Hydroelectric generating set fault diagnosis method and system based on digital twinning

Abstract

The invention discloses a hydroelectric generating set fault diagnosis method and system based on digital twinning, and relates to the technical field of digital twinning. Related data of a hydroelectric generating set are collected and preprocessed; constructing a three-dimensional twinborn model; constructing a fault diagnosis model by combining simulation data generated by the three-dimensional twin model, the preprocessed related data of the hydroelectric generating set and historical fault data; and carrying out analogue simulation and accident inversion, predicting the operation trend of the hydroelectric generating set equipment by combining the output of the fault diagnosis model, and carrying out hydroelectric generating set fault early warning based on the prediction result. According to the method, the comprehensive unbalance amount, the machine learning predicted value and the expert experience fusion result are integrated and calculated, the three-dimensional twinborn model is introduced, a multi-dimensional and multi-level fault diagnosis model is constructed, the fault diagnosis model output and the three-dimensional twinborn model simulation result are combined, a scientific basis is provided for formulating a preventive maintenance strategy, and the method is suitable for popularization and application. And the reliability and the operation efficiency of the equipment are obviously improved.

Description

Technical Field

[0001] The present invention relates to the technical field of digital twins, and in particular to a method and system for diagnosing faults of hydropower units based on digital twins. Background Art

[0002] With the use of online monitoring systems, hydropower units collect a large amount of signal data during operation and store it in the database server. How to extract useful knowledge from these complex and seemingly unrelated data to guide the safe and stable operation of power generation equipment has become a research topic, and data mining technology has emerged. The various types and dimensions of data obtained from online monitoring, regular monitoring and irregular monitoring are monitored through data mining technology to monitor their internal regular characteristics, and deeply analyze the operating data of hydropower unit components under different working conditions, including collecting historical operating data of hydropower unit components within a preset period of time. According to different working conditions, the historical operating data is divided into several sets, and the data is graded and classified through the model algorithm and data fusion is performed, and the fixed threshold diagnosis is changed, so as to achieve real-time dynamic fault diagnosis of hydropower equipment and better reflect the actual situation of the unit.

[0003] The failure of a hydro-turbine generator set is characterized by complex mechanisms, polymorphism, and fuzzy and uncertain logic. The traditional fault diagnosis method relies on the work experience of maintenance personnel, and obtains available information about the object being tested and extracts characteristic parameters to determine the equipment status. This results in slow fault diagnosis feedback, a lack of three-dimensional dynamic visualization of the equipment health status display, and inability to guarantee accuracy, making it difficult to directly analyze the cause of the failure.

[0004] As an intelligent and information-based technology, digital twin technology currently creates virtual models of physical entities in a digital way, simulates the behavior of physical entities in the real environment with the help of data, and can achieve dynamic and three-dimensional perception of the actual state of physical entities through virtual-real two-way mapping and interactive feedback, and realize functions such as state detection and fault diagnosis of physical entity equipment, making up for the lack of real-time collaboration and unfriendly human-computer interaction in the traditional fault diagnosis result feedback process, and providing solid technical support for remote digital operation and maintenance and intelligent maintenance decision-making of hydropower units. Summary of the invention

[0005] In view of the problems that the existing fault diagnosis method of hydropower units relies on the work experience of maintenance personnel, the accuracy cannot be guaranteed, and there is a lack of three-dimensional visualization, the present invention is proposed.

[0006] Therefore, the problem to be solved by the present invention is to construct a three-dimensional twin model and a fault diagnosis model, perform simulation and accident inversion, combine the fault diagnosis results, realize three-dimensional visualization, and perform accurate fault warning and trend analysis.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In the first aspect, an embodiment of the present invention provides a hydropower unit fault diagnosis method based on digital twins, which includes collecting and preprocessing relevant data of the hydropower unit; constructing a three-dimensional twin model based on the preprocessed relevant data of the hydropower unit; constructing a fault diagnosis model by combining the simulation data generated by the three-dimensional twin model, the preprocessed relevant data of the hydropower unit and the historical fault data; using the three-dimensional twin model to perform simulation and accident inversion, and combining the output of the fault diagnosis model to predict the operation trend of the hydropower unit equipment, and providing hydropower unit fault warning based on the prediction results.

[0009] As a preferred solution of the hydropower unit fault diagnosis method based on digital twin of the present invention, the construction of the three-dimensional twin model includes mapping the pre-processed hydropower unit related data into the three-dimensional space, and the mapping formula is as follows:

[0010]

[0011] Where F(·) represents the multivariate mapping function, α i represents the weighting coefficient, represents the standardized data of the hydropower unit, x, y, z are spatial coordinates, u, v, w are integral variables, σ ui ,σ vi ,σ wi are the standard deviations of the data of each subcomponent in the mapping function; the formula for generating a three-dimensional twin model through a multivariate mapping function is as follows:

[0012]

[0013] Among them, T(x,y,z) represents the output of the three-dimensional twin model, x, y, z are spatial coordinates, and x′, y′, z′ all represent integral variables.

[0014] As an optimal solution of the hydropower unit fault diagnosis method based on digital twin of the present invention, wherein: constructing a fault diagnosis model includes calculating the comprehensive unbalance amount, machine learning prediction value and expert experience fusion results, integrating the calculation of the comprehensive unbalance amount, machine learning prediction value and expert experience fusion results, and introducing a three-dimensional twin model to obtain a fault diagnosis model.

[0015] As a preferred solution of the hydropower unit fault diagnosis method based on digital twin of the present invention, the comprehensive unbalance amount, machine learning prediction value and expert experience fusion result are integrated and calculated, and a three-dimensional twin model is introduced to obtain a fault diagnosis model including obtaining magnetic pull unbalance data at different time points, rotor mass unbalance data at different time points and hydraulic unbalance data at different time points, calculating the magnetic pull unbalance amount, rotor mass unbalance amount and hydraulic unbalance amount, normalizing the three unbalance amounts, taking the square of the average value, and calculating the square root to obtain the comprehensive unbalance amount. The calculation formula of the comprehensive unbalance amount is as follows:

[0016]

[0017] Where V represents the comprehensive unbalance, N represents the number of magnetic pull unbalance data, and k mi represents the proportionality coefficient of the magnetic pull unbalance data of the i-th line, B 1i Indicates the first magnetic induction intensity in the i-th data, B 2i The second magnetic induction intensity in the i-th data, M represents the number of rotor mass imbalance data, k rj represents the proportionality factor of the jth rotor mass imbalance data, m j represents the mass in the jth rotor mass imbalance data, e j represents the eccentricity in the jth rotor mass imbalance data, P represents the number of hydraulic imbalance data, k hk represents the proportionality coefficient of the kth hydraulic imbalance data, P 1k Indicates the first pressure in the kth data, P 2k Indicates the second pressure in the kth data; standardizes the historical data feature vector using the standard deviation, sets weights for the historical data feature vector, calculates the influence value of each feature using the attenuation parameter, calculates the loss function gradient of the machine learning model, weights the standardized feature influence values, and calculates the fault prediction value. The calculation formula for the fault prediction value is as follows:

[0018]

[0019] in, represents the fault prediction value, N represents the number of features, and w i represents the weight of the i-th feature, σ(x i ) represents the standard deviation normalization function of the i-th feature, x i represents the value of the i-th feature, α i represents the decay parameter of the i-th feature, M represents the number of features for gradient calculation in the machine learning model, and f j represents the function value of the jth feature, represents the gradient of the loss function of the parameter corresponding to the jth feature, and C represents the regularization constant; collect expert experience rules, determine the output value of the rule and the corresponding weight, calculate the influence value of each rule, obtain the time feature, use the attenuation parameter to calculate the influence value of the time feature, and sum the weighted influence values ​​of the rule and the time feature to obtain the expert experience fusion result. The calculation formula of the expert experience fusion result is as follows:

[0020]

[0021] Among them, E represents the result of expert experience fusion, N represents the number of expert rules, and R i represents the weight of the i-th rule, E i is the output of the i-th rule, β i represents the parameter of the i-th rule, M represents the number of time-related features, and γ i represents the weight of the jth time feature, ω j represents the frequency of the jth time feature, t j represents the time of the jth time feature, κ j represents the attenuation parameter of the jth time feature; based on the fusion results of comprehensive imbalance, fault prediction value and expert experience, the three-dimensional twin model is introduced, and the formula of the fault diagnosis model is as follows:

[0022]

[0023] Where D represents the fault diagnosis result, V represents the comprehensive unbalance value, represents the fault prediction value, E represents the expert experience fusion result, T(x, y, z) represents the output of the 3D twin model, N represents the number of data points, α i represents the attenuation parameter of the ith data point, t i The time of the i-th data point.

[0024] As a preferred solution of the hydropower unit fault diagnosis method based on digital twin of the present invention, wherein: using the three-dimensional twin model for simulation and accident inversion includes simulating the normal operating state of the hydropower unit in the three-dimensional twin model, and establishing the baseline data of the hydropower unit in the normal operating state; adjusting the parameters of the three-dimensional twin model, and observing the response of the three-dimensional twin model; simulating the operating state of the three-dimensional twin model under different working conditions, recording and analyzing the simulation results, and simulating various fault conditions in the three-dimensional twin model based on historical fault data; adjusting fault-related parameters, analyzing the impact of the fault on various parts of the system, and recording the fault characteristics and development process.

[0025] As a preferred solution of the hydropower unit fault diagnosis method based on digital twin of the present invention, wherein: combining the output of the fault diagnosis model, predicting the operation trend of the hydropower unit equipment, including obtaining the output of the fault diagnosis model, judging the current operation status and fault type of the hydropower unit equipment according to the output of the fault diagnosis model; comparing the output of the fault diagnosis model with the benchmark data of the current equipment in the normal state to determine the severity of the fault; collecting the output of the fault diagnosis model within a set time, analyzing the changing trend of the output of the fault diagnosis model, judging the development direction of the fault, and completing the prediction of the operation trend of the hydropower unit equipment.

[0026] As a preferred solution of the hydropower unit fault diagnosis method based on digital twin of the present invention, wherein: hydropower unit fault warning based on the prediction results includes setting warning thresholds for the faults of each equipment of the hydropower unit based on the 3σ criterion; comparing the prediction results of the operation trend of the hydropower unit equipment with the warning threshold, judging whether the prediction results exceed the set warning value, and determining the warning level; generating corresponding warning information according to the warning level, and displaying the warning information in three-dimensional visualization; formulating response strategies according to the warning level and fault type, and sending the warning information and response strategies to relevant personnel and departments.

[0027] In the second aspect, an embodiment of the present invention provides a hydropower unit fault diagnosis system based on digital twins, which includes an acquisition module for collecting and preprocessing hydropower unit related data; a first model construction module for constructing a three-dimensional twin model based on the preprocessed hydropower unit related data; a second model construction module for combining the simulation data generated by the three-dimensional twin model, the preprocessed hydropower unit related data and historical fault data to construct a fault diagnosis model; an early warning module for using the three-dimensional twin model to perform simulation and accident inversion, combined with the output of the fault diagnosis model, to predict the operation trend of the hydropower unit equipment, and to provide a hydropower unit fault early warning based on the prediction results.

[0028] In a third aspect, an embodiment of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of the hydropower unit fault diagnosis method based on digital twins as in the first aspect of the present invention are implemented.

[0029] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, the steps of the hydropower unit fault diagnosis method based on digital twins as in the first aspect of the present invention are implemented.

[0030] The beneficial effects of the present invention are as follows: by integrating and calculating the comprehensive imbalance quantity, the machine learning prediction value and the fusion result of expert experience, and introducing the three-dimensional twin model, a multi-dimensional and multi-level fault diagnosis model is constructed, which fully utilizes the advantages of data-driven and knowledge-driven, significantly improves the accuracy and reliability of fault diagnosis, and can more comprehensively capture various potential faults. By combining the output of the fault diagnosis model and the simulation results of the three-dimensional twin model, using data mining and machine learning algorithms, an accurate prediction of the future operating status of the equipment is achieved, so that the operation and maintenance personnel can predict potential problems in advance, providing a scientific basis for the formulation of preventive maintenance strategies, and significantly improving the reliability and operating efficiency of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0032] Figure 1 This is a flow chart of the fault diagnosis method for hydropower units based on digital twins.

[0033] Figure 2 Digital twin data flow diagram of the hydropower unit fault diagnosis method based on digital twin.

[0034] Figure 3 This is the system architecture diagram of the hydropower unit fault diagnosis method based on digital twin.

[0035] Figure 4 It is the application layer of the hydropower unit fault diagnosis method based on digital twin. DETAILED DESCRIPTION

[0036] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0037] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0038] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or selective embodiment that is mutually exclusive with other embodiments.

[0039] Example 1

[0040] Reference Figure 1 to Figure 4 , which is the first embodiment of the present invention, and provides a fault diagnosis method for a hydropower unit based on digital twins, comprising:

[0041] S1: Collect and pre-process relevant data of hydropower units.

[0042] Specifically, hardware equipment and sensors are arranged, and the hardware equipment includes water-guided bearings, flow-through components, speed control system, excitation system, braking system, stator and rotor, upper guide bearings, upper frame, thrust bearings and lower frame, etc.

[0043] Furthermore, the collected hydropower unit related data include fixed parameter data, variable data sets (such as the set of current, voltage, and frequency during the start-up and shutdown processes; high-speed sampling data sets such as oscillation waveforms), real-time data, historical data, and recorded data.

[0044] Furthermore, the collected data is stored in the modes of structured data, unstructured data, message data, platform basic data, etc.

[0045] Furthermore, redundant data from multiple systems are integrated, duplicate data is eliminated, and merge rules are set to merge data from multiple data sources to ensure data consistency.

[0046] S2: Construct a three-dimensional twin model based on the preprocessed data of the hydropower unit.

[0047] Specifically, digital twin technology is used to carry out three-dimensional modeling of hydropower equipment, and to develop a three-dimensional image engine and a twin algorithm engine.

[0048] Furthermore, a three-dimensional digital twin model of the hydropower unit is constructed to realize real-time feedback simulation of the digital twin on the physical entity, while realizing real-time interaction and mutual mapping between the digital twin and the physical entity.

[0049] Furthermore, constructing the three-dimensional twin model includes mapping the preprocessed hydropower unit related data into the three-dimensional space. The mapping formula is as follows:

[0050]

[0051] Where F(·) represents the multivariate mapping function, αi represents the weighting coefficient, represents the standardized data of the hydropower unit, x, y, z are spatial coordinates, u, v, w are integral variables, σ ui ,σ vi ,σ wi are the standard deviations of each subcomponent data in the mapping function.

[0052] Preferably, high-dimensional data is mapped to three-dimensional space for intuitive understanding and analysis. The use of multivariate mapping functions can capture the nonlinear relationship between data, and the weighting coefficients can be used to highlight important features and suppress minor features.

[0053] Furthermore, the formula for generating a three-dimensional twin model through a multivariate mapping function is as follows:

[0054]

[0055] Among them, T(x,y,z) represents the output of the three-dimensional twin model, x, y, z are spatial coordinates, and x′, y′, z′ all represent integral variables.

[0056] Preferably, the output of the three-dimensional twin model is directly generated to facilitate subsequent analysis and visualization; through integration, the continuity and smoothness of the three-dimensional twin model in space are ensured.

[0057] S3: Use the simulation data generated by the 3D twin model, combined with the processed hydropower unit-related data and historical fault data, to build a fault diagnosis model.

[0058] Specifically, constructing a fault diagnosis model includes calculating the comprehensive imbalance amount, machine learning prediction values, and expert experience fusion results, integrating the calculation of the comprehensive imbalance amount, machine learning prediction values, and expert experience fusion results, and introducing a three-dimensional twin model to obtain a fault diagnosis model.

[0059] Furthermore, the magnetic pull imbalance data at different time points, the rotor mass imbalance data at different time points and the hydraulic imbalance data at different time points are obtained, and the magnetic pull imbalance, rotor mass imbalance and hydraulic imbalance are calculated. After the three imbalances are normalized, the square sum of the average values ​​is taken, and the square root is calculated to obtain the comprehensive imbalance. The calculation formula of the comprehensive imbalance is as follows:

[0060]

[0061] Where V represents the comprehensive unbalance, N represents the number of magnetic pull unbalance data, and k mi represents the proportionality coefficient of the magnetic pull unbalance data of the i-th line, B 1i Indicates the first magnetic induction intensity in the i-th data, B 2iThe second magnetic induction intensity in the i-th data, M represents the number of rotor mass imbalance data, k rj represents the proportionality factor of the jth rotor mass imbalance data, m j represents the mass in the jth rotor mass imbalance data, e j represents the eccentricity in the jth rotor mass imbalance data, P represents the number of hydraulic imbalance data, k hk represents the proportionality coefficient of the kth hydraulic imbalance data, P 1k Indicates the first pressure in the kth data, P 2k Indicates the second pressure in the kth data.

[0062] Preferably, three unbalance factors, namely magnetic pull, rotor mass and hydraulic force, are considered simultaneously, which is helpful to comprehensively evaluate the operating status and potential problems of the hydropower unit; data at different time points are considered, which can reflect the dynamic changes of imbalance.

[0063] Furthermore, the historical data is converted into a historical data feature vector, the historical data feature vector is standardized using the standard deviation, and a weight is set for the historical data feature vector. The influence value of each feature is calculated using the attenuation parameter, the loss function gradient of the machine learning model is calculated, the standardized feature influence values ​​are weighted and summed, and the fault prediction value is calculated. The calculation formula of the fault prediction value is as follows:

[0064]

[0065] in, represents the fault prediction value, N represents the number of features, and w i represents the weight of the i-th feature, σ(x i ) represents the standard deviation normalization function of the i-th feature, x i represents the value of the i-th feature, α i represents the decay parameter of the i-th feature, M represents the number of features for gradient calculation in the machine learning model, and f j represents the function value of the jth feature, represents the gradient of the loss function of the parameter corresponding to the jth feature, and C represents the regularization constant.

[0066] Preferably, multiple features and the output of the machine learning model are combined, and the training information of the model is introduced using the gradient of the loss function. By considering feature weights, normalization and attenuation parameters, possible failures can be predicted more accurately.

[0067] Furthermore, expert experience rules are collected, the output values ​​and corresponding weights of the rules are determined, the influence value of each rule is calculated, the time feature is obtained, the influence value of the time feature is calculated using the attenuation parameter, and the influence values ​​of the rules and time features are weighted and summed to obtain the expert experience fusion result. The calculation formula of the expert experience fusion result is as follows:

[0068]

[0069] Among them, E represents the result of expert experience fusion, N represents the number of expert rules, and R i represents the weight of the i-th rule, E i is the output of the i-th rule, β i represents the parameter of the i-th rule, M represents the number of time-related features, and γ i represents the weight of the jth time feature, ω j represents the frequency of the jth time feature, t j represents the time of the jth time feature, κ j Represents the attenuation parameter of the j-th time feature.

[0070] Preferably, the calculation formula of the expert experience fusion result combines expert rules and time characteristics, and integrates expert experience and time factors into fault diagnosis by weighted summation, which can effectively improve the accuracy and reliability of diagnosis.

[0071] Furthermore, based on the fusion results of comprehensive imbalance, fault prediction value and expert experience, a three-dimensional twin model is introduced, and the formula of the fault diagnosis model is obtained as follows:

[0072]

[0073] Where D represents the fault diagnosis result, V represents the comprehensive unbalance value, represents the fault prediction value, E represents the expert experience fusion result, T(x, y, z) represents the output of the 3D twin model, N represents the number of data points, α i represents the attenuation parameter of the ith data point, t i The time of the i-th data point.

[0074] The optimal fault diagnosis model comprehensively considers the imbalance, fault prediction value, expert experience fusion results and three-dimensional twin model, considers the change of diagnosis results over time through time series analysis, introduces attenuation parameters to give higher weight to recent data, and captures the interaction between various factors through nonlinear combination. Taking multiple factors into consideration, it provides comprehensive and accurate fault diagnosis results.

[0075] S4: Use the three-dimensional twin model for simulation and accident inversion, combine it with the output of the fault diagnosis model, predict the operation trend of the hydropower unit equipment, and provide hydropower unit fault warning based on the prediction results.

[0076] Specifically, the use of the three-dimensional twin model for simulation and accident inversion includes: simulating the normal operating status of the hydropower unit in the three-dimensional twin model, and establishing baseline data for the normal operating status of the hydropower unit; adjusting the parameters of the three-dimensional twin model and observing the response of the three-dimensional twin model; simulating the operating status of the three-dimensional twin model under different working conditions, recording and analyzing the simulation results, and simulating various fault conditions in the three-dimensional twin model based on historical fault data; adjusting fault-related parameters, analyzing the impact of the fault on various parts of the system, and recording the fault characteristics and development process.

[0077] Furthermore, combined with the output of the fault diagnosis model, the operating trend of the hydropower unit equipment is predicted, including obtaining the output of the fault diagnosis model, judging the current operating status and fault type of the hydropower unit equipment according to the output of the fault diagnosis model; comparing the output of the fault diagnosis model with the benchmark data of the current equipment in a normal state to determine the severity of the fault; collecting the output of the fault diagnosis model within a set time, analyzing the changing trend of the output of the fault diagnosis model, judging the development direction of the fault, and completing the prediction of the operating trend of the hydropower unit equipment.

[0078] Furthermore, the fault warning of the hydropower unit based on the prediction results includes setting a warning threshold for each equipment failure of the hydropower unit based on the 3σ criterion; comparing the prediction results of the hydropower unit equipment operation trend with the warning threshold to determine whether the prediction results exceed the set warning value and determine the warning level; generating corresponding warning information according to the warning level, and displaying the warning information in a three-dimensional visualization; formulating a response strategy according to the warning level and the fault type, and sending the warning information and response strategy to relevant personnel and departments.

[0079] Furthermore, the 3σ criterion is an important concept in statistics, which is mainly used to describe the distribution of data in a normal distribution (Gaussian distribution). This criterion describes the probability of data falling within different standard deviation ranges in a normal distribution: about 68% of the data fall within one standard deviation (1σ) of the mean; about 95% of the data fall within two standard deviations (2σ) of the mean; and about 99.7% of the data fall within three standard deviations (3σ) of the mean.

[0080] Furthermore, the three-dimensional visualization display includes providing a visual human-computer interaction interface and displaying the results through legends, reports, three-dimensional and statistical analysis.

[0081] Furthermore, status monitoring is performed, thresholds of performance parameters of hydropower units are set, simulation data are compared, and the results are reflected in the monitoring and early warning interface.

[0082] Furthermore, energy efficiency analysis is conducted to analyze the core energy efficiency indicators of the unit (turbine efficiency, water consumption rate, guide vane leakage rate, water diversion loss rate, and water energy utilization rate). Energy efficiency evaluation rules for the unit are established to evaluate the energy efficiency status of the unit, and to calculate and analyze the energy indicators, stability indicators, and temperature indicators of the unit.

[0083] Furthermore, operation optimization is carried out, unit safety operation analysis is performed, unit economic operation optimization is performed, and the priority operation sequence of the hydropower units is determined.

[0084] Furthermore, maintenance decisions are made, and screening decisions are made on the maintenance plans generated under fault conditions. Based on the results of equipment performance analysis and health analysis, risk assessment is conducted on the impact of equipment performance degradation and health status degradation, and maintenance decision recommendations are given.

[0085] Furthermore, the information visualization system with virtual reality display and video interaction provides the operators with the latest operating status characteristics of the unit, ensuring that production duty, maintenance personnel and relevant leaders can immediately receive accident-related information when an accident occurs.

[0086] Furthermore, this embodiment also provides a hydropower unit fault diagnosis system based on digital twins, comprising:

[0087] The acquisition module is used to collect and pre-process the relevant data of the hydropower unit;

[0088] The first model building module is used to build a three-dimensional twin model based on the pre-processed hydropower unit related data;

[0089] The second model building module is used to build a fault diagnosis model by combining the simulation data generated by the three-dimensional twin model, the pre-processed hydropower unit related data and the historical fault data;

[0090] The early warning module is used to use the three-dimensional twin model for simulation and accident inversion, combine the output of the fault diagnosis model, predict the operating trend of the hydropower unit equipment, and provide hydropower unit fault early warning based on the prediction results.

[0091] This embodiment also provides a computer device, which is suitable for the case of a hydropower unit fault diagnosis method based on digital twins, including a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the hydropower unit fault diagnosis method based on digital twins as proposed in the above embodiment.

[0092] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0093] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for diagnosing faults of hydropower units based on digital twins as proposed in the above embodiment is implemented.

[0094] Example 2

[0095] Reference Figure 1 to Figure 4 , which is the second embodiment of the present invention, and this embodiment provides a hydropower unit fault diagnosis method based on digital twins. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0096] Specifically, a six-month experiment was conducted at a large hydropower station. The hydropower station has an installed capacity of 1,200MW and a total of four 300MW hydropower generating units. The four units were continuously monitored and analyzed. The following table shows the key parameters and fault diagnosis results of one of the units at different time points;

[0097] Table 1 Experimental data of fault diagnosis of unit 1 of the hydro-turbine generator set

[0098]

[0099] Furthermore, from the data in Table 1, it can be seen that from time points T1 to T6, the comprehensive unbalance value showed a clear upward trend, increasing from 0.452 to 0.785, an increase of 73.7%. This indicates that the balance state of the unit is deteriorating and there may be potential faults. The temperature of the upper guide bearing increased from 62.3℃ to 75.6℃, an increase of 21.3%. The continuous increase in temperature may be related to the deterioration of the bearing lubrication condition or the increase in load, which requires attention.

[0100] Furthermore, the stator current increased from 8750A to 9010A, while the speed decreased from 136.4rpm to 133.9rpm. This change pattern may indicate an increase in unit load or a decrease in efficiency, which may be related to turbine blade wear or generator stator winding problems. The fault prediction value and the expert experience value show good consistency, and the difference between the two is always kept within 0.04. This proves the effectiveness of the machine learning model and expert experience fusion method in the present invention.

[0101] Furthermore, at time T3, the fault diagnosis result reached 0.27, exceeding the preset warning threshold of 0.25. This issued an early warning about 48 hours earlier than the traditional method, providing sufficient time for operation and maintenance personnel to take preventive measures.

[0102] Furthermore, at time T6, the fault diagnosis result reached 0.88, indicating that the unit was at a high risk of failure. Subsequent maintenance confirmed that the thrust bearing of the unit was severely worn, which was highly consistent with the diagnosis result.

[0103] Furthermore, by integrating the comprehensive imbalance, machine learning predictions, and expert experience, this method gives a higher failure risk assessment of 0.44 at time point T4, identifying potential problems earlier than relying solely on a single indicator such as comprehensive imbalance or temperature.

[0104] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A fault diagnosis method for hydropower units based on digital twins, characterized by: include, Collect relevant data of hydropower units and pre-process them; Constructing a three-dimensional twin model according to the preprocessed data related to the hydropower unit; Building a fault diagnosis model by combining the simulation data generated by the three-dimensional twin model, the pre-processed data related to the hydropower unit and the historical fault data; The three-dimensional twin model is used to perform simulation and accident inversion, and combined with the output of the fault diagnosis model, the operation trend of the hydropower unit equipment is predicted, and the hydropower unit fault warning is performed based on the prediction results.

2. The hydropower unit fault diagnosis method based on digital twin according to claim 1, characterized in that: The constructing of the three-dimensional twin model comprises: The pre-processed data related to the hydropower unit is mapped to the three-dimensional space, and the mapping formula is as follows: Where F(·) represents the multivariate mapping function; α i represents the weighting coefficient; represents the standardized hydropower unit related data; x, y, z are spatial coordinates; u, v, w are integral variables; σ ui ,σ vi ,σ wi are the standard deviations of each subcomponent data in the mapping function; The formula for generating a three-dimensional twin model through a multivariate mapping function is as follows: Among them, T(x,y,z) represents the output of the three-dimensional twin model, x, y, z are spatial coordinates, and x′, y′, z′ all represent integral variables.

3. The hydropower unit fault diagnosis method based on digital twin according to claim 2 is characterized in that: The construction of the fault diagnosis model includes calculating the comprehensive imbalance amount, the machine learning prediction value and the expert experience fusion result, integrating the calculation of the comprehensive imbalance amount, the machine learning prediction value and the expert experience fusion result, and introducing a three-dimensional twin model to obtain the fault diagnosis model.

4. The hydropower unit fault diagnosis method based on digital twin according to claim 3 is characterized in that: The integrated calculation of the comprehensive imbalance, the machine learning prediction value and the expert experience fusion result, and the introduction of the three-dimensional twin model, obtains the fault diagnosis model including: Obtain the magnetic force imbalance data at different time points, the rotor mass imbalance data at different time points, and the hydraulic imbalance data at different time points, calculate the magnetic force imbalance, rotor mass imbalance, and hydraulic imbalance, and after normalizing the three imbalances, take the square sum of the average values ​​and calculate the square root to obtain the comprehensive imbalance. The calculation formula for the comprehensive imbalance is as follows: Where V represents the comprehensive unbalance value, N represents the number of magnetic pull unbalance data, and k mi represents the proportionality coefficient of the ith magnetic pull unbalance data, B 1i Indicates the first magnetic induction intensity in the i-th data, B 2i The second magnetic induction intensity in the i-th data, M represents the number of rotor mass imbalance data, k rj Represents the proportionality coefficient of the jth rotor mass imbalance data, m j represents the mass in the jth rotor mass imbalance data, e j represents the eccentricity in the jth rotor mass imbalance data, P represents the number of hydraulic imbalance data, k hk represents the proportionality coefficient of the kth hydraulic imbalance data, P 1k Indicates the first pressure in the kth data, P 2k Indicates the second pressure in the k-th data; Standardize the historical data feature vectors and set weights for them. Use the attenuation parameter to calculate the impact value of each feature, calculate the loss function gradient of the machine learning model, and sum the weighted impact values ​​of the standardized features to calculate the fault prediction value. The calculation formula for the fault prediction value is as follows: in, represents the fault prediction value, N represents the number of features, and w i represents the weight of the i-th feature, σ(x i ) represents the standard deviation normalization function of the i-th feature, x i represents the value of the i-th feature, α i represents the decay parameter of the i-th feature, M represents the number of features for gradient calculation in the machine learning model, and f j represents the function value of the jth feature, represents the gradient of the loss function of the parameter corresponding to the jth feature, and C represents the regularization constant; Collect expert experience rules, determine the output value and corresponding weight of the rules, calculate the influence value of each rule, obtain time features, use the attenuation parameter to calculate the influence value of time features, and sum the weighted influence values ​​of rules and time features to obtain the expert experience fusion result. The calculation formula of the expert experience fusion result is as follows: Among them, E represents the result of expert experience fusion, N represents the number of expert rules, and R i represents the weight of the i-th rule, E i is the output of the i-th rule, β i represents the parameter of the i-th rule, M represents the number of time-related features, and γ i represents the weight of the jth time feature, ω j represents the frequency of the jth time feature, t j represents the time of the jth time feature, κ j represents the attenuation parameter of the j-th time feature; Based on the fusion results of comprehensive imbalance, fault prediction value and expert experience, the three-dimensional twin model is introduced, and the formula of the fault diagnosis model is obtained as follows: Where D represents the fault diagnosis result, V represents the comprehensive unbalance value, represents the fault prediction value, E represents the expert experience fusion result, T(x, y, z) represents the output of the 3D twin model, N represents the number of data points, α i represents the attenuation parameter of the ith data point, t i The time of the i-th data point.

5. The fault diagnosis method for hydropower units based on digital twins according to claim 4 is characterized in that: The simulation and accident inversion using the three-dimensional twin model include: Simulate the normal operation of the hydropower unit in the three-dimensional twin model and establish the benchmark data of the normal operation of the hydropower unit; Adjust the parameters of the 3D twin model and observe the response of the 3D twin model; Simulate the operating status of the 3D twin model under different working conditions, record and analyze the simulation results, and simulate various fault conditions in the 3D twin model based on historical fault data; Adjust fault-related parameters, analyze the impact of the fault on various parts of the system, and record the fault characteristics and development process.

6. The method for fault diagnosis of a hydropower unit based on digital twins according to claim 5 is characterized in that: Combined with the output of the fault diagnosis model, the operation trend of the hydropower unit equipment is predicted, including: Obtain the output of the fault diagnosis model, and determine the current operating status and fault type of the hydropower unit equipment according to the output of the fault diagnosis model; Compare the fault diagnosis model output with the baseline data under the current normal state of the equipment to determine the severity of the fault; Collect the fault diagnosis model output within the set time, analyze the changing trend of the fault diagnosis model output, determine the direction of fault development, and complete the prediction of the operation trend of the hydropower unit equipment.

7. The hydropower unit fault diagnosis method based on digital twin according to claim 6 is characterized in that: The hydropower unit fault warning based on the prediction result includes: Based on the 3σ criterion, set warning thresholds for failures of various equipment in hydropower units; Compare the prediction results of the operation trend of the hydropower unit equipment with the warning threshold, determine whether the prediction results exceed the set warning value, and determine the warning level; Generate corresponding warning information according to the warning level, and display the warning information in three-dimensional visualization; Develop response strategies based on warning levels and fault types, and send warning information and response strategies to relevant personnel and departments.

8. A hydropower unit fault diagnosis system based on digital twin, based on the hydropower unit fault diagnosis method based on digital twin according to any one of claims 1 to 7, characterized in that: Also includes, The acquisition module is used to collect and pre-process the relevant data of the hydropower unit; A first model building module is used to build a three-dimensional twin model based on the preprocessed relevant data of the hydropower unit; A second model building module is used to build a fault diagnosis model by combining the simulation data generated by the three-dimensional twin model, the pre-processed relevant data of the hydropower unit and the historical fault data; The early warning module is used to use the three-dimensional twin model to perform simulation and accident inversion, combine the output of the fault diagnosis model, predict the operation trend of the hydropower unit equipment, and provide early warning of hydropower unit faults based on the prediction results.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the hydropower unit fault diagnosis method based on digital twins as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the hydropower unit fault diagnosis method based on digital twins described in any one of claims 1 to 7 are implemented.

Images