics) and their amplitudes. This can be done by using spectral line analysis technology, for example, applying statistical methods to calculate the mean value, standard deviation, etc. to identify the spectrum characteristics under normal working conditions. Based on the results of the spectrum analysis, the baseline spectrum waveform data of the generator is established, including parameters such as the fundamental frequency, harmonic frequency, and their relative amplitude and phase. These baseline data will serve as a reference for subsequent abnormality detection.

[0136] Step S333: extracting peak spectrum waveform features from the generator vibration spectrum diagram to obtain peak spectrum waveform data, and performing reference spectrum waveform detection on the peak spectrum waveform data according to the generator vibration reference spectrum waveform data to obtain generator vibration abnormal waveform data;

[0137] In this embodiment, high peak frequency features are identified and extracted from the spectrum graph. This includes applying a peak detection algorithm (such as threshold-based detection or a more complex peak detection algorithm) to the spectrum graph to identify significant peaks in the spectrum. These peaks typically represent specific vibration modes or potential faults of the system. The extracted high peak spectrum waveform data is compared with the baseline spectrum waveform data. Difference calculations (such as root mean square error, kurtosis difference, etc.) can be used to identify deviations from the baseline data. This step requires the use of a pattern recognition algorithm to determine whether the change in vibration characteristics exceeds the normal range. Based on the baseline detection results, the abnormal vibration waveform data is marked and recorded. These data will show vibration characteristics that are significantly different from the baseline spectrum waveform, indicating a system failure or abnormality.

[0138] Step S234: performing frequency statistics according to the generator vibration abnormal waveform data, thereby obtaining the generator vibration high-frequency abnormal waveform data;

[0139] In this embodiment, the generator vibration abnormal waveform data is statistically analyzed, focusing on the high-frequency abnormal part. Statistical methods (such as spectrum distribution analysis, spectrum distribution density function, etc.) are used to calculate the frequency distribution, especially the high-frequency component. Abnormal waveform data in the high-frequency component is identified, including high-frequency noise, vibration spikes, etc. These abnormal waveforms are usually caused by equipment failure, mechanical wear or other problems. The identified high-frequency abnormal waveform data is recorded to provide a basis for further vibration mode recognition. These data usually need to be cleaned and preprocessed to ensure accuracy.

[0140] Step S235: performing abnormal vibration mode recognition on the generator vibration spectrum diagram according to the generator vibration high-frequency abnormal waveform data, thereby obtaining vibration mode abnormal data.

[0141] In this embodiment, machine learning or statistical pattern recognition technology is used to identify abnormal patterns in the vibration spectrum. Algorithms such as support vector machines (SVM), neural networks, or other classification algorithms can be applied to identify whether the vibration pattern is significantly different from the normal pattern. Based on the results of pattern recognition, vibration pattern abnormality data is generated. These data include the type of abnormal pattern identified, the degree of abnormality, and related frequency information. The abnormal pattern is verified to confirm whether it truly reflects the fault or potential problem of the generator. This can be done by checking and comparing with the actual equipment status, for example, using a fault diagnosis tool or further physical inspection.

[0142] Optionally, step S3 specifically includes:

[0143] Step S31: acquiring generator sensor data, and extracting temperature sensor data and inductive air gap sensor data according to the generator sensor data, thereby obtaining temperature sensor data and inductive air gap sensor data;

[0144] In this embodiment, the generator sensing data is obtained, including the readings of the temperature sensor and the inductive air gap sensor. In a specific implementation, a temperature sensor (such as a thermocouple or a resistance temperature detector (RTD)) can be installed near the generator winding and bearing, and the inductive air gap sensor can be installed between the rotor and the stator. The data of these sensors are recorded regularly through a data acquisition system (such as a programmable logic controller (PLC) or a data acquisition card), and the data is extracted through a preset time interval or trigger condition (such as the temperature exceeds a threshold). The data needs to be preliminarily processed and calibrated to ensure accurate measurement.

[0145] Step S32: performing generator temperature anomaly detection on the online monitoring data of the hydropower unit to be processed according to the temperature sensing data, thereby obtaining generator online temperature anomaly data;

[0146] In this embodiment, the acquired temperature sensing data is used to detect temperature anomalies. A temperature threshold can be set (for example, exceeding 90°C is considered abnormal), and the actual measured temperature data is compared with the preset threshold through a real-time monitoring system (such as a SCADA system (Supervisory Control And Data Acquisition)). If the actual temperature exceeds the normal range, an alarm is triggered and recorded as abnormal online temperature data of the generator. The anomaly detection algorithm may include a moving average method, standard deviation analysis, etc. to improve accuracy and reliability.

[0147] Step S33: performing generator air gap anomaly detection on the online monitoring data of the hydropower unit to be processed according to the inductive air gap sensor data, thereby obtaining generator online air gap anomaly data;

[0148] In this embodiment, inductive air gap sensor data is used to detect air gap anomalies. In a specific implementation, the normal range of the air gap sensor data can be established as a baseline through historical data. By comparing the current air gap data with the baseline, a detection algorithm (such as threshold detection, data fitting, etc.) is used to identify air gap anomalies (for example, air gap changes exceed a preset tolerance range), and generate generator online air gap anomaly data. This can help determine whether the relative position of the rotor and the stator is offset.

[0149] Step S34: integrating abnormal sensor data according to the generator online temperature abnormality data and the generator online air gap abnormality data, thereby obtaining the generator abnormality sensor data;

[0150] In this embodiment, the online temperature anomaly data and air gap anomaly data of the generator are integrated. This step includes summarizing and analyzing the anomaly information of the two data sources. The temperature and air gap anomaly data can be combined using data fusion techniques (such as weighted average, principal component analysis) to form a comprehensive anomaly report. The purpose of integrating abnormal sensor data is to provide more comprehensive anomaly identification so as to more accurately diagnose potential faults of the generator.

[0151] Step S35: Perform environmental correction on the abnormal vibration data according to the abnormal generator sensing data, so as to obtain abnormal vibration correction data of the generator.

[0152] In this embodiment, the abnormal sensor data of the generator is subjected to environmental correction to obtain abnormal vibration correction data. In specific implementation, the abnormal sensor data is compared with the vibration data, and the influence of environmental factors (such as temperature changes and load fluctuations) on the vibration data is eliminated through an environmental correction algorithm (such as regression analysis and filtering technology). The corrected data can more accurately reflect the actual vibration state of the generator, thereby providing reliable vibration abnormality diagnosis information.

[0153] Optionally, step S32 is specifically:

[0154] Step S321: Acquire ambient temperature monitoring data, and perform temperature statistics on the ambient temperature monitoring data, thereby obtaining high temperature environment data;

[0155] In this embodiment, temperature data is collected by ambient temperature sensors deployed at key locations. These sensors include thermocouples, RTDs (resistance temperature detectors), etc., which have high accuracy and a wide measurement range. The data is uploaded to the data repository through a real-time acquisition system. The collected ambient temperature data is statistically analyzed, such as calculating the daily maximum temperature, minimum temperature and its changing trend, to identify and extract high temperature environment data. Specifically, statistical analysis tools such as Python's Pandas library can be used for data processing to filter out high temperature data that exceeds a preset threshold.

[0156] Step S322: extracting the generator stator temperature monitoring data and the generator rotor temperature monitoring data from the temperature sensing data, thereby obtaining the generator stator temperature monitoring data and the generator rotor temperature monitoring data;

[0157] In this embodiment, temperature monitoring is performed using temperature sensors built into the generator, targeting the stator and rotor parts respectively. The stator temperature sensor is arranged near the stator winding, and the rotor temperature sensor is arranged near the rotor winding. The sensor data is collected by the monitoring system and uploaded to the central control system. These data are extracted and classified, and the stator temperature monitoring data and the rotor temperature monitoring data are extracted respectively. The data can be connected to the data processing system through a data acquisition card (DAQ), and the system automatically identifies and stores them in categories.

[0158] Step S323: extracting the core temperature feature of the generator stator temperature monitoring data, thereby obtaining the core temperature data;

[0159] In this embodiment, the extracted generator stator temperature data is deeply analyzed to extract the temperature characteristics of the core. This involves modeling the specific location and heat conduction characteristics of the stator core. The core temperature is calculated using algorithms such as heat conduction models (Fourier conduction equations) or numerical simulation methods to obtain specific temperature data of the core. These calculations can be performed using software such as MATLAB or ANSYS, and the obtained data can be compared with actual measured values ​​to verify the accuracy of the model.

[0160] Step S324: extracting rotor winding temperature characteristics from the generator rotor temperature monitoring data, thereby obtaining rotor winding temperature data;

[0161] In this embodiment, the temperature data of the generator rotor is analyzed to extract the temperature characteristics of the rotor winding. This requires the establishment of a thermal model of the rotor winding by modeling the thermal conductivity and current load of the winding material. Thermal simulation tools such as computational fluid dynamics (CFD) software are used to simulate the temperature distribution of the rotor winding. The acquired rotor winding temperature data is used for subsequent analysis and compared with the actual data to ensure the accuracy of feature extraction.

[0162] Step S325: constructing a generator thermal coupling model according to the core temperature data and the rotor winding temperature data, thereby obtaining a generator thermal coupling model;

[0163] In this embodiment, a thermal coupling model of the generator is constructed based on the extracted core temperature data and rotor winding temperature data. The model should include the thermal coupling relationship between the stator and the rotor, that is, the transfer of heat between the stator and the rotor. Finite element analysis (FEA) software, such as COMSOL Multiphysics, can be used to establish a coupling model. The model needs to consider the thermal conductivity and interface thermal resistance of each part to ensure the accuracy of the simulation. This model is used for subsequent high temperature simulation and abnormal analysis.

[0164] Step S326: performing a high temperature simulation of the generator according to the generator thermal coupling model, thereby obtaining high temperature simulation data of the generator;

[0165] In this embodiment, the constructed thermal coupling model is used to simulate the generator in a high temperature environment. The high temperature environment data and the operating parameters of the generator are input, and the simulation software (such as ANSYS or Fluent) is run to observe the thermal behavior and temperature distribution of the generator under extreme high temperature conditions. The temperature change data of each key component should be recorded during the simulation process to ensure the accuracy of the simulation data and generate the corresponding temperature distribution map.

[0166] Step S327: performing time distribution statistics on the generator high temperature simulation data and the high temperature environment data, so as to obtain the generator high temperature simulation time distribution data and the high temperature environment time distribution data;

[0167] In this embodiment, the time distribution statistics of the high temperature simulation data and the actual high temperature environment data of the generator are performed. First, the simulation data and the actual environment data are sorted in time series, and the temperature distribution in different time periods is calculated. Statistical analysis tools (such as R language or Python's SciPy library) can be used to perform time distribution analysis and generate time distribution data. These data help identify the high temperature conditions and environmental changes of the generator at different time points.

[0168] Step S328: Calculate the time distribution overlap according to the generator high temperature simulation time distribution data and the high temperature environment time distribution data, so as to obtain the generator temperature abnormality data;

[0169] In this embodiment, the overlap of the time distribution is calculated by comparing the generator high temperature simulation time distribution data and the high temperature environment time distribution data. The overlap calculation method, such as the Jaccard index or the correlation coefficient, can be used to evaluate the similarity of the two time distribution data sets. The calculation result is used to identify whether the generator is in an abnormally high temperature state. This can be implemented using the NumPy library in Python, and an overlap chart is drawn for easy analysis.

[0170] Step S329: comparing the generator temperature anomaly data with the online monitoring data of the hydropower unit to be processed, thereby obtaining the generator online temperature anomaly data.

[0171] In this embodiment, the identified abnormal temperature data of the generator is compared with the online monitoring data of the hydropower unit to be processed. During the comparison process, the historical data, operating status and real-time monitoring data of the generator should be taken into account. Use data comparison tools (such as Python's Pandas library) to perform detailed analysis, find abnormal data and generate a report. The report should include the time point, temperature value and potential cause of the abnormality, providing decision support for maintenance personnel.

[0172] Optionally, step S33 is specifically:

[0173] Step S331: Acquire environmental humidity monitoring data, and perform humidity statistics on the environmental humidity monitoring data, thereby obtaining high humidity environmental data;

[0174] In this embodiment, in the monitoring system of the power equipment, a high-precision humidity sensor (for example, based on a capacitive or resistive humidity sensor) is used to collect environmental humidity data in real time. The data collection frequency of the sensor is set to once an hour. The collected humidity data is uploaded to the central data processing system through the data collection module. The system groups the data by day and calculates the daily average humidity value and the standard deviation of the humidity. The humidity statistics process includes grouping the humidity data by time period (for example, 24 hours) and calculating the humidity average for each time period. If the humidity value exceeds the set threshold (for example, 80% relative humidity), it is marked as high humidity environmental data. The high humidity environmental data will be stored in the database for subsequent analysis.

[0175] Step S332: extracting generator air gap monitoring data from the online monitoring data of the hydropower unit to be processed, thereby obtaining the generator air gap monitoring data;

[0176] In this embodiment, in the monitoring system of the hydropower unit, the air gap state of the generator is monitored in real time by an air gap sensor (such as an inductive or Hall effect sensor) installed on the generator. The air gap sensor will regularly record the voltage or current change data of the air gap and send these data to the data acquisition module. The data acquisition module classifies the received data according to the timestamp and extracts the real-time monitoring data of the generator air gap. The system will collect data at a set time interval (for example, every 15 minutes) to ensure the continuity and real-time nature of the data. The extracted data will be stored in the form of CSV or database records and used for subsequent analysis.

[0177] Step S333: performing humidity statistics on the generator air gap monitoring data, thereby obtaining the generator air gap monitoring high humidity data;

[0178] In this embodiment, the extracted generator air gap monitoring data is subjected to humidity analysis. First, the generator air gap data is matched with the ambient humidity data to determine the ambient humidity value at the time of each air gap data record. After these data are correlated, the air gap data under high humidity environment is calculated. Through statistical analysis, the air gap data under high humidity (for example, humidity exceeding 80%) is screened out to form the generator air gap monitoring high humidity data. Data analysis tools (such as MATLAB or Python's Pandas library) are used to process the data, extract the air gap data under high humidity conditions, and store them as a special data set.

[0179] Step S334: performing material dielectric loss factor analysis according to the generator air gap monitoring high humidity data, thereby obtaining material dielectric loss factor data;

[0180] In this embodiment, a material dielectric loss factor analysis method is used to process the high humidity data of the generator air gap monitoring. First, determine the calculation model of the dielectric loss factor of the material, such as by conducting an experiment through a dielectric loss measuring instrument. In a high humidity environment, the electrical characteristics of the generator air gap are tested and the dielectric loss factor data of the material is recorded. By analyzing the changes in the dielectric loss factor in the air gap, the loss factor data related to humidity is obtained. Professional testing equipment (such as an LCR meter or a dielectric loss meter) can be used for measurement, and the results can be statistically analyzed and recorded for further analysis.

[0181] Step S335: performing air gap anomaly correlation on the generator air gap monitoring high humidity data according to the material dielectric loss factor data, thereby obtaining the generator high humidity air gap anomaly data;

[0182] In this embodiment, the air gap abnormality analysis is performed on the high humidity data of the generator air gap monitoring. First, the material dielectric loss factor data is used to establish a discrimination model for air gap abnormality, such as using a threshold discrimination method or an abnormality detection algorithm based on machine learning. By comparing the loss factor data with the air gap data, those abnormal data points that are beyond the normal range are identified. Algorithms (such as K-means clustering or isolation forest) are applied to classify the air gap abnormalities to obtain the generator high humidity air gap abnormal data. The abnormal data will be marked and output in the form of a report to facilitate engineers to perform fault analysis and maintenance.

[0183] Step S336: performing time distribution statistics according to the high humidity environment data and the generator high humidity air gap abnormality data, so as to obtain the high humidity environment time distribution data and the generator high humidity air gap abnormality time distribution data;

[0184] In this embodiment, a time distribution statistical analysis is performed on the high humidity environment data and the generator air gap abnormality data. First, the high humidity environment data is counted by time period (for example, every hour) to generate high humidity environment time distribution data. Similarly, the generator air gap abnormality data is summarized by time period to generate generator high humidity air gap abnormality time distribution data. Use statistical software (such as Excel or R language) to draw a time distribution chart to analyze the frequency and distribution of humidity and abnormality within the time period. By analyzing the time distribution data, find out the law of high humidity environment and air gap abnormality, and record it in the database.

[0185] Step S337: Calculate the time distribution overlap according to the high humidity environment time distribution data and the generator high humidity air gap abnormality time distribution data, so as to obtain the generator air gap abnormality data.

[0186] In this embodiment, the time distribution overlap calculation is performed to analyze the relationship between the high humidity environment and the generator air gap anomaly. First, the high humidity environment time distribution data and the generator air gap anomaly time distribution data are imported into the overlap calculation tool. The overlap of the two time distribution data sets is calculated using an intersection analysis method (such as Jaccard similarity or correlation coefficient). By comparing the time distribution of the high humidity environment with the time distribution of the air gap anomaly, the correlation between them is determined. For example, the SciPy library in Python is used to perform a correlation analysis to obtain the time distribution overlap results of humidity and air gap anomaly. The calculation results are used to identify the impact of high humidity on air gap anomaly and generate a detailed analysis report.

[0187] Optionally, step S4 is specifically:

[0188] Step S41: extracting the generator amplitude data and the generator vibration frequency data from the generator abnormal vibration correction data, thereby obtaining the generator amplitude data and the generator vibration frequency data;

[0189] In this example, raw vibration data is extracted from the vibration monitoring system of the generator. This includes amplitude data and vibration frequency data. Data extraction is performed by a data acquisition system, which usually contains time series data collected by vibration sensors. Amplitude data refers to the intensity of vibration in millimeters (mm), while vibration frequency data is expressed in Hertz (Hz). Assuming that we extract one hour of vibration data, recorded once every second, this will generate 3600 vibration data points.

[0190] Step S42: performing amplitude range statistics on the generator amplitude data, thereby obtaining the generator amplitude range data;

[0191] In this embodiment, the extracted amplitude data is range counted. First, all amplitude data points are divided into several intervals (for example: 0-0.5mm, 0.5-1mm, 1-1.5mm, etc.), and then the number of data points in each interval is calculated. In this way, the distribution of the amplitude data and its frequency range can be determined. For example, if there are 1500 data points in the interval of 0-0.5mm, the amplitude range data of this interval is 1500.

[0192] Step S43: performing period statistics on the generator vibration frequency data, thereby obtaining the generator vibration frequency period data;

[0193] In this embodiment, period statistics are performed on the vibration frequency data. This includes calculating the average period and frequency distribution of the frequency data. The statistical method may include performing a Fourier transform on the vibration frequency data to identify the main frequency components and their periods. For example, if the main frequency is found to be 50 Hz in the analysis, its period is 1 / 50 second, i.e. 0.02 second.

[0194] Step S44: constructing a generator vibration monitoring model according to the generator amplitude range data and the generator vibration frequency period data, thereby obtaining a generator vibration monitoring model;

[0195] In this embodiment, a generator vibration monitoring model is established, and the model can be based on statistical analysis methods (such as regression analysis) or machine learning techniques (such as support vector machines). The goal of the model is to identify normal and abnormal vibration modes. For example, the model can output a prediction result of whether abnormal vibration exists by inputting amplitude range and frequency period data.

[0196] Step S45: performing generator online abnormal vibration detection on the online monitoring data of the hydropower unit according to the generator vibration monitoring model, thereby obtaining generator online abnormal vibration data.

[0197] In this embodiment, the monitoring model is applied to the real-time data stream. The monitoring system continuously acquires the online vibration data of the generator and uses the model to detect anomalies. If the model detects that the vibration pattern is significantly different from the normal pattern, it is marked as abnormal. For example, if the model identifies a high amplitude signal that deviates significantly from the normal operating frequency, the system will issue an alarm, indicating that further inspection is needed.

[0198] Optionally, step S45 is specifically:

[0199] Step S451: performing generator online abnormal vibration detection on the online monitoring data of the hydropower unit according to the generator vibration monitoring model, thereby obtaining generator abnormal vibration data and generator normal vibration data;

[0200] In this embodiment, a vibration sensor is installed on the generator of the hydropower unit to collect its vibration data in real time. These data are transmitted to the monitoring center through the data acquisition system, and after preprocessing, they are compared with historical data. The monitoring model is based on machine learning algorithms, such as support vector machines (SVM) or convolutional neural networks (CNN), to analyze real-time vibration data and detect abnormal vibrations. Abnormal vibration data will be marked with its corresponding vibration frequency, amplitude and direction, and normal vibration data will be recorded as a comparison benchmark. During the implementation process, the sensitivity of the sensor is set to 1μm and the sampling frequency is 1kHz to ensure high-precision data acquisition.

[0201] Step S452: performing a normal vibration parameter comparison based on the abnormal vibration data of the generator and the normal vibration data of the generator, thereby obtaining a vibration error parameter set;

[0202] In this embodiment, the normal vibration data is compared with the abnormal vibration data. The normal vibration data is statistically analyzed to calculate its mean, standard deviation and spectral characteristics. These statistical characteristics are then used as reference values ​​for the normal state and compared with the corresponding characteristics of the abnormal vibration data. The degree of vibration abnormality is evaluated by calculating a vibration error parameter set, such as frequency deviation, amplitude deviation, and phase difference. These parameter sets include frequency error ranges (such as ±5Hz), amplitude variation ranges (such as ±10%), and generate error reports.

[0203] Step S453: extracting abnormal vibration speed features from the vibration error parameter set, thereby obtaining abnormal vibration speed data;

[0204] In this embodiment, vibration velocity feature extraction is performed based on the obtained vibration error parameter set. The vibration error parameter data is converted from the time domain to the frequency domain by Fourier transform (FFT) to obtain the spectrum characteristics of the vibration velocity. Feature extraction algorithms, such as principal component analysis (PCA) or independent component analysis (ICA), are applied to extract the main velocity features from the spectrum. For example, the main abnormal velocity frequency band (such as 1-10Hz) is extracted, and its amplitude and frequency characteristics are recorded. These features are used to identify potential vibration problems and generate a detailed velocity abnormality data report.

[0205] Step S454: Compare the online vibration speed of the generator with the online monitoring data of the hydropower unit according to the abnormal vibration speed data, so as to obtain the online abnormal vibration data of the generator.

[0206] In this embodiment, the vibration abnormal velocity data is used to perform a vibration velocity comparison analysis on the generator of the hydropower unit. The real-time monitored vibration velocity data is compared with the vibration abnormal velocity data to identify similar abnormal patterns. The comparison process includes matching the velocity characteristics of the real-time data with the abnormal velocity characteristics, for example, by calculating the similarity coefficient or using a classification algorithm (such as a K-nearest neighbor algorithm) for comparison. If it is found that the similarity exceeds a preset threshold (such as 85%), it is marked as abnormal vibration. These comparison results are used in the real-time monitoring system to promptly alarm and take corresponding maintenance measures. For example: Assume that the normal vibration data collected by the vibration monitoring system of a hydropower unit generator has a mean of 0.5μm, a standard deviation of 0.05μm, and a frequency range of 20-100Hz. After comparison, it is found that the mean of the abnormal vibration data is 1.0μm, the standard deviation is 0.2μm, and the frequency deviation is ±8Hz. In the vibration error parameter set, the amplitude of the abnormal velocity characteristic of the 3Hz frequency band is significantly higher than the normal range. The comparison results show that the amplitude of the 3Hz frequency band in the real-time data is 1.2μm, which is higher than the normal value. The system alarms and recommends inspection.

[0207] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0208] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for analyzing online monitoring data of a hydropower unit, characterized in that: The following steps are involved: Step S1: acquiring online monitoring data of a hydropower unit, and performing data preprocessing on the online monitoring data of the hydropower unit, thereby obtaining online monitoring data of the hydropower unit to be processed; Step S2: extracting generator vibration sensor monitoring data according to the online monitoring data of the hydropower unit to be processed, thereby obtaining generator vibration sensor monitoring data; performing vibration anomaly analysis according to the generator vibration sensor monitoring data, thereby obtaining vibration anomaly data; Step S3: acquiring the generator sensor data, and performing environmental correction on the abnormal vibration data according to the generator sensor data, thereby obtaining the generator abnormal vibration correction data; Step S4: constructing a generator vibration monitoring model according to the generator abnormal vibration correction data, thereby obtaining a generator vibration monitoring model; performing generator online abnormal vibration detection on the hydropower unit online monitoring data according to the generator vibration monitoring model, thereby obtaining generator online abnormal vibration data; Step S5: Perform vibration risk assessment based on the online abnormal vibration data of the generator to obtain the generator vibration risk data, and upload it to the hydropower unit management platform to execute the generator risk warning task.

2. The method for analyzing online monitoring data of a hydropower unit according to claim 1, characterized in that: Step S1 is specifically as follows: Step S11: obtaining online monitoring data of the hydropower unit; Step S12: performing data cleaning on the online monitoring data of the hydropower unit, thereby obtaining online monitoring cleaning data; Step S13: Fill missing values ​​in the online monitoring cleaning data to obtain online monitoring filling data; Step S14: normalize the online monitoring filling data to obtain the online monitoring data of the hydropower unit to be processed.

3. The method for analyzing online monitoring data of a hydropower unit according to claim 1, characterized in that: Step S2 is specifically as follows: Step S21: extracting generator vibration sensor monitoring data according to the online monitoring data of the hydropower unit to be processed, thereby obtaining generator vibration sensor monitoring data; Step S22: performing vibration displacement anomaly analysis on the generator vibration sensor monitoring data, thereby obtaining vibration displacement anomaly data; Step S23: performing vibration mode abnormality analysis on the generator vibration sensor monitoring data, thereby obtaining vibration mode abnormality data; Step S24: integrating the vibration abnormality data according to the vibration displacement abnormality data and the vibration mode abnormality data, so as to obtain the vibration abnormality data.

4. The method for analyzing online monitoring data of a hydropower unit according to claim 3, characterized in that: Step S22 is specifically as follows: Step S221: performing vibration displacement time domain analysis on the generator vibration sensor monitoring data, thereby obtaining vibration displacement time domain data; Step S222: performing frequency domain conversion according to the vibration displacement time domain data, thereby obtaining a vibration displacement spectrum diagram; Step S223: performing vibration displacement threshold statistics according to the vibration displacement spectrum diagram, thereby obtaining vibration displacement threshold data; Step S224: extracting vibration displacement monitoring data from the generator vibration sensor monitoring data, thereby obtaining vibration displacement monitoring data; Step S225: performing vibration displacement detection on the vibration displacement monitoring data according to the vibration displacement threshold data, thereby obtaining vibration displacement abnormality data.

5. The method for analyzing online monitoring data of a hydropower unit according to claim 3, characterized in that: Step S23 is specifically as follows: Step S231: Performing Fourier transform on the generator vibration sensor monitoring data to obtain a generator vibration spectrum diagram; Step S232: performing a reference spectrum waveform analysis on the generator vibration spectrum diagram, thereby obtaining generator vibration reference spectrum waveform data; Step S333: extracting peak spectrum waveform features from the generator vibration spectrum diagram to obtain peak spectrum waveform data, and performing reference spectrum waveform detection on the peak spectrum waveform data according to the generator vibration reference spectrum waveform data to obtain generator vibration abnormal waveform data; Step S234: performing frequency statistics according to the generator vibration abnormal waveform data, thereby obtaining the generator vibration high-frequency abnormal waveform data; Step S235: performing abnormal vibration mode recognition on the generator vibration spectrum diagram according to the generator vibration high-frequency abnormal waveform data, thereby obtaining vibration mode abnormal data.

6. The method for analyzing online monitoring data of a hydropower unit according to claim 1, characterized in that: Step S3 is specifically as follows: Step S31: acquiring generator sensor data, and extracting temperature sensor data and inductive air gap sensor data according to the generator sensor data, thereby obtaining temperature sensor data and inductive air gap sensor data; Step S32: performing generator temperature anomaly detection on the online monitoring data of the hydropower unit to be processed according to the temperature sensing data, thereby obtaining generator online temperature anomaly data; Step S33: performing generator air gap anomaly detection on the online monitoring data of the hydropower unit to be processed according to the inductive air gap sensor data, thereby obtaining generator online air gap anomaly data; Step S34: integrating abnormal sensor data according to the generator online temperature abnormality data and the generator online air gap abnormality data, thereby obtaining the generator abnormality sensor data; Step S35: Perform environmental correction on the abnormal vibration data according to the abnormal generator sensing data, so as to obtain abnormal vibration correction data of the generator.

7. The method for analyzing online monitoring data of a hydropower unit according to claim 6, characterized in that: Step S32 is specifically as follows: Step S321: Acquire ambient temperature monitoring data, and perform temperature statistics on the ambient temperature monitoring data, thereby obtaining high temperature environment data; Step S322: extracting the generator stator temperature monitoring data and the generator rotor temperature monitoring data from the temperature sensing data, thereby obtaining the generator stator temperature monitoring data and the generator rotor temperature monitoring data; Step S323: extracting the core temperature feature of the generator stator temperature monitoring data, thereby obtaining the core temperature data; Step S324: extracting rotor winding temperature characteristics from the generator rotor temperature monitoring data, thereby obtaining rotor winding temperature data; Step S325: constructing a generator thermal coupling model according to the core temperature data and the rotor winding temperature data, thereby obtaining a generator thermal coupling model; Step S326: performing a high temperature simulation of the generator according to the generator thermal coupling model, thereby obtaining high temperature simulation data of the generator; Step S327: performing time distribution statistics on the generator high temperature simulation data and the high temperature environment data, so as to obtain the generator high temperature simulation time distribution data and the high temperature environment time distribution data; Step S328: Calculate the time distribution overlap according to the generator high temperature simulation time distribution data and the high temperature environment time distribution data, so as to obtain the generator temperature abnormality data; Step S329: comparing the generator temperature anomaly data with the online monitoring data of the hydropower unit to be processed, thereby obtaining the generator online temperature anomaly data.

8. The method for analyzing online monitoring data of a hydropower unit according to claim 6, characterized in that: Step S33 is specifically as follows: Step S331: Acquire environmental humidity monitoring data, and perform humidity statistics on the environmental humidity monitoring data, thereby obtaining high humidity environmental data; Step S332: extracting generator air gap monitoring data from the online monitoring data of the hydropower unit to be processed, thereby obtaining the generator air gap monitoring data; Step S333: performing humidity statistics on the generator air gap monitoring data, thereby obtaining the generator air gap monitoring high humidity data; Step S334: performing material dielectric loss factor analysis according to the generator air gap monitoring high humidity data, thereby obtaining material dielectric loss factor data; Step S335: performing air gap anomaly correlation on the generator air gap monitoring high humidity data according to the material dielectric loss factor data, thereby obtaining the generator high humidity air gap anomaly data; Step S336: performing time distribution statistics according to the high humidity environment data and the generator high humidity air gap abnormality data, so as to obtain the high humidity environment time distribution data and the generator high humidity air gap abnormality time distribution data; Step S337: Calculate the time distribution overlap according to the high humidity environment time distribution data and the generator high humidity air gap abnormality time distribution data, so as to obtain the generator air gap abnormality data.

9. The method for analyzing online monitoring data of a hydropower unit according to claim 1, characterized in that: Step S4 is specifically as follows: Step S41: extracting the generator amplitude data and the generator vibration frequency data from the generator abnormal vibration correction data, thereby obtaining the generator amplitude data and the generator vibration frequency data; Step S42: performing amplitude range statistics on the generator amplitude data, thereby obtaining the generator amplitude range data; Step S43: performing period statistics on the generator vibration frequency data, thereby obtaining the generator vibration frequency period data; Step S44: constructing a generator vibration monitoring model according to the generator amplitude range data and the generator vibration frequency period data, thereby obtaining a generator vibration monitoring model; Step S45: performing generator online abnormal vibration detection on the online monitoring data of the hydropower unit according to the generator vibration monitoring model, thereby obtaining generator online abnormal vibration data.

10. The method for analyzing online monitoring data of a hydropower unit according to claim 9, characterized in that: Step S45 is specifically as follows: Step S451: performing generator online abnormal vibration detection on the online monitoring data of the hydropower unit according to the generator vibration monitoring model, thereby obtaining generator abnormal vibration data and generator normal vibration data; Step S452: performing a normal vibration parameter comparison based on the abnormal vibration data of the generator and the normal vibration data of the generator, thereby obtaining a vibration error parameter set; Step S453: extracting abnormal vibration speed features from the vibration error parameter set, thereby obtaining abnormal vibration speed data; Step S454: Compare the online vibration speed of the generator with the online monitoring data of the hydropower unit according to the abnormal vibration speed data, so as to obtain the online abnormal vibration data of the generator.

Images