A device fault detection method and system based on digital twin

Abstract

The present invention discloses a device fault detection method and system based on digital twins, which belongs to the field of data analysis technology, and specifically includes: collecting and preprocessing multi-source heterogeneous operation data of inverters, using a pre-built digital twin model to perform real-time data synchronization and simulation, and dynamically adjusting simulation deviations through adaptive algorithms; with the help of deep learning algorithms, automatically identifying fault modes and predicting inverter fault information; based on the prediction results, using predictive maintenance algorithms to formulate and simulate maintenance plans to ensure the effectiveness of maintenance measures; when a fault is about to occur, the system promptly warns and automatically generates a response plan, and at the same time displays fault prediction, maintenance simulation and response results through a visual interface, thereby improving data quality and reliability.

Description

Technical Field

[0001] The present invention belongs to the technical field of data analysis, and specifically relates to a device fault detection method and system based on digital twins. Background Art

[0002] With the rapid development of the manufacturing industry and the improvement of the level of intelligence, equipment fault diagnosis plays an increasingly important role in modern industrial production and management. However, traditional equipment fault diagnosis methods mainly rely on manual experience and expert knowledge, which are subjective and limited. They are not only inefficient and costly, but also difficult to meet the needs of large-scale production lines and remote monitoring.

[0003] In this context, equipment fault detection methods based on digital twins have emerged. Digital twin technology is a technology that maps the information and behavior of physical systems into virtual space. Through real-time data collection and analysis of physical systems, accurate simulation and prediction of the state of physical systems can be achieved. This technology provides more comprehensive and accurate data support and analysis methods for equipment fault diagnosis, which can overcome the limitations of traditional methods and improve the accuracy and efficiency of fault diagnosis. However, while the existing digital twin equipment fault detection methods bring many advantages, they also face challenges and shortcomings in terms of technology maturity, data quality, model accuracy, technology integration, privacy security, dependence and reliability.

[0004] For example, the patent with publication number CN117608971A discloses a fault detection method, apparatus, storage medium and equipment based on digital twins, including: by constructing an initial twin digital model of the target system and a performance analysis model of the target system, using the performance analysis model to predict the operating data of the target system at the next moment based on the acquired operating data of the target system, and substituting the predicted result into the initial twin digital model, by analyzing the output result of the initial twin digital model, it is possible to predict whether the target system will fail at the next moment, so as to timely discover the impending failure of the target system before the failure occurs, so that the user can make targeted adjustments to the operating conditions of the target system to prevent the occurrence of failures, thereby improving the reliability of the target system.

[0005] For example, the patent with the authorization announcement number CN113505655B discloses an intelligent bearing fault diagnosis method for a digital twin system, including: the digital twin system monitors the bearing vibration signal of the target bearing in real time; the current bearing vibration signal of the target bearing is input into the bearing fault diagnosis network, and the current fault detection result of the target bearing is obtained based on the output of the bearing fault diagnosis network. This technical solution is used for the data imbalance phenomenon that occurs in the actual scene of bearing diagnosis, and can improve the diagnosis effect of fault data in the actual scene where the normal and fault data are unbalanced. The bearing fault diagnosis network set up by using this technical solution in the digital twin system can monitor the health status of the bearing equipment in real time.

[0006] The above existing technologies all have the following problems: 1) limitation of application scenarios, lack of adaptability and dynamics; 2) mainly focus on fault prediction, but do not mention how to formulate maintenance plans or response plans after predicting a fault; 3) lack of generalization ability of fault diagnosis network; 4) lack of maintenance plans and response plans. Summary of the invention

[0007] In view of the shortcomings of the prior art, the present invention proposes a device fault detection method and system based on digital twins, which collects and preprocesses multi-source heterogeneous operation data of the inverter, uses a pre-built digital twin model to perform real-time data synchronization and simulation, and dynamically adjusts the simulation deviation through an adaptive algorithm; with the help of a deep learning algorithm, it automatically identifies fault modes and predicts inverter fault information; based on the prediction results, a predictive maintenance algorithm is used to formulate and simulate maintenance plans to ensure the effectiveness of maintenance measures; when a fault is about to occur, the system promptly issues an early warning and automatically generates a response plan, and at the same time displays fault prediction, maintenance simulation and response results through a visual interface, thereby improving data quality and reliability.

[0008] To achieve the above object, the present invention provides the following technical solutions:

[0009] A device fault detection method based on digital twins, comprising:

[0010] Step S1: Collecting multi-source heterogeneous operation data of the inverter and performing preprocessing;

[0011] Step S2: Load the pre-built digital twin model of the inverter, synchronize the pre-processed multi-source heterogeneous operation data to the digital twin model through the real-time data synchronization mechanism for simulation, calculate the deviation between the simulation result of the digital twin model and the real-time multi-source heterogeneous operation data, and introduce an adaptive algorithm to dynamically adjust the deviation calculation parameters;

[0012] Step S3: Analyze the deviation data using a deep learning algorithm, automatically identify the fault mode, and predict the fault information of the inverter according to the fault mode;

[0013] Step S4: According to the fault information predicted in step S3, a predictive maintenance algorithm is used to formulate an inverter maintenance plan, and the inverter maintenance plan is simulated through a digital twin model;

[0014] Step S5: When an inverter failure is predicted, the system sends out a warning signal and automatically generates a fault response plan based on the simulation results of predictive maintenance. At the same time, the fault prediction results, predictive maintenance simulation results and fault response results are visualized.

[0015] Specifically, the specific steps of step S2 include:

[0016] S2.1: Load the pre-built digital twin model of the inverter and synchronize the real-time multi-source heterogeneous operation data. , simulate the operating status and performance of the inverter, and monitor the changes of various parameters within the digital twin model, including: Indicates the ath current data of the inverter, Indicates the bth voltage data of the inverter, Indicates the cth temperature data of the inverter, Indicates the dth load data of the inverter, represents the nth multi-source heterogeneous operation data, a, b, c, d represent the number of current, voltage, temperature, and load data of the inverter respectively, n represents the number of multi-source heterogeneous operation data, and ;

[0017] S2.2: Perform multi-scale deviation analysis on the simulation results of the digital twin model and the real-time multi-source heterogeneous operation data, and calculate the deviation value between the two .

[0018] Specifically, the specific steps of step S2 also include:

[0019] S2.3: Use an adaptive algorithm to dynamically adjust the deviation calculation parameters according to the size and change trend of the deviation value, and feed the dynamically adjusted parameters back to the digital twin model, re-simulate and calculate to obtain the dynamically adjusted deviation value ;

[0020] S2.4: Setting the dynamic error threshold , iteratively execute steps S2.1 to S2.3;

[0021] like , then continue to iterate;

[0022] like , then the deviation data is obtained;

[0023] in, express The average value of express The standard deviation of , k represents the threshold width coefficient.

[0024] Specifically, the specific steps of S2.2 include:

[0025] S2.21: Acquisition Based on the existing digital twin model, input the corresponding parameters and conditions, perform simulation operation, and obtain simulation results ,in, Represents the simulation result of the nth multi-source heterogeneous operation data;

[0026] S2.22: and Matching in time and space, and comparing and analyzing simulation results with real-time multi-source heterogeneous operation data at different time and space scales;

[0027] S2.23: According to the relative deviation calculation method, calculate and The deviation between them is as follows:

[0028] ;

[0029] in, represents the deviation value of the nth multi-source heterogeneous running data, w represents the data importance weight, represents a non-zero positive number, represents the time-sensitive factor, Indicates the time span, Indicates the reference time, represents the system status adjustment factor, and t represents the current time.

[0030] Specifically, the specific steps of step S4 include:

[0031] S4.1: Obtain the fault information predicted by the deep learning algorithm in step S3, and analyze the fault information;

[0032] S4.2: Develop a preliminary maintenance plan based on the fault analysis results, combined with the inverter maintenance history and maintenance manual;

[0033] S4.3: Use the predictive maintenance algorithm to further optimize the maintenance plan, input the optimized maintenance plan into the digital twin model, and use the digital twin model to simulate the operating status of the inverter before and after the implementation of the maintenance plan;

[0034] S4.4: Analyze the simulation results of the digital twin model to evaluate the effectiveness and feasibility of the maintenance plan. If the simulation results show that the maintenance plan has deficiencies or needs to be adjusted, return to step S4.2 for adjustment;

[0035] S4.5: After simulation verification, confirm the maintenance plan and organize and implement the inverter maintenance work according to the maintenance plan.

[0036] Specifically, the specific steps of the predictive maintenance algorithm in S4.3 include:

[0037] S4.31: Load the pre-trained predictive maintenance model, set the multi-source heterogeneous operation data and environmental variables as input variables, and set the fault prediction results and maintenance suggestions as output variables;

[0038] S4.32: Use the current maintenance plan as a baseline and use the pre-trained predictive maintenance model to simulate and evaluate different maintenance plans and calculate the cost-effectiveness ratio of each maintenance plan ,in, represents the reduction in failure losses after implementing the i-th maintenance plan, represents the increased maintenance cost of implementing the i-th maintenance plan;

[0039] S4.33: Based on the evaluation results, The corresponding maintenance plan is used as the optimized maintenance plan, and the maintenance work is carried out according to the optimized maintenance plan, and the data and feedback during the implementation process are recorded, among which, express The maximum value of .

[0040] A digital twin-based equipment fault detection system, comprising: a data processing module, a deviation calculation module, a fault identification module, a predictive maintenance module, and an early warning module;

[0041] The data processing module is used to collect multi-source heterogeneous operation data of the inverter and perform pre-processing;

[0042] The deviation calculation module uses a pre-built digital twin model of the inverter to synchronize pre-processed multi-source heterogeneous operation data in real time, calculates deviation data, and introduces an adaptive algorithm to dynamically adjust parameters;

[0043] The fault identification module uses a deep learning algorithm to analyze the deviation data, automatically identifies the fault mode of the inverter, and predicts future fault information based on the fault mode;

[0044] The predictive maintenance module uses a predictive maintenance algorithm to formulate an inverter maintenance plan based on the predicted fault information, and optimizes the maintenance plan through simulation evaluation results of the digital twin model;

[0045] The early warning module is used to predict inverter failures, and the system will issue an early warning and automatically generate a response plan, while visually displaying the prediction, simulation and maintenance results.

[0046] Specifically, the deviation calculation module includes: a digital twin model unit, a data synchronization unit, a simulation calculation unit, a deviation calculation unit, and an adaptive adjustment unit;

[0047] The digital twin model unit is used to load the digital twin model of the inverter into the system;

[0048] The data synchronization unit is used to synchronize the multi-source heterogeneous operation data collected in real time with the digital twin model;

[0049] The simulation calculation unit is used to perform simulation calculations in the digital twin model to predict the behavior of the inverter under the current working conditions;

[0050] The deviation calculation unit is used to compare the difference between the simulation result and the actual multi-source heterogeneous operation data, and calculate the deviation data;

[0051] The adaptive adjustment unit adjusts the parameters of the digital twin model according to the deviation data.

[0052] Specifically, the fault identification module includes: a fault mode identification unit and a fault prediction unit;

[0053] The fault mode identification unit uses deep learning and other algorithms to analyze deviation data and identify the fault mode of the inverter;

[0054] The fault prediction unit predicts the fault information of the inverter according to the fault mode and the real-time data.

[0055] A computer-readable storage medium stores computer instructions, which, when executed, execute the steps of a device fault detection method based on digital twins.

[0056] Compared with the prior art, the present invention has the following beneficial effects:

[0057] 1. The present invention proposes an equipment fault detection system based on digital twins, and optimizes and improves the architecture, operation steps and processes. The system has the advantages of simple process, low investment and operation costs, and low production work costs.

[0058] 2. The present invention proposes a device fault detection method based on digital twins. It collects multi-source heterogeneous operation data of the inverter, constructs a digital twin model for real-time simulation and deviation analysis, and combines the deep learning algorithm to automatically identify fault modes, thereby improving the accuracy and timeliness of inverter fault detection. An adaptive algorithm is introduced to dynamically adjust the deviation calculation parameters, thereby enhancing the adaptability and stability of the system, enabling it to better adapt to complex and changeable operating environments.

[0059] 3. The present invention proposes a device fault detection method based on digital twins. After predicting a fault, a predictive maintenance plan can be quickly formulated, and simulation verification can be performed through a digital twin model to ensure the feasibility and effectiveness of the maintenance plan; at the same time, a fault response plan is automatically generated and visualized, providing users with an intuitive and convenient fault management tool, which helps users respond to faults quickly, reduce downtime, and improve equipment utilization and production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 This is a schematic diagram of a device fault detection method based on digital twins of the present invention;

[0061] Figure 2 This is a principle flow chart of a device fault detection method based on digital twins of the present invention;

[0062] Figure 3 This is a deviation data calculation flow chart of a device fault detection method based on digital twins of the present invention;

[0063] Figure 4 This is an architecture diagram of a device fault detection system based on digital twins in the present invention. DETAILED DESCRIPTION

[0064] Example 1

[0065] See also Figure 1-Figure 3 , an embodiment of the present invention provides: a device fault detection method based on digital twins, comprising the following steps:

[0066] Step S1: collecting multi-source heterogeneous operation data of the inverter and preprocessing the operation data;

[0067] Among them, sensors and monitoring systems are used to collect equipment operation status data in real time. Multi-source heterogeneous operation data include current, voltage, temperature, and load data; preprocessing includes data verification, cleaning, and denoising.

[0068] Step S2: Load the pre-built digital twin model of the inverter, synchronize the pre-processed multi-source heterogeneous operation data to the digital twin model through the real-time data synchronization mechanism for simulation, calculate the deviation between the simulation result of the digital twin model and the real-time multi-source heterogeneous operation data, and introduce an adaptive algorithm to dynamically adjust the deviation calculation parameters;

[0069] Step S3: Analyze the deviation data using a deep learning algorithm, automatically identify the fault mode, and predict the fault information of the inverter according to the fault mode;

[0070] Furthermore, the specific steps of step S3 include:

[0071] S3.1: Collect deviation data, clean and standardize the deviation data to ensure data quality;

[0072] S3.2: According to the specific requirements of inverter fault prediction, build a deep learning model architecture, including input layer, hidden layer and output layer;

[0073] S3.3: Use the prepared deviation data to train the deep learning model. During the training process, adjust the deep learning model parameters and the gradient descent optimization algorithm so that the deep learning model can accurately identify the fault mode and predict the fault information, wherein the deep learning model parameters include the learning rate, the batch size, and the number of iterations, and the calculation process of the gradient descent optimization algorithm is the prior art content in this field, and is not the inventive solution of the present application, and will not be described in detail here;

[0074] S3.4: Use the validation set to evaluate the trained deep learning model and check its accuracy in identifying failure modes and predicting failure information;

[0075] S3.5: Tune the model based on the evaluation results, including adjusting the model structure and optimizing parameter settings;

[0076] S3.6: Deploy the trained and tuned deep learning model to the inverter fault prediction system, receive deviation data in real time, and use the deep learning model to perform fault pattern recognition and fault information prediction.

[0077] Step S4: According to the fault information predicted in step S3, a predictive maintenance algorithm is used to formulate an inverter maintenance plan, and the inverter maintenance plan is simulated through a digital twin model;

[0078] Step S5: When an inverter failure is predicted, the system sends out a warning signal and automatically generates a fault response plan based on the simulation results of predictive maintenance. At the same time, the fault prediction results, predictive maintenance simulation results and fault response results are visualized.

[0079] Among them, the fault response plan includes fault type and impact, emergency response measures, repair steps, maintenance plan and resource allocation and coordination; moreover, emergency response measures include shutdown or isolation, temporary repair; maintenance plan includes regular inspection, replacement of wearing parts, and optimization of operating environment.

[0080] The specific steps of step S2 include:

[0081] S2.1: Load the pre-built digital twin model of the inverter and synchronize the real-time multi-source heterogeneous operation data. , simulate the operating status and performance of the inverter, and monitor the changes of various parameters within the digital twin model, including: Indicates the ath current data of the inverter, Indicates the bth voltage data of the inverter, Indicates the cth temperature data of the inverter, Indicates the dth load data of the inverter, represents the nth multi-source heterogeneous operation data, a, b, c, d represent the number of current, voltage, temperature, and load data of the inverter respectively, n represents the number of multi-source heterogeneous operation data, and ;

[0082] Furthermore, the specific steps of S2.1 include:

[0083] S2.11: Load a pre-built digital twin model of the inverter from a storage medium, and verify the integrity and consistency of the digital twin model to ensure that the digital twin model can be correctly loaded and run, wherein the storage medium refers to a database, a file system or a cloud storage. At the same time, the construction process of the digital twin model of the inverter is the prior art content in this field, and is not the inventive solution of the present application, and is not described in detail here;

[0084] S2.12: Setting an interface connected to the data acquisition system to ensure that multi-source heterogeneous operation data can be received in real time, and configuring a data parsing and conversion module to convert the received data into a digital twin model input format, wherein the digital twin model input format refers to a 3D Tiles format, and the specific conversion process is a prior art content in this field, and is not an inventive solution of the present application, and will not be elaborated here;

[0085] S2.13: Set up a real-time data synchronization mechanism to ensure that multi-source heterogeneous operation data can be transmitted to the digital twin model in real time and accurately to handle delays and packet loss problems during data synchronization and ensure data integrity and real-time performance;

[0086] S2.14: In the digital twin model, the operating status and performance of the inverter are simulated based on the real-time synchronized data, including calculating the output power, efficiency, and temperature parameters of the inverter, and simulating its dynamic change process. The formula is:

[0087] ;

[0088] in, Indicates the output power, represents the input power, Indicates efficiency, represents the temperature changing with time t, represents the initial temperature, represents the heat generated by the heat source at time l, represents the temperature change at time l, represents the specific heat capacity at constant pressure, m represents the mass, and C represents the heat capacity;

[0089] S2.15: Design a monitoring mechanism to track the changes of various parameters within the digital twin model in real time and record the historical data of key parameters for subsequent analysis and optimization;

[0090] S2.16: During the monitoring process, set up anomaly detection rules to identify possible faults or abnormal conditions of the inverter. Once an abnormality is detected, the alarm mechanism is immediately triggered to notify the staff to handle it.

[0091] S2.2: Perform multi-scale deviation analysis on the simulation results of the digital twin model and the real-time multi-source heterogeneous operation data, and calculate the deviation value between the two ;

[0092] S2.3: Use an adaptive algorithm to dynamically adjust the deviation calculation parameters according to the size and change trend of the deviation value, and feed the dynamically adjusted parameters back to the digital twin model, re-simulate and calculate to obtain the dynamically adjusted deviation value ;

[0093] Furthermore, the specific steps of dynamically adjusting the deviation calculation parameters using the adaptive algorithm include:

[0094] S2.31: Real-time acquisition during system operation , and conduct statistical analysis on the deviation values, including calculating the mean value, standard deviation, and change trend of the deviation;

[0095] S2.32: According to the size and change trend of the deviation value and the update rule of the adaptive algorithm, the deviation calculation parameters are dynamically adjusted. In the present invention, the adaptive algorithm selects the minimum mean square error algorithm, and the update rule of the adaptive algorithm uses the size of the deviation value to adjust the parameters, aiming to reduce the deviation and improve the system stability and response speed. However, the specific adjustment process is the prior art content in this field, which is not the inventive solution of the present application and will not be described in detail here. At the same time, the calculation formula of the minimum mean square error algorithm is the prior art content in this field, which is not the inventive solution of the present application and will not be described in detail here.

[0096] S2.33: Verify the performance of the adaptive algorithm through simulation, and adjust and optimize the adaptive algorithm and parameter adjustment strategy based on the verification results.

[0097] S2.4: Setting the dynamic error threshold , iteratively execute steps S2.1 to S2.3;

[0098] like , then continue to iterate;

[0099] like , then the deviation data is obtained;

[0100] in, express The average value of express The standard deviation of , k represents the threshold width coefficient.

[0101] The specific steps of S2.2 include:

[0102] S2.21: Acquisition Based on the existing digital twin model, input the corresponding parameters and conditions, perform simulation operation, and obtain simulation results ,in, Represents the simulation result of the nth multi-source heterogeneous operation data, wherein the specific process of the digital twin model simulating after inputting the data is the prior art content in this field, which is not the inventive solution of the present application and will not be described in detail here;

[0103] S2.22: and Matching in time and space, and comparing and analyzing simulation results with real-time multi-source heterogeneous operation data at different time and space scales;

[0104] Among them, the time scale can include different time units such as seconds, minutes, hours, and days; the spatial scale can include different levels of equipment level, system level, entire factory or network.

[0105] S2.23: According to the relative deviation calculation method, calculate and The deviation between them is expressed as follows:

[0106] ;

[0107] in, represents the deviation value of the nth multi-source heterogeneous operation data, and w represents the data importance weight, which is used to indicate the importance of different data points to the overall deviation assessment. It can be determined based on the data source, reliability, and historical deviation record factors. Represents a non-zero positive number, used to prevent division by zero errors when the real-time data is close to zero. It is usually a very small positive number. Represents the time sensitivity factor, which is used to indicate the sensitivity of the deviation to time changes. When , it means that the deviation increases with time. , indicating that the deviation decreases with time, and The value can be determined based on system characteristics and data update frequency. Represents the time span, which is used to normalize the time factor so that the time sensitive factor varies between [0,1]. It can be the time period for data collection or the operation period of the system. Indicates the reference time, which indicates the starting time of the deviation assessment or a specific benchmark time point. Represents the system state adjustment factor, which is used to adjust the deviation assessment according to the current state of the system. When the system is in a stable state, , when the system is in transition state or fault state, or , to reflect the impact of system status on deviation assessment, and t represents the current time.

[0108] It should be noted that, in the present invention, by introducing multiple parameters and variables, the calculation of relative deviation is made more comprehensive and flexible, which can more accurately reflect the difference between simulation results and real-time data, and take into account the importance of data, time factors and the influence of system status. At the same time, it also serves as a complex evaluation system combining multiple factors.

[0109] The specific steps of step S4 include:

[0110] S4.1: Obtain the fault information predicted by the deep learning algorithm in step S3, and analyze the fault information to determine the fault type, severity, and impact range;

[0111] S4.2: Based on the fault analysis results, combined with the inverter maintenance history and maintenance manual, formulate a preliminary maintenance plan, which includes maintenance time, maintenance content, required resources and expected results;

[0112] S4.3: Use the predictive maintenance algorithm to further optimize the maintenance plan, and input the optimized maintenance plan into the digital twin model. Use the digital twin model to simulate the operating status of the inverter before and after the implementation of the maintenance plan, including simulating the equipment status changes during the maintenance process and the performance recovery after maintenance;

[0113] S4.4: Analyze the simulation results of the digital twin model to evaluate the effectiveness and feasibility of the maintenance plan. If the simulation results show that the maintenance plan has deficiencies or needs to be adjusted, return to step S4.2 for adjustment. The specific process of evaluating the maintenance plan is the prior art content in this field and is not the inventive solution of the present application, so it will not be described in detail here.

[0114] S4.5: After simulation verification, confirm the maintenance plan and organize and implement the inverter maintenance work according to the maintenance plan, including preparing the required resources, arranging maintenance personnel, and performing maintenance tasks.

[0115] The specific steps of the predictive maintenance algorithm in S4.3 include:

[0116] S4.31: Load the pre-trained predictive maintenance model, set the multi-source heterogeneous operation data and environmental variables as input variables, and set the fault prediction results and maintenance suggestions as output variables;

[0117] S4.32: Use the current maintenance plan as a baseline and use the pre-trained predictive maintenance model to simulate and evaluate different maintenance plans and calculate the cost-effectiveness ratio of each maintenance plan ,in, represents the reduction in failure losses after implementing the i-th maintenance plan, represents the increased maintenance cost of implementing the i-th maintenance plan;

[0118] S4.33: Based on the evaluation results, The corresponding maintenance plan is used as the optimized maintenance plan, and the maintenance work is carried out according to the optimized maintenance plan, and the data and feedback during the implementation process are recorded, among which, express The maximum value of .

[0119] Example 2

[0120] See also Figure 4 Another embodiment provided by the present invention is a device fault detection system based on digital twins, comprising:

[0121] Data processing module, deviation calculation module, fault identification module, predictive maintenance module, early warning module;

[0122] The data processing module is used to collect the multi-source heterogeneous operation data of the inverter and perform pre-processing to ensure the quality and consistency of the data;

[0123] The deviation calculation module uses the pre-built digital twin model of the inverter to synchronize the pre-processed multi-source heterogeneous operation data in real time, calculate the deviation data, and introduce an adaptive algorithm to dynamically adjust the parameters to improve the accuracy and timeliness of the simulation;

[0124] The fault identification module uses deep learning algorithms to analyze deviation data, automatically identify inverter fault modes, and predict future fault information based on the fault modes;

[0125] The predictive maintenance module uses the predictive maintenance algorithm to develop an inverter maintenance plan based on the predicted fault information, and optimizes the maintenance plan through the simulation evaluation effect of the digital twin model;

[0126] The early warning module is used to predict inverter failures. When the system issues an early warning, it automatically generates a response plan and visually displays the prediction, simulation and maintenance results to facilitate timely response.

[0127] The deviation calculation module includes: a digital twin model unit, a data synchronization unit, a simulation calculation unit, a deviation calculation unit, and an adaptive adjustment unit;

[0128] A digital twin model unit, used to load the digital twin model of the inverter into the system;

[0129] The data synchronization unit is used to synchronize the multi-source heterogeneous operation data collected in real time with the digital twin model to ensure that the model can reflect the current real status of the equipment;

[0130] A simulation calculation unit is used to perform simulation calculations in the digital twin model to predict the behavior of the inverter under the current operating conditions;

[0131] A deviation calculation unit is used to compare the difference between the simulation results and the actual multi-source heterogeneous operation data and calculate the deviation data;

[0132] The adaptive adjustment unit adjusts the parameters of the digital twin model according to the deviation data to better match the behavior of the actual equipment and improve the accuracy of the model.

[0133] The fault identification module includes: a fault mode identification unit and a fault prediction unit;

[0134] Fault mode recognition unit, which uses deep learning and other algorithms to analyze deviation data and identify inverter failure modes;

[0135] The fault prediction unit predicts the future fault information of the inverter, including the fault type and occurrence time, based on the fault mode and real-time data.

[0136] The predictive maintenance module includes: predictive maintenance unit, planning simulation unit, and strategy optimization unit;

[0137] Predictive maintenance unit, which designs and generates maintenance plans based on predicted failure information to prevent potential failures;

[0138] The plan simulation unit simulates the implementation effect of the predictive maintenance plan in the digital twin model to evaluate its feasibility and effectiveness;

[0139] The strategy optimization unit optimizes the maintenance plan and maintenance strategy based on the simulation results to improve the maintenance efficiency and effect.

[0140] The early warning module includes: fault early warning unit, solution generation unit, and visualization display unit;

[0141] A fault warning unit is used to send out a warning signal when it is predicted that the inverter is about to fail;

[0142] The solution generation unit automatically generates detailed fault response plans based on the fault prediction results, including emergency measures and maintenance suggestions;

[0143] The visualization display unit is used to visualize the fault prediction results, the simulation results of predictive maintenance and the fault response plan in the form of charts and reports.

[0144] Example 3

[0145] A computer-readable storage medium stores computer instructions on the medium, which execute the steps of a device fault detection method based on digital twins when the computer instructions are executed, wherein the storage medium can be a volatile or non-volatile computer-readable storage medium.

[0146] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in the field may also change, modify, replace and modify the above-mentioned embodiments without departing from the purpose and scope of protection of the present invention, and all of these are within the protection of the present invention.

Claims

1. A device fault detection method based on digital twins, characterized in that: include: Step S1: Collect multi-source heterogeneous operation data of the inverter and perform preprocessing; Step S2: Load the pre-built digital twin model of the inverter, synchronize the pre-processed multi-source heterogeneous operation data to the digital twin model through the real-time data synchronization mechanism for simulation, calculate the deviation between the simulation result of the digital twin model and the real-time multi-source heterogeneous operation data, and introduce an adaptive algorithm to dynamically adjust the deviation calculation parameters; Step S3: Analyze the deviation data using a deep learning algorithm, automatically identify the fault mode, and predict the fault information of the inverter according to the fault mode; Step S4: According to the fault information predicted in step S3, a predictive maintenance algorithm is used to formulate an inverter maintenance plan, and the inverter maintenance plan is simulated through a digital twin model; Step S5: When a fault of the inverter is predicted, the system sends out an early warning signal and automatically generates a fault response plan based on the simulation results of the predictive maintenance. At the same time, the fault prediction results, the simulation results of the predictive maintenance and the fault response results are visualized; The specific steps of step S2 include: S2.1: Load the pre-built digital twin model of the inverter and synchronize the real-time multi-source heterogeneous operation data. , simulate the operating status and performance of the inverter, and monitor the changes of various parameters within the digital twin model, including: Indicates the ath current data of the inverter, Indicates the bth voltage data of the inverter, Indicates the cth temperature data of the inverter, Indicates the dth load data of the inverter, represents the nth multi-source heterogeneous operation data, a, b, c, d represent the number of current, voltage, temperature, and load data of the inverter respectively, n represents the number of multi-source heterogeneous operation data, and ; S2.2: Perform multi-scale deviation analysis on the simulation results of the digital twin model and the real-time multi-source heterogeneous operation data, and calculate the deviation value between the two ; The specific steps of S2.2 include: S2.21: Acquisition Based on the existing digital twin model, input the corresponding parameters and conditions, perform simulation operation, and obtain simulation results ,in, Represents the simulation result of the nth multi-source heterogeneous operation data; S2.22: and Matching in time and space, and comparing and analyzing simulation results with real-time multi-source heterogeneous operation data at different time and space scales; S2.23: According to the relative deviation calculation method, calculate and The deviation between them is expressed as follows: ； in, represents the deviation value of the nth multi-source heterogeneous running data, w represents the data importance weight, represents a non-zero positive number, represents the time-sensitive factor, Indicates the time span, Indicates the reference time, represents the system status adjustment factor, and t represents the current time.

2. A device fault detection method based on digital twins as claimed in claim 1, characterized in that: The specific steps of step S2 also include: S2.3: Use an adaptive algorithm to dynamically adjust the deviation calculation parameters according to the size and change trend of the deviation value, and feed the dynamically adjusted parameters back to the digital twin model, re-simulate and calculate to obtain the dynamically adjusted deviation value ; S2.4: Setting the dynamic error threshold , iteratively execute steps S2.1 to S2.3; like , then continue to iterate; like , then the deviation data is obtained; in, express The average value of express The standard deviation of , k represents the threshold width coefficient.

3. A device fault detection method based on digital twins as claimed in claim 2, characterized in that: The specific steps of step S4 include: S4.1: Obtain the fault information predicted by the deep learning algorithm in step S3, and analyze the fault information; S4.2: Develop a preliminary maintenance plan based on the fault analysis results, combined with the inverter maintenance history and maintenance manual; S4.3: Use the predictive maintenance algorithm to further optimize the maintenance plan, input the optimized maintenance plan into the digital twin model, and use the digital twin model to simulate the operating status of the inverter before and after the implementation of the maintenance plan; S4.4: Analyze the simulation results of the digital twin model to evaluate the effectiveness and feasibility of the maintenance plan. If the simulation results show that the maintenance plan has deficiencies or needs to be adjusted, return to step S4.2 for adjustment; S4.5: After simulation verification, confirm the maintenance plan and organize and implement the inverter maintenance work according to the maintenance plan.

4. A device fault detection method based on digital twins as claimed in claim 3, characterized in that: The specific steps of the predictive maintenance algorithm in S4.3 include: S4.31: Load the pre-trained predictive maintenance model, set the multi-source heterogeneous operation data and environmental variables as input variables, and set the fault prediction results and maintenance suggestions as output variables; S4.32: Use the current maintenance plan as a baseline and use the pre-trained predictive maintenance model to simulate and evaluate different maintenance plans and calculate the cost-effectiveness ratio of each maintenance plan ,in, represents the reduction in failure losses after implementing the i-th maintenance plan, represents the increased maintenance cost of implementing the i-th maintenance plan; S4.33: Based on the evaluation results, The corresponding maintenance plan is used as the optimized maintenance plan, and the maintenance work is carried out according to the optimized maintenance plan, and the data and feedback during the implementation process are recorded, among which, express The maximum value of .

5. A device fault detection system based on digital twins, which is used to implement a device fault detection method based on digital twins as claimed in any one of claims 1 to 4, characterized in that: include: Data processing module, deviation calculation module, fault identification module, predictive maintenance module, early warning module; The data processing module is used to collect multi-source heterogeneous operation data of the inverter and perform pre-processing; The deviation calculation module uses a pre-built digital twin model of the inverter to synchronize pre-processed multi-source heterogeneous operation data in real time, calculates deviation data, and introduces an adaptive algorithm to dynamically adjust parameters; The fault identification module uses a deep learning algorithm to analyze the deviation data, automatically identifies the fault mode of the inverter, and predicts future fault information based on the fault mode; The predictive maintenance module uses a predictive maintenance algorithm to formulate an inverter maintenance plan based on the predicted fault information, and optimizes the maintenance plan through simulation evaluation results of the digital twin model; The early warning module is used to predict inverter failures, and the system will issue an early warning and automatically generate a response plan, while visually displaying the prediction, simulation and maintenance results.

6. The device fault detection system based on digital twins according to claim 5, characterized in that: The deviation calculation module includes: a digital twin model unit, a data synchronization unit, a simulation calculation unit, a deviation calculation unit, and an adaptive adjustment unit; The digital twin model unit is used to load the digital twin model of the inverter into the system; The data synchronization unit is used to synchronize the multi-source heterogeneous operation data collected in real time with the digital twin model; The simulation calculation unit is used to perform simulation calculations in the digital twin model to predict the behavior of the inverter under the current working conditions; The deviation calculation unit is used to compare the difference between the simulation result and the actual multi-source heterogeneous operation data, and calculate the deviation data; The adaptive adjustment unit adjusts the parameters of the digital twin model according to the deviation data.

7. The device fault detection system based on digital twins according to claim 6, characterized in that: The fault identification module includes: a fault mode identification unit and a fault prediction unit; The fault mode identification unit uses a deep learning algorithm to analyze the deviation data and identify the fault mode of the inverter; The fault prediction unit predicts the fault information of the inverter according to the fault mode and the real-time data.

8. A computer-readable storage medium, characterized in that: Computer instructions are stored thereon, and when the computer instructions are executed, the steps of a device fault detection method based on digital twins as described in any one of claims 1 to 4 are executed.

Images