Health monitoring methods, devices and storage media for heavy-duty gas turbine turning gear

Abstract

This invention relates to the field of gas turbine health monitoring technology, and in particular to a method, device, and storage medium for health monitoring of the turning gear of a heavy-duty gas turbine. The invention synchronously collects multi-dimensional operating status data of the turning gear bearings; identifies operating condition stages based on speed and torque signals, and dynamically adjusts the weighting coefficients of each dimension of data; uses the adjusted weighting coefficients for adaptive fusion processing, combining bearing fault characteristic frequency analysis and torque fluctuation characteristic analysis to generate health status assessment results and remaining life prediction values; and generates maintenance strategy instructions based on the assessment results, prediction values, operating condition urgency, and spare parts inventory information. This invention, through multi-dimensional synchronous perception and adaptive weighting fusion, solves the problems of single perception dimensions and detachment from actual operating conditions in existing technologies, improves fault identification accuracy and prediction accuracy, and realizes closed-loop management from passive alarm to proactive decision-making.

Description

Technical Field

[0001] This invention relates to the field of gas turbine health monitoring technology, and in particular to a method, apparatus, equipment, and computer storage medium for health monitoring of a heavy-duty gas turbine turning gear. Background Technology

[0002] As a core component of electric and aerospace power systems, heavy-duty gas turbines utilize their turning gears to eliminate thermal deformation during start-up, shutdown, and maintenance by rotating the rotor at low speeds, thereby ensuring unit safety. In existing technologies, health monitoring of the turning gears typically employs a basic condition sensing system built through the coordinated operation of vibration sensing, temperature detection, and lubrication monitoring, covering the entire process from data acquisition to threshold alarms.

[0003] However, existing monitoring methods rely on independent analysis of single physical parameters without establishing multi-parameter spatiotemporal coupling models and adaptive operating condition mechanisms. This leads to frequent misjudgments and missed detections under conditions such as cold start preload changes or thermal stress during hot shutdown. Furthermore, the lack of simultaneous monitoring of key related dimensions like lubrication flow and transmission torque makes it impossible to effectively identify complex faults. This fragmented perception and predictive model detached from actual operating conditions result in a lack of actionable strategies for maintenance decisions, low response efficiency, and a high risk of unplanned downtime, severely impacting the operational reliability of large power equipment. Summary of the Invention

[0004] Therefore, the technical problem to be solved by the present invention is to overcome the problems in the prior art, such as low fault identification accuracy, life prediction deviating from actual working conditions, and lack of executable strategies to support operation and maintenance decisions, due to the single and fragmented perception dimension and lack of working condition adaptive mechanism.

[0005] To address the aforementioned technical problems, this invention provides a method for health monitoring of a heavy-duty gas turbine turning gear, comprising:

[0006] Synchronously collect multi-dimensional operating status data of the bearings of the turning gear device. The multi-dimensional operating status data includes at least three-dimensional vibration signals, bearing temperature signals, lubrication circuit oil status signals, and torque and speed coordination signals of the drive shaft. Based on the real-time collected speed and torque signals, the current operating condition stage of the turning gear is identified, and the weight coefficients of each dimension of operating status data in the health assessment are dynamically adjusted according to the identified operating condition stage. The adjusted weighting coefficients are used to adaptively fuse multidimensional operating status data. Combined with bearing fault characteristic frequency analysis and torque fluctuation characteristic analysis, bearing health status assessment results and remaining life prediction values ​​are generated. Based on the bearing health status assessment results, remaining life prediction values, current operating condition urgency, and spare parts inventory information, an operation and maintenance strategy instruction containing the operation and maintenance level and specific operation steps is generated.

[0007] Preferably, the synchronous acquisition of multi-dimensional operating status data of the bearings of the turning gear includes: The three-dimensional vibration signal is simultaneously acquired in the horizontal, vertical and axial directions by a three-dimensional integrated vibration sensor, and the bearing temperature signal is acquired in real time by a temperature sensor. The flow-pressure composite sensor installed at the oil outlet of the lubrication circuit acquires the oil status signal of the lubrication circuit. Instantaneous speed and instantaneous torque are collected synchronously by a torque-speed co-sensor to obtain the torque and speed co-sensor signal of the drive shaft; The collected signals are processed through a multi-channel synchronous sampling mechanism to achieve time alignment, thereby obtaining spatiotemporally correlated multidimensional operational status data.

[0008] Preferably, the synchronous acquisition of triaxial vibration signals via the triaxial integrated vibration sensor further includes: Perform kurtosis calculation on the horizontal vibration signal to generate horizontal kurtosis values ​​for analyzing shock sensitivity; Based on the bearing's preset structural parameters and shaft rotation frequency, the bearing's fault characteristic frequencies are calculated. The fault characteristic frequencies include at least the outer ring fault frequency, inner ring fault frequency, cage fault frequency, and rolling element fault frequency. The original vibration signal is bandpass filtered to select the high-frequency resonance region signal. The envelope signal is obtained by Hilbert transform, and the envelope signal is Fourier transformed to obtain the envelope spectrum. The fault characteristic frequency and its harmonic components are extracted from the envelope spectrum to obtain vibration analysis data containing fault characteristic frequency information.

[0009] Preferably, the acquisition of the torque and speed coordinated signal of the drive shaft via the torque-speed coordinated sensor includes: Collect instantaneous torque and instantaneous angular velocity, and calculate instantaneous power and average torque; The torque fluctuation coefficient, which characterizes the degree of load and friction unevenness, is calculated using the instantaneous torque and the average torque. The covariance and standard deviation of instantaneous torque and instantaneous speed are calculated, and then the torque-speed correlation coefficient is calculated. When the absolute value of the torque-speed correlation coefficient decreases, it is determined that there is a fault or a sign of jamming, thus obtaining analytical data that includes torque fluctuation characteristics and cooperative features.

[0010] Preferably, the current operating condition stage of the turning gear device based on real-time acquired speed and torque signals includes: Call the pre-stored typical operating condition feature library, which includes at least static turning gear start-up operating condition features and dynamic turning gear start-up operating condition features. The real-time collected speed-torque curve is matched with the typical operating condition feature library. When the static turning start operating condition feature is matched, the current stage is determined to be the turning start stage. When the stable speed range feature is matched, the current stage is determined to be the stable operation stage. When the torque decay feature is detected, the current stage is determined to be the turning stop stage, thus obtaining the current operating condition stage identification result.

[0011] Preferably, the step of dynamically adjusting the weighting coefficients of each dimension of operating status data in the health assessment based on the identified operating condition stage includes: When it is determined that the current stage is the start-up phase of the turning gear, the basic weight of the temperature change rate is increased, and the weights of vibration kurtosis and torque are reduced proportionally to obtain the weight coefficient group for the start-up phase. When the current operation is determined to be in a stable operation phase, the basic weight of vibration kurtosis is increased, and the weights of temperature change rate and torque are reduced proportionally to obtain the weight coefficient group for the stable operation phase. When it is determined that the current stage is the turning stop stage, the weights of temperature change rate and vibration kurtosis are both reduced, and the remaining weights are allocated to torque weights to obtain the weight coefficient group for the turning stop stage.

[0012] Preferably, the adaptive fusion processing of multidimensional operating status data using adjusted weighting coefficients includes: The vibration kurtosis value, temperature change rate and torque fluctuation coefficient are obtained as basic evaluation parameters. The corresponding weight coefficients in the weight coefficient group determined according to the operating condition stage are applied to the basic evaluation parameters respectively, and a weighted summation operation is performed to generate a comprehensive alarm coefficient. The comprehensive alarm coefficient is compared with the preset fault monitoring parameter threshold system. Based on the comparison results, combined with the bearing fault characteristic frequency analysis and torque fluctuation characteristic analysis, a bearing health status assessment result indicating the bearing health status level and remaining life prediction value is generated.

[0013] Preferably, the bearing health status assessment results that generate the bearing health status level and remaining life prediction value include: When the comprehensive alarm coefficient triggers the warning threshold and the absolute value of the torque-speed correlation coefficient decreases, the fault is determined to be poor lubrication, bearing wear, or uneven assembly load. When the vibration kurtosis value exceeds the preset threshold and the outer ring fault frequency or its harmonic component appears in the envelope spectrum, the fault is determined to be early pitting or cracking of the rolling element. When the triaxial composite vibration exceeds the preset threshold and the cage failure frequency is significant, the fault is determined to be a comprehensive anomaly, a multi-coupling fault, or a cage anomaly. Based on the identified fault location and current operating time, the predicted remaining life of the bearing is calculated using a model that incorporates historical fault data.

[0014] The present invention also provides a health monitoring device for a heavy-duty gas turbine turning gear, comprising: A multi-dimensional synchronous sensing module is used to synchronously collect multi-dimensional operating status data of the bearing of the turning gear device. The multi-dimensional operating status data includes at least three-dimensional vibration signals, bearing temperature signals, lubrication circuit oil status signals, and torque and speed coordination signals of the drive shaft. The operating condition identification and weight adjustment module is used to identify the current operating condition stage of the turning gear based on the real-time collected speed and torque signals, and dynamically adjust the weight coefficients of each dimension of operating status data in the health assessment according to the identified operating condition stage. The health assessment and life prediction module is used to adaptively fuse multi-dimensional operating status data using adjusted weighting coefficients, and combine bearing failure characteristic frequency analysis and torque fluctuation characteristic analysis to generate bearing health status assessment results and remaining life prediction values. The operation and maintenance strategy generation module is used to generate operation and maintenance strategy instructions containing operation and maintenance levels and specific operation steps based on the bearing health status assessment results, remaining life prediction values, current operating condition urgency, and spare parts inventory information.

[0015] The present invention also provides a health monitoring device for a heavy-duty gas turbine turning gear, comprising: Memory, used to store computer programs; A processor is used to implement the steps of the above-described method for health monitoring of a heavy-duty gas turbine turning gear when executing the computer program.

[0016] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described method for health monitoring of a heavy-duty gas turbine turning gear.

[0017] The technical solution of the present invention has the following advantages compared with the prior art: The health monitoring method for heavy-duty gas turbine turning gears described in this invention overcomes the problem of single and fragmented perception dimensions in existing technologies by simultaneously collecting multi-dimensional operating status data, including three-dimensional vibration signals, bearing temperature signals, lubrication circuit oil status signals, and torque and speed coordination signals of the drive shaft. It also overcomes the problem of lifespan prediction deviating from actual operating conditions due to the lack of an adaptive mechanism in existing technologies by identifying the current operating condition stage based on real-time collected speed and torque signals and dynamically adjusting the weight coefficients of each dimension. Furthermore, it significantly improves fault identification accuracy and prediction accuracy by using the adjusted weight coefficients for adaptive fusion processing and combining bearing fault characteristic frequency analysis and torque fluctuation characteristic analysis to generate health status assessment results and remaining lifespan prediction values. Finally, it generates maintenance strategy instructions containing maintenance levels and specific operating steps based on the health status assessment results, remaining lifespan prediction values, the urgency of the current operating condition, and spare parts inventory information, achieving closed-loop management from passive alarm to proactive decision-making, reducing the risk of unplanned downtime and improving maintenance response efficiency. Attached Figure Description

[0018] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings, wherein: Figure 1 This is a flowchart illustrating the implementation of a health monitoring method for a heavy-duty gas turbine turning gear provided by the present invention. Figure 2 This is a structural block diagram of a health monitoring device for a heavy-duty gas turbine turning gear provided in an embodiment of the present invention. Detailed Implementation

[0019] The core of this invention is to provide a method, device, equipment, and computer storage medium for health monitoring of heavy-duty gas turbine turning gears, which effectively solves the problems in the prior art, such as low fault identification accuracy, life prediction deviating from actual operating conditions, and lack of executable strategies to support operation and maintenance decisions, caused by the single and fragmented perception dimension and the lack of an adaptive mechanism for operating conditions.

[0020] To enable those skilled in the art to better understand the present invention, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely 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.

[0021] Please refer to Figure 1. Figure 1 The flowchart illustrates the implementation of a health monitoring method for a heavy-duty gas turbine turning gear provided by this invention; the specific operation steps are as follows: S101: Synchronously collect multi-dimensional operating status data of the bearing of the turning gear device. The multi-dimensional operating status data includes at least three-dimensional vibration signal, bearing temperature signal, lubrication circuit oil status signal, and torque and speed coordination signal of the transmission shaft. S102: Based on the real-time acquired speed and torque signals, identify the current operating condition stage of the turning gear, and dynamically adjust the weight coefficients of each dimension of operating status data in the health assessment according to the identified operating condition stage. S103: Adaptive fusion processing of multidimensional operating status data is performed using the adjusted weighting coefficients. Combined with bearing fault characteristic frequency analysis and torque fluctuation characteristic analysis, bearing health status assessment results and remaining life prediction values ​​are generated. S104: Based on the bearing health status assessment results, remaining life prediction values, current operating condition urgency, and spare parts inventory information, generate operation and maintenance strategy instructions that include operation and maintenance levels and specific operation steps.

[0022] In some embodiments, the synchronous acquisition of multi-dimensional operating status data of the turning gear bearing in step S101 aims to solve the problem of misjudgment due to coupling anomalies caused by the single and asynchronous sensing dimensions in traditional monitoring. This step constructs a distributed sensing system, deploying multiple types of sensors at key transmission nodes of the turning gear to synchronously acquire a comprehensive set of parameters reflecting the bearing's mechanical characteristics, thermal state, lubrication conditions, and transmission load with millisecond-level time accuracy. The multi-dimensional operating status data logically encompasses triaxial vibration signals characterizing rolling element wear and structural deformation, bearing temperature signals reflecting the actual thermal state of the bearing's inner ring, lubrication circuit oil state signals directly characterizing the working efficiency of the lubrication circuit, and torque and speed coordination signals of the transmission shaft used to identify precursors of transmission jamming. Through a multi-channel synchronous sampling mechanism, strict alignment of the above heterogeneous data in the spatiotemporal dimensions is ensured, thereby providing a highly correlated original data foundation for subsequent multi-parameter fusion analysis. This step, through the synchronous collection of multi-dimensional data, overcomes the limitations of single-parameter monitoring, comprehensively covers all factors affecting bearing life, effectively identifies complex fault modes where single parameters are normal but multiple parameters are abnormally coupled, and provides high-fidelity, strongly correlated data support for subsequent adaptive analysis of operating conditions and accurate life prediction, significantly improving the perception dimension and diagnostic reliability of the health monitoring system.

[0023] In other embodiments, step S102 aims to address the technical problem of traditional health monitoring life prediction models being detached from actual operating conditions. Its core lies in establishing a dynamic mapping mechanism between operating conditions and assessment parameter weights. Specifically, this method first extracts time-domain or frequency-domain features characterizing load changes and motion states based on real-time acquired transmission shaft torque and speed signals. These features are then matched and identified with preset typical operating condition feature patterns to determine the specific operating condition stage of the turning gear. Subsequently, based on the identified operating condition stage, an adaptive adjustment strategy for the weight coefficients is implemented. That is, according to the sensitivity differences of the impact of various dimensions of operating state data on bearing health under different operating conditions, the contribution ratio of each dimension signal in subsequent health assessment is dynamically changed. Through the above technical means, this step can overcome the limitations of the fixed-weight assessment mode, enabling the health assessment model to have operating condition adaptive capabilities, effectively eliminating assessment bias caused by operating condition switching, and significantly improving the accuracy of bearing fault identification and the practicality of remaining life prediction.

[0024] In one specific embodiment, step S103 uses adjusted weighting coefficients to adaptively fuse multidimensional operating status data, aiming to construct a comprehensive health assessment model that can dynamically respond to changes in operating conditions. The core of this step lies in linearly or nonlinearly weighting and aggregating the preprocessed multidimensional physical quantities according to the characteristic weights of the current operating stage, thereby eliminating the limitations of a single parameter under specific operating conditions and forming a comprehensive index characterizing the overall health of the bearing. Based on this, frequency domain analysis and time domain statistical characteristics are further combined to delve deeper into the fault mechanisms behind the data. This step, through adaptive fusion mechanisms and multi-feature joint analysis, effectively solves the assessment bias problem caused by differences in operating conditions in traditional monitoring, significantly improving the accuracy of fault identification and the practicality of life prediction, achieving a technological leap from single-parameter alarms to multidimensional accurate diagnosis.

[0025] Specifically, step S104 aims to address the technical problem of the disconnect between early warning signals and operation and maintenance (O&M) decisions in traditional health monitoring. Its core lies in constructing an intelligent decision-making mechanism based on multi-source information fusion. This mechanism automatically generates executable O&M strategy instructions by comprehensively considering multi-dimensional constraints such as bearing health status assessment results, remaining life prediction values, the urgency of current operating conditions, and spare parts inventory information, using a preset decision logic model. These O&M strategy instructions cover two levels: O&M level determination and specific operation step planning. The O&M level characterizes the urgency of fault handling, while the specific operation steps provide standardized execution guidelines, thereby achieving closed-loop control from status perception to action execution. Through these technical means, this step can transform abstract health assessment data into concrete and implementable O&M action plans, effectively eliminating the lag and uncertainty caused by manual experience-based judgment, significantly improving O&M response efficiency, and realizing a fundamental shift in equipment health management from passive alarm to proactive predictive maintenance.

[0026] It should be noted that the above steps, through multi-dimensional parameter synchronous perception and operating condition adaptive weight fusion, effectively solve the problems of misjudgment in single-parameter monitoring and life prediction deviating from actual operating conditions, significantly improving the accuracy of fault identification and prediction. At the same time, combined with the intelligent operation and maintenance strategy automatic generation mechanism, a closed-loop management from passive alarm to proactive decision-making is realized, reducing the risk of unplanned downtime and improving the efficiency of operation and maintenance response.

[0027] Based on the above embodiments, in some embodiments, the synchronous acquisition of multi-dimensional operating status data of the turning gear bearings includes: The three-dimensional vibration signal is simultaneously acquired in the horizontal, vertical and axial directions by a three-dimensional integrated vibration sensor, and the bearing temperature signal is acquired in real time by a temperature sensor. The flow-pressure composite sensor installed at the oil outlet of the lubrication circuit acquires the oil status signal of the lubrication circuit. Instantaneous speed and instantaneous torque are collected synchronously by a torque-speed co-sensor to obtain the torque and speed co-sensor signal of the drive shaft; The collected signals are processed through a multi-channel synchronous sampling mechanism to achieve time alignment, thereby obtaining spatiotemporally correlated multidimensional operational status data.

[0028] In one specific embodiment, the synchronous acquisition of multi-dimensional operating status data of the turning gear bearing is achieved through four types of dedicated sensors deployed at key locations. First, a three-dimensional integrated vibration sensor serves as the input source, employing a composite fixed structure to synchronously acquire horizontal vibration signals. Vertical vibration signal and axial vibration signal The number of sampling points is The sampling frequency is Secondly, a fiber optic temperature sensor acquires the bearing temperature signal in real time, reflecting the actual temperature of the bearing's inner ring. A micro-flow-pressure composite sensor is installed at the oil outlet of the lubrication circuit to directly monitor the oil flow rate and pressure passing through the bearing to obtain the oil status signal of the lubrication circuit. Furthermore, a modular torque-speed sensor acquires instantaneous torque without altering the drive shaft structure. With instantaneous angular velocity Finally, the above four types of signals are time-aligned through an 8-channel synchronous sampling mechanism to ensure that the time error of each channel signal is less than or equal to 1 millisecond, and finally output spatiotemporally correlated multidimensional operating status data for subsequent working condition identification.

[0029] Specifically, when the triaxial vibration signals are synchronously acquired using the triaxial integrated vibration sensor, the processing actions include performing kurtosis calculation on the horizontal signal to analyze impact sensitivity and generating a horizontal kurtosis value. Its calculation formula is ,in, : Kurtosis value of the horizontal vibration signal, used to characterize the signal's impact sensitivity; N: Number of sampling points. : The horizontal vibration amplitude at the i-th sampling point The arithmetic mean of the horizontal vibration signal; and based on a preset number of rolling elements. Pitch circle diameter , rolling element diameter Contact angle and shaft rotation frequency Calculate the outer ring fault frequency respectively Inner ring failure frequency Cage failure frequency and rolling element failure frequency The assembled torque-speed co-sensor collects instantaneous torque. With instantaneous angular velocity Then, calculate the instantaneous power. And based on the number of sampling points Calculate the average torque ,in, : Instantaneous torque value at the i-th sampling point.

[0030] It should be noted that this implementation method effectively solves the problem of misjudgment caused by single-parameter monitoring by using multi-dimensional sensor collaboration and high-precision time alignment. Combined with specific fault characteristic frequency algorithms and torque fluctuation analysis, it significantly improves the identification accuracy of early-stage minor bearing faults and the reliability of condition assessment.

[0031] Based on the above embodiments, in some embodiments, the synchronous acquisition of triaxial vibration signals by the triaxial integrated vibration sensor further includes: Perform kurtosis calculation on the horizontal vibration signal to generate horizontal kurtosis values ​​for analyzing shock sensitivity; Based on the bearing's preset structural parameters and shaft rotation frequency, the bearing's fault characteristic frequencies are calculated. The fault characteristic frequencies include at least the outer ring fault frequency, inner ring fault frequency, cage fault frequency, and rolling element fault frequency. The original vibration signal is bandpass filtered to select the high-frequency resonance region signal. The envelope signal is obtained by Hilbert transform, and the envelope signal is Fourier transformed to obtain the envelope spectrum. The fault characteristic frequency and its harmonic components are extracted from the envelope spectrum to obtain vibration analysis data containing fault characteristic frequency information.

[0032] In one specific embodiment, bandpass filtering is performed on the original vibration signal to select the high-frequency resonance region signal, and the envelope signal is obtained using Hilbert transform. ,in. Vibration signal after bandpass filtering Imaginary unit, : Filtered signal The Hilbert transform of the envelope signal is applied, and a fast Fourier transform is performed on the envelope signal to obtain the envelope spectrum. The outer ring fault frequency is then located in the spectrum. and inner ring failure frequency It also includes its harmonic components, outputting vibration analysis data containing fault characteristic frequency information. Among them, the outer ring fault frequency... The calculation formula is ,in, : Frequency of bearing outer ring failure Number of rolling elements Shaft rotation frequency : Rolling element diameter, Pitch circle diameter, Contact angle, inner ring failure frequency The calculation formula is cage failure frequency The calculation formula is ,in, : Bearing cage failure frequency Shaft rotation frequency : Rolling element diameter, Pitch circle diameter, Contact angle, rolling element failure frequency The calculation formula is ,in, The number of rolling elements. The diameter of the pitch circle. The diameter of the rolling element, Contact angle, The rotational frequency of the axis.

[0033] Specifically, the kurtosis value of a normal bearing Approximately 3, when peeling or cracking occurs, The value will increase significantly. To extract fault characteristic frequencies, the system performs bandpass filtering on the original signal to select high-frequency resonance signals, and uses Hilbert transform to obtain the envelope signal. Then, a fast Fourier transform is performed on the envelope signal to obtain the envelope spectrum, and the spectrum is searched. And its harmonic components, thereby accurately locating the fault type.

[0034] It should be noted that this implementation method, through specific kurtosis calculation and envelope spectrum analysis, can accurately identify early pitting, cracks and other faults in rolling elements, significantly improving the accuracy of fault identification and early warning capability.

[0035] Based on the above embodiments, in some embodiments, acquiring the torque and speed coordination signal of the drive shaft through a torque-speed coordination sensor includes: Collect instantaneous torque and instantaneous angular velocity, and calculate instantaneous power and average torque; The torque fluctuation coefficient, which characterizes the degree of load and friction unevenness, is calculated using the instantaneous torque and the average torque. The covariance and standard deviation of instantaneous torque and instantaneous speed are calculated, and then the torque-speed correlation coefficient is calculated. When the absolute value of the torque-speed correlation coefficient decreases, it is determined that there is a fault or a sign of jamming, thus obtaining analytical data that includes torque fluctuation characteristics and cooperative features.

[0036] In one specific embodiment, instantaneous torque is collected. With instantaneous angular velocity Calculate instantaneous power Based on the number of sampling points Calculate the average torque ; Utilizing the instantaneous torque With the average torque Calculate the torque ripple coefficient To characterize the degree of load and friction unevenness; to calculate instantaneous torque. With instantaneous speed covariance and their respective standard deviations and Then, the torque-speed co-correlation coefficient was calculated. When the absolute value of the torque-speed correlation coefficient decreases significantly, it is determined that there is a fault or a sign of impending jamming, and analytical data containing torque fluctuation characteristics and correlation features are obtained.

[0037] Specifically, under normal conditions, the absolute value of the torque-speed correlation coefficient. Approaching 1; when signs of malfunction or jamming occur, The damage will decrease significantly. Based on the above vibration and torque analysis results, the system makes judgments according to a preset fault monitoring parameter threshold system, thereby achieving precise identification of various faults such as poor lubrication, bearing wear, assembly misalignment, ball pitting, and cage abnormalities.

[0038] It should be noted that this implementation method, through the joint analysis of torque fluctuation coefficient and co-correlation coefficient, can effectively identify the precursors of bearing jamming and abnormal states of uneven load, providing key transmission characteristic data support for the prediction of remaining life.

[0039] Based on the above embodiments, in some embodiments, the current operating condition stage of the turning gear device based on the real-time acquired speed and torque signals includes: Call the pre-stored typical operating condition feature library, which includes at least static turning gear start-up operating condition features and dynamic turning gear start-up operating condition features. The real-time collected speed-torque curve is matched with the typical operating condition feature library. When the static turning start operating condition feature is matched, the current stage is determined to be the turning start stage. When the stable speed range feature is matched, the current stage is determined to be the stable operation stage. When the torque decay feature is detected, the current stage is determined to be the turning stop stage, thus obtaining the current operating condition stage identification result.

[0040] In one specific embodiment, step S102, which identifies the current operating condition stage of the turning gear based on the real-time acquired speed and torque signals, is specifically achieved by calling a typical operating condition feature library pre-stored in the edge computing acquisition unit's memory. The input source for this typical operating condition feature library is the statistical analysis results of historical operating data. It internally stores at least two types of core data templates: static turning gear start-up operating condition features and dynamic turning gear start-up operating condition features. The static turning gear start-up operating condition feature is defined as a process curve where the speed linearly increases from 0 r / min to 120 r / min, accompanied by a gradual increase in torque. The dynamic turning gear start-up operating condition feature is defined as a process curve where the speed decreases from 180 r / min to 120 r / min. The processing action involves performing sliding window matching calculations on the speed-torque curve collected in real time by the torque-speed co-sensor and the templates in the typical operating condition feature library. When the slope and trend of the real-time curve match the static turning gear start-up operating condition features, the current operating condition stage identification result is output as the turning gear start-up stage; when the real-time speed is maintained at around 120 r / min and the fluctuation is less than the preset threshold, it is determined to be the stable operation stage; when the torque is detected to show a continuous decay characteristic, it is determined to be the turning gear stop stage.

[0041] Specifically, during periods of drastic torque changes, the system increases its sensitivity to vibration anomalies. Simultaneously, the unit sets operating condition-related thresholds for each parameter. Once a parameter is detected to exceed its corresponding threshold, a high-frequency acquisition mode is immediately triggered, increasing the sampling frequency to 10kHz and automatically marking abnormal data segments for subsequent fault tracing analysis.

[0042] It should be noted that this implementation method, by establishing a precise operating condition feature matching mechanism, can accurately identify different stages such as turning gear start-up, stable operation and shutdown, providing a reliable basis for subsequent dynamic weight adjustment.

[0043] Based on the above embodiments, in some embodiments, dynamically adjusting the weighting coefficients of various dimensions of operational status data in health assessment according to the identified operational condition stages includes: When it is determined that the current stage is the start-up phase of the turning gear, the basic weight of the temperature change rate is increased, and the weights of vibration kurtosis and torque are reduced proportionally to obtain the weight coefficient group for the start-up phase. When the current operation is determined to be in a stable operation phase, the basic weight of vibration kurtosis is increased, and the weights of temperature change rate and torque are reduced proportionally to obtain the weight coefficient group for the stable operation phase. When it is determined that the current stage is the turning stop stage, the weights of temperature change rate and vibration kurtosis are both reduced, and the remaining weights are allocated to torque weights to obtain the weight coefficient group for the turning stop stage.

[0044] In one specific embodiment, the weighting coefficients of each dimension of operating status data in the health assessment are dynamically adjusted based on the identified operating condition stage. The input sources are the operating condition stage identification results output from the previous sub-step and preset basic weight parameters, where the basic weight for the temperature change rate is denoted as... The basic weight of vibration kurtosis is denoted as The basic weight of torque is denoted as If the current stage is determined to be the start-up phase of the turning gear, the processing action is to adjust the basic weight of the temperature change rate. Increase by 30% to obtain the temperature weight during the start-up phase. At the same time, according to The increase is in and The original ratio is deducted, that is This ensures the total weight is 1, and the weight coefficient set for the startup phase is output. If the current phase is determined to be stable, the processing action is to adjust the basic weight of the vibration kurtosis. Increase by 40% to obtain the oscillation weight in the stable phase The weights of temperature change rate and torque are adjusted proportionally. If the current condition is determined to be a turning gear stop phase, the action is to reduce both the weight of temperature change rate and vibration kurtosis by 30%, i.e., the weight of temperature change rate during the turning gear stop phase is reduced accordingly. The remaining weights are then allocated to the torque weights to calculate the torque weights for the stopping phase. The final output is a set of weight coefficients for the stopping phase of the rotary locomotive, which is then used for subsequent adaptive fusion processing.

[0045] Specifically, during the turning gear start-up phase, given the significant impact of temperature changes on bearing clearance, the system increases the weight of the temperature change rate; during the stable operation phase, focusing on wear-related faults, the system increases the weight of vibration kurtosis; and during the turning gear stop phase, focusing on monitoring torque decay, the system increases the weight of torque.

[0046] It should be noted that this implementation method solves the problem of evaluation deviation caused by different focus points in different start-up and shutdown stages of traditional monitoring methods by establishing a precise dynamic weight adjustment algorithm, and significantly improves the accuracy and adaptability of bearing health status assessment under cold and hot state switching and load change conditions.

[0047] Based on the above embodiments, in some embodiments, adaptive fusion processing of multidimensional operational status data using adjusted weighting coefficients includes: The vibration kurtosis value, temperature change rate and torque fluctuation coefficient are obtained as basic evaluation parameters. The corresponding weight coefficients in the weight coefficient group determined according to the operating condition stage are applied to the basic evaluation parameters respectively, and a weighted summation operation is performed to generate a comprehensive alarm coefficient. The comprehensive alarm coefficient is compared with the preset fault monitoring parameter threshold system. Based on the comparison results, combined with the bearing fault characteristic frequency analysis and torque fluctuation characteristic analysis, a bearing health status assessment result indicating the bearing health status level and remaining life prediction value is generated.

[0048] In one specific embodiment, the specific implementation process of adaptively fusing multidimensional operating state data using the adjusted weighting coefficients in step S103 is as follows: First, the input source is the preprocessed vibration kurtosis value. Temperature change rate and torque fluctuation coefficient The processing action involves calling the corresponding weighted summation operation based on the operating condition stage identified in step S102. Specifically, if the stage is identified as the turning gear start-up stage, the temperature change rate weight is used. Vibration kurtosis weight and torque weight If it is in a stable operation phase, then vibration kurtosis weighting is used. Temperature change rate weight and torque weight If it is the turning stop phase, then torque weighting is used. Vibration kurtosis weight and the weight of the rate of change of temperature Through the formula Calculate and generate comprehensive alarm coefficient As part of the output, : Rate of temperature change.

[0049] Specifically, the comprehensive alarm coefficient is compared with a preset fault monitoring parameter threshold system. This system defines the torque fluctuation coefficient as normal if it is less than 5%, as a warning if it is 5% to 10%, and as an alarm if it is greater than 10%. Similarly, the vibration velocity RMS is defined as normal if it is less than 2.3 mm / s, as a warning if it is 2.3 to 4.5 mm / s, and as an alarm if it is greater than 4.5 mm / s.

[0050] It should be noted that this implementation method effectively solves the problem of misjudgment caused by single parameter monitoring by dynamically adjusting weights and comparing multiple parameter thresholds. It can accurately identify a variety of specific fault types and significantly improve the accuracy of remaining life prediction and the pertinence of operation and maintenance decisions.

[0051] Based on the above embodiments, in some embodiments, the bearing health status assessment results that indicate the bearing health status level and the predicted remaining life include: When the comprehensive alarm coefficient triggers the warning threshold and the absolute value of the torque-speed correlation coefficient decreases, the fault is determined to be poor lubrication, bearing wear, or uneven assembly load. When the vibration kurtosis value exceeds the preset threshold and the outer ring fault frequency or its harmonic component appears in the envelope spectrum, the fault is determined to be early pitting or cracking of the rolling element. When the triaxial composite vibration exceeds the preset threshold and the cage failure frequency is significant, the fault is determined to be a comprehensive anomaly, a multi-coupling fault, or a cage anomaly. Based on the identified fault location and current operating time, the predicted remaining life of the bearing is calculated using a model that incorporates historical fault data.

[0052] In one specific embodiment, when the comprehensive alarm coefficient triggers the warning threshold and the absolute value of the torque-speed co-correlation coefficient is... During the descent, the fault is determined to be poor lubrication, bearing wear, or uneven assembly load; when the vibration kurtosis value $K$ is greater than 4.0 and an outer ring fault frequency appears in the envelope spectrum after performing an FFT on the envelope signal. When the frequency of the fault is found to be either early pitting or cracking of the ball bearing, the fault is determined to be either early pitting or cracking. When the triaxial combined vibration is greater than 5.0 and the cage fault frequency is... When significant, the fault is determined to be a comprehensive anomaly, a multi-coupled fault, or a cage anomaly. Finally, based on the determined specific fault location and the current operating time, the predicted remaining bearing life is calculated using historical fault data models.

[0053] Specifically, the system makes judgments based on a preset fault monitoring parameter threshold system: for example, when the torque fluctuation coefficient is less than 5%, it is normal; 5% to 10% is a warning; and greater than 10% is an alarm. When the vibration velocity RMS is less than 2.3 mm / s, it is normal; 2.3 to 4.5 mm / s is a warning; and greater than 4.5 mm / s is an alarm. When the bearing temperature is greater than 80℃ or the temperature rise rate is greater than 1.0℃ / min, the corresponding alarm is triggered. This enables precise identification of various faults such as poor lubrication, bearing wear, uneven assembly load, ball pitting, and cage abnormalities.

[0054] It should be noted that this implementation method, through multi-dimensional fault indication determination, can accurately identify specific fault types such as poor lubrication, ball pitting, and cage abnormalities, providing a reliable fault diagnosis basis for the generation of subsequent operation and maintenance strategies.

[0055] Based on the above embodiments, in some embodiments, the specific implementation process of step S104 generating the operation and maintenance strategy instruction is as follows: First, the system receives the bearing health status assessment result and remaining life prediction value from the digital twin processing unit, as well as the current operating condition signal from the gas turbine control system and spare parts inventory data from the enterprise resource planning system as input sources. Specifically, when processing the urgency of the current operating condition, the system compares the current speed and torque curve feature library in real time. If it identifies that the turning gear is in the critical startup stage of speed increasing from 0 to 120 r / min, the operation and maintenance priority flag is raised by one level. Subsequently, the system queries the spare parts inventory database using the bearing model as the index key to obtain the real-time inventory quantity of the corresponding bearing model. Then, the system performs branch judgment according to preset logic: when the remaining life prediction value is lower than the preset emergency threshold or the health status assessment result reaches the alarm state (such as vibration velocity RMS greater than 4.5 mm / s or kurtosis...), the system performs branch judgment. When the remaining life prediction value is greater than 4.0, the action is determined to be the highest risk level, and an instruction with an operation and maintenance level of "emergency replacement" is output. A task order containing specific steps such as immediately cutting off the power source and disengaging the turning gear is automatically generated by calling the shutdown operation procedure template. When the remaining life prediction value is within the planned maintenance range and the number of spare parts in stock is greater than zero, the action is determined to be routine maintenance, and an instruction with an operation and maintenance level of "planned replacement" is output. A task order containing the scheduled maintenance time is automatically generated based on the unit's idle time window. When the health status assessment result is in a warning state (e.g., vibration velocity RMS is in the range of 2.3~4.5mm / s) but has not reached the alarm threshold, the action is determined to be continuous monitoring, and an instruction with an operation and maintenance level of "observation operation" is output. A task order containing encrypted acquisition frequency parameters such as adjusting the edge computing acquisition unit sampling frequency to 10kHz is generated. The above sub-steps are closely linked through a data bus. The query and judgment results of the previous step directly serve as the trigger conditions for generating the task order in the next step, forming a complete decision-making closed loop.

[0056] In another specific embodiment, the method further includes redundant power supply and anti-interference protection steps: when the AC 220V main power supply is interrupted, it immediately switches to DC 24V backup battery power supply to ensure that the device can run continuously for more than 8 hours to avoid data loss; electromagnetic interference is isolated by double-layer shielded cables, and the shielding effectiveness of the processing unit shell is greater than or equal to 40dB by using electromagnetic sealing gaskets; when Ethernet communication is interrupted, it automatically switches to wireless communication mode to ensure that the signal transmission bit error rate is less than or equal to 0.01%, thereby ensuring the stability of multi-dimensional operating status data acquisition and operation and maintenance strategy command generation in an ambient temperature range of -30 degrees Celsius to 80 degrees Celsius.

[0057] It should be noted that this specific implementation deeply couples health assessment data with operational urgency and spare parts inventory, achieving a transformation from single alarms to tiered operation and maintenance decisions. This effectively avoids unplanned downtime caused by decision-making delays and optimizes the scheduling efficiency of spare parts resources. Redundant power supply and anti-interference design ensure stable system operation in harsh environments.

[0058] Please refer to Figure 2 , Figure 2 A structural block diagram of a health monitoring device for a heavy-duty gas turbine turning gear provided in an embodiment of the present invention; the specific device may include: The multi-dimensional synchronous sensing module 100 is used to synchronously collect multi-dimensional operating status data of the bearing of the turning gear device. The multi-dimensional operating status data includes at least three-dimensional vibration signals, bearing temperature signals, lubrication circuit oil status signals, and torque and speed coordination signals of the transmission shaft. The operating condition identification and weight adjustment module 200 is used to identify the current operating condition stage of the turning gear based on the real-time collected speed and torque signals, and dynamically adjust the weight coefficients of each dimension of operating status data in the health assessment according to the identified operating condition stage. The health assessment and life prediction module 300 is used to adaptively fuse multi-dimensional operating status data using adjusted weighting coefficients, and combine bearing fault characteristic frequency analysis and torque fluctuation characteristic analysis to generate bearing health status assessment results and remaining life prediction values. The operation and maintenance strategy generation module 400 is used to generate operation and maintenance strategy instructions containing operation and maintenance levels and specific operation steps based on the bearing health status assessment results, remaining life prediction values, current operating condition urgency and spare parts inventory information.

[0059] The heavy-duty gas turbine turning gear health monitoring device of this embodiment is used to implement the aforementioned heavy-duty gas turbine turning gear health monitoring method. Therefore, the specific implementation of the heavy-duty gas turbine turning gear health monitoring device can be found in the previous embodiment section of the heavy-duty gas turbine turning gear health monitoring method. For example, the multi-dimensional synchronous sensing module 100, the operating condition identification and weight adjustment module 200, the health assessment and life prediction module 300, and the operation and maintenance strategy generation module 400 are respectively used to implement steps S101, S102, S103, and S104 in the above-mentioned heavy-duty gas turbine turning gear health monitoring method. Therefore, its specific implementation can be referred to the description of the corresponding embodiments, which will not be repeated here.

[0060] In one specific embodiment, the system's hardware foundation is first built upon a multi-dimensional synchronous sensing module, deployed at key locations in the turning gear, aiming to overcome the limitations of traditional single-parameter acquisition. Specifically, a three-dimensional integrated vibration sensor is employed, using a composite fixing structure to balance installation stability and detachability, simultaneously acquiring horizontal, vertical, and axial vibration signals to capture high-frequency vibration characteristics caused by bearing ball wear and low-frequency vibration characteristics caused by outer ring deformation. Simultaneously, a fiber optic temperature sensor replaces the traditional PT100 sensor, utilizing its three-fold increase in thermal conductivity to reflect real-time temperature changes in the bearing's inner ring during start-up and shutdown. For lubrication monitoring, a micro-flow-pressure composite sensor is installed at the lubrication circuit outlet, directly monitoring the actual oil state flowing through the bearing, rather than the inlet state. Furthermore, a suite-type torque-speed co-sensor is used, simultaneously acquiring instantaneous speed and torque across the entire speed range without modifying the drive shaft structure, to identify the precursor to bearing jamming: "normal speed but sudden torque increase." The above four types of sensor signals are connected to the edge computing acquisition unit, which has the ability to synchronously sample 8 channels of signals with a time error controlled within 1ms, ensuring the spatiotemporal correlation of parameters such as vibration, temperature, and pressure, and laying the data foundation for subsequent multi-parameter fusion analysis.

[0061] Specifically, after completing synchronous data acquisition, the edge computing acquisition unit performs data preprocessing and operating condition adaptation logic. The unit internally stores a typical operating condition feature library, including static turning gear start-up features (speed increasing from 0 to 120 r / min, torque gradually increasing) and dynamic turning gear start-up features (speed decreasing from 180 r / min to 120 r / min). The system dynamically identifies the current operating condition stage by matching the real-time acquired speed-torque curve with the feature library. Regarding communication and power supply assurance, this embodiment employs redundant power supply and anti-interference design. When the AC220V main power supply is interrupted, the DC24V backup battery pack immediately switches, ensuring continuous operation for at least 8 hours. Strong electromagnetic interference is isolated through double-layer shielded cables and electromagnetic sealing gaskets with a shielding effectiveness of at least 40 dB. It automatically switches to wireless communication mode when the Ethernet connection is interrupted, ensuring a signal transmission bit error rate of less than or equal to 0.01%, and adapting to harsh environmental temperatures ranging from -30℃ to 80℃.

[0062] In one specific embodiment, the digital twin processing unit receives preprocessed multidimensional operational status data and then performs core algorithm analysis and health assessment. First, it processes the three-dimensional vibration signals, assuming the horizontal signal is... The vertical signal is axial signal is The number of sampling points is The sampling frequency is The system calculates the kurtosis of the horizontal signal. To analyze impact sensitivity, the calculation formula is as follows: ,in For the first Vibration amplitude at each sampling point This is the arithmetic mean of the vibration signal. The kurtosis value of a normal bearing. Approximately 3, when peeling or cracking occurs, The value will increase significantly. Furthermore, the system is based on a preset number of rolling elements. Pitch circle diameter , rolling element diameter Contact angle and shaft rotation frequency Calculate the characteristic frequencies of bearing failures.

[0063] Meanwhile, the digital twin processing unit performs in-depth analysis of the torque-speed coordinated signal. The system acquires instantaneous torque. With instantaneous angular velocity Calculate instantaneous power Based on the number of sampling points Calculate the average torque Then calculate the torque ripple coefficient. This characterizes the degree of unevenness in load and friction. The system also calculates instantaneous torque. With instantaneous speed Co-correlation coefficient ,in The covariance of instantaneous torque and instantaneous speed. These are the standard deviations of instantaneous torque and instantaneous speed, respectively. Under normal conditions, Approaching 1; when signs of malfunction or jamming occur, It will decrease significantly.

[0064] It should be noted that the health monitoring device for the heavy-duty gas turbine turning gear in this embodiment achieves closed-loop management from multi-dimensional data perception to intelligent operation and maintenance decision-making through a system architecture of "distributed perception + edge computing + cloud twin". This effectively solves the technical problems of single dimension, detachment from operating conditions and disconnection from decision-making in traditional monitoring, and significantly improves the reliability and maintenance efficiency of the heavy-duty gas turbine turning gear.

[0065] A specific embodiment of the present invention also provides a health monitoring device for a heavy-duty gas turbine turning gear, comprising: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of the above-described health monitoring method for a heavy-duty gas turbine turning gear.

[0066] A specific embodiment of the present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described method for health monitoring of a heavy-duty gas turbine turning gear.

[0067] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0068] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0069] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0070] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0071] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method for health monitoring of a heavy-duty gas turbine turning gear, characterized in that, include: The multi-dimensional operating status data of the bearing of the turning gear is collected synchronously. The multi-dimensional operating status data includes at least three-dimensional vibration signals, bearing temperature signals, lubrication circuit oil status signals, and torque and speed coordination signals of the transmission shaft. Among them, it includes: synchronously collecting three-dimensional vibration signals in the horizontal, vertical and axial directions through a three-dimensional integrated vibration sensor, performing kurtosis calculation on the horizontal vibration signal, and generating a horizontal kurtosis value for analyzing impact sensitivity. Based on the bearing's preset structural parameters and shaft rotation frequency, the bearing's fault characteristic frequencies are calculated. The fault characteristic frequencies include at least the outer ring fault frequency, inner ring fault frequency, cage fault frequency, and rolling element fault frequency. The original vibration signal is bandpass filtered to select the high-frequency resonance region signal. The envelope signal is obtained using Hilbert transform, and the envelope signal is Fourier transformed to obtain the envelope spectrum. The fault characteristic frequency and its harmonic components are extracted from the envelope spectrum to obtain vibration analysis data containing fault characteristic frequency information. The torque and speed signals of the drive shaft are obtained through a torque-speed co-sensor, including: Collect instantaneous torque and instantaneous angular velocity, and calculate instantaneous power and average torque; The torque fluctuation coefficient, which characterizes the degree of load and friction unevenness, is calculated using the instantaneous torque and the average torque. The covariance and standard deviation of instantaneous torque and instantaneous speed are calculated, and then the torque-speed correlation coefficient is calculated. When the absolute value of the torque-speed correlation coefficient decreases, it is determined that there is a fault or a sign of jamming. The analysis data containing torque fluctuation characteristics and coordination features are obtained. Based on the real-time collected speed and torque signals, the current operating condition stage of the turning gear is identified, and the weight coefficients of each dimension of operating status data in the health assessment are dynamically adjusted according to the identified operating condition stage. The adjusted weighting coefficients are used to adaptively fuse multidimensional operating status data. Combined with bearing fault characteristic frequency analysis and torque fluctuation characteristic analysis, bearing health status assessment results and remaining life prediction values ​​are generated. Based on the bearing health status assessment results, remaining life prediction values, current operating condition urgency, and spare parts inventory information, an operation and maintenance strategy instruction containing operation and maintenance levels and specific operation steps is generated. The adaptive fusion processing of multidimensional operational status data using adjusted weighting coefficients includes: The vibration kurtosis value, temperature change rate and torque fluctuation coefficient are obtained as basic evaluation parameters. The corresponding weight coefficients in the weight coefficient group determined according to the operating condition stage are applied to the basic evaluation parameters respectively, and a weighted summation operation is performed to generate a comprehensive alarm coefficient. The comprehensive alarm coefficient is compared with the preset fault monitoring parameter threshold system. Based on the comparison results, combined with the bearing fault characteristic frequency analysis and torque fluctuation characteristic analysis, a bearing health status assessment result indicating the bearing health status level and remaining life prediction value is generated.

2. The method according to claim 1, characterized in that, The synchronously acquired multi-dimensional operating status data of the bearings of the turning gear device includes: The bearing temperature signal is collected in real time using a temperature sensor; The flow-pressure composite sensor installed at the oil outlet of the lubrication circuit acquires the oil status signal of the lubrication circuit. Instantaneous speed and instantaneous torque are collected synchronously by a torque-speed co-sensor to obtain the torque and speed co-sensor signal of the drive shaft; The collected signals are processed through a multi-channel synchronous sampling mechanism to achieve time alignment, thereby obtaining spatiotemporally correlated multidimensional operational status data.

3. The method according to claim 1, characterized in that, The current operating condition stages of the turning gear device based on real-time acquired speed and torque signals include: Call the pre-stored typical operating condition feature library, which includes at least static turning gear start-up operating condition features and dynamic turning gear start-up operating condition features. The real-time collected speed-torque curve is matched with the typical operating condition feature library. When the static turning start operating condition feature is matched, the current stage is determined to be the turning start stage. When the stable speed range feature is matched, the current stage is determined to be the stable operation stage. When the torque decay feature is detected, the current stage is determined to be the turning stop stage, thus obtaining the current operating condition stage identification result.

4. The method according to claim 3, characterized in that, The dynamic adjustment of the weighting coefficients of each dimension of operational status data in the health assessment based on the identified operational condition stages includes: When it is determined that the current stage is the start-up phase of the turning gear, the basic weight of the temperature change rate is increased, and the weights of vibration kurtosis and torque are reduced proportionally to obtain the weight coefficient group for the start-up phase. When the current operation is determined to be in a stable operation phase, the basic weight of vibration kurtosis is increased, and the weights of temperature change rate and torque are reduced proportionally to obtain the weight coefficient group for the stable operation phase. When it is determined that the current stage is the turning stop stage, the weights of temperature change rate and vibration kurtosis are both reduced, and the remaining weights are allocated to torque weights to obtain the weight coefficient group for the turning stop stage.

5. The method according to claim 1, characterized in that, The bearing health status assessment results that generate indicators of bearing health status level and predicted remaining life include: When the comprehensive alarm coefficient triggers the warning threshold and the absolute value of the torque-speed correlation coefficient decreases, the fault is determined to be poor lubrication, bearing wear, or uneven assembly load. When the vibration kurtosis value exceeds the preset threshold and the outer ring fault frequency or its harmonic component appears in the envelope spectrum, the fault is determined to be early pitting or cracking of the rolling element. When the triaxial composite vibration exceeds the preset threshold and the cage failure frequency is significant, the fault is determined to be a comprehensive anomaly, a multi-coupling fault, or a cage anomaly. Based on the identified fault location and current operating time, the predicted remaining life of the bearing is calculated using a model that incorporates historical fault data.

6. A health monitoring device for a heavy-duty gas turbine turning gear, characterized in that, include: A multi-dimensional synchronous sensing module is used to synchronously collect multi-dimensional operating status data of the bearing of the turning gear device. The multi-dimensional operating status data includes at least three-dimensional vibration signals, bearing temperature signals, lubrication circuit oil status signals, and torque and speed coordination signals of the transmission shaft. It includes: synchronously collecting three-dimensional vibration signals in the horizontal, vertical and axial directions through a three-dimensional integrated vibration sensor, performing kurtosis calculation on the horizontal vibration signal, and generating a horizontal kurtosis value for analyzing impact sensitivity. Based on the bearing's preset structural parameters and shaft rotation frequency, the bearing's fault characteristic frequencies are calculated. The fault characteristic frequencies include at least the outer ring fault frequency, inner ring fault frequency, cage fault frequency, and rolling element fault frequency. The original vibration signal is bandpass filtered to select the high-frequency resonance region signal. The envelope signal is obtained using Hilbert transform, and the envelope signal is Fourier transformed to obtain the envelope spectrum. The fault characteristic frequency and its harmonic components are extracted from the envelope spectrum to obtain vibration analysis data containing fault characteristic frequency information. The torque and speed signals of the drive shaft are obtained through a torque-speed co-sensor, including: Collect instantaneous torque and instantaneous angular velocity, and calculate instantaneous power and average torque; The torque fluctuation coefficient, which characterizes the degree of load and friction unevenness, is calculated using the instantaneous torque and the average torque. The covariance and standard deviation of instantaneous torque and instantaneous speed are calculated, and then the torque-speed correlation coefficient is calculated. When the absolute value of the torque-speed correlation coefficient decreases, it is determined that there is a fault or a sign of jamming. The analysis data containing torque fluctuation characteristics and coordination features are obtained. The operating condition identification and weight adjustment module is used to identify the current operating condition stage of the turning gear based on the real-time collected speed and torque signals, and dynamically adjust the weight coefficients of each dimension of operating status data in the health assessment according to the identified operating condition stage. The health assessment and life prediction module is used to adaptively fuse multi-dimensional operating status data using adjusted weighting coefficients, and combine bearing failure characteristic frequency analysis and torque fluctuation characteristic analysis to generate bearing health status assessment results and remaining life prediction values. The operation and maintenance strategy generation module is used to generate operation and maintenance strategy instructions containing operation and maintenance levels and specific operation steps based on the bearing health status assessment results, remaining life prediction values, current operating condition urgency, and spare parts inventory information; The health assessment and life prediction module is used to obtain vibration kurtosis value, temperature change rate and torque fluctuation coefficient as basic assessment parameters, and apply the corresponding weight coefficients in the weight coefficient group determined according to the operating condition stage to the basic assessment parameters, and perform weighted summation to generate a comprehensive alarm coefficient. The comprehensive alarm coefficient is compared with the preset fault monitoring parameter threshold system. Based on the comparison results, combined with the bearing fault characteristic frequency analysis and torque fluctuation characteristic analysis, a bearing health status assessment result indicating the bearing health status level and remaining life prediction value is generated.

7. A health monitoring device for a heavy-duty gas turbine turning gear, characterized in that, include: Memory, used to store computer programs; A processor, configured to execute the computer program to implement the steps of the health monitoring method for a heavy-duty gas turbine turning gear as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the health monitoring method for a heavy-duty gas turbine turning gear as described in any one of claims 1 to 5.

Images