Pumped storage power station unit shafting safety early warning method and system based on digital twinning

Abstract

The invention discloses a digital twinning-based pumped storage power station unit shafting safety early warning method, which comprises the following steps of: S1, acquiring operation data of a pumped storage power station unit shafting, and preprocessing the operation data; s2, based on the preprocessed operation data, establishing a shafting digital twinborn model and a shafting attitude prediction model of the pumped storage power station unit; s3, comparing the difference between the output of the shafting digital twin model and the output of the attitude prediction model, and correcting the digital twin model; and S4, according to the corrected digital twinborn model and the shafting attitude prediction model, realizing safety early warning of the pumped storage power station unit. According to the invention, intelligent and real-time vibration monitoring and safety early warning of the set shaft system of the pumped storage power station can be realized.

Description

Technical Field

[0001] The present invention relates to the technical field related to shaft system safety warning of pumped storage power station units (pumped storage power station units). More specifically, the present invention relates to a shaft system safety warning method and system for pumped storage power station units based on digital twins. Background Art

[0002] With the increase in the proportion of new energy power generation such as wind power and photovoltaic power generation, the safety and stability of the power grid faces new challenges. As an important regulator, the functional positioning, operation mode and operation intensity of pumped storage have also undergone great changes. The frequency of unit start-up and shutdown has doubled, and the units often operate at the boundary of the stable zone. During the frequent start-up and shutdown and operating mode switching process, the shaft system of the pumped storage unit is easily affected by factors such as hydraulic shock, mechanical failure and electromagnetic imbalance, which may induce abnormal vibration, structural fatigue, electrical failure and other problems, seriously threatening the safe operation of the unit.

[0003] Against the backdrop of the rapid development of the digital economy, the deep integration of new technologies such as digital twins with various industries has provided a strong impetus for the digitalization, networking, and intelligent development of the industry. As a new generation of digital technology, digital twins have the potential to promote the development of the new economy. At present, the digital twin model and system of pumped-storage units are still in the initial planning stage. Although the establishment of a digital twin model of the unit is of great value, it is difficult to standardize due to geographical restrictions and system differences in different power stations. At the same time, the operation of pumped-storage units is greatly affected by the flow pattern of water, and the working conditions are complex, which increases the difficulty of modeling.

[0004] The main research difficulties include: there is an urgent need to break through the intelligent perception and monitoring technology of the vibration state of the shaft rotating equipment of the pumped storage unit. The current state monitoring technology cannot meet the needs of intelligent perception and information processing in the big data environment, resulting in a waste of computing and network resources, hindering the rationality and real-time performance of data processing; it is key to build a parameterized shaft rotor dynamics model and accurately describe its time-varying external excitation. The vibration of the pumped storage unit during operation is affected by multiple factors such as hydraulics, mechanics, and electromagnetics. The simulation analysis of a single factor is difficult to accurately describe the shaft movement; the implementation of the application of the digital twin integrated system of the pumped storage unit requires comprehensive and long-term experimental verification of cutting-edge scientific theories to ensure the robustness and stability of the method, thereby achieving comprehensive health management of the hydropower unit. Summary of the invention

[0005] One object of the present invention is to provide a pumped-storage power station unit shaft system safety early warning method and system based on digital twin, so as to realize intelligent and real-time vibration monitoring and safety early warning of the pumped-storage power station unit shaft system.

[0006] In order to achieve these objectives and other advantages of the present invention, according to one aspect of the present invention, the present invention provides a pumped-storage power station unit shaft system safety warning method based on digital twin, including: S1: obtaining the operating data of the pumped-storage power station unit shaft system and performing preprocessing; S2: based on the preprocessed operating data, establishing a pumped-storage power station unit shaft system digital twin model and a shaft system attitude prediction model; S3: comparing the difference between the output of the shaft system digital twin model and the output of the attitude prediction model, and correcting the digital twin model; S4: realizing the pumped-storage power station unit safety warning according to the corrected digital twin model and the shaft system attitude prediction model.

[0007] Furthermore, the formula of the shaft system digital twin model is as follows: Among them, M DT (t) is the output of the digital twin model of the shaft system at time t, w i is the weight of the i-th feature, x i (τ) is the value of the i-th feature after preprocessing at time t, τ is the half-height full width of the Lorentz distribution, λ is the time attenuation coefficient, m is the number of factors affecting the axis attitude, k j is the growth rate of the jth influencing factor, y i (t) is the value of the jth influencing factor at time t, y j0 is the reference value of the jth influencing factor, and n is the number of characteristics of the shaft system operation data.

[0008] Furthermore, the formula of the axis system attitude prediction model is as follows: Among them, M AP (t) is the output of the axis attitude prediction model at time t, N is the size of the historical data window, τ is the half-height full width of the Lorentz distribution, h is the bandwidth of the Gaussian kernel, and M DT (t) is the output of the digital twin model of the shaft system of the pumped storage unit at time t.

[0009] Further, in S3, the method for correcting the digital twin model includes: Acquire the attitude data of each key point of the axis system in real time from the digital twin model, and acquire the predicted attitude data of the corresponding time point from the axis system attitude prediction model; The two sets of data are aligned in time series, a difference analysis is performed, and a parameter space of the axis system digital twin model is established. The minimization of the difference score is used as the objective function to determine the target of parameter optimization; According to the goal of parameter optimization, a heuristic algorithm is used to search global parameters, and a gradient descent iterative optimization algorithm is used to search the local area to obtain the optimal parameters. The digital twin model is calibrated according to the optimal parameters.

[0010] Furthermore, the difference analysis includes the following steps: Calculate the difference between the output of the digital twin model and the output of the axis system attitude prediction model at each time point, determine the rate of change of the attitude data difference by the first-order derivative, and calculate the first-order difference, wherein the attitude data includes gradually changing data and rapidly changing data; For gradient data, position difference and shape difference are used to describe data differences; The fast-changing data are decomposed into trend terms and stationary terms using empirical modes; Based on the first-order difference, different difference evaluation indicators are used for trend terms and stable terms respectively; The difference evaluation index includes an amplitude difference evaluation index and a spectrum difference evaluation index; Various indicators are combined into a comprehensive difference score to reflect the overall degree of deviation.

[0011] Further, in S4, characteristic parameters are extracted from the corrected digital twin model, and corresponding prediction feature sequences are obtained from the shaft system attitude prediction model, wherein the characteristic parameters include amplitude, displacement and stress of each node of the shaft system; An early warning threshold is set for the characteristic parameter, and the real-time monitoring data is compared with the predicted data to determine whether it exceeds the early warning threshold.

[0012] Furthermore, the warning threshold includes a first-level warning threshold and a second-level warning threshold; When the bearing vibration amplitude is greater than the first-level warning threshold, an alarm message is generated and the bearing position is marked; If the operations staff does not respond within the specified time, the alarm is escalated and a notification is sent to higher-level management; When the vibration amplitude of the same bearing continues to rise and exceeds the second-level warning threshold, the fault mode library is called to compare the current vibration characteristics with typical fault modes to generate preliminary diagnostic opinions.

[0013] Furthermore, the formula for the first-level warning threshold is as follows: Among them, S 1 is the first-level warning threshold, T is the evaluation time window, κ 1 is the adjustment parameter of the first-level warning threshold, M DT (t) is the output of the digital twin model of the shaft system of the pumped storage unit at time t, MAP (t) is the output of the axis system attitude prediction model at time t; The specific formula for the second-level warning threshold is as follows: Among them, S 2 is the second-level warning threshold, T is the evaluation time window, κ 2 is the adjustment parameter of the second-level warning threshold, M DT (t) is the output of the digital twin model of the shaft system of the pumped storage unit at time t, M AP (t) is the output of the axis system attitude prediction model at time t.

[0014] According to another aspect of the present invention, a pumped-storage power station unit shaft system safety warning system based on digital twin is also provided, including: a collection module, used to obtain the operating data of the pumped-storage power station unit shaft system and perform preprocessing; an establishment module, used to establish a pumped-storage power station unit shaft system digital twin model and a shaft system attitude prediction model based on the preprocessed operating data; a correction module, used to compare the difference between the output of the shaft system digital twin model and the output of the attitude prediction model, and correct the digital twin model; an early warning module, used to implement pumped-storage power station unit safety early warning based on the corrected digital twin model and the shaft system attitude prediction model.

[0015] According to another aspect of the present invention, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, any of the digital twin pumped-storage power station unit shaft system safety warning methods is implemented.

[0016] The present invention has at least the following beneficial effects: The present invention adopts a multi-physics field coupled finite element full-order model, considers mechanical vibration, incorporates the influence of hydraulic and electromagnetic forces, improves the accuracy of the model, and reveals the typical fault mechanism of hydraulic imbalance; introduces Fourier transform for time-frequency analysis, extracts time domain and frequency domain features, captures the dynamic characteristics of the fault, and realizes early and accurate fault type identification; combines the real-time correction of the digital twin model of the shaft system of the pumped-storage unit and the shaft system attitude prediction model, sets a dynamic early warning threshold, improves the warning sensitivity, reduces false alarms and shutdown losses through a hierarchical response mechanism, and realizes safety early warning of the pumped-storage power station unit.

[0017] Other advantages, objectives and features of the present invention will be embodied in part through the following description, and in part will be understood by those skilled in the art through study and practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 This is a flow chart of the shaft system safety early warning method of the pumped-storage power station unit based on the digital twin of Example 1; Figure 2 The flowchart of the digital twin model of the shaft system of the pumped-storage power station unit of the digital twin pumped-storage power station unit shaft system safety early warning method of Example 1 is constructed; Figure 3 This is a correction flow chart of the shaft system safety warning method of the digital twin pumped-storage power station unit in Example 1. DETAILED DESCRIPTION

[0019] The present invention is further described in detail below in conjunction with the accompanying drawings so that those skilled in the art can implement the invention with reference to the description.

[0020] It should be understood that the terms such as "having", "including" and "comprising" used in the embodiments of the present application do not exclude the existence or addition of one or more other elements or their combinations. All directional indications (such as up, down, left, right, front, back...) in the embodiments of the present application are only used to explain the relative position relationship, movement, etc. between the components in a certain specific posture. If the specific posture changes, the directional indication will also change accordingly. When an element is referred to as "fixed on" or "set on" another element, it can be directly on the other element or there may be a centering element at the same time. When an element is referred to as "connecting" another element, it can be directly connected to another element or it can be indirectly connected to another element through a centering element. The description of "first", "second", etc. in the embodiments of the present application is only for descriptive purposes, and cannot be understood as indicating or implying its relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" can explicitly or implicitly include at least one of the features.

[0021] It should be noted that the technical solutions between the various embodiments of the present application can be combined with each other, but it must be based on the fact that ordinary technicians in the field can implement it. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0022] Embodiment 1: Reference Figure 1-3 , which is the first embodiment of the present application, and provides a shaft system safety early warning method for a pumped storage power station unit based on digital twins, including: S1: Collect the operating data of the shaft system of the pumped storage unit and perform preprocessing.

[0023] Specifically, sensors are installed at key positions of the shaft system of the pumped storage unit, such as the main bearing, guide bearing and generator bearing; three-axis accelerometers and displacement sensors are required at the bearing positions, where the accelerometer measures radial and axial vibrations, and the displacement sensor measures shaft displacement; T-type thermocouple temperature sensors are installed at the shaft system, rotor and stator, and a magnetoelectric tachometer speed sensor is installed at the shaft system position.

[0024] Furthermore, a data acquisition system is set up, using a high-speed data acquisition card with a sampling rate of at least twice the highest frequency of interest, and a GPS clock is used to synchronize the clocks of all sensors and acquisition devices to ensure that the timestamps of all data points are accurate to the millisecond level.

[0025] Furthermore, the collected operating data is preprocessed: data beyond the physical range is marked as abnormal, and statistical methods are used to detect outliers. Short-term missing data can be filled by linear interpolation, and long-term missing data can be filled by multivariate interpolation. A low-pass filter is applied to the vibration data to remove high-frequency noise, and a wavelet transform is used to perform multi-scale decomposition to remove noise in different frequency bands. Moving average or exponential smoothing can be used to reduce measurement noise for temperature data.

[0026] S2: Based on the preprocessed operating data, a digital twin model of the shaft system and a shaft system attitude prediction model of the pumped storage unit are established.

[0027] Specifically, the construction process of the digital twin model of the shaft system of the pumped storage unit includes the following steps: based on the finite element analysis method, a multi-physics field coupling finite element full-order model of the shaft system of the pumped storage unit is constructed; typical fault conditions are simulated according to the multi-physics field coupling finite element full-order model, sample data and classification results are obtained, the dynamic response under typical fault conditions is solved, and the vibration factors of the shaft system are analyzed and stored; by reducing the order of the multi-physics field coupling finite element full-order model, the time domain characteristics and frequency domain characteristics of the shaft system vibration response under the multi-physics field coupling of hydraulic and electromagnetic forces are obtained, and the calculation errors and time of each reduced-order model are compared; based on the time domain characteristics and frequency domain characteristics, the fault mechanism of the pumped storage unit is analyzed, and the parameters under the fault mechanism condition of the pumped storage unit are calculated; through each reduced-order model, the accuracy and timeliness of the reduced-order model are analyzed and verified, and a digital twin model of the shaft system of the pumped storage unit for real-time simulation of typical conditions is constructed. The specific formula is as follows: Among them, M DT (t) is the output of the digital twin model of the shaft system at time t, w i is the weight of the i-th feature, x i (τ) is the value of the i-th feature after preprocessing at time t, τ is the half-height full width of the Lorentz distribution, λ is the time attenuation coefficient, m is the number of factors affecting the axis attitude, k j is the growth rate of the jth influencing factor, y i (t) is the value of the jth influencing factor at time t, y j0 is the reference value of the jth influencing factor, and n is the number of characteristics of the shaft system operation data.

[0028] It should be noted that the digital twin model of the shaft system of the pumped storage unit includes large main shaft bending, large turning data, non-vertical main shaft and mirror plate, loose thrust head, non-horizontal thrust bearing, non-concentric bearing, bearing wear and rotation center offset. The formula of the digital twin model of the shaft system of the pumped storage unit integrates physical insights and data-driven, describes the attenuation effect of various characteristics of the shaft system over time based on the Lorentz distribution, and captures the time-varying characteristics and frequency distribution of the vibration amplitude; introduces the logarithmic growth model of the influencing factors of the shaft system attitude to reflect the nonlinear influence of hydraulic and electromagnetic forces on the dynamics of the shaft system; the optimization process of weights and parameters realizes the model's adaptation to actual working conditions, and the selection of characteristics and influencing factors reflects the understanding of complex fault mechanisms, providing real-time and accurate prediction of the shaft system status, revealing potential faults through its structure and parameter changes, and providing a solid foundation for predictive maintenance and safety warnings.

[0029] Furthermore, the relevant data of the shaft system attitude are obtained through the monitoring system of the pumped storage unit and preprocessed; the relevant data include vibration, temperature and speed; the preprocessed relevant data are segmented according to the time series alignment, and the time-frequency analysis of Fourier transform is performed to extract the time domain features and frequency domain features; the features of the non-stationary signal are extracted through the wavelet analysis algorithm, and the correlation analysis dimensionality reduction technology is used to screen the key features and construct a feature matrix containing the key features as the input data of the shaft system attitude prediction model; Furthermore, the shaft system attitude prediction model is lightweight; the feature matrix is ​​divided into training set, validation set and test set in proportion, the shaft system attitude prediction model is trained with the training set, and the parameters are adjusted; the shaft system attitude prediction model is evaluated with the validation set, and the generalization ability of the shaft system attitude prediction model is evaluated with the test set; the prediction results of the trained shaft system attitude prediction model on the training data are used as soft labels through the knowledge distillation strategy, and the evaluation results of the test set are used as hard labels, and the shaft system attitude prediction model is trained by combining hard labels and soft labels. The specific formula is as follows: Among them, M AP (t) is the output of the axis attitude prediction model at time t, N is the size of the historical data window, τ is the half-height full width of the Lorentz distribution, h is the bandwidth of the Gaussian kernel, and M DT (t) is the output of the digital twin model of the shaft system of the pumped storage unit at time t.

[0030] Specifically, the digital twin model of the shaft system of the pumped storage unit is corrected in real time according to the shaft system attitude prediction model.

[0031] It should be noted that the shaft system attitude prediction model formula integrates time series analysis and machine learning. The first part of the Lorentz distribution is based on capturing the changing trend of the shaft system attitude over time, which is particularly suitable for describing the attenuation of vibration amplitude. The Gaussian kernel is introduced to quantify the similarity between the output of the shaft system digital twin model of the pumped-storage unit and the output of the shaft system attitude prediction model, so as to realize the model's memory of historical data and attention to abnormal conditions; the historical window size, time influence range and similarity sensitivity are controlled respectively, and the model is optimized to adapt to different working conditions, providing accurate short-term attitude prediction, and continuously correcting and improving the performance of the entire early warning system through interaction with the shaft system digital twin model of the pumped-storage unit.

[0032] S3: By comparing the differences between the output of the digital twin model of the shaft system of the pumped-storage unit and the output of the posture prediction model, the model parameters are optimized and the digital twin model of the shaft system of the pumped-storage unit is corrected in real time.

[0033] Specifically, the correction includes the following steps: acquiring the attitude data of each key point of the shaft system from the digital twin model of the shaft system of the pumped storage unit in real time, and acquiring the predicted attitude data of the corresponding time point from the shaft system attitude prediction model; the attitude data includes displacement, velocity and acceleration; aligning the two sets of data in time series, performing difference analysis, and establishing the parameter space of the digital twin model of the shaft system of the pumped storage unit, taking minimizing the difference score as the objective function, and determining the target of parameter optimization; according to the target of parameter optimization, using a heuristic algorithm to perform global parameter search, and using a gradient descent iterative optimization algorithm to perform local search in the local area to obtain the optimal parameters; according to the optimal parameters, correcting the digital twin model of the shaft system of the pumped storage unit.

[0034] Furthermore, the parameter space includes adjustable parameters and the range of adjustable parameter values; the difference analysis includes the following steps: calculating the difference between the digital twin model output and the prediction model output at each time point, and determining the rate of change of the attitude data difference with the first-order derivative, and calculating the first-order difference; the attitude data includes gradual data and fast-changing data; the position difference and shape difference are used to characterize the data difference for the gradual data; the empirical mode is used to decompose the fast-changing data into trend terms and stable terms; based on the first-order difference, different difference evaluation indicators are used for the trend term and the stable term respectively; the difference evaluation indicators include amplitude difference evaluation indicators and spectrum difference evaluation indicators; various indicators are integrated into a comprehensive difference score to reflect the overall deviation degree and realize safety early warning of pumped storage power station units.

[0035] Furthermore, characteristic parameters are extracted from the corrected digital twin model, and the corresponding prediction feature sequence is obtained from the axis system attitude prediction model, and the time history and prediction trend of the characteristic parameters are displayed through visualization; the characteristic parameters include the amplitude, displacement and stress of each node of the axis system; according to the dynamic changes of the characteristic parameters, early warning thresholds are set, and the real-time monitoring data and the predicted data are compared to determine whether the warning thresholds are exceeded.

[0036] S4: Realize safety warning of pumped-storage power station units based on the calibrated digital twin model of the pumped-storage unit shaft system and the shaft system posture prediction model.

[0037] Specifically, the warning thresholds include the first-level warning threshold and the second-level warning threshold. When the bearing vibration amplitude is greater than the first-level warning threshold, an alarm message is generated and the bearing position is marked in a yellow area on the monitoring interface. If the operation and maintenance personnel do not respond within the specified time, the alarm will be upgraded and a notification will be sent to higher-level managers.

[0038] Furthermore, the specific formula for the first-level warning threshold is as follows: Among them, S 1 is the first-level warning threshold, T is the evaluation time window, κ 1 is the adjustment parameter of the first-level warning threshold, M DT (t) is the output of the digital twin model of the shaft system of the pumped storage unit at time t, M AP (t) is the output of the axis system attitude prediction model at time t; Furthermore, when the vibration amplitude of the same bearing continues to rise and is greater than the second-level warning threshold, the fault mode library is called to compare the current vibration characteristics and typical fault modes to generate preliminary diagnostic opinions; if the diagnosis result is main shaft misalignment, the main shaft load is reduced and closely monitored; if an abnormal peak appears in the vibration spectrum of the generator bearing, this frequency component is analyzed to determine the type of fault; if the shaft system vibration phase angle changes suddenly, the hydraulic parameters are abnormal and the guide vanes or impellers are checked; hydraulic parameters include head and flow; when periodic impact components appear in the shaft system vibration waveform, there is a bearing fault or gear fault; if the impact frequency matches the characteristic frequency of the bearing or gear, there is a defect in the inner ring or rolling element of the output bearing.

[0039] Specifically, the specific formula for the second-level warning threshold is as follows: Among them, S 2 is the second-level warning threshold, T is the evaluation time window, κ 2 is the adjustment parameter of the second-level warning threshold, M DT (t) is the output of the digital twin model of the shaft system of the pumped storage unit at time t, M AP(t) is the output of the axis system attitude prediction model at time t.

[0040] Furthermore, the present embodiment also provides a pumped-storage power station unit shaft system safety early warning system based on digital twin, including: a collection module, for collecting and preprocessing the operation data of the pumped-storage unit shaft system; An establishment module is used to establish a digital twin model of the shaft system of the pumped-storage unit and a shaft system attitude prediction model based on the preprocessed operating data; a correction module is used to optimize the model parameters and correct the digital twin model of the shaft system of the pumped-storage unit in real time by comparing the differences between the output of the digital twin model of the shaft system of the pumped-storage unit and the output of the attitude prediction model; an early warning module is used to realize safety early warning of the pumped-storage power station unit according to the corrected digital twin model of the shaft system of the pumped-storage unit and the shaft system attitude prediction model.

[0041] This embodiment also provides a computer device, which is suitable for the case of a digital twin pumped-storage power station unit shaft system safety early warning method, including a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the digital twin pumped-storage power station unit shaft system safety early warning method proposed in the above embodiment.

[0042] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covered on the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0043] In summary, the present invention adopts a multi-physics field coupled finite element full-order model, considers mechanical vibration, incorporates the influence of hydraulic and electromagnetic forces, improves the accuracy of the model, and reveals complex fault mechanisms such as hydraulic imbalance; introduces Fourier transform for time-frequency analysis, extracts time domain and frequency domain features, captures the dynamic characteristics of the fault, and realizes early and accurate fault type identification; combines the real-time correction of the digital twin model of the shaft system of the pumped-storage unit and the shaft system attitude prediction model, sets a dynamic warning threshold, improves the warning sensitivity, reduces false alarms and shutdown losses through a hierarchical response mechanism, and realizes safety warning of the pumped-storage power station unit.

[0044] Embodiment 2: Referring to Table 1, which is the second embodiment of the present application, this embodiment provides a shaft system safety warning method for a pumped-storage power station unit based on digital twins. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0045] Specifically, the present invention was tested on site for one month in a large pumped storage power station. The power station has a total of 6 pumped storage units, with a single unit capacity of 300MW. Before the test, engineers collected the operating data of the 6 units over the past year, including shaft vibration, bearing temperature, axial displacement and shaft power curve. At the same time, additional sensors were installed to collect environmental factor data, such as unit room temperature, humidity and cooling water flow.

[0046] Furthermore, the collected historical data is preprocessed. This includes data denoising, missing value interpolation and feature extraction. For the shaft vibration data, wavelet transform is used for denoising to effectively remove the influence of high-frequency noise. Some missing values ​​in the bearing temperature data are interpolated through Gaussian process regression to ensure the continuity of the data. For the axial displacement data, statistical features such as mean, variance and peak are extracted to capture the key characteristics of the shaft motion. The preprocessed data provides high-quality input for model building.

[0047] Furthermore, based on the preprocessed data, shaft system digital twin models M were established for the six units. DT (t) and the axis attitude prediction model M AP (t). In the process of model construction, it is found that traditional methods often only consider a single feature, such as the root mean square value of shaft system vibration. The model proposed in this invention can comprehensively consider the dynamic influence of multiple characteristics and environmental factors, and capture the time-varying nature of characteristic influences and the spectral characteristics of power fluctuations through parameters such as the time attenuation coefficient λ and the half-height full width of the Lorentz distribution τ. This makes the model's description of the shaft system state more accurate and comprehensive.

[0048] Specifically, in the model training phase, the objective function minimization method is used to optimize the model parameters. After the model is established, it enters the one-month real-time monitoring and early warning phase. Every 5 minutes, the system automatically collects real-time data and inputs it into the shaft system digital twin model M. DT (t) and the axis attitude prediction model M AP (t). By comparing the output differences between the two models, the shaft system digital twin model M is corrected in real time. DT (t). For example, when Unit 1 experiences a slight bearing temperature rise on the 15th day, M DT (t) quickly capture this change and AP (t), accurately predicting slight anomalies in the axis system attitude.

[0049] Furthermore, to quantify the performance of the model, the safety warning index is defined, and two-level warning thresholds and are set. During the test, the S values ​​of the six units and the warning conditions of the fixed threshold method based on the RMS value of the shaft system vibration were recorded.

[0050] Furthermore, as shown in Table 1, on the 15th day of Unit 1, the conventional method failed to capture the shaft system abnormality because the shaft system vibration (1.2 mm / s) did not exceed the fixed threshold (usually set at 1.8-2.0 mm / s). However, the safety warning index S of the present invention reached 0.42, exceeding the first-level warning threshold S 1 (0.35), successfully giving early warning. The present invention not only takes into account the shaft vibration, but also comprehensively considers factors such as the increase in bearing temperature (72°C) and the increase in power fluctuation (±8MW).

[0051] Specifically, on the 20th day of Unit 2, the traditional method triggered an early warning because the shaft vibration (1.8mm / s) exceeded the threshold. But in fact, the unit was still in good operating condition at this time, and the power fluctuation (±6MW) was within the normal range. The safety early warning index S of the present invention is 0.38, which is lower than the second-level early warning threshold S 2 (0.65), avoiding unnecessary downtime inspections. This shows that the present invention effectively reduces the false alarm rate by comprehensively evaluating multiple parameters.

[0052] Furthermore, on the 18th day, the safety warning index S of Unit 5 reached 0.51, triggering the second-level warning. Although the shaft vibration (1.3mm / s) is far below the traditional threshold, the present invention captures the abnormal growth trend of bearing temperature (73°C) and power fluctuation (±9MW). In fact, in subsequent inspections, it was found that the bearing oil film was slightly damaged, and timely preventive maintenance avoided more serious failures. This highlights the advantages of the present invention in early fault prevention.

[0053] Table 1 Experimental data table Furthermore, by comparing the data of Unit 3 on the 5th and 30th days, it can be found that although the shaft system parameters fluctuated slightly (such as the shaft system vibration increased from 0.9mm / s to 1.0mm / s), the safety warning index S did not change much (0.15 to 0.32). Unit 6 operated smoothly throughout the test period, and the safety warning index remained at a low level (such as 0.08 on the 8th day). This shows that the present invention not only performs well under abnormal conditions, but also maintains stability during normal operation, and will not frequently trigger warnings due to small fluctuations, which is conducive to the continuous and stable operation of the unit.

[0054] Although the embodiments of the present invention have been disclosed as above, they are not limited to the applications listed in the specification and the implementation modes, and they can be fully applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily implemented. Therefore, without departing from the general concept defined by the claims and the scope of equivalents, the present invention is not limited to the specific details and the illustrations shown and described herein.

Claims

1. A shaft system safety early warning method for pumped storage power station units based on digital twins, characterized in that: include: S1: Obtain the operating data of the shaft system of the pumped storage power station unit and perform preprocessing; S2: Based on the pre-processed operating data, a shaft system digital twin model and a shaft system attitude prediction model of the pumped storage power station unit are established; S3: comparing the difference between the output of the axis system digital twin model and the output of the attitude prediction model, and correcting the digital twin model; S4: Implementing safety warning of pumped-storage power station units based on the calibrated digital twin model and the shaft system posture prediction model.

2. The shaft system safety early warning method for pumped storage power station units based on digital twins according to claim 1 is characterized in that: The formula of the shaft system digital twin model is as follows: Among them, M DT (t) is the output of the digital twin model of the shaft system at time t, w i is the weight of the i-th feature, x i (τ) is the value of the i-th feature after preprocessing at time t, τ is the half-height full width of the Lorentz distribution, λ is the time attenuation coefficient, m is the number of factors affecting the axis attitude, k j is the growth rate of the jth influencing factor, y i (t) is the value of the jth influencing factor at time t, y j0 is the reference value of the jth influencing factor, and n is the number of characteristics of the shaft system operation data.

3. The shaft system safety early warning method for pumped storage power station units based on digital twins according to claim 1 is characterized in that: The formula of the axis system attitude prediction model is as follows: Among them, M AP (t) is the output of the axis attitude prediction model at time t, N is the size of the historical data window, τ is the half-height full width of the Lorentz distribution, h is the bandwidth of the Gaussian kernel, and M DT (t) is the output of the digital twin model of the shaft system of the pumped storage unit at time t.

4. The shaft system safety early warning method for pumped storage power station units based on digital twins according to claim 1 is characterized in that: In S3, the method for correcting the digital twin model includes: Acquire the attitude data of each key point of the axis system in real time from the digital twin model, and acquire the predicted attitude data of the corresponding time point from the axis system attitude prediction model; The two sets of data are aligned in time series, a difference analysis is performed, and a parameter space of the axis system digital twin model is established. The minimization of the difference score is used as the objective function to determine the target of parameter optimization; According to the goal of parameter optimization, a heuristic algorithm is used to search global parameters, and a gradient descent iterative optimization algorithm is used to search the local area to obtain the optimal parameters. The digital twin model is calibrated according to the optimal parameters.

5. The shaft system safety early warning method for pumped storage power station units based on digital twins according to claim 4 is characterized in that: The difference analysis includes the following steps: Calculate the difference between the output of the digital twin model and the output of the axis system attitude prediction model at each time point, determine the rate of change of the attitude data difference by the first-order derivative, and calculate the first-order difference, wherein the attitude data includes gradually changing data and rapidly changing data; For gradient data, position difference and shape difference are used to describe data differences; The fast-changing data are decomposed into trend terms and stationary terms using empirical modes; Based on the first-order difference, different difference evaluation indicators are used for trend terms and stable terms respectively; The difference evaluation index includes an amplitude difference evaluation index and a spectrum difference evaluation index; Various indicators are combined into a comprehensive difference score to reflect the overall degree of deviation.

6. The shaft system safety early warning method for pumped storage power station units based on digital twins according to claim 1 is characterized in that: In S4, characteristic parameters are extracted from the corrected digital twin model, and corresponding prediction feature sequences are obtained from the shaft system attitude prediction model, wherein the characteristic parameters include amplitude, displacement and stress of each node of the shaft system; An early warning threshold is set for the characteristic parameter, and the real-time monitoring data is compared with the predicted data to determine whether it exceeds the early warning threshold.

7. The shaft system safety early warning method for pumped storage power station units based on digital twins according to claim 6 is characterized in that: The warning threshold includes a first-level warning threshold and a second-level warning threshold; When the bearing vibration amplitude is greater than the first-level warning threshold, an alarm message is generated and the bearing position is marked; If the operations staff does not respond within the specified time, the alarm is escalated and a notification is sent to higher-level management; When the vibration amplitude of the same bearing continues to rise and exceeds the second-level warning threshold, the fault mode library is called to compare the current vibration characteristics with typical fault modes to generate preliminary diagnostic opinions.

8. The digital twin pumped storage power station unit shaft system safety early warning method according to claim 7, characterized in that: The formula for the first-level warning threshold is as follows: Among them, S1 is the first-level warning threshold, T is the evaluation time window, κ1 is the adjustment parameter of the first-level warning threshold, M DT (t) is the output of the digital twin model of the shaft system of the pumped storage unit at time t, M AP (t) is the output of the axis system attitude prediction model at time t; The specific formula for the second-level warning threshold is as follows: Among them, S2 is the second-level warning threshold, T is the evaluation time window, κ2 is the adjustment parameter of the second-level warning threshold, M DT (t) is the output of the digital twin model of the shaft system of the pumped storage unit at time t, M AP (t) is the output of the axis system attitude prediction model at time t.

9. The shaft system safety early warning system of pumped storage power station units based on digital twin is characterized by: include: The collection module is used to obtain the operating data of the shaft system of the pumped storage power station unit and perform preprocessing; Establishing a module for establishing a shaft system digital twin model and a shaft system attitude prediction model of a pumped storage power station unit based on the preprocessed operating data; A correction module, used for comparing the difference between the output of the axis system digital twin model and the output of the attitude prediction model, and correcting the digital twin model; The early warning module is used to realize safety early warning of the pumped-storage power station unit according to the corrected digital twin model and the shaft system posture prediction model.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that When the processor executes the computer program, it implements the digital twin pumped storage power station unit shaft system safety warning method as described in any one of claims 1-8.

Images