Simulation model verification method based on real installation and simulation combined test

Abstract

The invention provides a simulation model verification method based on a real installation and simulation combined test. The method comprises the steps of obtaining a real installation simulation model to be verified; dividing the real installation simulation model into a single-machine-level module and a subsystem-level module, wherein the subsystem-level module is formed by building a plurality of single-machine-level modules; performing uncertainty measurement on each single-machine-level module to obtain the uncertainty of each single-machine-level module; according to the uncertainty of each single-machine-level module, obtaining the uncertainty of each corresponding subsystem-level module; determining the uncertainty of the real installation simulation model according to the uncertainty of each subsystem level module; and when the uncertainty of the real installation simulation model does not exceed the preset range, the verification is passed. By implementing the method, the uncertainty is analyzed layer by layer, the uncertainty of the whole simulation model can be clearly distinguished, the key part and potential risk of model simulation can be more accurately identified, and the risk predicted by the model can be quantified so as to make a corresponding strategy.

Description

Technical Field

[0001] The invention belongs to the technical field of simulation verification, and in particular relates to a simulation model verification method based on a test combining actual installation with simulation. Background Art

[0002] Modeling and simulation have become an important research method in the field of engineering, playing an important role in predicting system performance, evaluating system design, and making decision support. However, due to various uncertainties in the real world, there is often a deviation between the output results of the actual simulation model and the output results of the actual equipment, resulting in inaccuracies in predicting system performance, evaluating system design, and making decision support. In related technologies, when verifying the model, the probability box method is generally used to quantify the total uncertainty of the model, which has low analysis transparency and cannot grasp the analysis of the key components of the simulation model. Summary of the invention

[0003] In view of this, the purpose of the present invention is to provide a simulation model verification method and device based on a combination of actual installation and simulation tests to meet the demand for improving analysis transparency.

[0004] In order to achieve the above object, the present invention provides the following technical solutions:

[0005] According to the first aspect, the present invention provides a simulation model verification method based on a combination of actual installation and simulation tests, including: obtaining an actual installation simulation model to be verified; dividing the actual installation simulation model into a single-machine level module and a subsystem level module, wherein the subsystem level module is composed of multiple single-machine level modules; performing uncertainty measurement on each single-machine level module to obtain the uncertainty of each single-machine level module; obtaining the uncertainty of each corresponding subsystem level module based on the uncertainty of each single-machine level module; determining the uncertainty of the actual installation simulation model based on the uncertainty of each subsystem level module; when the uncertainty of the actual installation simulation model does not exceed a preset range, the verification is passed.

[0006] Optionally, the construction process of the actual installation simulation model includes: collecting actual installation data and sensor data, the actual installation data is used to characterize the structural parameters of the simulation model construction; preprocessing the actual installation data and sensor data, including data sampling, data noise reduction processing, and invalid data elimination; building the physical structure of the model according to the preprocessed actual installation data; based on the preprocessed sensor data, combined with the physical structure performance characteristics, establishing a network structure mapping relationship between input data and output data; according to the trend of decreasing data dimension, adjusting the network structure mapping relationship to obtain the actual installation simulation model.

[0007] Optionally, uncertainty measurement is performed on each stand-alone-level module to obtain the uncertainty of each stand-alone-level module, including: determining the uncertainty factors of the stand-alone-level module according to a preset list of uncertainty factors; calculating the uncertainty of each uncertainty factor according to the uncertainty calculation rules of each uncertainty factor; and calculating the square root of the sum of the squares of the uncertainties of each uncertainty factor to obtain the uncertainty of the stand-alone-level module.

[0008] Optionally, the uncertainty of each uncertain factor is calculated according to the uncertainty calculation rules of each uncertain factor, including: when the uncertain factor is a random uncertain factor, multiple measurement values ​​are obtained by repeatedly measuring the target physical quantity that generates the random uncertain factor; an average value is determined based on the multiple measurement values, and the standard deviation of the single measurement is determined based on the single measurement value and the average value; a correction factor is obtained by looking up a table, and the uncertainty of the random uncertain factor is determined based on the correction factor and the standard deviation of the single measurement; when the uncertain factor is a fixed uncertain factor, the manufacturer's technical specification document or prior knowledge that generates the fixed uncertainty is obtained to determine the uncertainty of the fixed uncertainty.

[0009] Optionally, based on the uncertainty of each stand-alone module, the uncertainty of each corresponding subsystem-level module is obtained, including: determining the interface between each stand-alone module and the uncertainty propagation rule between each stand-alone module based on the connection relationship between each stand-alone module in the subsystem-level module; determining the uncertainty generated by the stand-alone module to the subsystem-level module based on the uncertainty propagation rule between each stand-alone module; determining the interface characteristics between each stand-alone module, and determining the uncertainty of the interface based on the interface characteristics; determining the uncertainty of each corresponding subsystem-level module based on the uncertainty generated by the stand-alone module to the subsystem-level module and the uncertainty of the interface.

[0010] Optionally, a simulation model verification method based on a combination of actual installation and simulation tests also includes: iteratively training the actual installation simulation model based on the test data set; verifying the model accuracy through a loss function or a metric to obtain high-confidence model parameters; verifying the accuracy of the actual installation simulation model with high-confidence model parameters through accuracy, precision, and recall rate to obtain a confidence interval of the prediction result under a target confidence level, and the target confidence level is determined based on a preset range of uncertainty.

[0011] Optionally, a simulation model verification method based on a combination of actual installation and simulation tests also includes: generating random numerical values ​​representing each uncertainty factor as a simulation model parameter according to a preset list of uncertainty factors and the uncertainty value range and distribution type of the corresponding factors; adjusting the actual installation simulation model according to each simulation model parameter, inputting target data, running the actual installation simulation model, and obtaining a first actual installation simulation model output; statistically analyzing the first actual installation simulation model output, and using a variance decomposition method to quantify the contribution of each uncertainty factor to the output variance; determining the degree of influence of each uncertainty factor on the actual installation simulation model according to the size of the contribution; sorting the factors from large to small according to the degree of influence, and selecting the target number of uncertainties. factors; for the uncertainty factors of the target quantity, random values ​​representing the corresponding uncertainty factors are generated in the corresponding uncertainty value range and distribution type to form multiple sets composed of random values ​​of multiple uncertain factors; the actual installation simulation model is adjusted according to the multiple sets composed of random values ​​of multiple uncertain factors, and the target data is input respectively, and the corresponding actual installation simulation model is run to obtain multiple groups of second actual installation simulation model outputs; the output when the target data is input to the actual installation is obtained; the output when the target data is input to the actual installation is compared with the multiple groups of second actual installation simulation model outputs to obtain the optimal combination of random values ​​of uncertain factors, and the optimal combination of random values ​​of uncertain factors is used as the uncertainty factor parameters corresponding to the actual installation simulation model.

[0012] According to the second aspect, an embodiment of the present invention provides a simulation model verification device based on a combination of actual installation and simulation tests, including: a model acquisition module, used to obtain an actual installation simulation model to be verified; a model division module, used to divide the actual installation simulation model into a single-machine level module and a subsystem level module, the subsystem level module is composed of a plurality of single-machine level modules; a single-machine level uncertainty module, used to measure the uncertainty of each single-machine level module to obtain the uncertainty of each single-machine level module; a subsystem level uncertainty module, used to obtain the uncertainty of each corresponding subsystem level module according to the uncertainty of each single-machine level module; a simulation model uncertainty module, used to determine the uncertainty of the actual installation simulation model according to the uncertainty of each subsystem level module; a verification module, used to verify that the uncertainty of the actual installation simulation model does not exceed a preset range.

[0013] According to the third aspect, an embodiment of the present invention provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the steps of the simulation model verification method based on a combination of actual installation and simulation tests as described in the first aspect or any implementation scheme of the first aspect.

[0014] According to the fourth aspect, an embodiment of the present invention provides a computer storage medium having computer instructions stored thereon, which, when executed by a processor, implement the steps of the simulation model verification method based on the combination of actual installation and simulation tests as described in the first aspect or any embodiment of the first aspect.

[0015] This method provides a simulation model verification method based on a combination of actual installation and simulation tests. The actual installation simulation model is divided into a single-machine level module and a subsystem level module, and the uncertainty is analyzed layer by layer. It can clearly distinguish the uncertainty of the entire simulation model, and more accurately identify the key parts and potential risks of the model simulation. In addition, by analyzing the uncertainty verification model of the simulation model, the credibility and reliability of the model can be improved, and the risks predicted by the model can be quantified to facilitate the formulation of corresponding strategies.

[0016] Other advantages, objectives and features of the present invention will be described in the following description and will be apparent to those skilled in the art to some extent, or those skilled in the art may be taught from the practice of the present invention. The objectives and other advantages of the present invention may be realized and obtained through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to make the purpose, technical solution and beneficial effects of the present invention clearer, the present invention provides the following drawings for illustration:

[0018] Figure 1 It is a specific flow chart of a simulation model verification method based on a test combining actual installation with simulation in the present invention;

[0019] Figure 2 It is a principle block diagram of a specific example of an electronic device in an embodiment of the present invention. DETAILED DESCRIPTION

[0020] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, it can also be the internal connection of two components, it can be a wireless connection, or it can be a wired connection. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0022] In addition, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0023] The embodiment of the present invention provides a simulation model verification method based on a combination of actual installation and simulation test, such as Figure 1 As shown, including:

[0024] S101, obtaining a real-installation simulation model to be verified;

[0025] S102, dividing the actual installation simulation model into a single-machine level module and a subsystem level module, wherein the subsystem level module is composed of a plurality of single-machine level modules;

[0026] S103, measuring the uncertainty of each stand-alone module to obtain the uncertainty of each stand-alone module;

[0027] S104, obtaining the uncertainty of each corresponding subsystem level module according to the uncertainty of each stand-alone level module;

[0028] S105, determining the uncertainty of the actual installation simulation model according to the uncertainty of each subsystem-level module;

[0029] S106, when the uncertainty of the actual simulation model does not exceed the preset range, the verification is passed.

[0030] Exemplarily, the actual installation simulation model refers to a simulation model of a simulated physical device running on the system, and the simulation model can simulate the behavior and performance of the physical device. The way to obtain the actual installation simulation model to be verified can be to obtain the corresponding simulation data from the simulation software. A complete simulation model is constructed from multiple parts. For example, when the physical device simulated by the simulation model is a reducer, the simulation model at least includes gears, shafts, housing modules, lubrication system modules, sealing system modules, sensors, controllers, and actuators. In order to reduce the difficulty of analysis, this embodiment decomposes the complex actual installation simulation model into stand-alone modules and subsystem modules. The stand-alone module represents the smallest module that can run independently, such as gears, oil pools, oil pipes, oil pumps, etc., while the subsystem module is a larger system composed of multiple stand-alone modules, such as a gear module, including various gears and related components, and a lubrication system module, including an oil pool, oil pipes, oil pumps, etc.

[0031] A specific method for dividing the actual simulation model into stand-alone-level modules and subsystem-level modules is to first obtain a list of stand-alone-level modules pre-divided according to the structural manual of the actual device and a mapping relationship file between each stand-alone-level module and subsystem-level module, which are pre-stored in the system, and then automatically divide the actual simulation model into stand-alone-level modules and subsystem-level modules according to the list of stand-alone-level modules and the mapping relationship file between each stand-alone-level module and subsystem-level module.

[0032] Since there are a large number of stand-alone-level modules in complex and large physical equipment, in order to further reduce the user's operating costs, this embodiment also provides a specific method for dividing the actual simulation model into stand-alone-level modules in the absence of a pre-divided list of stand-alone-level modules. The specific method can also be to use the smallest module that runs independently as a stand-alone-level module according to the structure manual of the actual device. Specifically, the corresponding modules of the simulation model can be determined according to the structure manual of the actual device, and each module can be divided independently. Automated tools can be used to assist in independence verification. When the independence verification passes, the module can be used as a stand-alone-level module. When the independence verification fails, the range of modules to be verified is expanded, and the expanded structure is re-verified for independence until it passes the independence verification.

[0033] Uncertainty may come from fluctuations in simulation performance caused by unpredictable and random factors, such as changes in material properties, or from fluctuations in simulation performance caused by fixed and predictable factors, such as model structure parameters. Statistical methods and sensitivity analysis can be used to evaluate the uncertainty of each stand-alone module.

[0034] Since the subsystem-level module is composed of multiple stand-alone-level modules, the uncertainty of the stand-alone-level module will propagate upward to the subsystem-level module. Therefore, when calculating the uncertainty of the subsystem-level module, it is necessary to consider the interaction and dependence between the modules. Similarly, the actual installation simulation model is composed of multiple subsystem-level modules. It is necessary to combine the uncertainties of all subsystem-level modules to obtain the uncertainty of the entire actual installation simulation model. Specifically, the uncertainty of the entire model can be evaluated through Monte Carlo simulation. When the uncertainty of the actual installation simulation model exceeds the preset range, it means that the uncertainty is too large, the reliability of the simulation model is low, and the simulation model parameters need to be readjusted. The size of the preset range is determined according to the actual simulation model and user needs, and this embodiment does not limit it. When the uncertainty of the actual installation simulation model does not exceed the preset range, it means that the reliability of the simulation model is high.

[0035] This embodiment provides a simulation model verification method based on a combination of actual installation and simulation tests, which divides the actual installation simulation model into single-machine level modules and subsystem level modules, analyzes the uncertainty layer by layer, can clearly distinguish the uncertainty of the entire simulation model, and more accurately identify the key parts and potential risks of the model simulation. In addition, by analyzing the uncertainty verification model of the simulation model, the credibility and reliability of the model can be improved, and the risks predicted by the model can be quantified to facilitate the formulation of corresponding strategies.

[0036] As an optional implementation method, a simulation model verification method based on a combination of actual installation and simulation tests, the construction process of the actual installation simulation model includes: collecting actual installation data and sensor data, the actual installation data is used to characterize the structural parameters of the simulation model construction; preprocessing the actual installation data and sensor data, including data sampling, data noise reduction processing, and invalid data elimination; building the model physical structure according to the preprocessed actual installation data; based on the preprocessed sensor data, combined with the physical structure performance characteristics, establishing a network structure mapping relationship between input data and output data; according to the trend of decreasing data dimension, adjusting the network structure mapping relationship to obtain the actual installation simulation model.

[0037] Exemplarily, the installed data represents the physical parameter measurements of the physical equipment, such as size, weight, material properties, etc. These data are used to construct the physical structure and parameters of the simulation model to ensure that the model can accurately reflect the characteristics of the physical equipment. Sensor data includes the operating status, performance indicators, environmental conditions, etc. of the equipment. These data are helpful for understanding the dynamic behavior and performance of the equipment. In order to reduce the amount of data, improve data quality and processing efficiency, this embodiment pre-processes the data, including sampling the data, and using filters, smoothing algorithms and other methods to remove noise in the data, identify and eliminate outliers, erroneous data or irrelevant data.

[0038] Use the pre-processed installed data, combined with the physical structure performance characteristics, to build a physical model of the physical equipment. In addition to building the three-dimensional structure of the physical equipment, a network structure mapping relationship between the input data and the output data is also required. Specifically, a polynomial response model, a Gaussian Kriging model, a chaotic polynomial model, a linear regression model, a neural network model, a radial basis function model, etc. can be used to establish a mapping relationship between input and output. This embodiment does not limit this, and those skilled in the art can determine it as needed. According to the trend of decreasing data dimension, the network structure is optimized, including feature engineering, model selection, parameter adjustment and other steps to improve the performance and generalization ability of the model.

[0039] As an optional implementation, uncertainty measurement is performed on each stand-alone-level module to obtain the uncertainty of each stand-alone-level module, including: determining the uncertainty factors of the stand-alone-level module according to a preset list of uncertainty factors; calculating the uncertainty of each uncertainty factor according to the uncertainty calculation rules of each uncertainty factor; and calculating the square root of the sum of the squares of the uncertainties of each uncertainty factor to obtain the uncertainty of the stand-alone-level module.

[0040] Exemplarily, according to a preset list of uncertain factors, all potential uncertain factors affecting the performance of a stand-alone module are identified. The list of uncertain factors may include changes in material properties, model structure parameters, etc. The specific list of uncertain factors may be set according to actual conditions, and this embodiment does not limit it. According to the uncertainty calculation rules of each uncertain factor, an uncertainty value is calculated for each uncertain factor, and the calculation rules may be determined based on analyzing historical data, expert review, or establishing a prediction model. When the uncertainty factor of the stand-alone module is a model structure parameter, the same physical quantity is measured multiple times, and its average value and standard deviation are calculated to estimate the model structure parameter error, and a statistical distribution (such as a normal distribution) is used to determine the possible error range of the measured value, and the measuring equipment is calibrated using a standard device of known accuracy to determine its error, and the calibration curve is obtained by a calibration experiment to evaluate the measurement error. The simulation model is simulated according to the data that generates the measurement error, and other model parameters remain unchanged to determine its error simulation result. The simulation model is simulated again according to the calibrated data, and other model parameters remain unchanged to determine its calibrated simulation result. The error simulation result is compared with the calibrated simulation result, and the corresponding uncertainty is quantified according to the difference interval. Specifically, a mapping relationship between the difference interval and the uncertainty can be given in advance based on expert judgment.

[0041] After determining the uncertainty of each uncertainty factor, it is necessary to integrate them to obtain the uncertainty D of the stand-alone module, which is obtained by the following calculation formula:

[0042]

[0043] Among them, α i represents the uncertainty of the ith uncertain factor, and N represents the total number of uncertain factors.

[0044] This embodiment provides a simulation model verification method based on a combination of actual installation and simulation tests. By presetting a list of uncertainty factors, it can ensure that all relevant uncertainties are taken into account during the evaluation process, thereby providing a comprehensive uncertainty evaluation. Different calculation methods are proposed for different uncertainties, which can adapt to different types of uncertainty distributions and calculation rules, making it suitable for a variety of different application scenarios.

[0045] As an optional implementation, the uncertainty of each uncertain factor is calculated according to the uncertainty calculation rules of each uncertain factor, including: when the uncertain factor is a random uncertain factor, multiple measurement values ​​are obtained by repeatedly measuring the target physical quantity that generates the random uncertain factor; an average value is determined based on the multiple measurement values, and the standard deviation of the single measurement is determined based on the single measurement value and the average value; a correction factor is obtained by looking up a table, and the uncertainty of the random uncertain factor is determined based on the correction factor and the standard deviation of the single measurement; when the uncertain factor is a fixed uncertain factor, the manufacturer's technical specification document or prior knowledge that generates the fixed uncertainty is obtained to determine the uncertainty of the fixed uncertainty.

[0046] Exemplarily, random uncertainties characterize random fluctuations or unpredictable factors. In order to evaluate random uncertainties, it is necessary to perform repeated measurements of the target physical quantity under the same conditions to reduce the influence of systematic effects. Then, all the measured values ​​are added up and divided by the number of measurements to obtain the average value, which can be used as the best estimate of the target physical quantity. The square of the difference between each measured value and the average value is calculated, then summed, and then divided by the number of measurements minus one (i.e., the degree of freedom), and finally the square root is taken to obtain the standard deviation of a single measurement. The standard deviation is an important indicator for measuring the degree of dispersion of the measured values, which reflects the size of the random uncertainties. After determining the standard deviation, the correction factor is determined according to a specific statistical table or standard. This correction factor is usually used to adjust the standard deviation so that it more accurately reflects the uncertainty of the random uncertainties. For example, for a specific confidence level (such as 95%), the correction factor is used to adjust the standard deviation to calculate the interval containing the uncertainty of the random uncertainties. The uncertainty of the random uncertainties is obtained by multiplying the correction factor by the standard deviation of a single measurement. Uncertainty represents the fluctuation range of the measured value that may be caused by random uncertainties at a given confidence level.

[0047] Fixed uncertainties are characterized by known, fixed factors, such as manufacturer specifications, design parameters, etc. For fixed uncertainties, the uncertainty can be determined based on the manufacturer's specification documents or based on prior knowledge. The manufacturer's specification documents or based on prior knowledge include the performance parameters of the equipment or system, such as maximum allowable error, stability, linearity, etc., as well as the accuracy grade of the equipment, calibration intervals, uncertainty statements on the calibration certificate, etc. The above data can be used to directly estimate the uncertainty of fixed uncertainties.

[0048] This embodiment provides a simulation model verification method based on a combination of actual installation and simulation tests, and proposes an uncertainty calculation method for random uncertainties and fixed uncertainties. For random uncertainties, the method of this embodiment is based on statistical principles and uses standard deviations to quantify the variability of random uncertainties. The correction factor obtained by table lookup can take into account the sample size and confidence level, making the uncertainty assessment more accurate and reliable. For fixed uncertainties, the use of technical specification documents or prior knowledge provided by the manufacturer can ensure that the assessment is more accurate and rigorous. The method of this embodiment distinguishes between the processing methods of random uncertainties and fixed uncertainties, which helps to adopt the most appropriate evaluation method for different types of uncertainties.

[0049] As an optional implementation, the uncertainty of each corresponding subsystem-level module is obtained based on the uncertainty of each stand-alone-level module, including: determining the interface between each stand-alone-level module and the uncertainty propagation rule between each stand-alone-level module based on the connection relationship between each stand-alone-level module in the subsystem-level module; determining the uncertainty generated by the stand-alone-level module to the subsystem-level module based on the uncertainty propagation rule between each stand-alone-level module; determining the interface characteristics between each stand-alone-level module, and determining the uncertainty of the interface based on the interface characteristics; determining the uncertainty of each corresponding subsystem-level module based on the uncertainty generated by the stand-alone-level module to the subsystem-level module and the uncertainty of the interface.

[0050] For example, for uncertain factors that affect data transmission of subsystem-level modules, in this embodiment, the interfaces between the individual stand-alone-level modules and the uncertainty propagation rules between the individual stand-alone-level modules can be determined according to the connection relationship between the individual stand-alone-level modules in the subsystem-level module, and the uncertainty of the subsystem-level module can be determined according to the propagation rules. Specific examples are as follows:

[0051] Assume that a subsystem-level module includes a stand-alone module A, a stand-alone module B, and a stand-alone module C, and that the stand-alone module A and the stand-alone module B are connected by a first interface, and the stand-alone module A and the stand-alone module C are connected by a second interface, and that the data flow direction is from the stand-alone module B to the stand-alone module A to the stand-alone module C. The uncertainty D of the stand-alone module B B It is propagated to the stand-alone module A through the first interface, and the propagation coefficient is k1. The uncertainty of the stand-alone module A due to the propagation of the stand-alone module B is D A =k1·D B , the uncertainty D of the single-machine module A A It is propagated to the stand-alone module C through the second interface, and the propagation coefficient is k2. The uncertainty of the stand-alone module C due to the propagation of the stand-alone module A is:

[0052] D C=k2·D A =k1·k2·D B ,

[0053] The total uncertainty caused by the stand-alone module to the subsystem module is determined as:

[0054] D As =D A +D B +D C =(1+k1+k1·k2)D B .

[0055] It should be noted that the propagation coefficient k1 and the propagation coefficient k2 can be determined through the mathematical relationship of data transmission between stand-alone modules. Specifically, the partial derivative between the input and output of the stand-alone module is calculated and the partial derivative is used as the propagation coefficient.

[0056] Determine the interface characteristics between each stand-alone module. According to the interface characteristics, the uncertainty of the interface is determined according to the fixed transmission loss of the interface. Specifically, the uncertainty of the first interface can be D int1 =L1·D B , the uncertainty of the second interface can be D int2 =L2·D A .

[0057] According to the uncertainty generated by the stand-alone module to the subsystem module and the uncertainty of the interface, the uncertainty of each corresponding subsystem module is determined, including: obtaining the cumulative sum of the uncertainty generated by the stand-alone module to the subsystem module and the uncertainties of multiple interfaces.

[0058] Similarly, according to the uncertainty of each subsystem-level module, the specific process of determining the uncertainty of the actual installation simulation model may include: determining the interface between each subsystem-level module and the uncertainty propagation rule between each subsystem-level module according to the connection relationship of the subsystem-level module in the actual installation simulation model; determining the uncertainty generated by the subsystem-level module to the actual installation simulation model according to the uncertainty propagation rule between each subsystem-level module; determining the interface characteristics between each subsystem-level module, and determining the uncertainty of the interface according to the interface characteristics; determining the uncertainty of the actual installation simulation model according to the uncertainty generated by the subsystem-level module to the actual installation simulation model and the uncertainty of the interface. Please refer to the above description for the specific content, which will not be repeated here.

[0059] This embodiment provides a simulation model verification method based on a combination of actual installation and simulation tests, which takes into account the propagation of uncertainty and the transmission loss of the interface during the propagation process, thereby improving the accuracy of uncertainty calculation of subsystem-level modules.

[0060] As an optional implementation, a simulation model verification method based on a combination of actual installation and simulation tests also includes: iteratively training the actual installation simulation model based on the test data set; verifying the model accuracy through a loss function or a metric to obtain high-confidence model parameters; verifying the accuracy of the actual installation simulation model with high-confidence model parameters through accuracy, precision, and recall rate to obtain a confidence interval of the prediction result under the target confidence level.

[0061] Exemplarily, this embodiment provides iterative training of the actual installation simulation model based on the test data set, and verifies the model accuracy through verification. First, collect the test data set required for the actual installation simulation, ensure that the data set covers all the input conditions and expected outputs required for the simulation model, and clean, standardize and normalize the data to ensure data quality. Then, select a suitable model architecture according to the simulation requirements, such as a neural network, a support vector machine, a decision tree, etc. Use the data set to iteratively train the model, adjust the model parameters to minimize the loss function, which can be mean square error (MSE), cross entropy, etc., and use optimization algorithms such as gradient descent and Adam to adjust the parameters.

[0062] After training, the validation set is used to verify the accuracy of the model to avoid overfitting, and the model performance is evaluated through loss functions or metrics (such as accuracy, precision, and recall). Then, the model parameters are adjusted according to the verification results, and high-confidence model parameters are obtained through multiple iterations of training and parameter adjustment. Then, the model is evaluated in detail using indicators such as accuracy, precision, and recall. The calculation methods of accuracy, precision, and recall are existing technologies and will not be repeated here.

[0063] The target confidence level can be preset by the user according to the desired accuracy level, such as 95%. The upper and lower bounds of the prediction results are calculated according to the preset target confidence level.

[0064] As an optional implementation, a simulation model verification method based on a combination of actual installation and simulation tests also includes: generating random numerical values ​​representing various uncertainty factors as simulation model parameters according to a preset list of uncertainty factors and the uncertainty value range and distribution type of the corresponding factors; adjusting the actual installation simulation model according to each simulation model parameter, inputting target data, running the actual installation simulation model, and obtaining a first actual installation simulation model output; statistically analyzing the first actual installation simulation model output, and using a variance decomposition method to quantify the contribution of various uncertainty factors to the output variance; determining the degree of influence of each uncertainty factor on the actual installation simulation model according to the size of the contribution; sorting the factors from large to small according to the degree of influence, and selecting the target number of uncertainties. Qualitative factors; for the uncertainty factors of the target quantity, a probability model is used to generate random values ​​representing the corresponding uncertainty factors in the corresponding uncertainty value range and distribution type, forming multiple sets composed of a combination of random values ​​of multiple uncertain factors; the actual installation simulation model is adjusted according to the multiple sets composed of a combination of random values ​​of multiple uncertain factors, and the target data is input respectively, and the corresponding actual installation simulation model is run to obtain multiple groups of second actual installation simulation model outputs; the output when the target data is input to the actual installation is obtained; the output when the target data is input to the actual installation is compared with the multiple groups of second actual installation simulation model outputs to obtain the optimal combination of random values ​​of uncertain factors, and the optimal combination of random values ​​of uncertain factors is used as the uncertainty factor parameter corresponding to the actual installation simulation model.

[0065] Exemplarily, for uncertainty factors whose uncertainty value range and distribution type are known, any random number can be selected within the uncertainty value range as the specific value of the uncertainty factor, and the specific values ​​of multiple uncertainty factors are used to adjust the corresponding structure or parameters of the simulation model. Input the target data into the adjusted simulation model for simulation. The target data can be any data consistent with the input of the physical equipment, which is used to drive the operation of the simulation model. After the simulation model obtains the output, the output of the simulation model can be statistically analyzed, such as calculating the mean, standard deviation and other statistics, and then using the variance decomposition method to quantify the contribution of each uncertainty factor to the output variance. Specifically, the Sobol method can be used. The contribution of the variance can be regarded as the degree of influence of the corresponding uncertainty factor on the actual simulation model.

[0066] Sort the influences from large to small, and select several uncertainty factors with large influences as the targets for subsequent calculations. For the uncertainty factors with large influences, generate random values ​​representing the corresponding uncertainty factors in the corresponding uncertainty value range and distribution type, and form multiple sets composed of random values ​​of multiple uncertainty factors. According to the multiple sets composed of random values ​​of multiple uncertainty factors, adjust the actual installation simulation models respectively to form multiple groups of simulation models, and input target data into the multiple groups of simulation models respectively. The target data can be any data consistent with the input of the physical equipment, which is used to drive the simulation model to run. The multiple groups of simulation models are run separately to obtain multiple groups of second actual installation simulation model outputs.

[0067] At the same time, the target data is input into the actual equipment to obtain the output of the actual equipment; the output of the actual equipment is compared with multiple sets of second actual simulation model outputs to find the simulation output result closest to the output of the actual equipment. The simulation result with the closest distance can indicate that the establishment of the simulation model is more in line with the actual situation. Therefore, the combination of multiple random values ​​of uncertain factors corresponding to the simulation output result with the closest distance is used as the uncertainty factor parameter corresponding to the actual simulation model. This method can make the simulation model more in line with the actual equipment and improve the accuracy of the simulation.

[0068] This embodiment provides a simulation model verification device based on a test combining actual installation and simulation, including:

[0069] The model acquisition module is used to obtain the actual assembly simulation model to be verified; please refer to the corresponding part of the above method embodiment for details, which will not be repeated here.

[0070] The model division module is used to divide the actual simulation model into a single-machine level module and a subsystem level module. The subsystem level module is composed of multiple single-machine level modules. Please refer to the corresponding part of the above method embodiment for details, which will not be repeated here.

[0071] The stand-alone level uncertainty module is used to measure the uncertainty of each stand-alone level module to obtain the uncertainty of each stand-alone level module; please refer to the corresponding part of the above method embodiment for details, which will not be repeated here.

[0072] The subsystem-level uncertainty module is used to obtain the uncertainty of each corresponding subsystem-level module according to the uncertainty of each stand-alone-level module; please refer to the corresponding part of the above method embodiment for details, which will not be repeated here.

[0073] The uncertainty module of the simulation model is used to determine the uncertainty of the actual simulation model according to the uncertainty of each subsystem-level module; please refer to the corresponding part of the above method embodiment for details, which will not be repeated here.

[0074] The verification module is used to verify that the verification fails when the uncertainty of the actual simulation model exceeds a preset range. For details, please refer to the corresponding part of the above method embodiment, which will not be repeated here.

[0075] The present application also provides an electronic device, such as Figure 2 As shown, a processor 501 and a memory 502, wherein the processor 501 and the memory 502 may be connected via a bus or other means.

[0076] The processor 501 may be a central processing unit (CPU). The processor 501 may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or a combination of the above chips.

[0077] The memory 502 is a non-transitory computer-readable storage medium that can be used to store non-transitory software programs, non-transitory computer executable programs and modules, such as program instructions / modules corresponding to the simulation model verification method based on the combined test of actual installation and simulation in the embodiment of the present invention. The processor executes various functional applications and data processing of the processor by running the non-transitory software programs, instructions and modules stored in the memory.

[0078] The memory 502 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created by the processor, etc. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory 502 may optionally include a memory remotely arranged relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0079] The one or more modules are stored in the memory 502, and when executed by the processor 501, the following is performed: Figure 1 The simulation model verification method based on the combined test of actual installation and simulation in the illustrated embodiment.

[0080] For details of the above electronic equipment, please refer to Figure 1 The corresponding related descriptions and effects in the illustrated embodiments can be understood and will not be repeated here.

[0081] This embodiment also provides a computer storage medium, which stores computer executable instructions, and the computer executable instructions can execute the simulation model verification method based on the combination of actual installation and simulation in any of the above method embodiments. Among them, the storage medium can be a disk, an optical disk, a read-only memory (ROM), a random access memory (RAM), a flash memory, a hard disk (HDD) or a solid-state drive (SSD), etc.; the storage medium can also include a combination of the above types of memory.

[0082] Finally, it should be noted that the above preferred embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail through the above preferred embodiments, those skilled in the art should understand that various changes can be made in form and details without departing from the scope defined by the claims of the present invention.

Claims

1. A simulation model verification method based on a combination of actual installation and simulation test, characterized in that: include: Obtain the actual simulation model to be verified; The actual installation simulation model is divided into a single-machine level module and a subsystem level module. The subsystem level module is composed of multiple single-machine level modules. The uncertainty of each stand-alone module is measured to obtain the uncertainty of each stand-alone module; According to the uncertainty of each stand-alone module, the uncertainty of each corresponding subsystem module is obtained; Determine the uncertainty of the actual installation simulation model based on the uncertainty of each subsystem-level module; When the uncertainty of the actual simulation model does not exceed the preset range, the verification is passed.

2. A simulation model verification method based on a combination of actual installation and simulation test according to claim 1, characterized in that: The construction process of the actual installation simulation model includes: Collect actual installation data and sensor data. The actual installation data is used to characterize the structural parameters of the simulation model. Pre-process the installed data and sensor data, including data sampling, data noise reduction, and invalid data removal; Construct the physical structure of the model based on the pre-processed actual installation data; Based on the preprocessed sensor data and combined with the physical structure performance characteristics, a network structure mapping relationship between input data and output data is established; According to the decreasing trend of data dimension, the network structure mapping relationship is adjusted to obtain the actual simulation model.

3. A simulation model verification method based on a combination of actual installation and simulation test according to claim 1, characterized in that: The uncertainty of each stand-alone module is measured to obtain the uncertainty of each stand-alone module, including: Determine the uncertainty factors of the stand-alone module according to the preset uncertainty factor list; Calculate the uncertainty of each uncertain factor according to the uncertainty calculation rules of each uncertain factor; The square root of the sum of the squares of the uncertainties of each uncertain factor is calculated to obtain the uncertainty of the stand-alone module.

4. A simulation model verification method based on a combination of actual installation and simulation test according to claim 3, characterized in that: According to the uncertainty calculation rules of each uncertain factor, calculate the uncertainty of each uncertain factor, including: When the uncertain factors are random uncertain factors, multiple measurement values ​​are obtained by repeatedly measuring the target physical quantity that generates the random uncertain factors; Determine the average value based on multiple measurements, and determine the standard deviation of a single measurement based on the single measurement value and the average value; The correction factor is obtained by looking up the table, and the uncertainty of the random uncertainty factor is determined based on the correction factor and the standard deviation of a single measurement; When the uncertainties are fixed uncertainties, the manufacturer's technical specification documents or prior knowledge that generate the fixed uncertainties are obtained to determine the uncertainty of the fixed uncertainties.

5. A simulation model verification method based on a combination of actual installation and simulation test according to claim 4, characterized in that: According to the uncertainty of each stand-alone module, the uncertainty of each corresponding subsystem module is obtained, including: According to the connection relationship between each stand-alone module in the subsystem level module, determine the interface between each stand-alone module and the uncertainty propagation rule between each stand-alone module; According to the uncertainty propagation rules between each stand-alone module, determine the uncertainty caused by the stand-alone module to the subsystem module; Determine the interface characteristics between each stand-alone module, and determine the uncertainty of the interface based on the interface characteristics; According to the uncertainty of the single-machine level module on the subsystem level module and the uncertainty of the interface, the uncertainty of each corresponding subsystem level module is determined.

6. A simulation model verification method based on a combination of actual installation and simulation test according to any one of claims 1 to 5, characterized in that: Also includes: Iteratively train the actual assembly simulation model based on the experimental data set; Verify the model accuracy through loss functions or metrics to obtain high-confidence model parameters; The accuracy of the actual simulation model with high-confidence model parameters is verified by accuracy, precision, and recall, and the confidence interval of the prediction result under the target confidence level is obtained. The target confidence level is determined according to the preset range of uncertainty.

7. A simulation model verification method based on a combination of actual installation and simulation test according to any one of claims 1 to 5, characterized in that: Also includes: According to the preset uncertainty factor list and the uncertainty value range and distribution type of the corresponding factors, random values ​​representing the uncertainty factors are generated as simulation model parameters; Adjusting the actual installation simulation model according to various simulation model parameters, inputting target data, running the actual installation simulation model, and obtaining a first actual installation simulation model output; Statistically analyze the output of the first real-life simulation model and use variance decomposition methods to quantify the contribution of each uncertainty factor to the output variance; According to the contribution, determine the influence of each uncertain factor on the actual installation simulation model; Select the uncertainty factors of the target quantity according to the order of influence from large to small; For the uncertainty factors of the target quantity, random values ​​representing the corresponding uncertainty factors are generated in the corresponding uncertainty value range and distribution type to form multiple sets composed of random values ​​of multiple uncertainty factors; Adjusting the actual installation simulation models respectively according to a plurality of sets composed of random values ​​of a plurality of uncertain factors, inputting target data respectively, running the corresponding actual installation simulation models, and obtaining a plurality of sets of second actual installation simulation model outputs; Get the output of inputting the target data to the implementation; The output when the target data is input to the actual installation is compared with multiple sets of second actual installation simulation model outputs to obtain the optimal combination of random values ​​of uncertain factors, and the optimal combination of random values ​​of uncertain factors is used as the uncertainty factor parameters corresponding to the actual installation simulation model.

8. A simulation model verification device based on a combination of actual installation and simulation test, characterized in that: include: A model acquisition module is used to obtain a real-installation simulation model to be verified; A model division module is used to divide the actual installation simulation model into a single-machine level module and a subsystem level module. The subsystem level module is composed of multiple single-machine level modules. The single-machine level uncertainty module is used to measure the uncertainty of each single-machine level module and obtain the uncertainty of each single-machine level module; The subsystem level uncertainty module is used to obtain the uncertainty of each corresponding subsystem level module according to the uncertainty of each stand-alone level module; The uncertainty module of the simulation model is used to determine the uncertainty of the actual simulation model according to the uncertainty of each subsystem level module; The verification module is used to verify that the uncertainty of the actual simulation model does not exceed the preset range.

9. An electronic device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the steps of the simulation model verification method based on the combined actual installation and simulation test as described in any one of claims 1 to 7.

10. A computer storage medium having computer instructions stored thereon, characterized in that: When the instruction is executed by the processor, the steps of the simulation model verification method based on the combined actual installation and simulation test described in any one of claims 1-7 are implemented.

Images