Vibration measurement system for end coil of generator

Abstract

The invention discloses a vibration measurement system for an end coil of a generator, and particularly relates to the technical field of driving mechanical state monitoring, comprising a measurement point planning module which divides an end coil monitoring area into a plurality of logic measurement point units according to a three-dimensional model of an end structure of the generator and electromagnetic field simulation data, each unit accurately corresponds to a nose end, a bevel edge and a linear segment key part of the coil; a non-contact laser vibration meter and a micro inertial sensing unit are adopted to form a hybrid sensing network; according to the method, a multi-source information fusion model of mechanical vibration, an electromagnetic field, thermodynamics and working condition data is constructed, dynamic correlation analysis is carried out on vibration characteristics, mechanical energy, load current and temperature parameters, three-dimensional evaluation of the health state of the end coil is achieved, and systematic improvement from single data monitoring to comprehensive diagnosis decision is achieved.

Description

Technical Field

[0001] This invention relates to the field of drive machinery condition monitoring technology, and more specifically, to a generator end coil vibration measurement system. Background Technology

[0002] As a core piece of equipment in modern power systems, the reliability of generators directly affects the stability of the power grid and the security of power supply. With the current vigorous development of high-efficiency and energy-saving power generation technologies, the requirements for operational stability of energy-saving generators and generator sets are increasingly stringent. During operation, the stator end coils are subjected to a strong rotating magnetic field and short-circuit electromotive force, generating harmful electromagnetic vibrations at twice the power frequency. This vibration not only threatens equipment safety but also causes additional mechanical energy dissipation, affecting the efficient operation of the generator set. Long-term continuous vibration can easily lead to loosening of the end coil fixing structure and insulation wear, potentially causing catastrophic failures such as inter-turn short circuits and main insulation breakdown, resulting in huge economic losses. Therefore, real-time and accurate online monitoring of the vibration state of the generator end coils is a key technical means to achieve predictive maintenance and ensure the continuous and efficient operation of energy-saving generator sets.

[0003] Traditional coil vibration measurement systems include a sensor module, a signal acquisition module, and a data management module. The sensor module consists of a piezoelectric accelerometer, which is directly fixed to the coil or support structure by adhesive or magnetic adsorption to sense vibration acceleration. The signal acquisition module is responsible for powering the sensor, receiving the raw analog signal from the sensor, and performing amplification and filtering preprocessing to eliminate some noise interference. The data management module performs analog-to-digital conversion on the conditioned signal, calculates the effective value of vibration velocity or displacement, and realizes data storage, historical curve display, and over-limit alarm in the host computer software.

[0004] However, in practical use, it still has some drawbacks, such as low measurement accuracy. Traditional systems use contact installation, and the mass of the sensor itself will have a mass load effect on the lightweight coil, changing the original vibration characteristics of the coil and causing the measurement results to be distorted. The fault diagnosis dimension is limited. Traditional systems mainly monitor the total amount of vibration, but cannot accurately separate the specific frequency components directly related to the coil loosening fault, making it difficult to support the refined operation and maintenance of energy-saving generators. The data acquisition is incomplete. Traditional solutions can usually only measure in a single direction, making it difficult to simultaneously acquire the vibration mode of the end coil in multiple directions, and making it impossible to comprehensively evaluate the tightness of the coil.

[0005] Therefore, there is an urgent need to provide a generator end coil vibration measurement system to solve the problems of low measurement accuracy, single fault diagnosis dimension, and incomplete data acquisition in existing coil vibration measurement systems. Summary of the Invention

[0006] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a generator end coil vibration measurement system, which solves the problems mentioned in the background art through the following solutions.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a generator end coil vibration measurement system, comprising: The measurement point planning module divides the end coil monitoring area into multiple logical measurement point units based on the three-dimensional model of the generator end structure and electromagnetic field simulation data. Each unit precisely corresponds to the nose, inclined side, and key parts of the straight section of the coil. A hybrid sensing network is formed by a non-contact laser vibration meter and a miniature inertial sensing unit. The signal acquisition module performs noise reduction filtering and trend term elimination processing on the acquired raw vibration signal to improve the signal-to-noise ratio. Then, it applies Fast Fourier Transform to convert the time-domain signal into a frequency-domain energy distribution, accurately separating the characteristic frequency components of the power frequency harmonics and its higher harmonics, and transforming the original waveform data into a set of frequency-domain characteristic parameters with clear physical meaning. The index calculation module converts the frequency domain feature parameter set into the data influence coefficients of the mechanical vibration layer, electromagnetic field layer, thermodynamic layer, and generator operating condition layer through multi-level physical modeling and normalization algorithms, and constructs a quantitative system for assessing health status. The comprehensive analysis module calculates the influence coefficients of mechanical vibration layer data, electromagnetic field layer data, thermodynamic layer data, and generator operating condition layer data through a multi-source influence coefficient fusion algorithm. It establishes a comprehensive health index based on a four-dimensional state space and establishes a quantitative mapping relationship between health status and the four influence coefficients, realizing the transformation from multi-dimensional monitoring to a unified health status. The coil vibration measurement module, based on the comprehensive health index output by the comprehensive analysis module, establishes a three-level early warning mechanism and a fault diagnosis rule base to realize fully automatic assessment and early warning of the end coil vibration status and generate a coil status assessment report. The human-computer interaction module transmits the comprehensive health index, four-dimensional influence coefficient, abnormal signs, and coil status assessment report to the user information terminal, providing reference data for making adjustment measures.

[0008] Preferably, the frequency domain feature parameter set includes mechanical vibration layer data, electromagnetic field layer data, thermodynamic layer data, and generator operating condition layer data.

[0009] Preferably, the mechanical vibration layer data includes axial vibration displacement, denoted as... Radial vibration displacement, denoted as Tangential vibration displacement, denoted as The amplitude of the 100Hz frequency vibration is denoted as... The vibration amplitude at a frequency of 200Hz is denoted as... Total vibration, denoted as RMS; Electromagnetic field layer data includes the 100Hz frequency vibration amplitude, denoted as... Stator A-phase current under 100Hz harmonic vibration, denoted as I; thermodynamic layer data includes thermally induced vibration deviation, denoted as The amplitude of the 100Hz frequency vibration is denoted as The generator operating condition data includes active power, denoted as P; cumulative start-stop count, denoted as N; and cumulative running time, denoted as T.

[0010] Preferably, the mechanical vibration layer data influence coefficient quantifies the concentration of vibration energy direction by calculating the ratio of the magnitude of the three-dimensional vibration vector to its Euclidean norm. At the same time, it combines the proportion of harmonic vibration energy in the total vibration energy to reflect the abnormal frequency response of the mechanical structure. The arithmetic average of the two realizes the fusion diagnosis of vibration mode and spectrum characteristics, which is used to characterize the degree of deterioration of the overall mechanical fastening state of the end coil.

[0011] Preferably, the electromagnetic field layer data influence coefficient quantifies the linear deviation of the vibration response under the current electromagnetic excitation by establishing the relative ratio of the 100Hz vibration generated by a unit current to the rated reference value. Its physical essence is to characterize the change in the efficiency of the electromagnetic force acting on the coil, and is used to diagnose the impedance matching state between the electromagnetic load and the mechanical system.

[0012] Preferably, the thermodynamic layer data influence coefficient is calculated by measuring the absolute ratio of the thermally induced vibration deviation to the current 100Hz vibration amplitude to quantify the relative contribution of non-thermal factors to the vibration state, thereby isolating the influence of temperature changes and realizing the extraction of the pure mechanical components of vibration.

[0013] Preferably, the generator operating condition data influence coefficient is obtained by geometrically averaging the square effect of instantaneous load, fatigue life consumption rate, and aging life consumption rate, and using the cube root to ensure dimensional normalization and balanced coupling of various contributions, so as to comprehensively reflect the normalized influence of operating conditions on the cumulative damage of the coil.

[0014] Preferably, the comprehensive health index achieves dimensional normalization and state fusion through the geometric mean of four-layer influence coefficients. The numerator represents the degree of loss of mechanical structural integrity, the first denominator quantifies the electromagnetic matching anomaly, the second denominator represents the thermal vibration deviation, and the third denominator reflects the cumulative damage of the working conditions. The fourth root of the four ratios ensures balanced coupling of the influence of each dimension, and finally, the health status is intuitively quantified by scaling on a percentage scale.

[0015] Preferably, the coil vibration measurement module specifically includes: S1. Real-time monitoring and classification of health status The system receives the Comprehensive Health Index (HI) value and classifies its status in real time using a preset three-level threshold system: when HI ≥ 85, it is determined to be in the safe operation zone, and the system maintains the normal monitoring frequency; when 70 ≤ HI < 85, it automatically enters the observation zone and starts the enhanced monitoring mode; when HI < 70, it immediately enters the early warning and maintenance zone and triggers the alarm response mechanism of the entire system. S2, Implementation of Tiered Response Strategy In the observation zone, the module automatically increases the data acquisition frequency from the usual 1Hz to 10Hz, and generates a list of operational optimization suggestions, including "It is recommended to limit the load fluctuation range to ±5%" and "Strengthen the inspection of the cooling system". In the early warning maintenance zone, the module immediately activates the audible and visual alarm device in the control room and initiates the fault diagnosis process by analyzing the combination mode of four influence coefficients: when Km increases significantly and Ke>1.2, it is diagnosed as a mechanical loosening fault; when Kt is abnormally high and Ko>0.8, it is diagnosed as an abnormal thermal stress. S3. Fault Location and Visualization The system calls upon the 3D model database at the generator end to map the health index and influence coefficient of each measuring point to the corresponding spatial location. A dynamic heat map is generated using a gradient coloring algorithm: the distribution of the health index of 85-70-60-50 is represented by a blue-green-yellow-red color transition, and the main direction of abnormal vibration is marked by vector arrows. The system automatically marks the three measuring points with the lowest health index as priority maintenance sites and highlights these key areas in the 3D model.

[0016] S4. Diagnostic Report Generation and Decision Support Integrating historical data trends over 24 consecutive hours, it automatically generates a structured diagnostic report, including a health index change curve, a pie chart of the contribution of each influencing coefficient, a fault probability analysis matrix, and a specific list of maintenance recommendations. S5, Adaptive Threshold Optimization The module performs monthly statistical analysis of historical operating data and dynamically adjusts the three-level threshold parameters based on the actual operating characteristics of the unit: if the health index remains stable above 90 for three consecutive months, the safe operating zone threshold is automatically raised from 85 to 88; if a certain influence coefficient is found to be continuously deteriorating, its weight coefficient in the fault diagnosis rule base of that model is adjusted accordingly, thereby realizing the self-evolution of the evaluation system.

[0017] The technical effects and advantages of this invention are as follows: 1. This invention constructs a multi-source information fusion model of mechanical vibration, electromagnetic field, thermodynamics and operating condition data, and performs dynamic correlation analysis between vibration characteristics and mechanical energy and load current and temperature parameters, realizing a three-dimensional assessment of the health status of the end coil, and achieving a systematic improvement from single data monitoring to comprehensive diagnostic decision-making. 2. This invention establishes a feature frequency extraction algorithm based on fast Fourier transform to accurately separate the key harmonic vibration components of 100Hz, and combines multi-physics field influence coefficients to perform fusion analysis of four-dimensional data influence relationships, monitor the transmission relationship between mechanical energy and coil vibration, and realize the dimensional expansion from total vibration monitoring to specific fault mechanism identification. 3. This invention, through a closed-loop diagnostic architecture from data acquisition and feature extraction to status assessment, transforms monitoring data into actionable operation and maintenance guidance and early warning of abnormal accumulation of mechanical energy in real time, thereby realizing visualized control and predictive maintenance of equipment status and improving the intelligence level and management efficiency of generator operation and maintenance. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the overall structure of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] As attached Figure 1 The generator end coil vibration measurement system shown includes: The measurement point planning module divides the end coil monitoring area into multiple logical measurement point units based on the three-dimensional model of the generator end structure and electromagnetic field simulation data. Each unit precisely corresponds to the nose, inclined side, and key parts of the straight section of the coil. A hybrid sensing network is formed by a non-contact laser vibration meter and a miniature inertial sensing unit.

[0021] In this embodiment, it should be specifically noted that the non-contact laser vibration meter is responsible for high-precision displacement measurement, while the latter captures the vibration of the structure. The spatial registration technology establishes the directional mapping relationship between the sensing unit and the logical measurement point, laying a spatial topological foundation for the subsequent three-dimensional vibration vector synthesis.

[0022] The signal acquisition module performs noise reduction filtering and trend term elimination processing on the acquired raw vibration signal to improve the signal-to-noise ratio. Then, it applies Fast Fourier Transform to convert the time-domain signal into a frequency-domain energy distribution, accurately separating the characteristic frequency components of the power frequency harmonics and its higher harmonics, and transforming the original waveform data into a set of frequency-domain characteristic parameters with clear physical meaning.

[0023] In this embodiment, it should be specifically noted that the frequency domain feature parameter set includes mechanical vibration layer data, electromagnetic field layer data, thermodynamic layer data, and generator operating condition layer data.

[0024] In this embodiment, it should be specifically noted that the mechanical vibration layer data includes axial vibration displacement, denoted as... Radial vibration displacement, denoted as Tangential vibration displacement, denoted as The amplitude of the 100Hz frequency vibration is denoted as... The vibration amplitude at a frequency of 200Hz is denoted as... The total vibration, denoted as RMS.

[0025] In this embodiment, it should be specifically noted that the electromagnetic field layer data includes the 100Hz frequency vibration amplitude, denoted as... The stator A-phase current under 100Hz harmonic vibration is denoted as I.

[0026] In this embodiment, it should be specifically noted that the thermodynamic layer data includes thermally induced vibration deviation, denoted as... The amplitude of the 100Hz frequency vibration is denoted as .

[0027] In this embodiment, it should be specifically explained that the thermally induced vibration deviation is extracted by measuring the coil surface temperature with a temperature sensor and collecting axial vibration displacement signals with a vibration meter. Specifically: , in This indicates the vibration amplitude at 100Hz. This represents the vibration displacement amplitude measured at 100Hz under rated operating conditions. Indicates the coefficient of thermal expansion at the vibration point. Indicates the surface temperature of the coil. This indicates the reference temperature of the coil.

[0028] In this embodiment, it should be specifically noted that the generator operating condition data includes active power, denoted as P; cumulative start-stop count, denoted as N; and cumulative running time, denoted as T.

[0029] The index calculation module converts the frequency domain feature parameter set into the data influence coefficients of the mechanical vibration layer, electromagnetic field layer, thermodynamic layer, and generator operating condition layer through multi-level physical modeling and normalization algorithms, and constructs a quantitative system for assessing health status.

[0030] In this embodiment, it should be specifically explained that the mechanical vibration layer data influence coefficient quantifies the concentration of vibration energy direction by calculating the ratio of the magnitude of the three-dimensional vibration vector to its Euclidean norm. Simultaneously, it reflects the frequency response anomalies of the mechanical structure by combining the proportion of harmonic vibration energy in the total vibration energy. The arithmetic average of these two factors achieves a fusion diagnosis of vibration mode and spectral characteristics, used to characterize the degree of deterioration of the overall mechanical fastening state of the end coil. Specifically: , in Indicates axial vibration displacement. Indicates radial vibration displacement. Indicates tangential vibration displacement; This represents the vibration amplitude at a frequency of 100Hz. It represents the vibration amplitude at a frequency of 200 Hz; RMS represents the total vibration.

[0031] In this embodiment, it should be specifically noted that the electromagnetic field layer data influence coefficient quantifies the linear deviation of the vibration response under the current electromagnetic excitation by establishing the relative ratio of the 100Hz vibration generated by a unit current to the rated reference value. Its physical essence is to characterize the change in the efficiency of the electromagnetic force acting on the coil, and it is used to diagnose the impedance matching state between the electromagnetic load and the mechanical system. Specifically: , in I represents the amplitude of the 100Hz frequency vibration, and I represents the stator A-phase current under 100Hz harmonic vibration. This represents the vibration displacement amplitude measured at 100Hz under rated operating conditions. This indicates the rated stator current.

[0032] In this embodiment, it should be specifically noted that the thermodynamic layer data influence coefficient is calculated by using the absolute ratio of the thermally induced vibration deviation to the current 100Hz vibration amplitude to quantify the relative contribution of non-thermal factors to the vibration state, thereby isolating the influence of temperature changes and realizing the extraction of the pure mechanical components of vibration. Specifically: , in Indicates thermally induced vibration deviation. It represents the vibration amplitude at a frequency of 100Hz.

[0033] In this embodiment, it should be specifically explained that the generator operating condition layer data influence coefficient is achieved by geometrically averaging the square effect of instantaneous load, fatigue life consumption rate, and aging life consumption rate, and using the cube root to ensure dimensional normalization and balanced coupling of various contributions. This comprehensively reflects the normalized influence of operating conditions on cumulative coil damage. Specifically: , Where P represents active power. This represents the rated power, and N represents the cumulative number of start-stop cycles. This indicates the number of start-stop cycles allowed by the design, where T represents the cumulative runtime. Indicates the design lifespan.

[0034] The comprehensive analysis module calculates the influence coefficients of mechanical vibration layer data, electromagnetic field layer data, thermodynamic layer data, and generator operating condition layer data through a multi-source influence coefficient fusion algorithm. It establishes a comprehensive health index based on a four-dimensional state space and establishes a quantitative mapping relationship between health status and the four influence coefficients, realizing the transformation from multi-dimensional monitoring to a unified health status.

[0035] In this embodiment, it should be specifically explained that the comprehensive health index achieves dimensional normalization and state fusion through the geometric mean of four-layer influence coefficients. The numerator represents the degree of mechanical structural integrity loss, the first denominator quantifies the electromagnetic matching anomaly, the second denominator represents the thermal vibration deviation, and the third denominator reflects the cumulative damage under operating conditions. The fourth root of the four ratios ensures balanced coupling of influences across dimensions. Finally, a percentage scaling method is used to achieve an intuitive quantification of the health status, specifically: , in This represents the influence coefficient of mechanical vibration layer data. This represents the influence coefficient of electromagnetic field layer data. This represents the influence coefficient of thermodynamic layer data. This represents the influence coefficient of the generator operating condition layer data.

[0036] The coil vibration measurement module, based on the comprehensive health index output by the comprehensive analysis module, establishes a three-level early warning mechanism and a fault diagnosis rule base to achieve fully automatic assessment and early warning of the end coil vibration status and generate a coil status assessment report.

[0037] In this embodiment, it should be specifically noted that the coil vibration measurement module includes: S1. Real-time monitoring and classification of health status The system receives the Comprehensive Health Index (HI) value and classifies its status in real time using a preset three-level threshold system: when HI ≥ 85, it is determined to be in the safe operation zone, and the system maintains the normal monitoring frequency; when 70 ≤ HI < 85, it automatically enters the observation zone and starts the enhanced monitoring mode; when HI < 70, it immediately enters the early warning and maintenance zone and triggers the alarm response mechanism of the entire system. S2, Implementation of Tiered Response Strategy In the observation zone, the module automatically increases the data acquisition frequency from the usual 1Hz to 10Hz, and generates a list of operational optimization suggestions, including "It is recommended to limit the load fluctuation range to ±5%" and "Strengthen the inspection of the cooling system". In the early warning maintenance zone, the module immediately activates the audible and visual alarm device in the control room and initiates the fault diagnosis process by analyzing the combination mode of four influence coefficients: when Km increases significantly and Ke>1.2, it is diagnosed as a mechanical loosening fault; when Kt is abnormally high and Ko>0.8, it is diagnosed as an abnormal thermal stress. S3. Fault Location and Visualization The system calls upon the 3D model database at the generator end to map the health index and influence coefficient of each measuring point to the corresponding spatial location. A dynamic heat map is generated using a gradient coloring algorithm: the distribution of the health index of 85-70-60-50 is represented by a blue-green-yellow-red color transition, and the main direction of abnormal vibration is marked by vector arrows. The system automatically marks the three measuring points with the lowest health index as priority maintenance sites and highlights these key areas in the 3D model.

[0038] S4. Diagnostic Report Generation and Decision Support Integrating historical data trends over 24 consecutive hours, it automatically generates a structured diagnostic report, including a health index change curve, a pie chart of the contribution of each influencing coefficient, a fault probability analysis matrix, and a specific list of maintenance recommendations. S5, Adaptive Threshold Optimization The module performs monthly statistical analysis of historical operating data and dynamically adjusts the three-level threshold parameters based on the actual operating characteristics of the unit: if the health index remains stable above 90 for three consecutive months, the safe operating zone threshold is automatically raised from 85 to 88; if a certain influence coefficient is found to be continuously deteriorating, its weight coefficient in the fault diagnosis rule base of that model is adjusted accordingly, thereby realizing the self-evolution of the evaluation system.

[0039] The human-computer interaction module transmits the comprehensive health index, four-dimensional influence coefficient, abnormal signs, and coil status assessment report to the user information terminal, providing reference data for making adjustment measures.

[0040] Secondly: The accompanying drawings of the embodiments disclosed in this invention only involve the structures involved in the embodiments disclosed in this invention. Other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of this invention can be combined with each other. In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A vibration measurement system for generator end coils, characterized in that, include: The measurement point planning module divides the end coil monitoring area into multiple logical measurement point units based on the three-dimensional model of the generator end structure and electromagnetic field simulation data. Each unit precisely corresponds to the nose, inclined side, and key parts of the straight section of the coil. A hybrid sensing network is formed by a non-contact laser vibration meter and a miniature inertial sensing unit. The signal acquisition module performs noise reduction filtering and trend term elimination processing on the acquired raw vibration signal to improve the signal-to-noise ratio. Then, it applies Fast Fourier Transform to convert the time-domain signal into a frequency-domain energy distribution, accurately separating the characteristic frequency components of the power frequency harmonics and its higher harmonics, and transforming the original waveform data into a set of frequency-domain characteristic parameters with clear physical meaning. The index calculation module converts the frequency domain feature parameter set into the data influence coefficients of the mechanical vibration layer, electromagnetic field layer, thermodynamic layer, and generator operating condition layer through multi-level physical modeling and normalization algorithms, and constructs a quantitative system for assessing health status. The comprehensive analysis module calculates the influence coefficients of mechanical vibration layer data, electromagnetic field layer data, thermodynamic layer data, and generator operating condition layer data through a multi-source influence coefficient fusion algorithm. It establishes a comprehensive health index based on a four-dimensional state space and establishes a quantitative mapping relationship between health status and the four influence coefficients, realizing the transformation from multi-dimensional monitoring to a unified health status. The coil vibration measurement module, based on the comprehensive health index output by the comprehensive analysis module, establishes a three-level early warning mechanism and a fault diagnosis rule base to realize fully automatic assessment and early warning of the end coil vibration status and generate a coil status assessment report. The human-computer interaction module transmits the comprehensive health index, four-dimensional influence coefficient, abnormal signs, and coil status assessment report to the user information terminal to provide reference data.

2. The generator end coil vibration measurement system according to claim 1, characterized in that: The frequency domain feature parameter set includes mechanical vibration layer data, electromagnetic field layer data, thermodynamic layer data, and generator operating condition layer data.

3. The generator end coil vibration measurement system according to claim 2, characterized in that: The mechanical vibration layer data includes axial vibration displacement, denoted as... Radial vibration displacement, denoted as Tangential vibration displacement, denoted as The amplitude of the 100Hz frequency vibration is denoted as... The vibration amplitude at a frequency of 200Hz is denoted as... Total vibration, denoted as RMS; Electromagnetic field layer data includes the 100Hz frequency vibration amplitude, denoted as... Stator A-phase current under 100Hz harmonic vibration, denoted as I; thermodynamic layer data includes thermally induced vibration deviation, denoted as The amplitude of the 100Hz frequency vibration is denoted as ; The generator operating condition data includes active power, denoted as P; cumulative start-stop count, denoted as N; and cumulative runtime, denoted as T.

4. The generator end coil vibration measurement system according to claim 1, characterized in that: The mechanical vibration layer data influence coefficient quantifies the concentration of vibration energy direction by calculating the ratio of the magnitude of the three-dimensional vibration vector to its Euclidean norm. At the same time, it reflects the frequency response anomaly of the mechanical structure by combining the proportion of harmonic vibration energy in the total vibration energy. The arithmetic average of the two realizes the fusion diagnosis of vibration mode and spectrum characteristics, which is used to characterize the degree of deterioration of the overall mechanical fastening state of the end coil.

5. The generator end coil vibration measurement system according to claim 1, characterized in that: The electromagnetic field layer data influence coefficient quantifies the linear deviation of the vibration response under the current electromagnetic excitation by establishing the relative ratio of the 100Hz vibration generated by a unit current to the rated reference value. Its physical essence is to characterize the change in the efficiency of the electromagnetic force acting on the coil, and it is used to diagnose the impedance matching status between the electromagnetic load and the mechanical system.

6. The generator end coil vibration measurement system according to claim 1, characterized in that: The thermodynamic layer data influence coefficient quantifies the relative contribution of non-thermal factors to the vibration state by calculating the absolute ratio of thermally induced vibration deviation to the current 100Hz vibration amplitude, thereby isolating the influence of temperature changes and realizing the extraction of the pure mechanical components of vibration.

7. The generator end coil vibration measurement system according to claim 1, characterized in that: The generator operating condition data influence coefficient is obtained by geometrically averaging the square effect of instantaneous load, fatigue life consumption rate, and aging life consumption rate, and using the cube root to ensure dimensional normalization and balanced coupling of various contributions, so as to comprehensively reflect the normalized influence of operating conditions on the cumulative damage of the coil.

8. The generator end coil vibration measurement system according to claim 1, characterized in that: The comprehensive health index achieves dimensional normalization and state fusion through the geometric mean of four-layer influence coefficients. The numerator represents the degree of loss of mechanical structural integrity, the first denominator quantifies the electromagnetic matching anomaly, the second denominator represents the thermal vibration deviation, and the third denominator reflects the cumulative damage under working conditions. The fourth root of the four ratios ensures balanced coupling of influences in each dimension, and finally, the health status is intuitively quantified by scaling on a percentage scale.

9. A generator end coil vibration measurement system according to claim 1, characterized in that: The coil vibration measurement module specifically includes: S1. Real-time monitoring and classification of health status The system receives the Comprehensive Health Index (HI) value and classifies its status in real time using a preset three-level threshold system: when HI ≥ 85, it is determined to be in the safe operation zone, and the system maintains the normal monitoring frequency; when 70 ≤ HI < 85, it automatically enters the observation zone and starts the enhanced monitoring mode; when HI < 70, it immediately enters the early warning and maintenance zone and triggers the alarm response mechanism of the entire system. S2, Implementation of Tiered Response Strategy In the observation zone, the module automatically increases the data acquisition frequency from the usual 1Hz to 10Hz, and generates a list of operational optimization suggestions, including "It is recommended to limit the load fluctuation range to ±5%" and "Strengthen the cooling system inspection". In the early warning maintenance zone, the module immediately activates the audible and visual alarm device in the control room and initiates the fault diagnosis process by analyzing the combination mode of four influence coefficients: when Km increases significantly and Ke>1.2, it is diagnosed as a mechanical loosening fault; when Kt is abnormally high and Ko>0.8, it is diagnosed as an abnormal thermal stress. S3. Fault Location and Visualization The system calls upon the 3D model database at the generator end to map the health index and influence coefficient of each measuring point to the corresponding spatial location. A dynamic heat map is generated using a gradient coloring algorithm: the distribution of the health index of 85-70-60-50 is represented by a blue-green-yellow-red color transition, and the main direction of abnormal vibration is marked by vector arrows. The system automatically marks the three measuring points with the lowest health index as priority maintenance sites and highlights these key areas in the 3D model. S4. Diagnostic Report Generation and Decision Support Integrating historical data trends over 24 consecutive hours, it automatically generates a structured diagnostic report, including a health index change curve, a pie chart of the contribution of each influencing coefficient, a fault probability analysis matrix, and a specific list of maintenance recommendations. S5, Adaptive Threshold Optimization The module performs monthly statistical analysis of historical operating data and dynamically adjusts the three-level threshold parameters based on the actual operating characteristics of the unit: if the health index remains stable above 90 for three consecutive months, the safe operating zone threshold is automatically raised from 85 to 88; if a certain influence coefficient is found to be continuously deteriorating, its weight coefficient in the fault diagnosis rule base of that model is adjusted accordingly, thereby realizing the self-evolution of the evaluation system.

Images