Digital twinning method and system based on industrial automation, computing device and storage medium

Abstract

The invention provides a digital twinning method and system based on industrial automation, computing equipment and a storage medium. The method comprises the following steps: acquiring real-time operation parameters from a physical system; based on the real-time operation parameters and a preset physical model, constructing a digital twinborn body of the physical system in a virtual environment; simulating the operation state of the physical system by using the digital twin, and predicting the performance of the physical system under different working conditions to generate a simulation result; comparing the simulation result with the obtained actual operation data of the physical system so as to evaluate the digital twin body and obtain an evaluation result; adjusting the physical model according to the evaluation result so as to update the digital twin; and performing fault prediction and / or optimizing operation parameters of the physical system by using the updated digital twin. According to the technical scheme provided by the invention, the digital twin can be continuously optimized and the physical model can be adjusted.

Description

Technical Field

[0001] Embodiments of the present application relate to the field of digital twin technology, and in particular, to a digital twin method, system, computing device, and storage medium based on industrial automation. Background Art

[0002] With the development of technology, digital twin technology has been widely used in the field of industrial automation. Digital twin is a multidisciplinary optimization dynamic simulation process that integrates data such as physical models, sensor updates, and operation history. It achieves real-time monitoring, prediction, and optimization of physical entities by building a virtual model of the physical system. In the field of industrial automation, digital twin technology can be used in predictive maintenance, performance optimization, fault diagnosis, and many other aspects.

[0003] Traditional industrial automation systems usually rely on static models and post-analysis. Although this method can achieve basic monitoring of the operating status of equipment to a certain extent, it is insufficient in terms of predictiveness and initiative. With the continuous improvement of production efficiency and product quality requirements, traditional industrial automation methods are increasingly unable to meet the needs of modern industrial production. Therefore, it is particularly important to develop new technologies that can monitor and predict the operating status of physical systems in real time.

[0004] Although existing digital twin technologies can improve the system's predictive capabilities and operating efficiency to a certain extent, they still have some limitations. For example, these technologies often require a large amount of historical data to train the model, and may not be able to make accurate predictions for new or atypical working conditions; in addition, model updates usually rely on manual intervention, and lack an automated mechanism to continuously optimize and adjust the model, which limits the application effect of digital twin technology in the field of industrial automation. Summary of the invention

[0005] The embodiments of the present application provide a digital twin method, system, computing device and storage medium based on industrial automation, which are used to solve the problem of lack of automated mechanism in the prior art to continuously optimize digital twins and adjust physical models.

[0006] In a first aspect, an embodiment of the present application provides a digital twin method based on industrial automation, including:

[0007] Obtain real-time operating parameters from physical systems;

[0008] Based on the real-time operating parameters and the preset physical model, constructing a digital twin of the physical system in a virtual environment;

[0009] Using the digital twin to simulate the operating state of the physical system and predict the performance of the physical system under different working conditions to generate simulation results;

[0010] Comparing the simulation result with the acquired actual operation data of the physical system to evaluate the digital twin and obtain an evaluation result;

[0011] Adjusting the physical model according to the evaluation result to update the digital twin;

[0012] The updated digital twin is used to predict faults and / or optimize operating parameters of the physical system.

[0013] Optionally, the acquiring of real-time operating parameters from the physical system includes:

[0014] Collect real-time sensor data from physical systems through IoT devices;

[0015] The real-time sensor data is preprocessed to generate real-time operating parameters; wherein the preprocessing includes at least: data cleaning, outlier detection and missing value filling.

[0016] Optionally, constructing a digital twin of the physical system in a virtual environment based on the real-time operating parameters and a preset physical model includes:

[0017] analyzing real-time operating parameters of the physical system using a machine learning algorithm to extract key features;

[0018] constructing a dynamic mathematical model of the physical system based on the key features;

[0019] The operating state of the physical system is simulated in the virtual environment, and is updated in real time through the dynamic mathematical model to complete the construction of the digital twin of the physical system.

[0020] Optionally, the using the digital twin to simulate the operating state of the physical system and predicting the performance of the physical system under different working conditions to generate simulation results includes:

[0021] simulating, using the digital twin, performance of the physical system under normal operating conditions;

[0022] Using the digital twin to simulate the performance of the physical system under abnormal operating conditions and predict the probability of occurrence of potential failures;

[0023] generating simulation results according to the performance of the physical system under normal operating conditions, the performance of the physical system under abnormal operating conditions, and the probability of occurrence of the potential failure;

[0024] The predicted probability of occurrence of a potential failure includes:

[0025] The probability of occurrence of a potential fault is predicted by a fault prediction model, wherein the fault prediction model calculates the probability of occurrence of a potential fault by the following formula:

[0026]

[0027] Where P(f) represents the probability of occurrence of potential failure, w0, w i 、v ij They are respectively represented as bias term, linear weight and cross term weight, x i Represents the value of the i-th running parameter, x j It is represented as the value of the jth operating parameter, where j>i and n represents the number of operating parameters.

[0028] Optionally, comparing the simulation result with the acquired actual operation data of the physical system to evaluate the digital twin and obtain the evaluation result includes:

[0029] Calculate the difference between the simulation result and the acquired actual operation data of the physical system using a statistical method;

[0030] Determining the accuracy of the digital twin through error analysis to obtain an evaluation result;

[0031] Determining the accuracy of the digital twin through error analysis to obtain an evaluation result includes:

[0032] By formula: Calculate the accuracy of the digital twin to obtain an evaluation result; where RMSE is expressed as the evaluation result, y i Represents actual operating data. is represented as the simulation result, and N is represented as the number of data points.

[0033] Optionally, adjusting the physical model according to the evaluation result to update the digital twin includes:

[0034] Optimizing model parameters in the physical model by a back propagation algorithm;

[0035] The back propagation algorithm updates the model parameters using the following formula:

[0036]

[0037] Among them, w j is represented as the jth model parameter, η is represented as the learning rate, Expressed as the loss function relative to the model parameters w j The partial derivative of , λ represents the weight of the regularization term, and L represents the loss function, where

[0038] Optionally, the using the updated digital twin to perform fault prediction and / or optimize operating parameters of the physical system includes:

[0039] Using the updated digital twin to predict faults, so as to take preventive measures according to the probability of occurrence of potential faults;

[0040] The operating parameters of the physical system are automatically adjusted by an optimization algorithm, wherein the operating parameter optimization calculates the optimal parameter set p of the physical system by the following formula: * :

[0041]

[0042] Among them, f(p) represents the objective function, p represents the parameter set, and p * is represented as the optimal parameter set, and λ is represented as the weight of the regularization term.

[0043] In a second aspect, an embodiment of the present application provides a digital twin system based on industrial automation, including:

[0044] An acquisition module, used to obtain real-time operating parameters from a physical system;

[0045] A construction module, used to construct a digital twin of the physical system in a virtual environment based on the real-time operating parameters and a preset physical model;

[0046] A generation module, used to simulate the operating state of the physical system using the digital twin, and predict the performance of the physical system under different working conditions to generate simulation results;

[0047] The acquisition module is further used to compare the simulation result with the acquired actual operation data of the physical system to evaluate the digital twin and obtain an evaluation result;

[0048] An updating module, configured to adjust the physical model according to the evaluation result to update the digital twin;

[0049] A processing module is used to use the updated digital twin to predict faults and / or optimize operating parameters of the physical system.

[0050] In a third aspect, an embodiment of the present application provides a computing device, comprising a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement the digital twin method based on industrial automation as described in the first aspect above.

[0051] In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program. When the computer program is executed by a computer, the digital twin method based on industrial automation as described in the first aspect above is implemented.

[0052] In an embodiment of the present application, real-time operating parameters are obtained from a physical system; based on the real-time operating parameters and a preset physical model, a digital twin of the physical system is constructed in a virtual environment; the operating state of the physical system is simulated using the digital twin, and the performance of the physical system under different working conditions is predicted to generate simulation results; the simulation results are compared with the actual operating data of the physical system obtained to evaluate the digital twin and obtain evaluation results; the physical model is adjusted according to the evaluation results to update the digital twin; and the updated digital twin is used to predict faults and / or optimize the operating parameters of the physical system.

[0053] The technical solution of this application has the following significant beneficial effects:

[0054] Improve prediction accuracy: Through real-time updated operating parameters and physical models, digital twins can more accurately simulate the operating status of physical systems and improve the accuracy of prediction results.

[0055] Enhance system adaptability: The automated model adjustment mechanism enables the digital twin to adapt to changes under different working conditions and improve the system’s adaptability.

[0056] Optimize operating parameters: By comparing simulation results with actual operating data, the operating parameters of the physical system can be discovered and optimized in a timely manner to improve the overall performance of the system.

[0057] Reduce maintenance costs: Through fault prediction, measures can be taken in advance to avoid equipment failures, reduce unplanned downtime, and reduce maintenance costs.

[0058] Improve production efficiency: Accurate prediction and optimization capabilities help improve production efficiency and ensure that equipment is in optimal operating condition.

[0059] Enhanced decision support: The prediction results and optimization suggestions provided by digital twins provide strong support for decision makers and help them make more scientific and reasonable decisions.

[0060] In summary, the solution of the present application not only improves the prediction accuracy of the digital twin, but also enhances the adaptability and optimization capabilities of the system, thereby effectively improving the overall performance and economic benefits of the industrial automation system.

[0061] These and other aspects of the present application will become more clearly understood in the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0063] Figure 1 A flowchart of a digital twin method based on industrial automation provided in an embodiment of the present application;

[0064] Figure 2 A schematic diagram of the structure of a digital twin system based on industrial automation provided in an embodiment of the present application;

[0065] Figure 3 A schematic diagram of the structure of a computing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0066] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0067] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0068] According to the background technology, in order to solve the defects of the existing solutions, this application proposes a digital twin method based on industrial automation, which obtains the real-time operating parameters of the physical system and builds a digital twin of the physical system in a virtual environment based on these parameters and a preset physical model. By simulating the operating state of the physical system and comparing it with the actual operating data, the physical model of the digital twin is continuously adjusted and optimized, thereby improving the accuracy of the prediction and realizing the optimization of fault prediction and operating parameters.

[0069] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0070] Figure 1 A flowchart of a digital twin method based on industrial automation is provided for an embodiment of the present application, such as Figure 1 As shown, the method includes:

[0071] 101. Obtain real-time operating parameters from the physical system;

[0072] Optionally, in an embodiment of the present application, step 101 may specifically include: collecting real-time sensor data of the physical system through an Internet of Things device; preprocessing the real-time sensor data to generate real-time operating parameters; wherein the preprocessing includes at least: data cleaning, outlier detection and missing value filling.

[0073] In this step, IoT devices refer to sensors or terminal devices that can connect to the Internet and transmit data wirelessly or wired. These devices can be installed on industrial equipment to monitor the status and environmental conditions of the equipment in real time.

[0074] Real-time sensor data: refers to various types of data collected by IoT devices from physical systems, such as temperature, pressure, vibration, etc. This data is usually collected continuously at a certain frequency.

[0075] Preprocessing: The processing steps performed on raw data before data analysis to improve the quality and usability of the data. Preprocessing usually includes steps such as data cleaning, outlier detection, and missing value filling.

[0076] Data cleaning: refers to removing or correcting errors and inconsistencies in data, such as removing duplicate data, correcting spelling errors, etc.

[0077] Outlier detection: Detecting data points in the data that deviate significantly from the rest of the data, which could be due to measurement errors or system failures, etc.

[0078] Missing value filling: For missing data points, use appropriate methods to fill them to ensure data integrity.

[0079] In an embodiment of the present application, consider an equipment monitoring system in an automated manufacturing plant, which is intended to monitor the operating status of production equipment in order to detect potential faults in advance and perform maintenance.

[0080] Collect real-time sensor data of physical systems through IoT devices: IoT devices include temperature sensors, pressure sensors, and vibration sensors installed on production equipment. These sensors send data to the central control system every 1 second. The data includes information such as the temperature, pressure, and vibration amplitude of the equipment.

[0081] The real-time sensor data is preprocessed to generate real-time operating parameters: Data cleaning: remove duplicate data records and correct occasional sensor reading errors. Outlier detection: use statistical methods (such as Z-score) to identify abnormal values ​​outside the normal range. For example, for temperature sensor data, if a reading is much higher or lower than the average temperature, then this reading is considered an outlier. Missing value filling: use the nearest neighbor method or moving average method to fill in occasional missing sensor data. For example, if a temperature sensor does not send data for a period of time, the average temperature value of the two previous and subsequent time points can be used to fill in the temperature data during this period.

[0082] Generate real-time operating parameters: After preprocessing, the generated real-time operating parameters include stable and accurate temperature, pressure and vibration data, which can be used to monitor the health of the equipment.

[0083] Through the above steps, the factory can monitor the operating status of the equipment in real time and predict potential failures by analyzing the changing trends of these parameters, so as to take necessary maintenance measures to ensure the efficient operation of the equipment.

[0084] 102. Based on the real-time operating parameters and the preset physical model, construct a digital twin of the physical system in a virtual environment;

[0085] Optionally, in the embodiment of the present application, step 102 may specifically include:

[0086] 1021. Analyze the real-time operating parameters of the physical system using a machine learning algorithm to extract key features;

[0087] 1022. Construct a dynamic mathematical model of the physical system based on the key features;

[0088] 1023. Simulate the operating state of the physical system in the virtual environment, and update it in real time through the dynamic mathematical model to complete the construction of the digital twin of the physical system.

[0089] In this step, real-time operating parameters: These parameters are real-time data collected from the physical system, including but not limited to temperature, pressure, flow, vibration, etc. These data are used to reflect the current working status of the physical system.

[0090] Preset physical model: refers to a mathematical model based on physical laws or empirical formulas to describe the dynamic characteristics of a physical system. For example, a thermodynamic model can be used to describe the heat exchange process of an engine.

[0091] Machine learning algorithm: is a statistical method used to automatically "learn" patterns and regularities from data. These algorithms can be used to identify key features in data, such as support vector machines, neural networks, etc.

[0092] Key features: Factors extracted from the data that have a significant impact on the behavior of the system. These features can help build more accurate models.

[0093] Dynamic mathematical model: A mathematical expression used to describe the behavior of a system over time. This model can be in the form of a system of differential equations, a state space model, etc.

[0094] Digital Twin: An exact replica of a physical entity in a virtual environment. It uses real-time data to simulate the performance of the physical entity and is able to predict future behavior.

[0095] Virtual environment: A computer-generated environment that can be used to simulate the behavior of real-world systems. In a virtual environment, different scenarios can be created and tested using software.

[0096] Simulation: Using a model on a computer to reproduce the behavior of a physical system. Simulation can be used to predict how a system will behave under different conditions.

[0097] In the embodiment of this application, it is assumed that we need to build a digital twin for monitoring and predicting the performance of wind turbines. The following are the specific steps:

[0098] Use machine learning algorithms to analyze the real-time operating parameters of the physical system to extract key features: Collect real-time operating parameters of wind turbines, such as wind speed, blade speed, generator temperature, etc. Use machine learning algorithms (such as cluster analysis, principal component analysis PCA, etc.) to analyze these data to identify key variables that have a significant impact on generator performance, such as the relationship between wind speed and blade speed.

[0099] Constructing a dynamic mathematical model of the physical system based on the key features: Based on the extracted key features, such as the relationship between wind speed and blade speed, a dynamic mathematical model is constructed to describe the operating characteristics of the generator. This model may include physical equations describing how wind power is converted into mechanical energy, and electrical equations describing how mechanical energy is converted into electrical energy.

[0100] The operating state of the physical system is simulated in the virtual environment, and the dynamic mathematical model is used to update the system in real time to complete the construction of the digital twin of the physical system: the operating state of the wind turbine is simulated in the virtual environment using the dynamic mathematical model. The wind speed data collected in real time is input into the simulation environment, and the parameters in the model are dynamically updated to reflect the real-time operating state of the wind turbine in the real world. The digital twin can predict possible problems with the generator based on real-time data and provide preventive maintenance recommendations.

[0101] Through the above steps, we can build a digital twin that can reflect the status of the wind turbine in real time, which helps to improve the operating efficiency of the generator and reduce maintenance costs.

[0102] It can be seen from the above embodiments that the construction of a digital twin is a process involving data collection, analysis, and modeling. In the embodiments of the present application, we analyzed the real-time operating parameters of the wind turbine using a machine learning algorithm, extracted key features, and built a dynamic mathematical model based on these features. Next, the operating state of the generator was simulated in a virtual environment, and the construction of the digital twin was completed by updating the model in real time. This method can be widely used in the monitoring and prediction of other physical systems.

[0103] 103. Using the digital twin to simulate the operating state of the physical system and predict the performance of the physical system under different working conditions to generate simulation results;

[0104] Optionally, in the embodiment of the present application, step 103 may specifically include:

[0105] 1031. Using the digital twin to simulate the performance of the physical system under normal operating conditions;

[0106] 1032. Using the digital twin to simulate the performance of the physical system under abnormal operating conditions and predict the probability of occurrence of potential failures;

[0107] 1033. Generate simulation results according to the performance of the physical system under normal operating conditions, the performance of the physical system under abnormal operating conditions, and the probability of occurrence of the potential fault;

[0108] The predicting of the probability of occurrence of a potential fault includes: predicting the probability of occurrence of a potential fault by a fault prediction model, wherein the fault prediction model calculates the probability of occurrence of a potential fault by the following formula:

[0109]

[0110] Where P(f) represents the probability of occurrence of potential failure, w0, wi 、v ij They are respectively represented as bias term, linear weight and cross term weight, x i Represents the value of the i-th running parameter, x j It is represented as the value of the jth operating parameter, where j>i and n represents the number of operating parameters.

[0111] In this step, digital twin: A digital twin is a virtual representation of a physical system that can reflect the state of the physical system in real time and can be updated based on the real-time data of the physical system.

[0112] Simulation: Reproducing the behavior of a physical system under given conditions in a digital twin.

[0113] Normal operating conditions: A state in which a physical system operates within its intended operating range, typically within design specifications and safety limits.

[0114] Abnormal operating condition: The operating state of a physical system when it is outside its normal range or when a fault occurs.

[0115] Probability of potential failure: Predict the possibility of system failure by analyzing the performance of the physical system under abnormal conditions.

[0116] Fault prediction model: A mathematical model that estimates the probability of a physical system failing. The model is usually trained based on historical data and current operating parameters.

[0117] Bias term: A constant term in a model that reflects the probability of a potential fault occurring even when all operating parameters are zero.

[0118] Linear weights: related to each running parameter (x i ) reflects the direct impact of this parameter on the probability of potential failure.

[0119] Cross-term weight: related to the two operating parameters (x i and x j ) reflects the influence of the interaction between these two parameters on the probability of potential failure.

[0120] In the embodiment of this application, it is assumed that we are developing a digital twin to simulate the operating state of a turbocharged engine and predict its performance under different working conditions. The following are the specific steps:

[0121] Use the digital twin to simulate the performance of the physical system under normal operating conditions: First, we collect real-time data of the engine under normal operating conditions, such as speed, temperature, pressure, etc. Then, we simulate the performance of the engine under these conditions in the digital twin to verify the accuracy of the model.

[0122] Use the digital twin to simulate the performance of the physical system under abnormal operating conditions and predict the probability of potential failures: Next, we change some operating parameters to simulate the operation of the engine under abnormal conditions. We use the fault prediction model to predict the probability of engine failure under these abnormal conditions.

[0123] Generate simulation results based on the performance of the physical system under normal operating conditions, the performance of the physical system under abnormal operating conditions, and the probability of occurrence of the potential faults: Finally, we combine the performance under normal operating conditions, the performance under abnormal conditions, and the probability of occurrence of potential faults to generate complete simulation results.

[0124] Here is a specific example:

[0125] Assume we have the following run parameters:

[0126] Speed ​​(x1);

[0127] Temperature (x2);

[0128] Pressure (x3);

[0129] And the following model parameters:

[0130] Bias term w0 = -2;

[0131] Linear weights w1=0.05, w2=0.1, w3=0.15;

[0132] Cross-term weight v 12 =0.001,v 13 =0.002,v 23 =0.003;

[0133] For a specific operating state, assume that the operating parameter value is:

[0134] Speed ​​x1 = 2000;

[0135] Temperature x2 = 90°C;

[0136] Pressure x3 = 15 bar;

[0137] The probability of occurrence of potential failure P(f) is calculated according to the formula:

[0138]

[0139] Substituting the above specific values:

[0140]

[0141] The calculation results are:

[0142]

[0143] Due to e -127.3 is a very small positive number, we can approximate P(f) to be 1, which means that in this case, the probability of the potential failure occurring is very high.

[0144] Through such calculations, we can predict the potential failure probability of the engine under abnormal operating conditions and take preventive measures accordingly.

[0145] 104. Compare the simulation result with the acquired actual operation data of the physical system to evaluate the digital twin and obtain an evaluation result;

[0146] Optionally, in the embodiment of the present application, step 104 may specifically include:

[0147] 1041. Calculate the difference between the simulation result and the acquired actual operation data of the physical system using a statistical method;

[0148] 1042. Determine the accuracy of the digital twin through error analysis to obtain an evaluation result;

[0149] Wherein, determining the accuracy of the digital twin through error analysis to obtain an evaluation result includes:

[0150] By formula: Calculate the accuracy of the digital twin to obtain an evaluation result; where RMSE is expressed as the evaluation result, y i Represents actual operating data. is represented as the simulation result, and N is represented as the number of data points.

[0151] In the above RMSE formula, RMSE is expressed as the root mean square error, which represents the average error between the digital twin simulation results and the actual operation data, y i Represented as the actual running data of the ith data point, is represented as the simulation result of the ith data point, and N is the total number of data points.

[0152] In this step, simulation results: the results obtained by simulating the performance of the physical system under different working conditions through the digital twin.

[0153] Actual operation data: Data collected during the actual operation of the physical system, which can be used to verify the simulation results of the digital twin.

[0154] Statistical Methods: A range of mathematical techniques used to analyse and interpret data to reveal relationships or trends among them.

[0155] Discrepancy: The gap between simulation results and actual operating data, used to evaluate the simulation accuracy of the digital twin.

[0156] Error analysis: Determine the accuracy of the digital twin by calculating the error between the simulation results and the actual data.

[0157] Root mean square error (RMSE): A commonly used error metric used to measure the average error between the digital twin simulation results and the actual operating data.

[0158] In the present embodiment, it is assumed that we are evaluating a digital twin that is used to simulate the performance of a wind turbine. Our goal is to evaluate the difference between the simulation results of the digital twin and the actual operating data.

[0159] Calculate the difference between the simulation results and the actual operating data of the physical system using statistical methods: Collect the actual operating data of the wind turbine, including wind speed, power output, etc. Use the digital twin to simulate the power output of the wind turbine at different wind speeds. Compare the simulation results with the actual operating data, and use statistical methods (such as calculating the difference, drawing a scatter plot, etc.) to evaluate the difference between the two.

[0160] The accuracy of the digital twin is determined by error analysis to obtain the evaluation results: the root mean square error (RMSE) is used to quantify the error between the simulation results and the actual operation data. The RMSE value is obtained by the calculation formula to evaluate the accuracy of the digital twin.

[0161] Here is a specific example:

[0162] Assume that we have collected 10 sets of actual power output data of wind turbines at different wind speeds and the corresponding simulation results, as shown in Table 1 below:

[0163] Table 1

[0164]

[0165] Based on Table 1 above, we use the RMSE formula to calculate the accuracy of the digital twin:

[0166]

[0167] Substitute the specific values ​​in Table 1 above:

[0168]

[0169] RMSE = 2

[0170] This means that the average error between the power output simulated by the digital twin and the actual power output is 2kW, and this average error is used as the evaluation result. By doing this calculation, we can evaluate the accuracy of the digital twin and make the necessary adjustments to improve its accuracy.

[0171] 105. Adjust the physical model according to the evaluation result to update the digital twin;

[0172] Optionally, in the embodiment of the present application, step 105 may specifically include: optimizing the model parameters in the physical model by a back propagation algorithm;

[0173] The back propagation algorithm updates the model parameters using the following formula:

[0174]

[0175] Among them, w j is represented as the jth model parameter, η is represented as the learning rate, Expressed as the loss function relative to the model parameters w j The partial derivative of , λ represents the weight of the regularization term, and L represents the loss function, where

[0176] In this step, the results are evaluated: evaluation indicators such as the root mean square error (RMSE) are obtained by comparing the simulation results of the digital twin with the actual operating data of the physical system.

[0177] Physical model: A mathematical model used to describe the behavior of a physical system, usually consisting of a set of equations and parameters.

[0178] Digital Twin: A virtual representation of a physical system that simulates the behavior of the physical system through a physical model.

[0179] Backpropagation algorithm: A supervised learning method for training neural networks that adjusts model parameters by minimizing a loss function.

[0180] Model parameters: Variables in a physical model that determine the behavior of the model.

[0181] Learning rate: A hyperparameter that controls the magnitude of the parameter updates at each iteration.

[0182] Loss function: A function used to quantify the difference between the predicted value and the true value, here it is the square of the root mean square error.

[0183] Partial derivative: The rate of change of the loss function relative to a certain parameter, which is used to guide the update direction of the parameter.

[0184] Regularization term: A penalty term used to prevent overfitting, which is achieved by adding the sum of squares of model parameters.

[0185] In the embodiment of the present application, it is assumed that we have a digital twin for simulating the performance of a wind turbine, and we need to adjust the parameters in the physical model according to the evaluation results to improve the accuracy of the digital twin.

[0186] Adjust the physical model according to the evaluation result: Use the previously calculated root mean square error (RMSE) as the evaluation result. Construct a loss function, which is the square of RMSE, i.e. L = RMSE 2 . Use the back-propagation algorithm to adjust the parameters in the physical model.

[0187] The model parameters in the physical model are optimized by back propagation algorithm: the learning rate η and the weight λ of the regularization term are set. The partial derivative of the loss function with respect to each parameter is calculated. The model parameters are updated to minimize the loss function.

[0188] Here is a specific example:

[0189] Assume that we have calculated the square of RMSE L=4 (based on the previous embodiment), and now we want to update the model parameters through the back propagation algorithm.

[0190] We set the learning rate to η = 0.1 and the weight of the regularization term to λ = 0.01.

[0191] Suppose we have two parameters w1 and w2 with initial values ​​2 and 3 respectively.

[0192] We need to calculate the partial derivative of the loss function with respect to each parameter and

[0193] To simplify the explanation, we assume and

[0194] Update parameters w1 and w2 according to the formula:

[0195]

[0196] w1=2-0.1(1+0.01·2)

[0197] w1=2-0.1(1.02)

[0198] w1=2-0.102

[0199] w1=1.898

[0200] Similarly, for w2:

[0201]

[0202] w2=3-0.1(2+0.01·3)

[0203] w2=3-0.1(2.03)

[0204] w2=3-0.203

[0205] w2=2.797

[0206] Therefore, the updated parameters are w1=1.898 and w2=2.797.

[0207] Through this iterative process, we can gradually optimize the parameters in the physical model and thus improve the accuracy of the digital twin. This process may require multiple iterations until satisfactory accuracy is achieved or the stopping condition is met.

[0208] 106. Use the updated digital twin to predict faults and / or optimize operating parameters of the physical system.

[0209] Optionally, in the embodiment of the present application, step 106 may specifically include:

[0210] 1061. Using the updated digital twin to predict faults, so as to take preventive measures according to the probability of occurrence of potential faults;

[0211] 1062. Automatically adjust the operating parameters of the physical system by using an optimization algorithm, wherein the operating parameter optimization calculates the optimal parameter set p of the physical system by the following formula: * :

[0212]

[0213] Among them, f(p) represents the objective function, p represents the parameter set, and p * is represented as the optimal parameter set, and λ is represented as the weight of the regularization term.

[0214] In this step, the updated digital twin is: a digital twin that more accurately reflects the behavior of the physical system after evaluation and parameter adjustment.

[0215] Failure prediction: Use digital twins to predict possible failures of physical systems so that preventive measures can be taken.

[0216] Probability of potential failure: The probability of a physical system failing under different operating conditions obtained through digital twin simulation.

[0217] Preventive measures: Take measures to reduce or avoid the occurrence of potential failures based on the probability of potential failures.

[0218] Operation parameter optimization: Adjust the operation parameters of the physical system through optimization algorithms to improve the performance of the system.

[0219] Optimization algorithm: an algorithm used to find the optimal solution, such as gradient descent, genetic algorithm, etc.

[0220] Objective function: A function used to quantify the impact of operating parameters on system performance. The goal of optimization is to find a set of parameters that minimizes the objective function.

[0221] Regularization term: A penalty term used to prevent overfitting, which is achieved by adding the sum of squares of the parameters.

[0222] Optimal parameter set: The parameter set that minimizes the objective function.

[0223] In this embodiment of the application, it is assumed that we are using a digital twin to monitor and optimize the performance of a wind turbine. The following are the specific steps:

[0224] Fault prediction using the updated digital twin: Using the digital twin to simulate the performance of the wind turbine under different operating conditions. Calculate the probability of potential faults through the fault prediction model. If the probability of potential faults exceeds a threshold, take preventive measures, such as adjusting the maintenance plan or replacing parts in advance.

[0225] Automatically adjust the operating parameters of the physical system by an optimization algorithm: define an objective function, such as maximum power output or minimum energy consumption. Use an optimization algorithm, such as gradient descent, to find the optimal set of operating parameters. Calculate the optimal parameter set p* by the optimization algorithm.

[0226] Here is a specific example:

[0227] Suppose we want to optimize the maximum power output of a wind turbine. We can do this by following these steps:

[0228] Fault prediction using the updated digital twin: Assume that we have calculated the probability of occurrence of a potential fault P(f). If P(f) is greater than a preset threshold, such as 0.5, it is considered that a potential fault may exist. Take corresponding preventive measures, such as increasing the frequency of maintenance inspections.

[0229] Automatically adjust the operating parameters of the physical system by using an optimization algorithm: define the objective function f(p) as the maximum power output of the wind turbine. Set the weight λ of the regularization term to 0.01. Use the gradient descent method or other optimization algorithms to calculate the optimal parameter set p*.

[0230] Specifically, suppose we have three operating parameters p1, p2, p3, and define the objective function f(p) as the maximum power output of the wind turbine. We want to find the optimal parameter set p* through an optimization algorithm.

[0231] Assume that we have defined the objective function f(p) and set the weight of the regularization term to λ = 0.01. We now want to find the optimal set of parameters p* to maximize the power output.

[0232] Calculate the optimal parameter set p* according to the formula:

[0233] To simplify the explanation, assume that we have a simple gradient descent method with initial parameters p1=1, p2=2, p3=3 and a learning rate η=0.1.

[0234] Assume that the gradient of the objective function f(p) is Right now:

[0235] Calculate the parameters after a gradient descent update:

[0236]

[0237] p1=1-0.1(1+0.01·1)

[0238] p1=1-0.1(1.01)

[0239] p1=0.899

[0240] Similarly, for p2 and p3:

[0241] p2=2-0.1(2+0.01·2)

[0242] p2=2-0.1(2.02)

[0243] p2=1.798

[0244] p3=3-0.1(3+0.01·3)

[0245] p3=3-0.1(3.03)

[0246] p3=2.697

[0247] Therefore, the updated parameters are p1=0.899, p2=1.798, p3=2.697.

[0248] Through this iterative process, we can gradually optimize the operating parameters of the physical system to achieve optimal performance. This process may require multiple iterations until satisfactory performance is achieved or the stopping condition is met. Through the above steps, we can use digital twins to predict faults and optimize operating parameters, thereby improving the reliability and efficiency of the physical system.

[0249] Figure 2 A structural diagram of a digital twin system based on industrial automation is provided for the embodiment of the present application, such as Figure 2 As shown, the system includes:

[0250] An acquisition module 21 is used to acquire real-time operating parameters from a physical system;

[0251] A construction module 22, configured to construct a digital twin of the physical system in a virtual environment based on the real-time operating parameters and a preset physical model;

[0252] A generation module 23, configured to use the digital twin to simulate the operating state of the physical system and predict the performance of the physical system under different working conditions to generate simulation results;

[0253] The acquisition module 21 is further used to compare the simulation result with the acquired actual operation data of the physical system to evaluate the digital twin and obtain an evaluation result;

[0254] An updating module 24, configured to adjust the physical model according to the evaluation result to update the digital twin;

[0255] The processing module 25 is used to use the updated digital twin to predict faults and / or optimize the operating parameters of the physical system.

[0256] Optionally, in an embodiment of the present application, the acquisition module 21 is specifically used to collect real-time sensor data of the physical system through an Internet of Things device; preprocess the real-time sensor data to generate real-time operating parameters; wherein the preprocessing includes at least: data cleaning, outlier detection and missing value filling.

[0257] Optionally, in an embodiment of the present application, the construction module 22 is specifically used to use a machine learning algorithm to analyze the real-time operating parameters of the physical system to extract key features; construct a dynamic mathematical model of the physical system based on the key features; simulate the operating state of the physical system in the virtual environment, and update it in real time through the dynamic mathematical model to complete the construction of the digital twin of the physical system.

[0258] Optionally, in an embodiment of the present application, the generation module 23 is specifically used to use the digital twin to simulate the performance of the physical system under normal operating conditions; use the digital twin to simulate the performance of the physical system under abnormal operating conditions, and predict the probability of occurrence of potential faults; generate simulation results according to the performance of the physical system under normal operating conditions, the performance of the physical system under abnormal operating conditions, and the probability of occurrence of the potential faults; the predicting the probability of occurrence of potential faults includes: predicting the probability of occurrence of potential faults by a fault prediction model, wherein the fault prediction model calculates the probability of occurrence of potential faults by the following formula:

[0259]

[0260] Where P(f) represents the probability of occurrence of potential failure, w0, w i 、v ij They are respectively represented as bias term, linear weight and cross term weight, x i Represents the value of the i-th running parameter, x j It is represented as the value of the jth operating parameter, where j>i and n represents the number of operating parameters.

[0261] Optionally, in the embodiment of the present application, the acquisition module 21 is specifically used to calculate the difference between the simulation result and the actual operation data of the physical system obtained by using a statistical method; determine the accuracy of the digital twin through error analysis to obtain an evaluation result; the determination of the accuracy of the digital twin through error analysis to obtain an evaluation result includes:

[0262] By formula: Calculate the accuracy of the digital twin to obtain an evaluation result; where RMSE is expressed as the evaluation result, y i Represents actual operating data. is represented as the simulation result, and N is represented as the number of data points.

[0263] Optionally, in the embodiment of the present application, the updating module 24 is specifically used to optimize the model parameters in the physical model by a back propagation algorithm; wherein the back propagation algorithm updates the model parameters by the following formula:

[0264]

[0265] Among them, w j is represented as the jth model parameter, η is represented as the learning rate, Expressed as the loss function relative to the model parameters w j The partial derivative of , λ represents the weight of the regularization term, and L represents the loss function, where

[0266] Optionally, in the embodiment of the present application, the processing module 25 is specifically used to use the updated digital twin to perform fault prediction, so as to take preventive measures according to the probability of occurrence of potential faults; automatically adjust the operating parameters of the physical system through an optimization algorithm, wherein the operating parameter optimization calculates the optimal parameter set p of the physical system through the following formula: * :

[0267]

[0268] Among them, f(p) represents the objective function, p represents the parameter set, and p * is represented as the optimal parameter set, and λ is represented as the weight of the regularization term.

[0269] Figure 2 The industrial automation-based digital twin system can perform Figure 1 The implementation principle and technical effect of the digital twin method based on industrial automation described in the illustrated embodiment will not be described in detail. The specific manner in which each module and unit performs operations in the digital twin system based on industrial automation in the above embodiment has been described in detail in the embodiment of the method, and will not be elaborated here.

[0270] In one possible design, Figure 2 The industrial automation-based digital twin system of the illustrated embodiment can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0271] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0272] The processing component 32 is used to: obtain real-time operating parameters from the physical system; construct a digital twin of the physical system in a virtual environment based on the real-time operating parameters and a preset physical model; use the digital twin to simulate the operating state of the physical system and predict the performance of the physical system under different working conditions to generate simulation results; compare the simulation results with the actual operating data of the physical system obtained to evaluate the digital twin and obtain evaluation results; adjust the physical model according to the evaluation results to update the digital twin; use the updated digital twin to predict faults and / or optimize the operating parameters of the physical system.

[0273] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above method.

[0274] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0275] Of course, the computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0276] The input / output interface provides an interface between the processing component and the peripheral interface module, which may be an output device, an input device, etc.

[0277] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0278] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0279] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 The illustrated embodiment is a digital twin method based on industrial automation.

[0280] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0281] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0282] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0283] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A digital twin approach based on industrial automation, It is characterized in that include: Obtain real-time operating parameters from physical systems; Based on the real-time operating parameters and the preset physical model, constructing a digital twin of the physical system in a virtual environment; Using the digital twin to simulate the operating state of the physical system and predict the performance of the physical system under different working conditions to generate simulation results; Comparing the simulation result with the acquired actual operation data of the physical system to evaluate the digital twin and obtain an evaluation result; Adjusting the physical model according to the evaluation result to update the digital twin; The updated digital twin is used to predict faults and / or optimize operating parameters of the physical system.

2. The method according to claim 1, It is characterized in that The obtaining of real-time operating parameters from the physical system includes: Collect real-time sensor data from physical systems through IoT devices; The real-time sensor data is preprocessed to generate real-time operating parameters; wherein the preprocessing includes at least: data cleaning, outlier detection and missing value filling.

3. The method according to claim 1, It is characterized in that The step of constructing a digital twin of the physical system in a virtual environment based on the real-time operating parameters and the preset physical model includes: analyzing real-time operating parameters of the physical system using a machine learning algorithm to extract key features; constructing a dynamic mathematical model of the physical system based on the key features; The operating state of the physical system is simulated in the virtual environment, and is updated in real time through the dynamic mathematical model to complete the construction of the digital twin of the physical system.

4. The method according to claim 3, It is characterized in that The using the digital twin to simulate the operating state of the physical system and predict the performance of the physical system under different working conditions to generate simulation results includes: simulating, using the digital twin, performance of the physical system under normal operating conditions; Using the digital twin to simulate the performance of the physical system under abnormal operating conditions and predict the probability of occurrence of potential failures; generating simulation results according to the performance of the physical system under normal operating conditions, the performance of the physical system under abnormal operating conditions, and the probability of occurrence of the potential failure; The predicted probability of occurrence of a potential failure includes: The probability of occurrence of a potential fault is predicted by a fault prediction model, wherein the fault prediction model calculates the probability of occurrence of a potential fault by the following formula: Where P(f) represents the probability of occurrence of potential failure, w 0 、w i 、v ij They are respectively represented as bias term, linear weight and cross term weight, x i Represents the value of the i-th operating parameter, x j It is represented as the value of the jth operating parameter, where j>i and n represents the number of operating parameters.

5. The method according to claim 1, It is characterized in that The comparing the simulation result with the acquired actual operation data of the physical system to evaluate the digital twin and obtain the evaluation result includes: Calculate the difference between the simulation result and the acquired actual operation data of the physical system using a statistical method; Determining the accuracy of the digital twin through error analysis to obtain an evaluation result; Determining the accuracy of the digital twin through error analysis to obtain an evaluation result includes: By formula: Calculate the accuracy of the digital twin to obtain an evaluation result; where RMSE is expressed as the evaluation result, y i Represents actual operating data. is represented as the simulation result, and N is represented as the number of data points.

6. The method according to claim 1, It is characterized in that The adjusting the physical model according to the evaluation result to update the digital twin includes: Optimizing model parameters in the physical model by a back propagation algorithm; The back propagation algorithm updates the model parameters using the following formula: Among them, w j is represented as the jth model parameter, η is represented as the learning rate, Expressed as the loss function relative to the model parameters w j The partial derivative of , λ represents the weight of the regularization term, and L represents the loss function, where 7. The method according to claim 1, It is characterized in that The using the updated digital twin to predict faults and / or optimize operating parameters of the physical system includes: Using the updated digital twin to predict faults, so as to take preventive measures according to the probability of occurrence of potential faults; The operating parameters of the physical system are automatically adjusted by an optimization algorithm, wherein the operating parameter optimization calculates the optimal parameter set p of the physical system by the following formula: * : Among them, f(p) represents the objective function, p represents the parameter set, and p * is represented as the optimal parameter set, and λ is represented as the weight of the regularization term.

8. A digital twin system based on industrial automation, It is characterized in that include: An acquisition module, used to obtain real-time operating parameters from a physical system; A construction module, used to construct a digital twin of the physical system in a virtual environment based on the real-time operating parameters and a preset physical model; A generation module, used to simulate the operating state of the physical system using the digital twin, and predict the performance of the physical system under different working conditions to generate simulation results; The acquisition module is further used to compare the simulation result with the acquired actual operation data of the physical system to evaluate the digital twin and obtain an evaluation result; An updating module, configured to adjust the physical model according to the evaluation result to update the digital twin; A processing module is used to use the updated digital twin to predict faults and / or optimize operating parameters of the physical system.

9. A computing device, It is characterized in that It comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement the digital twin method based on industrial automation as described in any one of claims 1 to 7.

10. A computer storage medium, It is characterized in that A computer program is stored, and when the computer program is executed by a computer, the digital twin method based on industrial automation as described in any one of claims 1 to 7 is implemented.

Images