Pumped storage group top cover state monitoring and early warning alarm system

Abstract

The invention relates to a pumped storage group top cover state monitoring and early warning alarm system, which relates to the technical field of fault monitoring and comprises a data acquisition unit, a working condition identification unit, a digital twin evaluation engine, a virtual-real mapping self-checking unit and an intelligent early warning unit. According to the method, space mapping comparison is carried out by utilizing a theoretical reference generated by digital twinning and an actually measured data set, zero drift, signal failure or installation looseness of the sensor can be accurately identified, the accuracy of an early warning instruction is ensured, point-shaped sensing data is converted into a planar stress distribution cloud picture through a three-dimensional finite element model, and the accuracy of the early warning instruction is ensured. Local stress concentration of a non-measuring-point area can be effectively captured, dynamic safety threshold values are automatically switched according to different working conditions such as water pumping, power generation and transient transition, and the problem of traditional static alarm normalization false alarm caused by sudden load change in the transition processes such as start-stop and load shedding of a unit is effectively solved. And scientific data support is provided for refined operation and maintenance and life prediction of the unit.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fault monitoring, in particular to a pumped storage unit top cover state monitoring and early warning alarm system. BACKGROUND

[0002] The water turbine of a pumped storage power station is one of the core devices of this type of power station, mainly used for regulating the load of the power grid, playing a role in balancing power demand. The pumped storage power station usually includes two reservoirs: an upper reservoir and a lower reservoir, and through the bidirectional operation of the water turbine and the generator set, the storage and release of energy are realized;

[0003] The top cover of the pumped storage power station is one of the core flow components of the unit, which has long been subjected to huge water pressure, complex radial and axial loads, and unit vibration. In the prior art, the pumped storage power station water turbine fixes the unit top cover through the top cover bolts to realize accurate monitoring of the top cover state, which is crucial for preventing bolt fracture, top cover displacement, and water leakage and other serious accidents;

[0004] However, the prior art has the following defects in monitoring the state of the pumped storage unit top cover:

[0005] First, the existing monitoring mostly uses discrete point arrangement. The top cover is a giant cast forging with a diameter of several meters and an extremely complex structure. Only 4-8 displacement meters or vibration sensors in the circumferential direction cannot capture the stress concentration of the local structure. When micro-cracks or local stiffness degradation occur inside the top cover due to fatigue, the discrete sensors may not be sensitive at all, resulting in the monitoring system failing to give an early warning before the overall structure fails.

[0006] Second, the pumped storage unit frequently switches between pumping and power generation and performs load rejection operations. The top cover bears extremely fluctuating non-steady-state loads. The existing system is mostly based on static or broad alarm thresholds. In the transient process, the conventional threshold alarm either produces false positives because it is too sensitive or masks the instantaneous structural damage risk because the threshold is too high.

[0007] Third, the working environment of the top cover is extremely harsh. Traditional electrical signal sensors are prone to zero drift or sealing failure. The existing monitoring system lacks a metering self-checking mechanism. It is often difficult for maintenance personnel to distinguish whether the abnormal reading is due to a problem with the top cover or the sensor itself, which may lead to incorrect shutdown decisions.

[0008] In view of the above technical defects, the present application provides a solution. SUMMARY

[0009] The purpose of the present application is to: by using the spatial mapping comparison between the theoretical benchmark generated by digital twinning and the measured data set, the zero point drift, signal failure or installation looseness of the sensor can be accurately identified, the accuracy of the early warning instruction is ensured, and through the three-dimensional finite element model, the point-shaped sensing data is converted into a planar stress distribution cloud diagram, the local stress concentration in the non-measuring point area can be effectively captured, the dynamic safety threshold is automatically switched according to different working conditions such as pumping, power generation and transient transition, the problem of traditional static alarm normalization false alarm caused by load change in the transition process of unit start-stop, load shedding and the like is effectively solved, and scientific data support is provided for fine operation and maintenance and life prediction of the unit.

[0010] In order to achieve the above purpose, the present application adopts the following technical scheme: a pumped storage unit top cover state monitoring and early warning alarm system, comprising a data acquisition unit, a working condition recognition unit, a digital twinning evaluation engine, a virtual-real mapping self-checking unit and an intelligent early warning unit, wherein:

[0011] The data acquisition unit is used to obtain bolt prestress data, top cover vibration data and pressure pulsation data in the flow passage adjacent to the top cover through a sensor group, and after preprocessing, the data are integrated into a top cover real-time state data set and sent to the virtual-real mapping self-checking unit;

[0012] The working condition recognition unit is electrically connected with the pumped storage unit monitoring system, is used to acquire real-time operation parameters of the pumped storage unit in real time, and recognizes the working condition mode currently in based on standard operation parameters, the working condition mode includes pumping mode, power generation mode, phase modulation mode and transient transition mode, and the recognition result is sent to the digital twinning evaluation engine;

[0013] The digital twinning evaluation engine pre-stores a three-dimensional finite element model established based on the physical structure of the top cover, according to the working condition mode and the real-time operation parameters, the theoretical stress distribution cloud diagram and the theoretical displacement value of the top cover under the corresponding working condition model are solved in real time, and are sent to the virtual-real mapping self-checking unit;

[0014] The virtual-real mapping self-checking unit is used to acquire the top cover real-time state data set and the theoretical stress distribution cloud diagram and the theoretical displacement value, perform spatial mapping comparison, calculate data consistency residual, and determine the working state of the sensor in the sensor group according to the residual distribution, to generate a sensor self-checking signal and a sensor approval signal and send them to the intelligent early warning alarm unit;

[0015] The intelligent early warning alarm unit is used to send a sensor self-checking signal to issue a warning reminder;

[0016] After the sensor approval signal is acquired, according to the effective data after self-checking, combined with the dynamic safety threshold under the current working condition mode, a multi-level early warning is performed on the local stress concentration and the structural health state of the pumped storage unit top cover.

[0017] Further, the sensor group includes a plurality of top cover bolt pre-tightening force sensors, a plurality of vibration sensors, and a plurality of pressure pulsation sensors, wherein:

[0018] The top cover bolt pre-tightening force sensor is installed on the bolt of the top cover of the pumped storage unit, and is used for real-time detection of the pre-tightening force of each bolt;

[0019] The vibration sensor is installed on the surface of the top cover or a structural member rigidly connected with the top cover, and is used for real-time detection of the vibration signal of the top cover;

[0020] The pressure pulsation sensor is installed in the flow channel near the top cover or at a position communicated with the flow channel, and is used for real-time detection of the pressure pulsation in the flow channel.

[0021] Further, the specific process of identifying the current working condition mode based on the standard operating parameters is as follows:

[0022] S101, real-time operating parameters are obtained through real-time data interface connection with the pumped storage unit monitoring system, and the real-time operating parameters include guide vane opening, active power, unit speed, and main circuit breaker state;

[0023] S102, the characteristic parameter matrix of each standard working condition, i.e., the standard operating parameter, is obtained, and the accurate identification of the current working condition mode is realized through the following logical process:

[0024] S1021, logical determination of the steady-state working condition mode: through multi-dimensional parameter combination triggering logic, it is determined whether the pumped storage unit enters a specific steady-state operating mode:

[0025] Power generation mode recognition: when it is detected that the main circuit breaker is in the closed state, the unit speed is in the rated synchronous speed range, the active power P is greater than the minimum power threshold for power generation grid connection, and the guide vane opening is in the power generation adjustment range, it is determined that the pumped storage unit is in the power generation mode;

[0026] Pumped storage mode recognition: when it is detected that the main circuit breaker is closed, the unit speed is the rated speed and the rotating direction is opposite to the power generation direction, the active power is less than the rated consumption power, and the guide vane opening is opened to the pumped storage working condition preset value, it is determined that the pumped storage unit is in the pumped storage mode;

[0027] Phase modulation mode recognition: when the pumped storage unit is in the grid-connected state, the speed is the rated value, but the active power only maintains the no-load loss, and the guide vane opening is in the fully closed state, and it is detected that there is a gas pressure modulation signal in the flow channel, it is determined that the pumped storage unit is in the phase modulation mode;

[0028] S1022, logical capture of transient transition mode:

[0029] When the change rate of the guide vane opening in the real-time parameters is greater than a preset change rate threshold;

[0030] When the active power undergoes a zero-point jump;

[0031] When the unit speed deviates from the rated speed range;

[0032] Then it is determined that all pumped storage units are locked in transient transition mode.

[0033] S103. After obtaining the real-time operating parameters, normalize the real-time operating parameters to the standard parameter coordinate system, and then calculate the Euclidean distance between the normalized parameter vector and the cluster center in the standard operating space.

[0034] If the calculated distance is less than the preset range, the working condition identification is considered valid.

[0035] If the parameter is at the edge of multiple operating conditions, time series inference will be performed by combining historical sequence data.

[0036] Furthermore, the specific process for calculating the theoretical stress distribution cloud map and theoretical displacement value of the top cover under the corresponding working condition model is as follows:

[0037] S201. Obtain a three-dimensional finite element model based on the physical structure of the roof, divide the stress concentration sensitive area according to the roof structure distribution, perform non-uniform mesh refinement on the stress concentration sensitive area, obtain the material properties and constraints of the roof, and define them as adjustable parameters.

[0038] S202. Receive the identification result from the working condition identification unit and the pressure pulsation data from the data acquisition unit, and execute the following load mapping logic:

[0039] Hydraulic load mapping: Based on the identified operating mode, the corresponding water pressure distribution function in the flow channel is automatically retrieved. Combined with the measured pressure pulsation data in the adjacent flow channel, the point pressure value is extended into a dynamic load surface acting on the flow surface of the top cover through the interpolation algorithm.

[0040] Structural constraint compensation: The collected measured bolt prestress data is used as the initial stress input in the three-dimensional finite element model to correct the deviation of the model boundary stiffness caused by bolt loosening and changes in preload.

[0041] S204. During the initialization phase, the displacement and stress response under multiple standard working conditions are pre-calculated to form a basis vector space. When the real-time working condition parameters are input, the real-time load vector is projected to the low-dimensional space, and the linear combination coefficients are solved to obtain the displacement vector and strain components of each node.

[0042] S205. The solved displacement vectors and strain components of each node are converted into the von Mises equivalent stress field. Through coordinate mapping, a spatial data matrix covering the entire top cover area is generated, namely the theoretical stress distribution cloud map.

[0043] S205, preset a virtual monitoring point completely corresponding to the physical sensor installation position in the three-dimensional finite element model, and extract a theoretical displacement value of the virtual monitoring point.

[0044] Further, the specific process of generating the sensor self-check signal and the sensor approval signal is as follows:

[0045] S301, obtain a pre-stored sensor layout coordinate matrix, align the installation position of the sensor group with the corresponding finite element node in the three-dimensional finite element model in coordinate, and specifically: input the measured data set {S1, S2, …, Sn} to obtain the theoretical calculation data set {V1, V2, …, Vn};

[0046] S302, calculate the real-time residual error of each group of mapped data using a sliding average weighting algorithm:

[0047] Wherein, Ri(t) is the consistency residual error of the i-th measuring point at time t, and ω is a preset weight coefficient;

[0048] S303, according to the spatial distribution and evolution trend of the residual error, the following hierarchical judgment logic is executed:

[0049] If the single-point residual error exceeds the standard, the adjacent points are normal, and if only the residual error of the measuring point is greater than the preset precision threshold, and the residual errors of other sensors at the spatial adjacent position are within the normal range, it is determined that the sensor fault occurs at the measuring point, and a sensor self-check signal for the sensor is generated;

[0050] If the bolt prestress, the top cover displacement and the vibration data in a certain area deviate from the theoretical value at the same time, and the deviation trend conforms to the mechanical logic, it is determined that the sensor group is normal, but the top cover structure appears unexpected local damage, a sensor approval signal is generated, it is determined that the data is real and reliable, and the deviation amount is taken as a structure abnormal signal;

[0051] If the global residual error is within the tolerance range, it is determined that the sensor state is good and the system runs stably, and a sensor approval signal is generated.

[0052] Further, the specific process of multi-level early warning of local stress concentration and structure health state of the top cover of the pumped storage unit is as follows:

[0053] S401, obtain a pre-stored dynamic safety threshold matrix T, T={Vth, Dth, Sth, Pth}, the dynamic safety threshold matrix corresponds to the working condition mode one by one:

[0054] In the power generation mode and the pumping mode, set a low-amplitude vibration amplitude threshold Vth and a displacement deviation threshold Dth;

[0055] In phase modulation mode and transient transition mode, it automatically switches to the wideband oscillation threshold, but at the same time tightens the monitoring of the bolt prestress change rate;

[0056] S402. Obtain the theoretical stress distribution cloud map, extract the high-stress region from the theoretical stress distribution cloud map, and compare it with the measured sensor data near the high-stress region:

[0057] If the measured stress value exceeds the theoretical threshold under this working condition and shows a continuous upward trend, it is determined to be an abnormal local stress concentration.

[0058] S403. Using the spatial coordinates of the digital twin model, the stress concentration point is accurately located and highlighted in real time on the visualization interface;

[0059] S404. The structural health index Uy of the roof is calculated in real time using a multi-physical quantity weighted algorithm.

[0060] Where α, β, and γ are preset proportional coefficients, Vth is the vibration amplitude threshold, Vl is the real-time vibration data, Dth is the displacement deviation threshold, ΔD is the displacement deviation data, Sth is the bolt stress loss threshold, and ΔS is the bolt stress loss value.

[0061] S405. Execute the following multi-level early warning procedures:

[0062] When the sensor self-test signal shows that an individual sensor is drifting, a first-level early warning is triggered, the abnormality is recorded, and a notice is issued on the human-machine interface.

[0063] When the local stress concentration factor exceeds the safety factor, a level-two early warning is triggered, which in turn triggers an abnormal status monitoring alarm, reminding maintenance personnel to increase the frequency of unit inspections.

[0064] When multiple approved data simultaneously exceed the dynamic safety threshold under the corresponding operating conditions, a Level 3 early warning is triggered, which in turn triggers a severe structural risk alarm, reminding power plant operators to reduce the load.

[0065] When the bolt prestress drops and the displacement data exceeds the structural safety limit, a Level IV warning is triggered, an emergency shutdown recommendation signal is issued, and high-speed historical data before and after the accident is automatically locked.

[0066] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0067] This pumped storage unit's top cover status monitoring and early warning alarm system utilizes a theoretical benchmark generated by digital twins and a spatial mapping comparison with measured datasets to accurately identify sensor zero-point drift, signal failure, or loose installation, ensuring the accuracy of early warning commands. Simultaneously, by using a three-dimensional finite element model to transform point-like sensor data into surface-like stress distribution cloud maps, it can effectively capture local stress concentrations in non-measuring areas. It automatically switches dynamic safety thresholds for different operating conditions such as pumping, power generation, and transient transitions, effectively solving the problem of common false alarms in traditional static alarms caused by drastic load changes during unit start-up, shutdown, and load shedding. This provides scientific data support for the unit's refined operation and maintenance and lifespan prediction. Attached Figure Description

[0068] Figure 1 A schematic diagram of the overall structure of the present invention is shown. Detailed Implementation

[0069] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0070] Example:

[0071] like Figure 1 As shown, the pumped storage unit roof status monitoring and early warning alarm system includes a data acquisition unit, an operating condition identification unit, a digital twin evaluation engine, a virtual-real mapping self-test unit, and an intelligent early warning unit, wherein:

[0072] The data acquisition unit is used to acquire bolt prestress data, top cover vibration data and pressure pulsation data in the flow channel near the top cover through the sensor group, and after preprocessing, integrates them into a top cover real-time status dataset and sends it to the virtual-real mapping self-test unit.

[0073] The sensor group includes multiple top cover bolt preload sensors, multiple vibration sensors, and multiple pressure pulsation sensors, among which:

[0074] The preload force sensor for the top cover bolts is installed on the bolts of the pumped storage unit's top cover to detect the preload force of each bolt in real time.

[0075] Vibration sensors are installed on the surface of the top cover or on structural components rigidly connected to the top cover to detect vibration signals of the top cover in real time.

[0076] The pressure pulsation sensor is installed in the flow channel near the top cover or in a position connected to the flow channel to detect pressure pulsations in the flow channel in real time.

[0077] The operating condition identification unit is electrically connected to the pumped storage unit monitoring system to obtain the real-time operating parameters of the pumped storage unit and identify the current operating condition mode based on the standard operating parameters. The operating condition modes include pumping mode, power generation mode, phase adjustment mode and transient transition mode. The identification results are sent to the digital twin evaluation engine.

[0078] The specific process for identifying the current operating mode based on standard operating parameters is as follows:

[0079] S101. Real-time operating parameters are obtained by connecting with the real-time data interface of the pumped storage unit monitoring system. The real-time operating parameters include guide vane opening, active power, unit speed, and main circuit breaker status.

[0080] S102. Obtain the pre-stored feature parameter matrix of each standard operating condition, i.e., the standard operating parameters, and achieve accurate identification of the current operating condition mode through the following logical process:

[0081] S1021. Logic determination of steady-state operating mode: The logic is triggered by a combination of multi-dimensional parameters to determine whether the pumped storage unit has entered a specific steady-state operating mode.

[0082] Power generation mode identification: When the main circuit breaker is detected to be in the closed state, the unit speed is within the rated synchronous speed range, the active power P is greater than the minimum power threshold for grid connection, and the guide vane opening is within the power generation adjustment range, the pumped storage unit is determined to be in power generation mode.

[0083] Pumping mode identification: When the main circuit breaker is closed, the unit speed is at the rated speed and the rotation direction is opposite to the power generation direction, the active power is less than the rated power consumption, and the guide vane opening is opened to the preset value of pumping operation, the pumped storage unit is determined to be in pumping mode.

[0084] Phase modulation mode identification: When the pumped storage unit is in grid-connected state, the speed is at the rated value, but the active power only maintains the no-load loss, and the guide vane opening is in the fully closed state, and at the same time the compressed air phase modulation signal is detected in the flow channel, it is determined that the pumped storage unit is in phase modulation mode.

[0085] S1022, Logic capture of transient transition mode:

[0086] When the rate of change of the guide vane opening in the real-time parameters exceeds the preset rate of change threshold;

[0087] When the active power undergoes a zero-point jump;

[0088] When the unit speed deviates from the rated speed range;

[0089] Then it is determined that all pumped storage units are locked in transient transition mode.

[0090] S103. After obtaining the real-time operating parameters, normalize the real-time operating parameters to the standard parameter coordinate system, and then calculate the Euclidean distance between the normalized parameter vector and the cluster center in the standard operating space.

[0091] If the calculated distance is less than the preset range, the working condition identification is considered valid.

[0092] If the parameter is on the edge of multiple operating condition characteristics, time-series inference will be performed by combining historical sequence data. (If the preceding sequence is pumping mode, and the current guide vane closes rapidly and the power drops sharply, it is inferred to be a transient transition from pumping to shutdown). The identification result includes the operating condition code and operating condition confidence level, and will be sent to the digital twin evaluation engine to ensure the instantaneous switching of the simulation model.

[0093] The digital twin evaluation engine has a pre-stored three-dimensional finite element model based on the physical structure of the top cover. According to the working mode and real-time operating parameters, it calculates the theoretical stress distribution cloud map and theoretical displacement value of the top cover under the corresponding working mode model in real time and sends them to the virtual-real mapping self-test unit.

[0094] The specific process for calculating the theoretical stress distribution cloud map and theoretical displacement value of the top cover under the corresponding working condition model is as follows:

[0095] S201. Obtain a three-dimensional finite element model based on the physical structure of the top cover, divide the stress concentration sensitive area according to the distribution of the top cover structure, perform non-uniform mesh refinement on the stress concentration sensitive area, obtain the material properties (elastic modulus, Poisson's ratio) and constraint conditions (seat ring fixed end constraint, bolt preload load step) of the top cover, and define them as adjustable parameters.

[0096] S202. Receive the identification result from the working condition identification unit and the pressure pulsation data from the data acquisition unit, and execute the following load mapping logic:

[0097] Hydraulic load mapping: Based on the identified operating mode, the corresponding water pressure distribution function in the flow channel is automatically retrieved. Combined with the measured pressure pulsation data in the adjacent flow channel, the point pressure value is extended into a dynamic load surface acting on the flow surface of the top cover through the interpolation algorithm.

[0098] Structural constraint compensation: The collected measured bolt prestress data is used as the initial stress input in the three-dimensional finite element model to correct the deviation of the model boundary stiffness caused by bolt loosening and changes in preload.

[0099] S204. During the initialization phase, the displacement and stress response under multiple standard working conditions are pre-calculated to form a basis vector space. When the real-time working condition parameters are input, the real-time load vector is projected to the low-dimensional space, and the linear combination coefficients are solved to obtain the displacement vector and strain components of each node.

[0100] S205. The solved displacement vectors and strain components of each node are converted into the von Mises equivalent stress field. Through coordinate mapping, a spatial data matrix covering the entire top cover area is generated, namely the theoretical stress distribution cloud map.

[0101] S205. In the three-dimensional finite element model, preset virtual monitoring points that completely correspond to the installation positions of physical sensors, and extract the theoretical displacement values ​​of the virtual monitoring points for subsequent virtual-real mapping comparison.

[0102] The virtual-real mapping self-test unit is used to acquire the real-time status dataset of the top cover and the theoretical stress distribution cloud map and theoretical displacement value, perform spatial mapping comparison, calculate the data consistency residual, and determine the working status of the sensors in the sensor group based on the residual distribution, so as to generate sensor self-test signals and sensor verification signals and send them to the intelligent early warning alarm unit.

[0103] The specific process for generating sensor self-test signals and sensor verification signals is as follows:

[0104] S301. Obtain the pre-stored sensor deployment coordinate matrix and align the installation positions of the sensor group with the corresponding finite element nodes in the three-dimensional finite element model. Specifically: input the measured dataset {S1, S2, ..., Sn} to obtain the theoretical solution dataset {V1, V2, ..., Vn}.

[0105] S302. Calculate the real-time residual for each set of mapped data using a moving average weighted algorithm:

[0106] , where Ri(t) is the consistency residual of the i-th measurement point at time t, and ω is the preset weight coefficient;

[0107] S303. Based on the spatial distribution and evolution trend of the residuals, execute the following hierarchical judgment logic:

[0108] If the residual at a single point exceeds the standard, while the residual at neighboring points is normal, and if only the residual at a measuring point is greater than the preset accuracy threshold, while the residuals of other sensors at its spatially adjacent locations are within the normal range, then it is determined that a sensor fault has occurred at the measuring point, and a sensor self-test signal for that sensor is generated.

[0109] If the bolt prestress, top cover displacement and vibration data in a certain area deviate from the theoretical value at the same time, and the deviation trend is in line with mechanical logic, then the sensor group is judged to be normal. However, if the top cover structure has unexpected local damage, then a sensor verification signal is generated to determine that the data is true and reliable, and this deviation is used as a structural abnormality signal.

[0110] If the global residual is within the tolerance range, the sensor is considered to be in good condition, the system is operating stably, and a sensor approval signal is generated.

[0111] The intelligent early warning and alarm unit is used to acquire sensor self-test signals and issue early warning reminders;

[0112] After acquiring the sensor verification signal, based on the valid data after self-testing and combined with the dynamic safety threshold under the current operating condition, it provides multi-level early warning of local stress concentration and structural health status of the pumped storage unit top cover.

[0113] The specific process for multi-level early warning of local stress concentration and structural health status of pumped storage unit roof is as follows:

[0114] S401. Obtain the pre-stored dynamic safety threshold matrix T, T={Vth, Dth, Sth, Pth}, where the dynamic safety threshold matrix corresponds one-to-one with the operating mode:

[0115] In both power generation and pumping modes, low-amplitude vibration amplitude thresholds Vth and displacement deviation thresholds Dth are set.

[0116] In phase modulation mode and transient transition mode, it automatically switches to wideband oscillation threshold, allowing for large pressure pulsations and displacement fluctuations in a short period of time, but at the same time tightens the monitoring of bolt prestress change rate to prevent instantaneous fatigue damage caused by impact.

[0117] S402. Obtain the theoretical stress distribution cloud map, extract the high-stress region from the theoretical stress distribution cloud map, and compare it with the measured sensor data near the high-stress region:

[0118] If the measured stress value exceeds the theoretical threshold under this working condition and shows a continuous upward trend, it is determined to be an abnormal local stress concentration.

[0119] S403. Using the spatial coordinates of the digital twin model, the stress concentration point is accurately located and highlighted in real time on the visualization interface;

[0120] S404. The structural health index Uy of the roof is calculated in real time using a multi-physical quantity weighted algorithm.

[0121] Where α, β, and γ are preset proportional coefficients, Vth is the vibration amplitude threshold, Vl is the real-time vibration data, Dth is the displacement deviation threshold, ΔD is the displacement deviation data, Sth is the bolt stress loss threshold, and ΔS is the bolt stress loss value.

[0122] S405. Execute the following multi-level early warning procedures:

[0123] When the sensor self-test signal shows that an individual sensor is drifting, a first-level early warning is triggered, the abnormality is recorded, and a notice is issued on the human-machine interface.

[0124] When the local stress concentration factor exceeds the safety factor, a level-two early warning is triggered, which in turn triggers an abnormal status monitoring alarm, reminding maintenance personnel to increase the frequency of unit inspections.

[0125] When multiple approved data simultaneously exceed the dynamic safety threshold under the corresponding operating conditions, a Level 3 early warning is triggered, which in turn triggers a severe structural risk alarm, reminding power plant operators to reduce the load.

[0126] When the bolt prestress drops and the displacement data exceeds the structural safety limit, a Level IV warning is triggered, an emergency shutdown recommendation signal is issued, and high-speed historical data before and after the accident is automatically locked.

[0127] This invention utilizes a theoretical benchmark generated by digital twins to spatially map and compare with measured datasets, enabling precise identification of sensor zero-point drift, signal failure, or loose installation, ensuring the accuracy of early warning commands. Simultaneously, by using a three-dimensional finite element model to transform point-like sensor data into surface-like stress distribution cloud maps, it effectively captures localized stress concentrations in non-measuring areas. It automatically switches dynamic safety thresholds for different operating conditions such as pumping, power generation, and transient transitions, effectively solving the problem of common false alarms in traditional static alarms caused by drastic load changes during unit start-up, shutdown, and load shedding. This provides scientific data support for refined operation and maintenance and lifespan prediction of the units.

[0128] The size of the interval and threshold is set to facilitate comparison. The size of the threshold depends on the amount of sample data and the number of bases set by those skilled in the art for each set of sample data; as long as it does not affect the ratio between the parameter and the quantized value.

[0129] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0130] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A pumped storage unit top cover status monitoring and early warning alarm system, characterized in that, It includes a data acquisition unit, a working condition identification unit, a digital twin evaluation engine, a virtual-real mapping self-testing unit, and an intelligent early warning unit, among which: The data acquisition unit is used to acquire bolt prestress data, top cover vibration data and pressure pulsation data in the flow channel near the top cover through the sensor group, and after preprocessing, integrates them into a top cover real-time status dataset and sends it to the virtual-real mapping self-test unit. The operating condition identification unit is electrically connected to the pumped storage unit monitoring system to obtain the real-time operating parameters of the pumped storage unit in real time, and identify the current operating condition mode based on the standard operating parameters. The operating condition mode includes pumping mode, power generation mode, phase adjustment mode and transient transition mode, and sends the identification result to the digital twin evaluation engine. The digital twin evaluation engine has a three-dimensional finite element model pre-stored based on the physical structure of the top cover. According to the working condition mode and real-time operating parameters, it calculates the theoretical stress distribution cloud map and theoretical displacement value of the top cover under the corresponding working condition model in real time and sends them to the virtual-real mapping self-test unit. The virtual-real mapping self-test unit is used to acquire the real-time status dataset of the top cover and the theoretical stress distribution cloud map and theoretical displacement value, perform spatial mapping comparison, calculate the data consistency residual, and determine the working status of the sensors in the sensor group based on the residual distribution, so as to generate sensor self-test signals and sensor verification signals and send them to the intelligent early warning alarm unit. The intelligent early warning and alarm unit is used to acquire sensor self-test signals and issue early warning reminders; After acquiring the sensor verification signal, based on the valid data after self-testing and combined with the dynamic safety threshold under the current operating condition, it provides multi-level early warning of local stress concentration and structural health status of the pumped storage unit top cover.

2. The pumped storage unit top cover status monitoring and early warning alarm system according to claim 1, characterized in that, The sensor group includes multiple top cover bolt preload sensors, multiple vibration sensors, and multiple pressure pulsation sensors, wherein: The preload force sensor for the top cover bolts is installed on the bolts of the pumped storage unit's top cover to detect the preload force of each bolt in real time. Vibration sensors are installed on the surface of the top cover or on structural components rigidly connected to the top cover to detect vibration signals of the top cover in real time. The pressure pulsation sensor is installed in the flow channel near the top cover or in a position connected to the flow channel to detect pressure pulsations in the flow channel in real time.

3. The pumped storage unit top cover status monitoring and early warning alarm system according to claim 1, characterized in that, The specific process for identifying the current operating mode based on standard operating parameters is as follows: S101. Obtain real-time operating parameters by connecting to the real-time data interface of the pumped storage unit monitoring system. The real-time operating parameters include guide vane opening, active power, unit speed, and main circuit breaker status. S102. Obtain the pre-stored feature parameter matrix of each standard operating condition, i.e., the standard operating parameters, and achieve accurate identification of the current operating condition mode through the following logical process: S1021. Logic determination of steady-state operating mode: The logic is triggered by a combination of multi-dimensional parameters to determine whether the pumped storage unit has entered a specific steady-state operating mode. Power generation mode identification: When the main circuit breaker is detected to be in the closed state, the unit speed is within the rated synchronous speed range, the active power P is greater than the minimum power threshold for grid connection, and the guide vane opening is within the power generation adjustment range, the pumped storage unit is determined to be in power generation mode. Pumping mode identification: When the main circuit breaker is closed, the unit speed is at the rated speed and the rotation direction is opposite to the power generation direction, the active power is less than the rated power consumption, and the guide vane opening is opened to the preset value of pumping operation, the pumped storage unit is determined to be in pumping mode. Phase modulation mode identification: When the pumped storage unit is in grid-connected state, the speed is at the rated value, but the active power only maintains the no-load loss, and the guide vane opening is in the fully closed state, and at the same time the compressed air phase modulation signal is detected in the flow channel, it is determined that the pumped storage unit is in phase modulation mode. S1022, Logic capture of transient transition mode: When the rate of change of the guide vane opening in the real-time parameters exceeds the preset rate of change threshold; When the active power undergoes a zero-point jump; When the unit speed deviates from the rated speed range; Then it is determined that all pumped storage units are locked in transient transition mode. 4.S103 After obtaining the real-time operating parameters, normalize the real-time operating parameters to the standard parameter coordinate system, and then calculate the Euclidean distance between the normalized parameter vector and the cluster center in the standard operating space. If the calculated distance is less than the preset range, the working condition identification is considered valid. If the parameter is at the edge of multiple operating conditions, time series inference will be performed by combining historical sequence data.

5. The pumped storage unit top cover status monitoring and early warning alarm system according to claim 1, characterized in that, The specific process for calculating the theoretical stress distribution cloud map and theoretical displacement value of the top cover under the corresponding working condition model is as follows: S201. Obtain a three-dimensional finite element model based on the physical structure of the roof, divide the stress concentration sensitive area according to the roof structure distribution, perform non-uniform mesh refinement on the stress concentration sensitive area, obtain the material properties and constraints of the roof, and define them as adjustable parameters. S202. Receive the identification result from the working condition identification unit and the pressure pulsation data from the data acquisition unit, and execute the following load mapping logic: Hydraulic load mapping: Based on the identified operating mode, the corresponding water pressure distribution function in the flow channel is automatically retrieved. Combined with the measured pressure pulsation data in the adjacent flow channel, the point pressure value is extended into a dynamic load surface acting on the flow surface of the top cover through the interpolation algorithm. Structural constraint compensation: The collected measured bolt prestress data is used as the initial stress input in the three-dimensional finite element model to correct the deviation of the model boundary stiffness caused by bolt loosening and changes in preload. S204. During the initialization phase, the displacement and stress response under multiple standard working conditions are pre-calculated to form a basis vector space. When the real-time working condition parameters are input, the real-time load vector is projected to the low-dimensional space, and the linear combination coefficients are solved to obtain the displacement vector and strain components of each node. S205. The solved displacement vectors and strain components of each node are converted into the von Mises equivalent stress field. Through coordinate mapping, a spatial data matrix covering the entire top cover area is generated, namely the theoretical stress distribution cloud map. S205. In the three-dimensional finite element model, preset virtual monitoring points that completely correspond to the installation positions of the physical sensors, and extract the theoretical displacement values ​​of the virtual monitoring points.

6. The pumped storage unit top cover status monitoring and early warning alarm system according to claim 1, characterized in that, The specific process for generating sensor self-test signals and sensor verification signals is as follows: S301. Obtain the pre-stored sensor deployment coordinate matrix and align the installation positions of the sensor group with the corresponding finite element nodes in the three-dimensional finite element model. Specifically: input the measured dataset {S1, S2, ..., Sn} to obtain the theoretical solution dataset {V1, V2, ..., Vn}. S302. Calculate the real-time residual for each set of mapped data using a moving average weighted algorithm: , where Ri(t) is the consistency residual of the i-th measurement point at time t, and ω is the preset weight coefficient; S303. Based on the spatial distribution and evolution trend of the residuals, execute the following hierarchical judgment logic: If the residual at a single point exceeds the standard, while the residual at neighboring points is normal, and if only the residual at a measuring point is greater than the preset accuracy threshold, while the residuals of other sensors at its spatially adjacent locations are within the normal range, then it is determined that a sensor fault has occurred at the measuring point, and a sensor self-test signal for that sensor is generated. If the bolt prestress, top cover displacement and vibration data in a certain area deviate from the theoretical value at the same time, and the deviation trend is in line with mechanical logic, then the sensor group is judged to be normal. However, if the top cover structure has unexpected local damage, then a sensor verification signal is generated to determine that the data is true and reliable, and this deviation is used as a structural abnormality signal. If the global residual is within the tolerance range, the sensor is considered to be in good condition, the system is operating stably, and a sensor approval signal is generated.

7. The pumped storage unit top cover status monitoring and early warning alarm system according to claim 1, characterized in that, The specific process for multi-level early warning of local stress concentration and structural health status of pumped storage unit roof is as follows: S401. Obtain the pre-stored dynamic safety threshold matrix T, T={Vth, Dth, Sth, Pth}, wherein the dynamic safety threshold matrix corresponds one-to-one with the operating mode: In both power generation and pumping modes, low-amplitude vibration amplitude thresholds Vth and displacement deviation thresholds Dth are set. In phase modulation mode and transient transition mode, it automatically switches to the wideband oscillation threshold, but at the same time tightens the monitoring of the bolt prestress change rate; S402. Obtain the theoretical stress distribution cloud map, extract the high-stress region from the theoretical stress distribution cloud map, and compare it with the measured sensor data near the high-stress region: If the measured stress value exceeds the theoretical threshold under this working condition and shows a continuous upward trend, it is determined to be an abnormal local stress concentration. S403. Using the spatial coordinates of the digital twin model, the stress concentration point is accurately located and highlighted in real time on the visualization interface; S404. The structural health index Uy of the roof is calculated in real time using a multi-physical quantity weighted algorithm. Where α, β, and γ are preset proportional coefficients, Vth is the vibration amplitude threshold, Vl is the real-time vibration data, Dth is the displacement deviation threshold, ΔD is the displacement deviation data, Sth is the bolt stress loss threshold, and ΔS is the bolt stress loss value. S405. Execute the following multi-level early warning procedures: When the sensor self-test signal shows that an individual sensor is drifting, a first-level early warning is triggered, the abnormality is recorded, and a notice is issued on the human-machine interface. When the local stress concentration factor exceeds the safety factor, a level-two early warning is triggered, which in turn triggers an abnormal status monitoring alarm, reminding maintenance personnel to increase the frequency of unit inspections. When multiple approved data simultaneously exceed the dynamic safety threshold under the corresponding operating conditions, a Level 3 early warning is triggered, which in turn triggers a severe structural risk alarm, reminding power plant operators to reduce the load. When the bolt prestress drops and the displacement data exceeds the structural safety limit, a Level IV warning is triggered, an emergency shutdown recommendation signal is issued, and high-speed historical data before and after the accident is automatically locked.

Images