Defect diagnosis and early warning method and device for steam turbine generator

Abstract

The invention relates to the technical field of fault monitoring, and provides a steam turbine generator defect diagnosis and early warning method and device. The method comprises the following steps: acquiring operation data of a to-be-monitored steam turbine generator; according to the operation data, identifying the current operation condition of the steam turbine generator to be monitored, and generating a current condition label; calling a preset baseline model corresponding to the current working condition label, wherein the preset baseline model corresponding to the current working condition label is a mathematical model which is established based on historical operation data and is used for calculating a theoretical value of the defect diagnosis index under the current operation working condition; based on the preset baseline model and the operation data corresponding to the current working condition label, defect diagnosis analysis is carried out on the steam turbine generator to be monitored, a defect diagnosis result is generated, and the defect diagnosis result is at least one of normal operation or stator winding temperature abnormity, rotor turn-to-turn short circuit and water-soluble hydrogen water electric connector leakage. According to the method, accurate diagnosis of three types of typical defects is realized through dynamic working condition recognition and a working condition self-adaptive baseline model.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fault monitoring, in particular to a steam turbine generator defect diagnosis and early warning method and device. BACKGROUND

[0002] With the advancement of new power system construction, the demand for flexibility of the power system is continuously increasing, and the steam turbine generator is frequently in variable load and variable operating condition, and the unit performance degradation is significantly accelerated, which becomes an important factor causing fault shutdown. A large number of operation experiences show that stator winding temperature anomaly, rotor inter-turn short circuit, and hydrogen-water electrical joint leakage are typical defects in the operation process of large generators. If not discovered in the early stage in time, it may cause insulation burnout, vibration intensification, cross contamination of hydrogen and water, and serious threat to the safe operation of power plants.

[0003] However, the existing defect monitoring method has significant limitations under complex operating conditions. Traditional monitoring mainly relies on periodic power-off offline testing, which cannot capture early dynamic signals of defects in flexible operation in real time, often missing the best intervention opportunity, and making slight degradation evolve into serious failure. To address this challenge, current online diagnosis methods have been proposed, but generally rely on fixed thresholds for fault determination. Under the condition of frequent start-stop or load fluctuation, the operating parameters often deviate from the threshold for a short time, thereby causing false positives or false negatives. SUMMARY

[0004] The technical problem to be solved by the present application is to provide a steam turbine generator defect diagnosis and early warning method and device to solve the above technical problems.

[0005] The technical solution of the present application to solve the above technical problems is as follows: a steam turbine generator defect diagnosis and early warning method, comprising: obtaining operation data of a steam turbine generator to be monitored; identifying the current operating condition of the steam turbine generator to be monitored according to the operation data, and generating a current condition label; calling a preset baseline model corresponding to the current condition label, the preset baseline model corresponding to the current condition label being a mathematical model established based on historical operation data for calculating the theoretical value of a defect diagnosis index under the current operating condition; based on the preset baseline model corresponding to the current condition label and the operation data, performing defect diagnosis analysis on the steam turbine generator to be monitored, and generating a defect diagnosis result, the defect diagnosis result being at least one of normal operation, stator winding temperature anomaly, rotor inter-turn short circuit, and water-soluble hydrogen-water electrical joint leakage.

[0006] The present application has the beneficial effects that: the present application realizes the accurate diagnosis of three typical defects of the turbine generator stator winding temperature anomaly, the rotor inter-turn short circuit and the water-soluble hydrogen water joint leakage through the dynamic working condition recognition and the working condition self-adaptive baseline model. The present application breaks through the limitation of high false alarm rate of the traditional fixed threshold under variable working conditions through the dynamic calling of the baseline model, greatly improves the accuracy of defect recognition, and provides reliable guarantee for the safe, stable and flexible operation of the unit.

[0007] On the basis of the above technical scheme, the present application can also be improved as follows.

[0008] Further, the preset baseline model corresponding to the current working condition label is obtained by the following way: obtaining historical operation data corresponding to the current working condition label; performing linear regression fitting processing on the historical operation data to calibrate parameters in the pre-constructed initial baseline model to obtain the preset baseline model corresponding to the current working condition label; wherein the preset baseline model corresponding to the current working condition label includes a stator winding temperature rise baseline model, a rotor power deviation baseline model and a hydrogen-water pressure difference baseline model.

[0009] Further, the defect diagnosis analysis of the to-be-monitored turbine generator based on the preset baseline model corresponding to the current working condition label and the operation data includes: calculating a stator winding temperature rise theoretical value according to the stator winding temperature rise baseline model and the operation data; calculating a temperature deviation value according to the stator winding temperature rise theoretical value and the operation data, the temperature deviation value being the difference between a stator winding temperature rise measured value and the stator winding temperature rise theoretical value; calculating a temperature change rate according to the operation data; comparing the temperature deviation value with a temperature deviation dynamic threshold value and comparing the temperature change rate with a rate threshold value to analyze the stator winding temperature anomaly of the to-be-monitored turbine generator.

[0010] Further, the temperature deviation dynamic threshold value comprises a temperature deviation early warning threshold value and a temperature deviation emergency threshold value, and the rate threshold value comprises a rate early warning threshold value and a rate emergency threshold value; the comparison of the temperature deviation value with the temperature deviation dynamic threshold value and the comparison of the temperature change rate with the rate threshold value for the stator winding temperature abnormality analysis of the to-be-monitored steam turbine generator comprises: comparison of the temperature deviation value with the temperature deviation early warning threshold value and the temperature deviation emergency threshold value respectively, and comparison of the temperature change rate with the rate early warning threshold value and the rate emergency threshold value respectively; when the temperature deviation value is greater than the temperature deviation early warning threshold value and lasts for a first preset time length, and the temperature change rate is greater than the rate early warning threshold value and lasts for a first preset time length, it is determined that the stator winding temperature is abnormal and a first-level temperature early warning signal is generated; when the temperature deviation value is greater than the temperature deviation emergency threshold value and lasts for a first preset time length, and the temperature change rate is greater than the rate emergency threshold value and lasts for a first preset time length, it is determined that the stator winding temperature is abnormal and a second-level temperature early warning signal is generated.

[0011] Further, the defect diagnosis analysis of the to-be-monitored steam turbine generator based on the preset baseline model corresponding to the current working condition label and the operation data further comprises: calculating the generator active power theoretical value at each time within a preset time window according to the rotor power deviation baseline model and the operation data; calculating the power deviation value at each time according to the generator active power theoretical value at each time and the operation data, and the power deviation value at each time is the difference between the generator active power measured value at the time and the generator active power theoretical value at the time; performing distribution characteristic statistics according to the power deviation value at each time to obtain a statistical ratio, and the statistical ratio represents the proportion of the number of times when the corresponding power deviation value is greater than a preset deviation value to the number of all times within the preset time window; comparing the statistical ratio with a proportion threshold value to perform rotor inter-turn short circuit analysis on the to-be-monitored steam turbine generator.

[0012] Further, the proportion threshold value comprises a first proportion threshold value and a second proportion threshold value, the preset deviation value comprises a first deviation value and a second deviation value; the statistical ratio comprises a first ratio and a second ratio, the first ratio represents a proportion of a number of time points at which the corresponding power deviation value is greater than the first deviation value to a number of all time points in the preset time window, and the second ratio represents a proportion of a number of time points at which the corresponding power deviation value is greater than the second deviation value to the number of all time points in the preset time window; the comparison of the statistical ratio with the proportion threshold value to analyze the rotor inter-turn short circuit of the to-be-monitored turbogenerator comprises: comparison of the first ratio with the first proportion threshold value and comparison of the second ratio with the second proportion threshold value; when the first ratio is greater than the first proportion threshold value, it is determined that there is a rotor inter-turn short circuit and a first-level short circuit early warning signal is generated; when the second ratio is greater than the second proportion threshold value, it is determined that there is a rotor inter-turn short circuit and a second-level short circuit early warning signal is generated.

[0013] Further, the defect diagnosis analysis of the to-be-monitored turbogenerator based on the preset baseline model corresponding to the current working condition label and the operation data further comprises: calculating a hydrogen-water pressure difference theoretical value according to the hydrogen-water pressure difference baseline model and the operation data; calculating a pressure difference deviation value according to the hydrogen-water pressure difference theoretical value and the operation data, the pressure difference deviation value being a difference between a hydrogen-water pressure difference measured value and the hydrogen-water pressure difference theoretical value; calculating a leakage trend index according to the pressure difference deviation value and the operation data; comparing the leakage trend index with a leakage dynamic threshold value, and when the leakage trend index is greater than the leakage dynamic threshold value and lasts for a second preset time length, it is determined that there is a water-soluble hydrogen-water electrical joint leakage.

[0014] Further, the defect diagnosis analysis of the to-be-monitored turbogenerator based on the preset baseline model corresponding to the current working condition label and the operation data further comprises: calculating a hydrogen saturation concentration theoretical value based on Henry's law according to the operation data; calculating a concentration deviation value according to the hydrogen saturation concentration theoretical value and the operation data, the concentration deviation value being a difference between a dissolved hydrogen concentration measured value and the hydrogen saturation concentration theoretical value; performing hydrogen leakage amount estimation according to the concentration deviation value and the operation data to obtain a leakage intensity; when it is determined that there is a water-soluble hydrogen-water electrical joint leakage, generating a hydrogen leakage early warning signal according to the leakage intensity.

[0015] Further, the current operating condition of the to-be-monitored steam turbine generator is identified according to the operation data, and a current condition label is generated, including: constructing a condition feature vector according to the operation data, the condition feature vector being a vector reflecting the operating state of the to-be-monitored steam turbine generator; determining a condition cluster center closest to the condition feature vector from a plurality of pre-generated condition cluster centers, and taking the condition label corresponding to the condition cluster center as the current condition label.

[0016] To solve the above technical problems, the application further provides a steam turbine generator defect diagnosis and early warning device, comprising: a data acquisition module configured to acquire operation data of a to-be-monitored steam turbine generator; a condition identification module configured to identify a current operating condition of the to-be-monitored steam turbine generator according to the operation data, and generate a current condition label; a model calling module configured to call a preset baseline model corresponding to the current condition label, the preset baseline model corresponding to the current condition label being a mathematical model established based on historical operation data, and used to calculate a theoretical value of a defect diagnosis index under the current operating condition; a defect diagnosis module configured to perform defect diagnosis analysis on the to-be-monitored steam turbine generator based on the preset baseline model corresponding to the current condition label and the operation data, and generate a defect diagnosis result, the defect diagnosis result being at least one of normal operation, abnormal stator winding temperature, rotor inter-turn short circuit, and water-soluble hydrogen water joint leakage. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 a flowchart of a steam turbine generator defect diagnosis and early warning method of the application; Figure 2 a schematic diagram of a steam turbine generator defect diagnosis and early warning device of the application. DETAILED DESCRIPTION

[0018] The principles and features of the application are described below, and the examples are only used to explain the application and not to limit the scope of the application.

[0019] Existing multi-defect monitoring systems usually require additional installation of special sensors, and the cost of single-unit transformation is high and complex, and it will also affect the power plant power generation plan and economic benefits. In addition, although the distributed control system (DCS) / safety instrumented system (SIS) of the power plant can continuously collect and store a large amount of operation data, the existing method has not established a quantitative correlation model between defects and conventional operation parameters, resulting in that these massive data cannot be fully utilized, and the early warning capability of defects is insufficient.

[0020] Therefore, the prior art has the problems of offline diagnosis lag, strong hardware dependence, poor dynamic adaptability, and insufficient data utilization under flexible operating conditions. There is an urgent need for an online diagnosis method based on existing online data of power plants, without the need for additional hardware, capable of adapting to dynamic operating conditions and achieving accurate identification and active early warning of typical defects, to effectively improve the safety and reliability of the operation of the steam turbine generator. N Embodiment one Based on this, as Figure 1 shown, the embodiment provides a steam turbine generator defect diagnosis and early warning method, comprising: S101, obtaining operation data of a steam turbine generator to be monitored.

[0022] S102, identifying the current operating condition of the steam turbine generator to be monitored according to the operation data, and generating a current condition label.

[0023] S103, calling a preset baseline model corresponding to the current condition label, the preset baseline model corresponding to the current condition label being a mathematical model established based on historical operation data and used to calculate the theoretical value of the defect diagnosis index under the current operating condition.

[0024] S104, performing defect diagnosis analysis on the steam turbine generator to be monitored based on the preset baseline model corresponding to the current condition label and the operation data, and generating a defect diagnosis result, the defect diagnosis result being at least one of normal operation, stator winding temperature anomaly, rotor inter-turn short circuit, and water-soluble hydrogen water-electric joint leakage.

[0025] The method realizes accurate diagnosis of three typical defects of the steam turbine generator, i.e., stator winding temperature anomaly, rotor inter-turn short circuit, and water-soluble hydrogen water-electric joint leakage, through dynamic operating condition identification and operating condition self-adaptive baseline model. The method breaks through the limitation of high false alarm rate of traditional fixed threshold under variable operating conditions by dynamically calling the baseline model, greatly improves the accuracy of defect identification, and provides reliable protection for safe, stable, and flexible operation of the unit.

[0026] The multi-dimensional parameters related to the three types of defects during the operation of the generator are collected, i.e., the operation data is obtained, which specifically includes: Stator winding temperature anomaly related data: stator current (three-phase current, A), winding hot spot temperature (℃), cooling water inlet and outlet temperature (℃), cooling water flow (t / h), cooler inlet air temperature T air (℃), stator DC resistance Ra (Ω, design value); Rotor inter-turn short circuit related data: field current (A) Active power (MW), reactive power (MVar), terminal voltage (kV), Synchronous Reactance Xd (pu, design value), Stator DC Resistance Rs (Ω, design value); Data related to water-soluble hydrogen electrical connector leakage: hydrogen pressure (MPa), hydrogen-water pressure difference ΔP (MPa), cooling water flow rate Qw (m³ / h), dissolved hydrogen concentration (μg / L), hydrogen water temperature Tw (°C).

[0027] The collected operational data is preprocessed, including parameter mapping (unifying parameter names and sources), outlier filtering (removing sensor malfunctions or data transmission errors), and unit unification, before being stored in a time-series database to ensure parameter consistency in subsequent calculations.

[0028] Optionally, in an embodiment, based on the operating data, the current operating condition of the turbine generator to be monitored is identified, and a current operating condition label is generated, including: constructing an operating condition feature vector based on the operating data, wherein the operating condition feature vector is a vector reflecting the operating state of the turbine generator to be monitored; among multiple pre-generated operating condition cluster centers, determining the operating condition cluster center that is closest to the operating condition feature vector, and using the operating condition label corresponding to the operating condition cluster center as the current operating condition label.

[0029] Specifically, the first step is to acquire historical operating data. Based on this data, a time-series operating condition feature vector is constructed. This feature vector includes multi-dimensional parameters reflecting the core operating state of the generator, specifically expressed as follows: ; in, The active power (MW) at time t reflects the unit load level; The excitation current (A) at time t reflects the magnetic field strength; The reactive power at time t (MVar) reflects the reactive power support of the power grid. The temperature difference between hydrogen and water at time t (°C) reflects the state of the cooling system. The power factor reflects the ratio of active to reactive power.

[0030] The active power change rate (MW / min) reflects the dynamics of the operating conditions (stable / fluctuating). The excitation current change rate (A / min) reflects the speed of magnetic field adjustment.

[0031] These parameters together constitute a complete feature space describing the "load level + dynamic change + energy balance + cooling state" of the unit. By clustering algorithm, the feature vectors of each working condition constructed based on the historical period operation data are clustered. The clustering algorithm divides the operating state into several stable working condition categories by identifying the similarity of the feature vectors. Clustering is a multi-dimensional working condition feature vector of the real-time operation of the generator. The purpose is to classify similar operating states into the same category to provide a "same working condition benchmark" for subsequent baseline modeling.

[0032] In this embodiment, the clustering adopts the K-means algorithm (the most commonly used unsupervised clustering method), the core of which is to realize the optimal division of the feature vectors by minimizing the "within-cluster sum of squares". The objective function formula is as follows: ; Wherein, is the objective function (within-cluster sum of squares), the smaller the value, the higher the similarity of the same samples; is the preset number of working condition categories (usually 5 to 8 categories according to the actual operation data of the power plant); is the th working condition cluster (a set of feature vectors of the same working condition); is a certain feature vector belonging to the cluster ; is the center of the cluster (the mean of all feature vectors of the cluster, representing the "typical features" of the working condition); is the Euclidean distance square of the feature vector and the cluster center (measuring the similarity of the sample and the center).

[0033] The clustering process is as follows: Step 1, randomly initialize k cluster centers, i.e. , , .

[0034] Step 2, calculate the distance between each feature vector x(t) and all cluster centers, and assign x(t) to the nearest cluster; Step 3, recalculate the center of each cluster (take the mean of all vectors in the cluster); Step 4, repeat steps 2-3 until the cluster center no longer changes or J converges (change <0.001).

[0035] According to the above process, a plurality of working condition cluster centers can be obtained. According to the characteristics of each working condition cluster center, the working conditions can be divided into multiple categories. The working condition label is a symbolic identification of the clustering result, that is, a unique label is assigned to each working condition cluster Cc divided by clustering (such as ), parameter call for subsequent baseline model and differentiated adaptation of diagnostic logic.

[0036] In this embodiment, according to the typical operation scenario of a large steam turbine generator, the clustered operating condition labels and definitions are as shown in Table 1: Table 1 Operating condition classification table When the unit has been running for a long time (e.g., every 3 months), the newly collected operating condition feature vector may deviate from the original clustering boundary. At this time, the system will automatically re-execute K-means clustering to adjust the partitioning criteria of the operating condition labels (e.g., the upper limit of the load of C1 is adjusted from 30%P N to 25%P N ), ensuring the matching of the labels and the actual operating state.

[0037] During real-time diagnosis, the baseline model corresponding to the current operating condition label C(t) is automatically called. For example, when the unit switches from C2 to C3, the baseline model completes parameter switching within 10 seconds, ensuring the matching degree of the baseline and the actual operating state.

[0038] The method proposed in this embodiment breaks through the limitations of traditional fixed thresholds and can maintain high accuracy in unit load changes by recognizing operating conditions based on multi-dimensional feature vectors and dynamically adjusting the baseline and thresholds.

[0039] Optionally, in the embodiment, the preset baseline model corresponding to the current operating condition label is obtained by: obtaining historical operating data corresponding to the current operating condition label; performing linear regression fitting processing on the historical operating data to calibrate parameters in the pre-constructed initial baseline model, thereby obtaining the preset baseline model corresponding to the current operating condition label; wherein the preset baseline model corresponding to the current operating condition label includes a stator winding temperature rise baseline model, a rotor power deviation baseline model, and a hydrogen-water pressure difference baseline model.

[0040] Baseline modeling is to construct a baseline range of normal operating parameters for each type of operating condition, i.e., to fit the theoretical value (baseline) of the core parameter under this operating condition through historical normal data (defect-free period) as a reference for judging whether the real-time parameter is abnormal.

[0041] Specifically, the historical operating data (operating data during the historical defect-free period) corresponding to each type of operating condition is obtained, and the preset baseline model is constructed by the following methods according to the historical operating data corresponding to each type of operating condition, respectively, to obtain the stator winding temperature rise baseline model, the rotor power deviation baseline model, and the hydrogen-water pressure difference baseline model corresponding to each type of operating condition.

[0042] Specifically, the stator winding temperature rise baseline calculates the theoretical temperature rise of the stator winding under the current working condition (temperature reference in normal state), which is used to judge whether the measured temperature is abnormal: The temperature rise of the stator winding is mainly determined by the temperature difference between the inlet and outlet water, the load (stator current), and the cooling condition (cooling water flow, inlet air temperature). Under normal circumstances, the temperature rise is approximately linearly related to the square of the stator current.

[0043] Calculate the stator copper loss: ; Cooling capacity calculation: ; Where, is the specific heat capacity of water (take ), is the cooling water flow, , are the cooling water inlet and outlet temperatures, respectively.

[0044] Based on the above formula, the winding temperature reference value model is constructed. Specifically, the copper loss is divided by the cooling capacity to obtain the reference temperature rise: ; Where, is the ambient reference temperature (take 25℃), , is the empirical coefficient (historical normal working condition regression calibration), is the cooler inlet air temperature.

[0045] Calculate the temperature deviation feature quantity, specifically compare the measured hot spot temperature with the baseline temperature: .

[0046] The rotor power deviation baseline calculates the theoretical active power of the rotor without inter-turn short circuit under the current working condition, which is compared with the measured power to judge whether there is a short circuit defect: Real-time calculation of operating power angle, i.e. calculation of the power angle (δ) reflecting the phase relationship between the rotor and the stator magnetic field: ; Where, power factor , , Xd is the synchronous reactance, and Rs is the stator direct current resistance.

[0047] Prediction power calculation, specifically: Under ideal conditions, the excitation current I f has a stable functional relationship with the generator no-load electromotive force: .

[0048] Constructing array (I f , ), the function of the no-load electromotive force about the excitation current can be obtained by curve fitting: .

[0049] The theoretical output power is calculated by using the terminal voltage, power angle and no-load electromotive force: .

[0050] The actual power deviation is calculated: ; Where, P meas is the measured active power of DCS.

[0051] The hydrogen-water pressure difference baseline calculates the theoretical pressure difference when the hydrogen-water system has no leakage under the current working condition. By comparing with the measured pressure difference, it can assist in judging whether the hydrogen-water joint is leaking: ; Where, is the theoretical hydrogen-water pressure difference (MPa) at time t; is the basic pressure difference (MPa) corresponding to the working condition , reflecting the inherent pressure difference under stable flow; is the flow variation coefficient corresponding to the working condition , reflecting the influence of flow rate change on pressure difference; is the cooling water flow rate change (m 3 / h 2 ), a key factor under dynamic working condition.

[0052] The specific method of parameter determination is: collect historical operation data under this working condition, specifically the data when there is no leakage, calculate and , by linear regression , and require the residual standard deviation .

[0053] The pressure difference deviation value calculation formula is: .

[0054] This method establishes a baseline model that combines physical mechanism (formula, law) and data-driven (historical data fitting), and the model parameters are dynamically corrected with working condition tags.

[0055] Independent diagnostic models are constructed for the three types of defects, and precise diagnosis is achieved through "parameter deviation + trend analysis".​

[0056] Optionally, in the embodiment, based on the preset baseline model corresponding to the current working condition label and the operation data, the defect diagnosis analysis is performed on the to-be-monitored steam turbine generator, including: calculating a stator winding temperature rise theoretical value according to the stator winding temperature rise baseline model and the operation data; calculating a temperature deviation value according to the stator winding temperature rise theoretical value and the operation data, the temperature deviation value being a difference between a stator winding temperature rise measured value and the stator winding temperature rise theoretical value; calculating a temperature change rate according to the operation data; comparing the temperature deviation value with a temperature deviation dynamic threshold value, and comparing the temperature change rate with a rate threshold value, so as to perform stator winding temperature abnormality analysis on the to-be-monitored steam turbine generator.

[0057] After obtaining the current working condition label, the baseline model corresponding to the current working condition is called. Based on the stator winding temperature rise baseline model corresponding to the current working condition, the temperature deviation value is calculated.

[0058] And the temperature change rate is calculated. In this embodiment, the temperature change rate is a difference between a measured temperature at a current time (t) and a measured temperature at 5 minutes ago (t-5min) divided by a time interval (5min), to obtain a temperature change rate per unit time, which is used to judge whether the temperature shows an abnormal rising trend. The specific calculation formula is: .

[0059] Optionally, in the embodiment, the temperature deviation dynamic threshold value includes a temperature deviation early warning threshold value and a temperature deviation emergency threshold value, and the rate threshold value includes a rate early warning threshold value and a rate emergency threshold value; the temperature deviation value is compared with the temperature deviation dynamic threshold value, and the temperature change rate is compared with the rate threshold value, so as to perform stator winding temperature abnormality analysis on the to-be-monitored steam turbine generator, including: the temperature deviation value is compared with the temperature deviation early warning threshold value and the temperature deviation emergency threshold value respectively, and the temperature change rate is compared with the rate early warning threshold value and the rate emergency threshold value respectively; when the temperature deviation value is greater than the temperature deviation early warning threshold value and lasts for a first preset time length, and the temperature change rate is greater than the rate early warning threshold value and lasts for the first preset time length, it is determined that the stator winding temperature is abnormal and a first-level temperature early warning signal is generated; when the temperature deviation value is greater than the temperature deviation emergency threshold value and lasts for the first preset time length, and the temperature change rate is greater than the rate emergency threshold value and lasts for the first preset time length, it is determined that the stator winding temperature is abnormal and a second-level temperature early warning signal is generated.

[0060] For the setting of the temperature deviation dynamic threshold value, specifically: Past 7-day normal working condition data is obtained, specifically the temperature deviation value under the past 7-day normal operation condition. According to the collected data, the temperature deviation value mean And the temperature deviation value standard deviation .

[0061] The temperature deviation early warning threshold is set as: The temperature deviation emergency threshold is In this embodiment, the rate early warning threshold is set as The rate emergency threshold is .

[0062] According to the above thresholds, the stator winding temperature abnormality is judged, and specifically: When and and lasts , it is determined that the temperature is abnormal early warning; when and and lasts , it is determined that the temperature is abnormal emergency warning.

[0063] In some embodiments, and can be dynamically adjusted according to the working condition label, for example, under fluctuating working conditions C4 / C5 relaxed to (avoid false alarm when load changes rapidly); under stable working conditions (C1 / C2 / C3) strictly (improve sensitivity under steady state).

[0064] Stator winding temperature abnormality deviation + trend double criteria: diagnose the condition in combination with the real-time deviation parameter change rate, which can capture the temperature abnormality trend earlier than single over-temperature alarm, and balance sensitivity and reliability.

[0065] Optionally, in the embodiment, based on the preset baseline model corresponding to the current working condition label and the operation data, the defect diagnosis analysis of the to-be-monitored steam turbine generator is further included: calculating the generator active power theoretical value at each time within a preset time window according to the rotor power deviation baseline model and the operation data; calculating the power deviation value at each time according to the generator active power theoretical value at each time and the operation data, the power deviation value at each time being the difference between the measured value of the generator active power at the time and the generator active power theoretical value at the time; performing distribution characteristic statistics according to the power deviation values at each time to obtain a statistical ratio, the statistical ratio representing the proportion of the number of times when the corresponding power deviation value is greater than a preset deviation value to the number of all times within the preset time window; and comparing the statistical ratio with a proportion threshold to analyze the rotor inter-turn short circuit of the to-be-monitored steam turbine generator.

[0066] After obtaining the current working condition label, the baseline model corresponding to the current working condition is called. Based on the rotor power deviation baseline model corresponding to the current working condition, the power deviation value at each time within a preset time window is calculated.

[0067] Optionally, in the embodiments, the proportional threshold includes a first proportional threshold and a second proportional threshold, and the preset deviation value includes a first deviation value and a second deviation value; the statistical ratio includes a first ratio and a second ratio, wherein the first ratio represents the proportion of the number of times when the corresponding power deviation value is greater than the first deviation value to the total number of times within the preset time window, and the second ratio represents the proportion of the number of times when the corresponding power deviation value is greater than the second deviation value to the total number of times within the preset time window; comparing the statistical ratio with the proportional threshold to perform rotor inter-turn short circuit analysis on the turbine generator under monitoring includes: comparing the first ratio with the first proportional threshold and comparing the second ratio with the second proportional threshold; when the first ratio is greater than the first proportional threshold, determining a rotor inter-turn short circuit and generating a first-level short circuit warning signal; when the second ratio is greater than the second proportional threshold, determining a rotor inter-turn short circuit and generating a second-level short circuit warning signal.

[0068] In this embodiment, the first proportional threshold is set to 30%, and the second proportional threshold is set to 70%. First deviation value. Set as ( (Rated power), second deviation value Set as .

[0069] Within a preset time window (such as 1 day or 1 week), the statistical power deviation value is calculated. Distribution characteristics: If the first ratio, i.e. If the proportion exceeds 30%, it is judged as a minor risk of inter-turn short circuit, and a level one short circuit warning signal is generated. like If the proportion exceeds 70%, it is judged as a serious risk of inter-turn short circuit, and a level 2 short circuit warning signal is generated.

[0070] The rotor-turn short-circuit power deviation statistical method in this embodiment introduces the deviation distribution analysis between theoretical power and measured power to improve the accuracy and sensitivity of short-circuit risk identification.

[0071] Optionally, in the embodiments, the defect diagnosis analysis of the turbine generator to be monitored, based on the preset baseline model and operating data corresponding to the current operating condition label, further includes: calculating the theoretical value of hydrogen-water pressure difference according to the hydrogen-water pressure difference baseline model and operating data; calculating the pressure difference deviation value according to the theoretical value of hydrogen-water pressure difference and operating data, wherein the pressure difference deviation value is the difference between the measured value of hydrogen-water pressure difference and the theoretical value of hydrogen-water pressure difference; calculating the leakage trend index according to the pressure difference deviation value and operating data; comparing the leakage trend index with the leakage dynamic threshold, and determining that the water-soluble hydrogen-water electrical connector is leaking when the leakage trend index is greater than the leakage dynamic threshold and continues for a second preset time.

[0072] Optionally, in this embodiment, the defect diagnosis analysis of the turbine generator to be monitored, based on the preset baseline model and operating data corresponding to the current operating condition label, further includes: calculating the theoretical value of hydrogen saturation concentration based on Henry's Law according to the operating data; calculating the concentration deviation value based on the theoretical value of hydrogen saturation concentration and the operating data, wherein the concentration deviation value is the difference between the measured value of dissolved hydrogen concentration and the theoretical value of hydrogen saturation concentration; estimating the amount of hydrogen leakage based on the concentration deviation value and the operating data to obtain the leakage intensity; and generating a hydrogen leakage warning signal based on the leakage intensity when it is determined that the water-soluble hydrogen water-electric connector is leaking.

[0073] Specifically, the hydrogen concentration deviation and leakage trend indicators are calculated to estimate the leakage amount and assess the risk.

[0074] The calculation of hydrogen concentration deviation is as follows: the solubility of hydrogen in water is related to temperature and pressure, and follows Henry's Law. The formula for calculating the theoretical saturation concentration is: ; in, The Henry's constant is temperature-dependent (e.g., at 25°C). ), For hydrogen pressure, The temperature of the hydrogen water.

[0075] Calculate the deviation of the measured concentration: ; in, This represents the measured dissolved hydrogen concentration.

[0076] If the joint leaks, the hydrogen concentration in the water sample will increase, which, combined with the cooling water flow rate Q, will affect the concentration of hydrogen in the sample. W Estimate the leakage amount and quantify the leakage intensity: ; in, Cooling water flow rate (converted to L / h, 1m) 3 / h=1000L / h), Units are .

[0077] Simultaneously, the leakage trend index is calculated considering the hydrogen-water pressure difference: ; in, (Weighting coefficient).

[0078] For leakage dynamic threshold The settings are as follows: Obtain normal operating data from the past 7 days, specifically the leakage trend indicators under normal operating conditions over the past 7 days. Based on the collected data, the mean of the leakage trend index was calculated. and a leakage trend index standard deviation .

[0079] Setting a leakage dynamic threshold is: .

[0080] According to the leakage dynamic threshold, the water-soluble hydrogen water electrical joint leakage is judged, and specifically: When the leakage trend index and lasts , it is determined that the hydrogen water joint leaks early warning, and The larger the leakage level is. The method is based on Henry's law and hydrogen water pressure difference double index, and for the first time realizes quantitative estimation and trend analysis of leakage.

[0081] After the diagnosis and analysis of the above three typical defects, the defect diagnosis result can be obtained. If the determination conditions of the above stator winding temperature anomaly, rotor turn-to-turn short circuit and water-soluble hydrogen water electrical joint leakage are not met, it is determined that the defect diagnosis result is normal. If one or more of the above determination conditions are met, the defect diagnosis result is generated according to the determination situation. For example, the defect diagnosis result is rotor turn-to-turn short circuit and water-soluble hydrogen water electrical joint leakage.

[0082] The method connects DCS / SIS system through OPCUA / Modbus protocol, real-time acquires multi-dimensional running parameters of three types of defects of generator, and the collection delay is less than 1s. It can completely rely on existing DCS / SIS online data, without the need for additional installation of monitoring sensors, reducing downtime modification cost, and greatly shortening the project implementation cycle. Through working condition recognition and dynamic baseline, in the load fluctuation (such as 10%-100% rated load) scene, false positives and false negatives are reduced. The synchronous diagnosis of three types of core defects of stator temperature anomaly, rotor turn-to-turn short circuit and water-soluble hydrogen leakage is realized, without deploying multiple independent systems, and the comprehensiveness of unit state evaluation is improved. Based on the early defects of small parameter deviation, the method can provide early warning several hours to several days earlier than the traditional method. Moreover, the algorithm is lightweight, which can be directly integrated into the power plant monitoring platform, and the results (such as charts, warning levels and expert suggestions) can be directly output, without the need for professional interpretation, easy to popularize and apply. Moreover, the diagnosis result data can be archived, with grading warning and self-learning ability, which can be randomly used for long-term operation and continuous optimization, ensuring adaptability and stability.

[0083] Embodiment Two As shown in Figure 2 , the embodiment provides a steam turbine generator defect diagnosis and early warning device 200, which comprises: A data acquisition module 201 is configured to acquire running data of a steam turbine generator to be monitored. The working condition recognition module 202 is configured to recognize a current working condition of the to-be-monitored steam turbine generator according to the operation data, and generate a current working condition label. The model calling module 203 is configured to call a preset baseline model corresponding to the current working condition label, the preset baseline model corresponding to the current working condition label being a mathematical model established based on historical operation data and used to calculate a theoretical value of a defect diagnosis index under the current working condition. The defect diagnosis module 204 is configured to perform defect diagnosis analysis on the to-be-monitored steam turbine generator based on the preset baseline model corresponding to the current working condition label and the operation data, and generate a defect diagnosis result, the defect diagnosis result being at least one of normal operation, stator winding temperature abnormality, rotor inter-turn short circuit, and water-soluble hydrogen water electric joint leakage.

[0084] Optionally, in an embodiment, the preset baseline model corresponding to the current working condition label is obtained by: obtaining historical operation data corresponding to the current working condition label; performing linear regression fitting processing on the historical operation data to calibrate parameters in a pre-constructed initial baseline model, and obtaining the preset baseline model corresponding to the current working condition label; wherein the preset baseline model corresponding to the current working condition label includes a stator winding temperature rise baseline model, a rotor power deviation baseline model, and a hydrogen water pressure difference baseline model.

[0085] Optionally, in an embodiment, the defect diagnosis module 204 includes: A first theoretical value calculation unit is configured to calculate a stator winding temperature rise theoretical value according to the stator winding temperature rise baseline model and the operation data. A first deviation value calculation unit is configured to calculate a temperature deviation value according to the stator winding temperature rise theoretical value and the operation data, the temperature deviation value being a difference between a stator winding temperature rise measured value and the stator winding temperature rise theoretical value. A rate calculation unit is configured to calculate a temperature change rate according to the operation data. A first diagnosis analysis unit is configured to compare the temperature deviation value with a temperature deviation dynamic threshold value, and compare the temperature change rate with a rate threshold value, to perform stator winding temperature abnormality analysis on the to-be-monitored steam turbine generator.

[0086] Optionally, in an embodiment, the temperature deviation dynamic threshold value includes a temperature deviation warning threshold value and a temperature deviation emergency threshold value, and the rate threshold value includes a rate warning threshold value and a rate emergency threshold value; the first diagnosis analysis unit includes: A first comparison subunit is configured to compare the temperature deviation value with the temperature deviation warning threshold value and the temperature deviation emergency threshold value respectively, and compare the temperature change rate with the rate warning threshold value and the rate emergency threshold value respectively. The first-level temperature early warning unit is configured to determine that the temperature of the stator winding is abnormal and generate a first-level temperature early warning signal when the temperature deviation value is greater than the temperature deviation early warning threshold and lasts for the first preset time length, and the temperature change rate is greater than the rate early warning threshold and lasts for the first preset time length. The second-level temperature early warning unit is configured to determine that the temperature of the stator winding is abnormal and generate a second-level temperature early warning signal when the temperature deviation value is greater than the temperature deviation emergency threshold and lasts for the first preset time length, and the temperature change rate is greater than the rate emergency threshold and lasts for the first preset time length.

[0087] Optionally, in the embodiments, the defect diagnosis module 204 further includes: The second theoretical value calculation unit is configured to calculate the generator active power theoretical value at each time within the preset time window according to the rotor power deviation baseline model and the operation data; The second deviation value calculation unit is configured to calculate the power deviation value at each time according to the generator active power theoretical value at each time and the operation data, the power deviation value at each time being a difference between the generator active power measured value at the time and the generator active power theoretical value at the time; The distribution statistical unit is configured to perform distribution characteristic statistics according to the power deviation values at the times to obtain a statistical ratio, the statistical ratio representing a proportion of a number of times when the corresponding power deviation value is greater than the preset deviation value to a number of all times within the preset time window; The second diagnostic analysis unit is configured to compare the statistical ratio with a proportion threshold to perform rotor inter-turn short circuit analysis on the monitored steam turbine generator.

[0088] Optionally, in the embodiments, the proportion threshold includes a first proportion threshold and a second proportion threshold, and the preset deviation value includes a first deviation value and a second deviation value; the statistical ratio includes a first ratio and a second ratio, the first ratio representing a proportion of a number of times when the corresponding power deviation value is greater than the first deviation value to a number of all times within the preset time window, and the second ratio representing a proportion of a number of times when the corresponding power deviation value is greater than the second deviation value to the number of all times within the preset time window; and the second diagnostic analysis unit includes: The second comparison subunit is configured to compare the first ratio with the first proportion threshold and compare the second ratio with the second proportion threshold; The first-level short circuit early warning subunit is configured to determine the rotor inter-turn short circuit and generate a first-level short circuit early warning signal when the first ratio is greater than the first proportion threshold; The second-level short circuit early warning subunit is configured to determine the rotor inter-turn short circuit and generate a second-level short circuit early warning signal when the second ratio is greater than the second proportion threshold.

[0089] Optionally, in the embodiments, the defect diagnosis module 204 further includes: A third theoretical value calculation unit is configured to calculate a hydrogen water pressure difference theoretical value according to the hydrogen water pressure difference baseline model and the operation data; A third deviation value calculation unit is configured to calculate a pressure difference deviation value according to the hydrogen water pressure difference theoretical value and the operation data, the pressure difference deviation value being a difference between the hydrogen water pressure difference measured value and the hydrogen water pressure difference theoretical value; A leakage trend index calculation unit is configured to calculate a leakage trend index according to the pressure difference deviation value and the operation data; A third diagnosis analysis unit is configured to compare the leakage trend index with a leakage dynamic threshold value, and determine that the water-soluble hydrogen water electrical joint is leaking when the leakage trend index is greater than the leakage dynamic threshold value and lasts for a second preset time length.

[0090] Optionally, in an embodiment, the defect diagnosis module 204 further includes: A fourth theoretical value calculation unit is configured to calculate a hydrogen gas saturation concentration theoretical value according to the operation data based on the Henry's law; A fourth deviation value calculation unit is configured to calculate a concentration deviation value according to the hydrogen gas saturation concentration theoretical value and the operation data, the concentration deviation value being a difference between the dissolved hydrogen concentration measured value and the hydrogen gas saturation concentration theoretical value; A leakage intensity calculation unit is configured to estimate a hydrogen gas leakage amount according to the concentration deviation value and the operation data, and obtain a leakage intensity; A hydrogen gas leakage early warning unit is configured to generate a hydrogen gas leakage early warning signal according to the leakage intensity when it is determined that the water-soluble hydrogen water electrical joint is leaking.

[0091] Optionally, in an embodiment, the working condition recognition module 202 includes: A vector construction unit is configured to construct a working condition feature vector according to the operation data, the working condition feature vector being a vector reflecting a running state of the to-be-monitored steam turbine generator; A working condition matching unit is configured to determine a working condition cluster center closest to the working condition feature vector from a plurality of pre-generated working condition cluster centers, and take a working condition label corresponding to the working condition cluster center as a current working condition label.

[0092] In some embodiments, the steam turbine generator defect diagnosis and early warning device 200 of the present application can be realized in a combination of software and hardware. For example, the steam turbine generator defect diagnosis and early warning device 200 of the present application can be a hardware decoding processor programmed to execute the steam turbine generator defect diagnosis and early warning method of the present application. For example, the hardware decoding processor can be one or more application specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field programmable gate arrays (FPGAs), or other electronic components.

[0093] The modules described in the embodiments of the present application can be implemented in software or hardware. In some cases, the names of the modules do not limit the modules themselves.

[0094] The above description is merely preferred embodiments of the present application and a description of the principles of the technology used. Those skilled in the art should understand that the disclosed range of the present application is not limited to the technical solutions formed by the specific combinations of the above technical features, and should also cover other technical solutions formed by any combination of the above technical features or equivalent features without departing from the disclosed concept. For example, the above features can be replaced with similar features disclosed in the present application (but not limited to) to form technical solutions.

[0095] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, and represent a specific order or sequence. In appropriate cases, the order of similar objects can be interchanged, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described.

[0096] Those skilled in the art know that the present application can be implemented as a system, a method or a computer program product, therefore, the present application can be embodied in the form of a complete hardware, a complete software (including firmware, resident software, microcode, etc.), or a combination of hardware and software, which is generally referred to as "circuit", "module" or "system". In addition, in some embodiments, the present application can also be implemented as a computer program product in one or more computer readable media, which contains computer readable program codes.

[0097] Although the embodiments of the present application have been shown and described above, it should be understood by those skilled in the art that the above-mentioned embodiments are exemplary and cannot be understood as limiting the present application, and those skilled in the art can make changes, modifications, replacements and variations to the above-mentioned embodiments within the scope of the present application.

Claims

1. A method for diagnosing and warning defects of a steam turbine generator, characterized by, The method comprises: acquiring operation data of a to-be-monitored steam turbine generator; identifying a current operation condition of the to-be-monitored steam turbine generator according to the operation data, and generating a current condition label; calling a preset baseline model corresponding to the current condition label, the preset baseline model corresponding to the current condition label being a mathematical model established based on historical operation data and used to calculate a theoretical value of a defect diagnosis index in the current operation condition; performing defect diagnosis analysis on the to-be-monitored steam turbine generator based on the preset baseline model corresponding to the current condition label and the operation data, and generating a defect diagnosis result, the defect diagnosis result being at least one of normal operation, abnormal stator winding temperature, rotor inter-turn short circuit, and water-soluble hydrogen water joint leakage.

2. The method of claim 1, wherein the method further comprises: The preset baseline model corresponding to the current condition label is obtained by: acquiring historical operation data corresponding to the current condition label; performing linear regression fitting processing on the historical operation data to calibrate parameters in a pre-constructed initial baseline model, and obtaining the preset baseline model corresponding to the current condition label; The preset baseline model corresponding to the current condition label includes a stator winding temperature rise baseline model, a rotor power deviation baseline model, and a hydrogen water pressure difference baseline model.

3. The method of claim 2, wherein the method further comprises: The defect diagnosis analysis on the to-be-monitored steam turbine generator based on the preset baseline model corresponding to the current condition label and the operation data comprises: calculating a stator winding temperature rise theoretical value according to the stator winding temperature rise baseline model and the operation data; calculating a temperature deviation value according to the stator winding temperature rise theoretical value and the operation data, the temperature deviation value being a difference between a measured value of the stator winding temperature rise and the stator winding temperature rise theoretical value; calculating a temperature change rate according to the operation data; comparing the temperature deviation value with a temperature deviation dynamic threshold value and comparing the temperature change rate with a rate threshold value to perform abnormal stator winding temperature analysis on the to-be-monitored steam turbine generator.

4. The method of claim 3, wherein the method further comprises: The temperature deviation dynamic threshold value includes a temperature deviation early warning threshold value and a temperature deviation emergency threshold value, and the rate threshold value includes a rate early warning threshold value and a rate emergency threshold value; the comparison of the temperature deviation value with the temperature deviation dynamic threshold value and the comparison of the temperature change rate with the rate threshold value to perform abnormal stator winding temperature analysis on the to-be-monitored steam turbine generator comprises: comparing the temperature deviation value with the temperature deviation early warning threshold value and the temperature deviation emergency threshold value respectively, and comparing the temperature change rate with the rate early warning threshold value and the rate emergency threshold value respectively; when the temperature deviation value is greater than the temperature deviation early warning threshold value and lasts for a first preset time length, and the temperature change rate is greater than the rate early warning threshold value and lasts for a first preset time length, determining abnormal stator winding temperature and generating a first-level temperature early warning signal; when the temperature deviation value is greater than the temperature deviation emergency threshold value and lasts for a first preset time length, and the temperature change rate is greater than the rate emergency threshold value and lasts for a first preset time length, determining abnormal stator winding temperature and generating a second-level temperature early warning signal.

5. The method of claim 2, wherein the method further comprises: The defect diagnosis analysis on the to-be-monitored steam turbine generator based on the preset baseline model corresponding to the current working condition label and the operation data further comprises: According to the rotor power deviation baseline model and the operation data, the active power theoretical value of the generator at each time within a preset time window is calculated; According to the active power theoretical value of the generator at each time and the operation data, the power deviation value at each time is calculated respectively, and the power deviation value at each time is the difference between the active power measured value of the generator at the time and the active power theoretical value of the generator at the time; According to the power deviation value at each time, the distribution characteristics are statistically analyzed to obtain a statistical ratio, and the statistical ratio represents the proportion of the number of times when the corresponding power deviation value is greater than a preset deviation value to the number of all times within the preset time window; The statistical ratio is compared with a proportion threshold value to analyze the rotor inter-turn short circuit of the to-be-monitored steam turbine generator.

6. The method of claim 5, wherein the method further comprises: The proportion threshold value includes a first proportion threshold value and a second proportion threshold value, and the preset deviation value includes a first deviation value and a second deviation value; the statistical ratio includes a first ratio and a second ratio, the first ratio represents the proportion of the number of times when the corresponding power deviation value is greater than the first deviation value to the number of all times within the preset time window, and the second ratio represents the proportion of the number of times when the corresponding power deviation value is greater than the second deviation value to the number of all times within the preset time window; The comparison of the statistical ratio with the proportion threshold value to analyze the rotor inter-turn short circuit of the to-be-monitored steam turbine generator comprises: The first ratio is compared with the first proportion threshold value, and the second ratio is compared with the second proportion threshold value; When the first ratio is greater than the first proportion threshold value, it is determined that there is a rotor inter-turn short circuit and a first-level short circuit early warning signal is generated; When the second ratio is greater than the second proportion threshold value, it is determined that there is a rotor inter-turn short circuit and a second-level short circuit early warning signal is generated.

7. The method of claim 2, wherein the method further comprises: The defect diagnosis analysis on the to-be-monitored steam turbine generator based on the preset baseline model corresponding to the current working condition label and the operation data further comprises: According to the hydrogen-water pressure difference baseline model and the operation data, a hydrogen-water pressure difference theoretical value is calculated; According to the hydrogen-water pressure difference theoretical value and the operation data, a pressure difference deviation value is calculated, and the pressure difference deviation value is the difference between the hydrogen-water pressure difference measured value and the hydrogen-water pressure difference theoretical value; According to the pressure difference deviation value and the operation data, a leakage trend index is calculated; The leakage trend index is compared with a leakage dynamic threshold value, and when the leakage trend index is greater than the leakage dynamic threshold value and lasts for a second preset time length, it is determined that there is a water-soluble hydrogen-water electrical joint leakage.

8. The method of claim 7, wherein the method further comprises: The defect diagnosis analysis on the to-be-monitored steam turbine generator based on the preset baseline model corresponding to the current working condition label and the operation data further comprises: According to the operation data, a hydrogen saturation concentration theoretical value is calculated based on Henry's law; According to the hydrogen saturation concentration theoretical value and the operation data, a concentration deviation value is calculated, the concentration deviation value being a difference between a measured value of the dissolved hydrogen concentration and the hydrogen saturation concentration theoretical value; According to the concentration deviation value and the operation data, hydrogen leakage amount estimation is performed to obtain a leakage intensity; When it is determined that the water-soluble hydrogen water-electric joint is leaking, a hydrogen leakage early warning signal is generated according to the leakage intensity.

9. The method of claim 1, wherein the method further comprises: The current operation condition of the to-be-monitored steam turbine generator is identified according to the operation data, and a current condition label is generated, including: A condition feature vector is constructed according to the operation data, the condition feature vector being a vector reflecting the operation state of the to-be-monitored steam turbine generator; Among a plurality of pre-generated condition cluster centers, a condition cluster center closest to the condition feature vector is determined, and a condition label corresponding to the condition cluster center is taken as the current condition label.

10. A steam turbine generator defect diagnosis and early warning device, characterized in that, It includes: A data acquisition module is configured to acquire operation data of a to-be-monitored steam turbine generator; A condition identification module is configured to identify a current operation condition of the to-be-monitored steam turbine generator according to the operation data, and generate a current condition label; A model calling module is configured to call a preset baseline model corresponding to the current condition label, the preset baseline model corresponding to the current condition label being a mathematical model established based on historical operation data and used to calculate a theoretical value of a defect diagnosis index under the current operation condition; A defect diagnosis module is configured to perform defect diagnosis analysis on the to-be-monitored steam turbine generator based on the preset baseline model corresponding to the current condition label and the operation data, and generate a defect diagnosis result, the defect diagnosis result being at least one of normal operation, abnormal stator winding temperature, rotor turn-to-turn short circuit, and water-soluble hydrogen water-electric joint leakage.

Images