A Functional Mapping Management Method for a Data Acquisition System and Simulation Software

Abstract

The present invention provides a functional mapping management method for a data acquisition system and a simulation software, which relates to the technical field of data processing. The method includes: collecting an original data set from a target device; performing noise filtering and outlier detection on the original data set; extracting a plurality of feature data corresponding to the simulation software model; generating an initial mapping parameter set according to the mapping rule between the feature data and the input parameters of the simulation software model; dynamically adjusting the initial mapping parameter set; converting the feature data into input format data required by the simulation software model; performing consistency verification; inputting into the simulation software; verifying the simulation result data. If the verification fails, feedback adjustment parameters are generated according to the difference degree between the simulation result data and the feature data; if the verification passes, the simulation result data is used as the output data for the operation evaluation of the target device. The present invention improves the autonomy and accuracy of functional mapping.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly to a method for managing the functional mapping between a data acquisition system and simulation software. Background Art

[0002] With the rapid development of industrial simulation and digital modeling technologies, data acquisition systems and simulation software have been widely used in multiple fields such as manufacturing, transportation, and energy. In the prior art, a data acquisition system collects the operation data of physical devices in real time through sensors or monitoring devices and converts it into input data available for simulation software, thereby realizing the prediction, optimization, and evaluation of the operation of complex systems. For example, in the manufacturing industry, a common practice is to collect operation data such as vibration and temperature of mechanical equipment and input it into simulation software to predict equipment failures. However, existing technical implementations usually require manually defining the mapping relationship between the collected data and the input parameters of the simulation software, and this manually defined method is prone to configuration errors in complex data scenarios, resulting in inaccurate data mapping.

[0003] In specific application scenarios, such as the status monitoring and simulation of wind power generation equipment, the acquisition system needs to map multi-dimensional data, such as wind speed, blade angle, and power generation, to the operation model of the simulation software. However, due to the lack of dynamic adaptation management capabilities for data acquisition and simulation software model parameters in the prior art, the mapping relationship may need to be frequently updated under different device states, and most existing methods rely on static mapping configurations. This static configuration may lead to the omission or incorrect mapping of key parameters when the device state changes, thereby affecting the accurate evaluation of the performance of the wind power system by the simulation software. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for managing the functional mapping between a data acquisition system and simulation software, aiming to solve the problems mentioned in the background art.

[0005] To solve the above technical problems, the technical solution of the present invention is as follows:

[0006] A method for managing the functional mapping between a data acquisition system and simulation software, the method comprising:

[0007] S100. Collect an original data set from a target device, where the original data set includes at least one data dimension;

[0008] S200. Perform noise filtering and outlier detection on the original data set to obtain a preprocessed data set;

[0009] S300. Extract multiple feature data corresponding to the simulation software model from the preprocessed data set to generate a comprehensive feature data set, where the feature data includes independent data sets for different physical attributes;

[0010] S400. Generate an initial mapping parameter set according to the mapping rules between the feature data and the input parameters of the simulation software model;

[0011] S500. Dynamically adjust the initial mapping parameter set according to the real-time operating state of the target device. The dynamic adjustment includes detecting the change amount of the operating state of the target device, selecting a corresponding preset adjustment strategy according to the change amount, and generating an adapted mapping parameter set;

[0012] S600. Convert the feature data into input format data required by the simulation software model according to the adapted mapping parameter set. The conversion includes format normalization processing for different types of simulation software models, and output a simulation input data set;

[0013] S700. Perform consistency verification on the adapted mapping parameter set and the simulation input data set, where:

[0014] If the verification passes, store the adapted mapping parameter set as the mapping model configuration; if the verification fails, adjust the adapted mapping parameter set according to the predefined correction rules and regenerate the simulation input data set;

[0015] S800. Input the simulation input data set into the simulation software to obtain simulation result data;

[0016] S900. Verify the simulation result data. If the verification fails, generate feedback adjustment parameters according to the difference degree between the simulation result data and the feature data, where:

[0017] Compare the key index differences between the simulation result data and the operating data of the target device to generate a set of difference factors;

[0018] Calculate the feedback adjustment weight according to the set of difference factors and update the adapted mapping parameter set;

[0019] Input the updated adapted mapping parameter set into step S500 for dynamic adjustment; if the verification passes, use the simulation result data as the output data for the operating evaluation of the target device.

[0020] Preferably, the extracting multiple feature data corresponding to the simulation software model according to the preprocessed data set includes:

[0021] According to the preprocessed data set, adopt a dimension-by-dimension analysis method to extract the trend features and periodic features in the time series to obtain a trend data set and a periodic data set;

[0022] Based on the trend data set, use the segmented statistical method to generate a key statistic data set for multiple intervals;

[0023] Perform spectral analysis on the periodic data set, extract frequency feature data and classify it into a frequency domain feature set;

[0024] Perform feature fusion on the trend data, statistic data and frequency feature data to generate a comprehensive feature data set.

[0025] Preferably, the step of generating a key statistic data set for multiple intervals based on the trend data set by using the segmented statistical method includes:

[0026] Divide the trend data set into multiple time intervals along the time axis, and the length of each time interval is dynamically set according to the operating cycle of the target device;

[0027] Perform statistical analysis on the trend data within each time interval, and calculate the average value, median, variance and maximum value of the interval;

[0028] According to the predefined key index screening rules, select the statistics with significant fluctuations or change trends as key statistics;

[0029] Normalize the key statistics within all time intervals to generate a key statistic data set for multiple intervals.

[0030] Preferably, the step of performing spectral analysis on the periodic data set, extracting frequency feature data and classifying it into a frequency domain feature set includes:

[0031] Perform a fast Fourier transform on the periodic data set to convert the time domain data into frequency domain data;

[0032] Extract the main frequency components in the spectrum, and calculate the corresponding amplitude, phase and power density;

[0033] According to the amplitude size of the frequency components, select the frequency components exceeding the preset amplitude threshold as the main frequency features;

[0034] Classify the amplitude, phase and power density of the main frequency components and generate a frequency domain feature set.

[0035] Preferably, the step of generating an initial mapping parameter set according to the mapping rules between the feature data and the input parameters of the simulation software model includes:

[0036] Group and correspond the comprehensive feature data set with the input parameter set of the simulation software model to establish a feature grouping mapping rule;

[0037] According to the feature grouping mapping rule, the correlation of feature data is weighted and calculated to generate an initial mapping parameter matrix; the generation formula of the initial mapping parameter matrix is:

[0038] , where is the parameter value of the th row and th column in the initial mapping parameter matrix, is the correlation coefficient between the th feature data and the th input parameter, and are adjustment coefficients, is the th eigenvalue in the feature dataset, is the nominal value of the th parameter in the set of input parameters of the simulation software model, is the total number of features in the feature dataset, is the base of the natural logarithm; where

[0039] , is the value of the th feature data in the th dimension, is the value of the th input parameter in the th dimension, is the mean value of the feature data, is the mean value of the input parameters, is the weight coefficient of the th dimension, is the total number of data dimensions;

[0040] According to the initial mapping parameter matrix, each parameter value is extracted and subjected to range normalization and structuring processing to generate an initial mapping parameter set.

[0041] Preferably, the initial mapping parameter set is dynamically adjusted according to the real-time operating state of the target device. The dynamic adjustment includes detecting the change amount of the operating state of the target device, selecting a corresponding preset adjustment strategy according to the change amount, and generating an adapted mapping parameter set, including:

[0042] According to the real-time operating state of the target device, the change amount of the operating state is extracted to form a state change index dataset;

[0043] According to the state change index dataset, the weight factor in the initial mapping parameter matrix is recalculated to generate an adjusted mapping parameter matrix; the generation formula of the adjusted mapping parameter matrix is:

[0044] , where is the parameter value of the th row and the th column in the adjusted mapping parameter matrix, is the parameter value of the th row and the th column in the mapping parameter matrix before adjustment, is the adjustment coefficient, is the change amount of the real-time operating state of the th characteristic data, , is the actual operating state of the th characteristic data, is the expected operating state of the th characteristic data, is the total number of state change amounts, is the change amount of the real-time operating state of the th characteristic data, is the adjustment coefficient;

[0045] Adjust the initial mapping parameter set according to the adjusted mapping parameter matrix to generate an adapted mapping parameter set.

[0046] Preferably, according to the adapted mapping parameter set, convert the characteristic data into input format data required by the simulation software model, where the conversion includes format normalization processing for different types of simulation software models, and output a simulation input data set, including:

[0047] Group and decompose the comprehensive characteristic data set according to the adapted mapping parameter set to form a formatted characteristic data set;

[0048] Perform unit normalization processing on each group of data in the formatted characteristic data set to generate a normalized characteristic data set;

[0049] Adjust the arrangement order in the normalized characteristic data set and encapsulate it into a simulation input data packet according to the input requirements of different simulation software models;

[0050] Output the simulation input data packet as the simulation input data set.

[0051] Preferably, the consistency verification of the adapted mapping parameter set and the simulation input data set includes:

[0052] Verify whether each parameter in the simulation input data packet meets the input requirements of the simulation software model;

[0053] Compare the key parameters in the adapted mapping parameter set with the constraint conditions of the simulation software model to judge their consistency;

[0054] If a parameter conflict is detected, a corrected data set is generated according to the type of the conflicting parameter, where:

[0055] For a parameter that exceeds the input range of the simulation software model, adjust its normalized value to the input range of the simulation software model;

[0056] For a missing parameter, find the closest alternative value from the adapted mapping parameter set to replace it;

[0057] For duplicate or redundant parameters, delete the conflicting items and maintain the uniqueness of the key parameters;

[0058] Update the adapted mapping parameter set according to the corrected data set and regenerate the simulation input data set.

[0059] Preferably, comparing the key index differences between the simulation result data and the target device operation data to generate a difference factor set, including:

[0060] Extract the key index values of the target device operation data, where the key indexes include the fluctuation range, statistical characteristic, and frequency characteristic of the time series parameter;

[0061] Extract the key index values of the simulation result data, and the key indexes correspond one by one to the key indexes in the target device operation data;

[0062] Compare the differences of each key index between the simulation result data and the target device operation data, including calculating the mean deviation, extreme value deviation, and deviation degree of the change trend of the corresponding indexes;

[0063] Based on the differences of the key indexes, generate a difference factor set according to the preset weight rule, and the preset weight rule is set according to the influence degree of the key indexes on the simulation result, where the influence degree of the key indexes is determined by the sensitivity analysis in the simulation software model.

[0064] Preferably, the calculation formula for the feedback adjustment weight is:

[0065] , where, is the adjusted feedback weight, expressed as the th parameter value in the th simulation result data, is the feedback weight before adjustment, expressed as the th parameter value in the th simulation result data, inherited from the previous round of adapted mapping parameter set, is the adjustment coefficient, is the th simulation result data and the The key index difference of the parameter values of the , is the th parameter value in the th simulation result data, the th parameter value in the th actual operation data, is the th weight coefficient of the th parameter value among the differential factors, is the number of differential factors participating in the adjustment calculation, is the key index difference between the th simulation result data and the th actual operation data in the th parameter value,

[0066] The above - mentioned solution of the present invention has at least the following beneficial effects:

[0067] By collecting the original data set from the target device, the present invention can obtain multi - dimensional data during the operation of the device in real time. These data cover information such as the operation state of the device, environmental parameters, and external interferences, providing comprehensive basic data support for subsequent simulations. Compared with the traditional method that relies on manual collection and collation, this method automatically collects data through sensors and transmits it in real time, not only significantly improving the data acquisition efficiency but also covering more key parameters in multiple dimensions, such as wind speed, blade angle, and power generation in wind power generation. For the noise and outliers in the original data, this method can effectively identify and eliminate invalid data through noise filtering and outlier detection techniques, thus ensuring the high quality and reliability of the data input.

[0068] Furthermore, by extracting multiple feature data related to the simulation software model from the pre - processed data set to generate a comprehensive feature data set, the present invention completely solves the problem in the prior art that complex multi - dimensional data cannot be effectively processed. For example, in the wind power generation simulation scenario, it is possible to extract the trend features of wind speed, the dynamic change features of blade angle, and the periodic characteristics of power generation by grouping to form a comprehensive feature data set. This way of feature extraction and integration ensures the comprehensiveness and accuracy of the input of the simulation software model, avoiding the problem of insufficient simulation accuracy caused by single - dimensional input data.

[0069] The core innovation of this method lies in the introduction of a dynamic adjustment and feedback mechanism. By dynamically adjusting the initial mapping parameter set according to the real-time operating state of the target device, this method can quickly respond to changes in the device state, solving the problem in the prior art that static mapping relationships are difficult to adapt to complex scenarios. For example, in a wind farm, when the wind speed suddenly changes or the blade angle is adjusted, the change amount of the operating state can be detected in real time, and the mapping parameter set can be recalculated through a preset adjustment strategy, enabling the input parameters to adapt to the requirements of the simulation software model in real time. This dynamic adjustment ability greatly improves the flexibility and adaptability of the simulation, avoiding the problems of missing key parameters or incorrect mapping.

[0070] In addition, through a consistency verification and feedback adjustment mechanism, this method establishes a closed-loop optimization process from the simulation results to the input parameters. In the step of verifying the simulation result data, the simulation result data can be compared and analyzed with the actual operating data of the target device, generating a feedback adjustment weight based on the set of difference factors, and updating the adapted mapping parameter set. This closed-loop feedback mechanism can continuously optimize the simulation input parameters, making the simulation results gradually approach the real operating state of the device. For example, in the performance evaluation of a wind power generation system, by comparing the generated power output from the simulation with the actual generated power data, a deviation correction factor is generated to adjust the mapping parameters, thereby improving the accuracy and reliability of the next round of simulation.

[0071] Compared with the prior art, the mapping management method of the present invention has significant advantages in multiple aspects. First, this method eliminates the error sources relying on manual configuration in the traditional method through automatic acquisition and dynamic adjustment, improving the accuracy and efficiency of data mapping. Second, this method constructs a full-process dynamic management system between data input and the simulation software model through feature extraction, dynamic adjustment, and feedback optimization, capable of adapting to changes in the operating states of different devices, significantly improving the flexibility and stability of the simulation software model. Finally, the closed-loop optimization process of this method enables the accuracy of the simulation results to gradually increase with the operation cycle, being particularly suitable for high-precision scenarios such as device performance evaluation and fault prediction.

[0072] Generally speaking, the present invention integrates data acquisition, dynamic adjustment, and feedback optimization technologies, solving the problems of inaccurate data mapping and insufficient adaptability caused by static configuration and manual intervention in the prior art, providing more intelligent, efficient, and reliable technical support for the simulation of complex systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 is a flowchart of a functional mapping management method for a data acquisition system and a simulation software provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0074] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be fully conveyed to those skilled in the art.

[0075] As Figure 1 shown, an embodiment of the present invention provides a method for managing the functional mapping between a data acquisition system and simulation software, the method comprising:

[0076] S100. Collect an original data set from a target device, where the original data set includes at least one data dimension;

[0077] S200. Perform noise filtering and outlier detection on the original data set to obtain a preprocessed data set;

[0078] S300. Extract multiple feature data corresponding to the simulation software model from the preprocessed data set to generate a comprehensive feature data set, where the feature data includes independent data sets for different physical attributes;

[0079] S400. Generate an initial mapping parameter set according to the mapping rules between the feature data and the input parameters of the simulation software model;

[0080] S500. Dynamically adjust the initial mapping parameter set according to the real-time operating state of the target device, where the dynamic adjustment includes detecting the change amount of the operating state of the target device, selecting a corresponding preset adjustment strategy according to the change amount, and generating an adapted mapping parameter set;

[0081] S600. Convert the feature data into the input format data required by the simulation software model according to the adapted mapping parameter set, where the conversion includes format normalization processing for different types of simulation software models, and output a simulation input data set;

[0082] S700. Perform consistency verification on the adapted mapping parameter set and the simulation input data set, where:

[0083] If the verification is passed, store the adapted mapping parameter set as a mapping model configuration; if the verification fails, adjust the adapted mapping parameter set according to a predefined correction rule and regenerate the simulation input data set;

[0084] S800. Input the simulation input data set into the simulation software to obtain simulation result data;

[0085] S900. Verify the simulation result data. If the verification fails, generate feedback adjustment parameters according to the difference degree between the simulation result data and the feature data, where:

[0086] Compare the key index differences between the simulation result data and the target device operation data to generate a set of difference factors;

[0087] Calculate the feedback adjustment weights according to the set of difference factors and update the set of adaptation mapping parameters;

[0088] Input the updated set of adaptation mapping parameters into step S500 for dynamic adjustment; if the verification passes, use the simulation result data as the output data for the target device operation evaluation.

[0089] In the embodiment of the present invention, by collecting the original data set from the target device and performing multi-step processing on it, the full-process optimization from data acquisition to simulation result evaluation can be achieved. First, data acquisition can capture multi-dimensional dynamic information during the operation of the target device, including time series characteristics, environmental parameters, and internal operation indicators of the device, etc. This multi-dimensional data acquisition method effectively adapts to the requirements of different devices and scenarios. For industrial scenarios, the collected data may include temperature, pressure, vibration amplitude, etc.; while in intelligent transportation applications, data acquisition may cover vehicle speed, position, and environmental conditions, etc.

[0090] After data acquisition, preprocess the original data through noise filtering and outlier detection steps. This process can significantly improve the quality and credibility of the data. By filtering out invalid data and outliers, the problem of inaccurate simulation results caused by data deviation can be avoided. For example, in equipment fault monitoring, sensors may record abnormally high values due to external interference. Through detection techniques, such data can be identified and removed, thus generating a reliable preprocessed data set.

[0091] Subsequently, extract various feature data related to the simulation software model, including trend characteristics, frequency characteristics, and other key characteristics. These feature data not only cover the static operation state of the target device but also reflect its dynamic change behavior. After extraction, the generation of the comprehensive feature data set significantly improves the accuracy of the simulation input, ensuring that the model input is more diverse and comprehensive. For example, in wind power generation simulation, the comprehensive feature data can accurately characterize the relationship between wind speed, power output, and blade vibration, thereby improving the prediction ability of the simulation software model.

[0092] An initial mapping parameter set is generated through the mapping relationship between the feature data and the input parameters of the simulation software. This process fully combines the complex correlation between the data characteristics and the requirements of the simulation software model. By combining the detection and dynamic adjustment of the real-time operating state change amount, the adaptive optimization of the initial mapping parameter set can be achieved. This dynamic adjustment is particularly important when the operating state of the device changes significantly. For example, during the process of the manufacturing device switching from the high-speed mode to the low-speed mode, the adjusted mapping parameters can timely reflect the change of the operating characteristics, ensuring the adaptability of the simulation software model.

[0093] In addition, by converting the mapping parameters into the input data of the simulation software and performing consistency verification, it can be ensured that the finally generated simulation input data packet meets the model requirements. In the case where the verification fails, new parameter inputs are generated through predefined correction rules, thus forming a closed-loop optimization process. This process greatly improves the adaptability of the data and the accuracy of the model calculation, avoiding simulation failures caused by inconsistent input data and model requirements.

[0094] Finally, by inputting the simulation input data into the simulation software model, the simulation result data is obtained, and through the comparison of the key indicators with the actual operating data, the mapping parameters can be further optimized. This dynamic feedback mechanism based on the adjustment of the difference factor provides strong support for the long-term optimization of the simulation software model. Through the complete process design and multi-level data processing and optimization techniques, the whole method not only improves the adaptability of the simulation, but also significantly enhances the accuracy and reliability of the simulation results.

[0095] In a preferred embodiment of the present invention, the extraction of multiple feature data corresponding to the simulation software model according to the preprocessed data set includes:

[0096] According to the preprocessed data set, the trend features and periodic features in the time series are extracted by using the dimension-by-dimension analysis method to obtain a trend data set and a periodic data set;

[0097] According to the trend data set, a multi-interval key statistic data set is generated by using the piecewise statistical method;

[0098] The periodic data set is subjected to spectral analysis, and the frequency feature data is extracted and classified into a frequency domain feature set;

[0099] Feature fusion is performed on the trend data, statistic data, and frequency feature data to generate a comprehensive feature data set.

[0100] In the embodiments of the present invention, by establishing a connection between the preprocessed data set and the feature data corresponding to the simulation software model, the logical classification and structural optimization of data can be efficiently achieved. The dimension-by-dimension analysis method can gradually extract key characteristics for different dimensions of data. For example, trend feature extraction enables capturing the long-term change patterns in the device operation data, while cycle feature extraction can highlight the repetitive patterns or fixed fluctuations in the data. This segmented analysis method ensures the integrity and accuracy of feature extraction.

[0101] Furthermore, by generating a multi-interval key statistic data set through the segmented statistical method, in-depth analysis of the trend features can be carried out. For example, key statistics such as mean, variance, and maximum value can be extracted for different intervals of time series data. These statistics can effectively summarize the operation patterns of each time period, thereby providing an accurate description for the input of the simulation software model. In addition, performing spectral analysis on the cycle data set can convert time-domain information into frequency-domain information and extract key spectral features such as frequency and amplitude. This frequency-domain information is particularly suitable for vibration analysis or periodic signal modeling, providing high-quality input data for the simulation software.

[0102] Integrating the above technologies, by integrating trend features, statistic features, and frequency features into a comprehensive feature data set through the feature fusion step, the adaptability of the simulation software model to complex systems can be significantly improved. For example, in the operation simulation of complex mechanical equipment, these comprehensive features can significantly enhance the robustness and reliability of the model, thereby achieving higher-precision prediction and optimization.

[0103] More specifically, the method of extracting trend features and cycle features in the time series according to the preprocessed data set by using the dimension-by-dimension analysis method to obtain a trend data set and a cycle data set includes:

[0104] The trend feature refers to the direction or pattern of long-term changes in the data. In this method, each dimension of data can be processed separately through the dimension-by-dimension analysis method. For example, for the temperature sequence data of a certain device, it can be observed that the temperature gradually rises over time. In specific implementation, the short-term fluctuations can be eliminated by using common methods in the prior art (such as the moving average method or the cumulative sum analysis method) to extract the overall change trend of the data.

[0105] For complex scenarios, the change points of the trend can also be identified by setting thresholds. For example, the time point when the temperature changes from rising to falling can be marked by threshold difference technology, and the segmented characteristics of different trend stages can be further extracted.

[0106] Periodic features refer to repetitive or periodic patterns existing in data. For example, in vibration monitoring, periodic fluctuations can indicate that the device is in normal operating condition. The extraction of periodic features can be achieved through frequency domain conversion techniques, such as the Fast Fourier Transform, which transforms time domain data into frequency domain data to identify the main frequency and harmonic frequencies of the signal.

[0107] In addition, repetitive changes can also be directly detected in the time domain through the sliding window method. For example, by applying window calculations to the change sequence of the vibration amplitude of a certain device, it is possible to discover the change in amplitude peaks within a fixed time interval, which can be used as a direct manifestation of periodic features.

[0108] By processing the data of each dimension in a dimension-by-dimension analysis manner, the respective trend features and periodic features can be extracted separately. The trend data set and the periodic data set respectively record the long-term change patterns and periodic characteristics of the target device in different dimensions, providing independent and detailed data support for subsequent feature fusion.

[0109] More specifically, the feature fusion of the trend data, statistical data, and frequency feature data to generate a comprehensive feature data set includes:

[0110] Trend data describes the overall change pattern of a time series, while statistical data provides the results of quantitative analysis, such as mean, variance, maximum value, minimum value, etc. The fusion of trend data and statistical data can be achieved by segmenting the trend data and calculating the statistical features of each time period. For example, after dividing the rising trend of temperature into multiple time periods, the average temperature, fluctuation range, etc. of each segment can be calculated, and these statistics can more accurately describe the quantitative characteristics of the trend.

[0111] Frequency feature data provides periodic information about the dynamic behavior of the device, such as the main frequency and power density, etc. These frequency features can be used as separate feature dimensions and combined with trend and statistical data. During fusion, the corresponding frequency features can be matched for each time period. For example, the fluctuations in temperature data can be combined with the main frequency components of the device operation to further analyze the relationship between temperature fluctuations and the operating state.

[0112] In specific implementation, the above features can be fused in the form of a table structure or a multi-dimensional matrix. For example, the data of each time period contains multiple fields of trend data, statistical data, and frequency feature data, and finally forms a comprehensive feature data set. For time periods of different durations, feature fusion can perform alignment processing on the data through interpolation methods to ensure the consistency between different features.

[0113] The generation of the comprehensive feature data set organically combines the static state and dynamic behavior of the device, providing a multi-level and comprehensive feature description for subsequent model input.

[0114] In a preferred embodiment of the present invention, based on the trend data set, a key statistic data set for multiple intervals is generated by using a piecewise statistical method, including:

[0115] The trend data set is divided into multiple time intervals along the time axis, and the length of each time interval is dynamically set according to the operating cycle of the target device;

[0116] Statistical analysis is performed on the trend data within each time interval, and the average value, median, variance, and maximum value of the interval are calculated;

[0117] According to the predefined key index screening rules, the statistics with significant fluctuations or change trends are selected as key statistics;

[0118] Normalization processing is performed on the key statistics within all time intervals to generate a key statistic data set for multiple intervals.

[0119] In the embodiment of the present invention, generating a key statistic data set for multiple intervals by using a piecewise statistical method can make feature extraction more refined. Dividing the trend data set into multiple time intervals along the time axis and dynamically adjusting the interval length according to the operating cycle of the target device can adapt to the characteristics of different devices and working conditions. For example, in short-cycle devices, the intervals can be divided more densely, while for long-term operating devices, the intervals can be set more loosely, so as to achieve adaptive processing of data characteristics.

[0120] Furthermore, by performing statistical analysis on the trend data within each time interval, key statistics reflecting the device state can be extracted, such as the mean value, median, variance, and maximum value, etc. These statistics can effectively describe the fluctuation range and change trend of the data. For example, in the monitoring of wind power generation equipment, the wind speed changes and power output fluctuations in different time periods can be effectively characterized by these statistics.

[0121] By predefined key index screening rules, the statistics with significant change trends or fluctuations can be extracted, so as to filter out irrelevant or noisy data. Finally, normalization processing of all key statistics can make them meet the input requirements of the simulation software model and avoid model calculation errors caused by different data dimensions. This process greatly improves the quality and consistency of the simulation input data, thus providing a reliable basis for the subsequent generation of mapping parameters.

[0122] More specifically, the step of dividing the trend data set into multiple time intervals along the time axis, where the length of each time interval is dynamically set according to the operating cycle of the target device, includes:

[0123] The operating cycle of a device is usually determined by the working characteristics of the device. For example, for industrial devices, the operating cycle may include stages such as startup, operation, and shutdown. For transportation devices (such as vehicles), the operating cycle may be related to driving speed or road condition changes. Therefore, the length of the time interval can be dynamically adjusted according to the operating state of the device. For example, during the stable stage of device operation, the time interval can be divided more loosely; while during the stage with drastic state changes, the time interval can be divided more densely.

[0124] In practical applications, dynamically adjusting the length of the time interval can be achieved by real-time monitoring of the device's state changes. For example, by monitoring the vibration frequency of the device, when the frequency remains constant, a longer time interval can be adopted; while when the frequency changes significantly, the time interval can be shortened to capture more details. This dynamic adjustment strategy can be combined with the sliding window method to adjust the interval length in real time during different time periods.

[0125] For example, in the simulation of a wind turbine generator, when the wind speed changes slightly, the time interval can be set to 1 minute; while when the wind speed changes drastically, it can be adjusted to 30 seconds, so as to more accurately reflect the operating characteristics of the device.

[0126] By dynamically dividing the time interval, the accuracy and adaptability of feature extraction can be significantly improved, ensuring that the data in each time period can accurately reflect the actual operating state of the device.

[0127] More specifically, according to the predefined key index screening rules, statistics with significant fluctuations or change trends are selected as key statistics, including:

[0128] The key index screening rules can be set according to the operating characteristics and simulation requirements of the target device. For example, for wind power simulation, the variances of wind speed and power generation may be important key indexes because they can reflect the volatility of the system; while for temperature monitoring of industrial devices, the maximum and minimum values may be more important because they can reflect the characteristics of extreme states.

[0129] First, calculate the statistics within each time interval, such as mean, variance, maximum value, minimum value, etc. Subsequently, compare these statistics with the predefined screening rules to screen out the qualified statistics. For example, set a fluctuation threshold. If the variance of a certain statistic exceeds this threshold, it is marked as a key statistic.

[0130] The screened key statistics will be used as the main data for the next step of processing, and the remaining non-key statistics can be ignored. This processing method can significantly reduce the data volume and at the same time increase the influence of key features on the simulation software model.

[0131] By screening key statistics, it is ensured that the input data can efficiently reflect the operating characteristics of the target device, avoiding simulation deviations caused by irrelevant or noisy data.

[0132] In a preferred embodiment of the present invention, the spectral analysis of the periodic data set, extracting frequency feature data and classifying it into a frequency domain feature set, includes:

[0133] Performing a fast Fourier transform on the periodic data set to convert the time-domain data into frequency-domain data;

[0134] Extracting the main frequency components in the spectrum and calculating the corresponding amplitude, phase, and power density;

[0135] According to the magnitude of the frequency components, selecting the frequency components exceeding a preset amplitude threshold as the main frequency features;

[0136] Classifying the amplitude, phase, and power density of the main frequency components and generating a frequency domain feature set.

[0137] In the embodiment of the present invention, performing spectral analysis on the periodic data set can effectively capture the periodic information in the data. By performing a fast Fourier transform to convert the time-domain data into frequency-domain data, the hidden frequency components, amplitude distribution, and phase characteristics in the data can be intuitively presented. For example, in the monitoring of rotating machinery, spectral analysis can reveal the main frequency and harmonic components of the device, providing a basis for fault diagnosis and performance optimization.

[0138] Extracting the main frequency components in the spectrum and calculating the corresponding amplitude, phase, and power density can deeply analyze the energy distribution and characteristic positions of the signal. Further, by setting a preset amplitude threshold, noise signals and unimportant frequency components can be filtered out to ensure that the extracted frequency features can accurately reflect the periodic behavior of the device.

[0139] By classifying the amplitude, phase, and power density of the main frequency components, a frequency domain feature set with clear physical meanings can be formed. These frequency domain features can not only be used as the input of the simulation software model but also be directly applied to the monitoring and evaluation of the device operating state. For example, in vibration analysis, frequency domain features can be used to determine whether there are abnormal operating states of the device, providing decision support for device maintenance and optimization.

[0140] In a preferred embodiment of the present invention, the generation of the initial mapping parameter set according to the mapping rule between the feature data and the input parameters of the simulation software model includes:

[0141] Grouping and corresponding the comprehensive feature data set with the input parameter set of the simulation software model to establish a feature grouping mapping rule;

[0142] According to the characteristic grouping mapping rule, calculate the correlation of the characteristic data with weights to generate an initial mapping parameter matrix; the generation formula of the initial mapping parameter matrix is:

[0143] , where is the parameter value of the th row and the th column in the initial mapping parameter matrix, is the correlation coefficient between the th characteristic data and the th input parameter, and are adjustment coefficients, is the th eigenvalue in the characteristic dataset, is the nominal value of the th parameter in the set of input parameters of the simulation software model, is the total number of characteristics in the characteristic dataset, is the base of the natural logarithm; where

[0144] , is the value of the th characteristic data in the th dimension, is the value of the th input parameter in the th dimension, is the mean value of the characteristic data, is the mean value of the input parameters, is the weight coefficient of the th dimension, is the total number of data dimensions;

[0145] Extract each parameter value according to the initial mapping parameter matrix, and perform range normalization and structuring processing on it to generate an initial mapping parameter set.

[0146] In the embodiment of the present invention, generating an initial mapping parameter set through the initial mapping parameter matrix can achieve efficient docking of characteristic data and input parameters of the simulation software model. This process is based on the characteristic grouping mapping rule, grouping and corresponding the comprehensive characteristic dataset with the model input parameters, making the design of the mapping rule more in line with the requirements of the actual application scenario. For example, for industrial equipment simulation, the mapping rule can be grouped according to physical characteristics such as temperature, pressure, and vibration, thereby improving the correlation between input parameters and characteristic data.

[0147] During the generation of the mapping parameter matrix, the correlation between the feature data and the input parameters is calculated through weighting to ensure that the generated parameter matrix can comprehensively reflect the data characteristics and model requirements. This process uses a mathematical model to accurately describe the non-linear relationship between the feature data and the input parameters, and quantifies these complex relationships through the weighting factors and adjustment coefficients of the mapping matrix. Through this mapping matrix, a large amount of scattered feature data can be integrated into a set of mapping parameters with clear logic and strong adaptability.

[0148] In addition, by performing range normalization on the parameter values of the initial mapping parameter matrix, it is possible to further ensure that all parameter values meet the unified input requirements. Range normalization not only improves the consistency of the mapping parameters but also enhances the stability of the input data in model calculations. Especially in the application of multi-dimensional data, normalization can avoid calculation errors caused by inconsistent dimensions or ranges. For example, when dealing with a data set involving different units (such as temperature and pressure), through normalization, the numerical values of the parameter matrix can be uniformly mapped to a specified range, thus ensuring the reliability of model calculations.

[0149] Finally, through structured processing, the generated initial set of mapping parameters can meet the input requirements of different types of simulation software models. The flexibility of this set makes it suitable for a variety of application scenarios. For example, in traffic simulation, this set may include vehicle dynamic characteristics; in industrial equipment monitoring, this set may include dynamic data of key performance indicators. The parameter set generated through this process provides a solid foundation for subsequent dynamic adjustment and simulation optimization, significantly enhancing the adaptability of the system and the calculation efficiency of the simulation software model.

[0150] More specifically, the grouping and corresponding of the comprehensive feature data set and the input parameter set of the simulation software model, and the establishment of the feature grouping mapping rule, include:

[0151] There is usually an inherent logical connection between each type of feature data in the comprehensive feature data set and the input parameter set of the simulation software. For example, in wind power generation simulation, wind speed and blade angle data correspond to wind load parameters, while power generation data corresponds to power generation efficiency parameters. Through the analysis of the feature data and input parameters, they can be grouped according to physical meaning or application requirements.

[0152] First, each feature in the comprehensive feature data set is labeled, including information such as the data source, physical meaning, and dimension. Subsequently, according to the classification of the input parameters of the simulation software, the feature data and the input parameters are grouped and corresponding. For example, in wind power generation simulation, features such as wind speed and wind direction can be grouped into one group corresponding to the wind load input parameters, while features such as temperature and humidity can be grouped into another group corresponding to the environmental parameter input.

[0153] After grouping is completed, mapping rules need to be defined for each group. The mapping rules include the following:

[0154] The mapping relationship between feature data and input parameters, such as direct mapping or mapping through weight combination.

[0155] The processing steps required during the mapping process, such as normalization, denoising, or dimension conversion.

[0156] The priority of mapping, such as which feature data has a higher weight in many-to-one mapping.

[0157] By establishing groups and mapping rules, the automated docking of the comprehensive feature dataset and simulation input parameters can be effectively achieved, avoiding logical errors in manual configuration.

[0158] More specifically, according to the initial mapping parameter matrix, each parameter value is extracted, and range normalization and structuring processing are performed on it to generate an initial mapping parameter set, including:

[0159] Each element of the initial mapping parameter matrix corresponds to a mapping parameter value, and these values are usually calculated from the relationship between feature data and input parameters. In actual operations, each value in the matrix can be directly read and stored as a list or array for subsequent processing.

[0160] The extracted parameter values may be distributed in different numerical ranges, and simulation software models usually have specific requirements for the range of input parameters. To ensure that these parameter values are within a unified range, normalization processing is required. For example, the parameter values can be adjusted to the range of 0 to 1 through linear mapping, or adjusted to other ranges (such as -1 to 1) according to the actual needs of the model. This processing step can effectively avoid simulation calculation errors caused by too large or too small parameter value ranges.

[0161] After completing the range normalization, these parameter values need to be reorganized into a structured format that meets the requirements of the simulation software model. For example, the parameter values can be rearranged according to the row and column structure of the original matrix, or organized into key-value pairs according to the grouping correspondence rules. If the simulation software requires an input multi-level data structure, the parameter values can be further divided into different levels.

[0162] The generated initial mapping parameter set ultimately serves as the initial input for the simulation software model, providing a reliable data basis for model calculation.

[0163] In a preferred embodiment of the present invention, the initial mapping parameter set is dynamically adjusted according to the real-time operating state of the target device. The dynamic adjustment includes detecting the change amount of the operating state of the target device, selecting a corresponding preset adjustment strategy according to the change amount, and generating an adapted mapping parameter set, including:

[0164] Extract the change amount of the operating state according to the real-time operating state of the target device to form a state change index data set;

[0165] According to the state change index data set, recalculate the weight factors in the initial mapping parameter matrix to generate an adjusted mapping parameter matrix; the generation formula of the adjusted mapping parameter matrix is:

[0166] , where is the parameter value of the th row and the th column in the adjusted mapping parameter matrix, is the parameter value of the th row and the th column in the mapping parameter matrix before adjustment, is the adjustment coefficient, is the real-time operating state change amount of the th feature data, , is the actual operating state of the th feature data, is the expected operating state of the th feature data, is the total number of state change amounts, is the real-time operating state change amount of the th feature data, is the adjustment coefficient;

[0167] According to the adjusted mapping parameter matrix, adjust the initial mapping parameter set to generate an adapted mapping parameter set.

[0168] In the embodiment of the present invention, dynamically adjusting the initial mapping parameter set according to the real-time operating state of the target device can significantly improve the adaptability and flexibility of the parameter set. In the dynamic adjustment process, extracting the change amount of the real-time operating state of the device and generating a state change index data set can capture the change characteristics of the device operating state. This process is particularly important for devices under complex working conditions. For example, on a production line with high-frequency changes, real-time monitoring of key indicators of the device (such as temperature, pressure, vibration amplitude, etc.) can dynamically reflect the current operating state of the device.

[0169] According to the status change index data set, recalculate the weight factors in the initial mapping parameter matrix and generate an adjusted mapping parameter matrix. Through this process of adjusting the weight factors, the mapping relationship can more accurately reflect the current state characteristics of the device. The recalculation of the weight factors incorporates the normalization of real-time change amounts, ensuring the stability and balance of the adjustment results. For example, for device parameters with large sudden changes, the adjustment amplitude of the weight factors will be greater, thereby enhancing the sensitivity to these parameters.

[0170] By adjusting the mapping parameter matrix, an adaptive mapping parameter set can be dynamically generated to better meet the requirements of the device's real-time operation. This dynamic adjustment method is not only applicable to the real-time optimization of a single device but also can improve the overall operation efficiency in scenarios of multi-device collaborative operation. For example, in a multi-machine joint simulation, the operating state of each device may affect the overall simulation result, and dynamic adjustment can achieve global optimization to ensure a high degree of consistency between the simulation result and the actual operating state.

[0171] More specifically, adjusting the initial mapping parameter set according to the adjusted mapping parameter matrix to generate an adaptive mapping parameter set includes:

[0172] The adjusted mapping parameter matrix is generated based on the real-time operating state of the target device. For example, by monitoring the vibration frequency, temperature, or other operating metrics of the device, the change amount of the device state can be calculated in real time. The adjusted mapping parameter matrix takes these change amounts as inputs and modifies the initial parameters according to certain rules.

[0173] Each element in the adjusted mapping parameter matrix represents the correction value for the initial parameter set. By superimposing the adjustment values onto the initial parameter set, an adaptive mapping parameter set can be obtained. For example, if a certain initial parameter needs to be adjusted according to the temperature change of the device, the adjustment value can be jointly calculated by the temperature change amount and a preset weight, and the adjusted parameter value reflects the optimal configuration in the current state.

[0174] The adjusted parameter set is stored in the form of an adaptive mapping parameter set for the next input to the simulation software model. Compared with the initial mapping parameter set, the adaptive mapping parameter set can better reflect the current operating state of the device, thereby improving the accuracy of the simulation result.

[0175] In a preferred embodiment of the present invention, converting the feature data into the input format data required by the simulation software model according to the adaptive mapping parameter set, where the conversion includes format normalization processing for different types of simulation software models, and outputting a simulation input data set, including:

[0176] Group and decompose the comprehensive feature dataset according to the adaptation mapping parameter set to form a formatted feature dataset;

[0177] Perform unit normalization processing on each group of data in the formatted feature dataset to generate a normalized feature dataset;

[0178] Adjust the arrangement order in the normalized feature dataset and encapsulate it into a simulation input data packet according to the input requirements of different simulation software models;

[0179] Output the simulation input data packet as a simulation input dataset.

[0180] In the embodiment of the present invention, using the adaptation mapping parameter set for the transformation of feature data can meet the normalization requirements of the input format of the simulation software model. In this process, first, the comprehensive feature dataset is grouped and decomposed, which can effectively reduce the complexity of the input data, so that each group of data can correspond to the specific input requirements of the simulation software model. The grouped data not only improves the efficiency of model input but also facilitates subsequent normalization and sorting processing.

[0181] Through unit normalization processing, the numerical range in the feature data can be adjusted to the unified standard required by the simulation software model. For example, for a model with strict input range requirements, the feature data can be adjusted to the range from zero to one to ensure the accuracy of model calculation. In addition, by adjusting the arrangement order of the normalized feature data and encapsulating it into a simulation input data packet, the input data can fully meet the format requirements of the simulation software model, avoiding calculation errors caused by data format problems.

[0182] The technical effect of this feature data transformation process significantly improves the quality and consistency of the simulation input data. For example, in the simulation of complex equipment, different input features may come from multiple data sources. Through this transformation step, seamless integration of data can be achieved to ensure the accuracy and efficiency of model input. At the same time, through data packet encapsulation, the security and stability of the simulation software model input can be effectively improved, especially in remote simulation applications, this encapsulation format is particularly important.

[0183] More specifically, the group and decomposition of the comprehensive feature dataset according to the adaptation mapping parameter set to form a formatted feature dataset includes:

[0184] The comprehensive feature dataset contains multi-dimensional and multi-type data, and the requirements of the simulation software model for input data may be more refined. Therefore, according to the rules in the adaptation mapping parameter set, the comprehensive feature dataset is grouped and decomposed according to different dimensions or categories. For example, temperature, pressure, and vibration features can be divided into independent groups to facilitate the simulation software model to process these data separately.

[0185] For each set of data, format it according to the input format requirements of the simulation software model. For example, if the simulation software model requires a two-dimensional array as input, each set of data needs to be organized in a fixed row-column structure. If the model requires hierarchical input of data, the characteristic data can be stored hierarchically through a nested structure.

[0186] The formatted characteristic data set meets the input requirements of the simulation software model and can be directly used for model calculations. For example, in multi-device collaborative simulation, the formatted characteristic data set can represent the characteristic data of each device respectively, while maintaining the overall logical structure for easy model processing.

[0187] In a preferred embodiment of the present invention, the consistency verification of the adaptation mapping parameter set and the simulation input data set includes:

[0188] Verify whether each parameter in the simulation input data packet meets the input requirements of the simulation software model;

[0189] Compare the key parameters in the adaptation mapping parameter set with the constraint conditions of the simulation software model to judge their consistency;

[0190] If a parameter conflict is detected, generate a correction data set according to the type of the conflicting parameter, where:

[0191] For a parameter that exceeds the input range of the simulation software model, adjust its normalization value to the input range of the simulation software model;

[0192] For a missing parameter, find the closest alternative value from the adaptation mapping parameter set to replace it;

[0193] For duplicate or redundant parameters, delete the conflicting items and maintain the uniqueness of the key parameters;

[0194] Update the adaptation mapping parameter set according to the correction data set and regenerate the simulation input data set.

[0195] In the embodiment of the present invention, the consistency verification of the adaptation mapping parameter set and the simulation input data set can ensure the accuracy and rationality of the input of the simulation software model. During the consistency verification process, first check whether each parameter in the simulation input data packet meets the input requirements of the simulation software model. For example, for a parameter that exceeds the input range, its normalization value can be adjusted to meet the model requirements again, and this operation can significantly reduce the impact of the input data on the simulation software model.

[0196] By comparing the key parameters in the set of adaptation mapping parameters with the constraints of the simulation software model, it is possible to further determine whether the parameters are consistent. During this process, if parameter conflicts are detected, such as parameter duplication, missing, or out-of-range, a correction data set can be generated according to the specific type of conflict. The generation of the correction data set can ensure the integrity and validity of all input parameters.

[0197] For example, in high-precision industrial simulations, different models may have extremely high requirements for the integrity of input parameters. Any inconsistency in a single parameter may lead to the invalidation of the simulation results. Through this verification process, problems with the input data can be promptly discovered and corrected, thereby improving the computational efficiency of the model and the reliability of the output results. Ultimately, by regenerating the simulation input data set with the updated set of adaptation mapping parameters, it is possible to ensure the final consistency of the inputs to the simulation software model, providing a reliable guarantee for subsequent simulation calculations.

[0198] More specifically, verifying whether each parameter in the verified simulation input data packet meets the input requirements of the simulation software model includes:

[0199] Background and Significance of Parameter Verification

[0200] A simulation input data packet usually consists of multiple parameters, including physical properties (such as temperature, pressure, speed, etc.), environmental variables (such as humidity, light intensity, etc.), and operating status data (such as equipment rotation speed, power output, etc.). The simulation software model may have specific input requirements for these parameters, such as data format, numerical range, or data integrity. If the input data does not meet these requirements, it may lead to simulation failure or distorted results.

[0201] Steps of Parameter Verification

[0202] Data format verification: Check whether each parameter conforms to the input format of the simulation software model. For example, some parameters may need to be input in matrix form, while others may need to be input in time series form. If a format mismatch is found, it can be corrected using a format conversion tool.

[0203] Numerical range verification: Check the value of each parameter to ensure that it falls within the valid range required by the model. For example, a certain physical property may require a value between 0 and 100, but the input value is negative, in which case it needs to be corrected.

[0204] Data integrity verification: Check whether there are missing or redundant parameters in the input data packet. For example, if the simulation software model requires 10 input parameters, but only 8 parameters are provided in the data packet, the missing parameters need to be supplemented according to the actual situation.

[0205] Verification Method

[0206] In actual operation, automated verification tools can be used to check the parameters one by one. For example, write a verification script to compare each parameter in the data packet with the requirements of the simulation software model one by one and output the verification results. If an anomaly is found, the error will be recorded and a correction suggestion will be prompted.

[0207] More specifically, compare the key parameters in the set of comparison adaptation mapping parameters with the constraint conditions of the simulation software model to judge their consistency, including:

[0208] Definition of key parameters and constraint conditions

[0209] Key parameters: refer to the parameters that have a greater impact on the results of the simulation software model. For example, in mechanical equipment simulation, they may include temperature, vibration frequency, power output, etc.

[0210] Constraint conditions: refer to the restrictions on the logical relationship or range of parameters by the simulation software model. For example, a certain model may require that the temperature and pressure satisfy a certain physical law, or the sum of multiple input parameters is equal to a certain constant.

[0211] Specific steps for consistency verification

[0212] Parameter range constraint: Check whether the value of the key parameter conforms to the defined range of the simulation software model. For example, check whether the temperature parameter meets the specified upper and lower limits, or whether the frequency parameter is within the effective frequency band of the model.

[0213] Verification of logical relationship between parameters: Check whether the logical constraints of the simulation software model are satisfied among multiple key parameters. For example, in wind power generation simulation, there may be a certain logical relationship between the power generation power, wind speed, and blade angle. If the wind speed is low but the power output is high, it indicates that the data may be abnormal.

[0214] Dynamic constraint verification: For the parameters generated in real time during operation, verify whether they meet the dynamic constraint conditions of the model. For example, in vehicle simulation, whether the change between speed and acceleration conforms to the actual physical law.

[0215] Implementation method of verification

[0216] Consistency verification usually needs to be combined with the constraint rules of the simulation software model and is implemented through a rule engine or a logical check tool. For example, the parameter set can be automatically verified through a predefined rule library. If an inconsistent situation is found, the system will prompt and mark the abnormal parameters.

[0217] More specifically, updating the set of adaptation mapping parameters according to the correction data set and regenerating the simulation input data set includes:

[0218] A corrected dataset is a set of data that supplements or replaces problematic parameters, mainly used to fix anomalies found during the verification process. For example:

[0219] Supplementation of missing parameters: Generate missing parameters based on historical data or model default values. For example, if a certain operating status parameter is found to be missing, historical values from similar scenarios can be extracted from the database for supplementation.

[0220] Correction of incorrect parameters: Correct parameters that are out of range. For example, if the value of an input parameter is below the lower limit, it can be adjusted to the lower limit value, or dynamically adjusted according to its relationship with other parameters.

[0221] Cleaning of redundant parameters: Remove redundant parameters that do not belong to the requirements of the simulation software model to reduce the computational burden.

[0222] Add the new parameters or corrected values in the corrected dataset to the adapted mapping parameter set to replace the original incorrect values. During the update process, it is necessary to ensure that the corrected parameter set still meets the constraints of the simulation software model. For example, the updated parameter set needs to undergo consistency verification again to ensure its logical integrity.

[0223] According to the updated adapted mapping parameter set, decompose and format the comprehensive feature dataset to generate a new simulation input dataset. The following factors need to be considered during this process:

[0224] Format requirements: Ensure that the regenerated dataset conforms to the input format of the simulation software.

[0225] Dynamic adjustment: Dynamically optimize the input data according to the real-time operating status. For example, enhance the influence of certain key parameters through a weighting method.

[0226] Efficient generation: Rapidly generate a new dataset through automated tools to avoid time delays caused by manual operations.

[0227] The regenerated simulation input dataset needs to undergo parameter verification and consistency checks again to ensure that the corrected data can be correctly recognized and used by the simulation software model. Finally, input the verified dataset into the simulation software model to complete one iteration of optimization.

[0228] In a preferred embodiment of the present invention, comparing the key index differences between the simulation result data and the target device operation data to generate a set of difference factors, including:

[0229] Extract the key index values of the target device operation data, where the key indexes include the fluctuation range of time series parameters, statistical characteristics, and frequency characteristics;

[0230] Extract the key index values of the simulation result data, where the key indexes correspond one by one to the key indexes in the target device operation data;

[0231] Compare the differences between the simulation result data and each key index in the target device operation data, including calculating the mean deviation, extreme value deviation, and deviation degree of the change trend of the corresponding indexes;

[0232] Based on the differences of the key indexes, generate a set of difference factors according to the preset weight rules, where the preset weight rules are set according to the influence degree of the key indexes on the simulation results, and the influence degree of the key indexes is determined by the sensitivity analysis in the simulation software model.

[0233] In the embodiment of the present invention, by comparing the differences of the key indexes between the simulation result data and the target device operation data, a set of difference factors is generated. This method can quantify the deviation between the simulation result and the actual operation state, and provide data support for subsequent feedback adjustment and simulation model optimization.

[0234] First of all, by extracting the key index values of the target device operation data and the simulation result data, this method can conduct a detailed comparison of the most core parameters during the device operation. These key indexes may include the fluctuation range of time series data, statistical characteristics (such as mean, variance, extreme value), and frequency characteristics (such as main frequency, harmonic amplitude, etc.). Through one-by-one corresponding comparison, the differences between the simulation result and the actual operation data can be comprehensively analyzed. For example, in the simulation of wind power generation equipment, key indexes such as blade angle, wind speed, and power generation can be extracted to accurately analyze the deviation degree of the simulation model in these parameters.

[0235] Secondly, through multi-dimensional calculation of the key index differences, this method can comprehensively quantify the matching degree between the simulation result and the target device operation state. For example, through mean deviation calculation, it can be reflected whether the simulation result is consistent with the actual operation in the overall trend; through extreme value deviation analysis, it can be revealed whether the simulation model accurately captures the extreme state of the operation data; through trend deviation analysis, it can be judged whether there is a lag or prediction error in the dynamic change of the simulation result. This multi-dimensional deviation calculation method can cover the possible simulation errors in different scenarios, laying a solid foundation for the generation of difference factors.

[0236] In addition, this method also weights the differences of key indicators through a preset weight rule to generate a set of difference factors. The setting of the weight rule is based on the sensitivity analysis of key indicators to the simulation results. For example, the impact of certain indicators (such as power output) on the simulation results may be much higher than that of other indicators (such as environmental temperature), so higher weights will be assigned when calculating the difference factors. Through this weighting process, this method can highlight the impact of differences in key indicators and ensure that the generated set of difference factors can truly reflect the main optimization directions of the simulation model.

[0237] The generated set of difference factors has important technical effects: on the one hand, it provides accurate input for subsequent feedback adjustments, can guide the optimization of simulation model parameters, and gradually reduce the difference between the simulation results and the actual operation data; on the other hand, the generation process of the set of difference factors is traceable, which can help analyze the sources of simulation errors and provide a basis for further model improvement. For example, in the simulation optimization of a wind farm, if the set of difference factors shows that the deviation of wind speed input has a greater impact on the simulation results, then the mapping parameters related to wind speed can be preferentially adjusted to quickly improve the accuracy of the simulation model.

[0238] Generally speaking, this method constructs a closed-loop optimization system from the simulation results to the actual operation state through the detailed analysis of the differences in key indicators and the quantitative generation of difference factors. It not only significantly improves the adaptability of simulation input parameters, but also enhances the dynamic adjustment ability of the simulation model, providing reliable technical support for the high-precision simulation of complex systems.

[0239] More specifically, the preset weight rule is set according to the degree of influence of key indicators on the simulation results, where the degree of influence of key indicators is determined through sensitivity analysis in the simulation software model of the simulation, including:

[0240] During the simulation process of complex systems, the degrees of influence of different key indicators on the simulation results may vary significantly. For example, in the simulation of wind power generation equipment, wind speed may have a greater impact on the prediction accuracy of power generation, while the impact of environmental temperature is smaller. Therefore, it is necessary to conduct sensitivity analysis on the degrees of influence of key indicators and set weight rules according to the analysis results to ensure that the importance of key indicators can be accurately reflected when calculating difference factors.

[0241] The weight rule assigns different weights to the differences of each key indicator, enabling the difference factors to highlight the key points, thereby guiding the optimization direction of the simulation model. For example, for some key indicators that have a greater impact on the simulation results, such as wind speed or vibration amplitude, the weights will be higher; while for indicators with smaller impacts, such as environmental temperature and humidity, the weights will be lower.

[0242] The concept and method of sensitivity analysis

[0243] Sensitivity analysis is a method of analyzing the degree of influence of input parameters on simulation results. By changing the values of input parameters, observing the degree of change in simulation results, and judging the sensitivity of each key indicator.

[0244] Sensitivity refers to the degree of influence of a certain key indicator on the simulation results. For example, when a small change in a certain key indicator causes a large change in the simulation results, the sensitivity of this indicator is relatively high; otherwise, the sensitivity is relatively low.

[0245] The goal of sensitivity analysis is to quantify the contribution of each key indicator to the simulation results, so as to provide a basis for subsequent weight setting.

[0246] Common methods of sensitivity analysis:

[0247] Local sensitivity analysis: By making a small change to a single indicator, such as increasing or decreasing a certain percentage, and observing the change range of the simulation results. This method is suitable for the sensitivity assessment of linear or quasi-linear models.

[0248] Global sensitivity analysis: By changing the values of multiple key indicators simultaneously and observing the overall change trend of the simulation results. This method is suitable for evaluating non-linear and multi-dimensional complex models.

[0249] Sensitivity analysis based on historical data: Analyze the actual impact of changes in each key indicator on the results in the past operation or simulation historical data. This method can directly obtain sensitivity information from actual data.

[0250] The process of setting weight rules

[0251] Determine the degree of influence of key indicators:

[0252] Through sensitivity analysis, obtain the contribution value of each key indicator to the simulation results. For example, in wind power generation simulation, by adjusting the wind speed, blade angle, and ambient temperature, observe the change trend of the power generation. If the change in wind speed causes a large change in the power generation, while the ambient temperature only causes a small fluctuation, it can be judged that the wind speed has a greater impact on the simulation results.

[0253] Classify the importance of key indicators:

[0254] According to the results of sensitivity analysis, classify the key indicators into three categories: high sensitivity, medium sensitivity, and low sensitivity. Indicators in each category can be assigned similar weight values. For example:

[0255] High sensitivity: Higher weight, for example, accounting for 50% to 60%.

[0256] Medium sensitivity: Moderate weight, for example, accounting for 20% to 30%.

[0257] Low sensitivity: The weight is relatively low, for example, less than 10%.

[0258] During the simulation process, the weight is dynamically adjusted according to the real-time operating state. For example, when the device is in a specific operating stage, such as when the wind speed fluctuates violently, the weight of the wind speed-related indicators can be temporarily increased, while the weight is reduced during the stable operating stage. This dynamic adjustment mechanism can be achieved through the real-time results of sensitivity analysis.

[0259] During the calculation of the difference factor, according to the classification results of the key indicators, the indicators with high sensitivity are assigned higher weights. For example, when calculating the difference of a certain key indicator, the corresponding weight value can be directly referenced according to the predefined weight table.

[0260] In a preferred embodiment of the present invention, the calculation formula for the feedback adjustment weight is:

[0261] , where is the adjusted feedback weight, expressed as the th parameter value in the th simulation result data, is the feedback weight before adjustment, expressed as the th parameter value in the th simulation result data, inherited from the adaptation mapping parameter set of the previous round, is the adjustment coefficient, is the th simulation result data and the th actual operation data in the th parameter value of the key indicator difference, , is the th parameter value in the th simulation result data, the th actual operation data in the th parameter value, is the th weight coefficient of the th parameter value in the th difference factor, is the th simulation result data and the th actual operation data in the th parameter value of the key indicator difference, is the total number of difference factors.

[0262] In the embodiments of the present invention, by calculating the feedback adjustment weights, the matching degree between the simulation results and the actual operation data can be dynamically optimized. In this process, by comparing the key index differences between the simulation result data and the actual operation data, a set of difference factors can be generated. This set quantifies the deficiencies of the simulation results and provides a basis for subsequent adjustments.

[0263] Calculating the feedback adjustment weights according to the set of difference factors can optimize the mapping relationship of the simulation software model by adjusting the parameter weights. For example, in the parameter dimension with a large difference, increasing its weight can significantly improve the adaptability of the model to these key parameters. Through this dynamic adjustment process, the simulation results can gradually approach the actual operation data, thereby improving the prediction accuracy of the simulation.

[0264] Finally, by inputting the updated set of adapted mapping parameters into the dynamic adjustment step, a closed-loop optimization process can be formed. This closed-loop optimization can continuously correct the parameters and inputs of the model within multiple simulation cycles, and ultimately enable the simulation results to achieve the goals of high precision and high reliability. This process is not only applicable to static simulations, but also can be widely applied to real-time dynamic simulations and multi-device collaborative simulation scenarios, providing strong support for the operation optimization of complex systems.

[0265] The above is the preferred embodiment of the present invention. It should be noted that for those of ordinary skill in the art of this technology, without departing from the principle described in the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. A functional mapping management method for a data acquisition system and simulation software, characterized in that The method includes: S100. Collect an original data set from a target device, where the original data set includes at least one data dimension; S200. Perform noise filtering and outlier detection on the original data set to obtain a preprocessed data set; S300. Extract multiple feature data corresponding to a simulation software model from the preprocessed data set to generate a comprehensive feature data set, where the feature data includes independent data sets for different physical attributes; S400. Generate an initial mapping parameter set according to the mapping rule between the comprehensive feature data set and the input parameters of the simulation software model; S500. Dynamically adjust the initial mapping parameter set according to the real-time operating state of the target device. The dynamic adjustment includes detecting the change amount of the operating state of the target device, selecting a corresponding preset adjustment strategy according to the change amount, and generating an adapted mapping parameter set; S600. Convert the feature data into input format data required by the simulation software model according to the adapted mapping parameter set. The conversion includes format normalization processing for different types of simulation software models, and output a simulation input data set; S700. Perform consistency verification on the adapted mapping parameter set and the simulation input data set, where: If the verification passes, store the adapted mapping parameter set as a mapping model configuration; if the verification fails, adjust the adapted mapping parameter set according to a predefined correction rule and regenerate the simulation input data set; S800. Input the simulation input data set into the simulation software model to obtain simulation result data; S900. Verify the simulation result data. If the verification fails, generate feedback adjustment parameters according to the difference degree between the simulation result data and the feature data, where: Compare the key index differences between the simulation result data and the operating data of the target device to generate a set of difference factors; Calculate the feedback adjustment weight according to the set of difference factors and update the adapted mapping parameter set; Input the updated adapted mapping parameter set into step S500 for dynamic adjustment; if the verification passes, use the simulation result data as the output data for the operating evaluation of the target device; The generating of the initial mapping parameter set according to the mapping rule between the comprehensive feature data set and the input parameters of the simulation software model includes: Group and correspond the comprehensive feature data set with the input parameter set of the simulation software model to establish a feature grouping mapping rule; According to the feature grouping mapping rule, perform weighted calculation on the correlation of the feature data to generate an initial mapping parameter matrix. The generation formula of the initial mapping parameter matrix is: , where is the parameter value at the -th row and -th column in the initial mapping parameter matrix, is the correlation coefficient between the -th feature data and the -th input parameter, and are adjustment coefficients, is the -th eigenvalue in the comprehensive feature dataset, is the nominal value of the -th parameter in the input parameter set of the simulation software model, is the total number of features in the comprehensive feature dataset, is the base of the natural logarithm; where , is the value of the th feature data on the th dimension, is the value of the th input parameter on the th dimension, is the mean value of the feature data, is the mean value of the input parameter, is the weight coefficient of the th dimension, is the total number of data dimensions; According to the initial mapping parameter matrix, extract each parameter value and perform range normalization and structuring processing on it to generate an initial mapping parameter set.

2. The functional mapping management method of a data acquisition system and a simulation software according to claim 1, characterized in that The extracting of multiple feature data corresponding to the simulation software model according to the preprocessed data set includes: According to the preprocessed data set, adopt a dimension-by-dimension analysis method to extract the trend features and periodic features in the time series to obtain a trend data set and a periodic data set; According to the trend data set, use a segmented statistical method to generate a set of key statistical quantity data for multiple intervals; Perform spectral analysis on the periodic data set, extract frequency feature data and classify it into a frequency domain feature set; Perform feature fusion on the trend data set, the key statistic data set and the frequency domain feature set to generate a comprehensive feature data set.

3. A functional mapping management method for a data acquisition system and a simulation software according to claim 2, characterized in that, According to the trend data set, generate a key statistic data set with multiple intervals using the segmented statistical method, including: Divide the trend data set into multiple time intervals along the time axis, and the length of each time interval is dynamically set according to the operating cycle of the target device; Perform statistical analysis on the trend data within each time interval, and calculate the average value, median, variance and maximum value of the time interval; According to the predefined key index screening rules, select the statistics with significant fluctuations or change trends as key statistics; Normalize the key statistics within all time intervals to generate a key statistic data set with multiple intervals.

4. A functional mapping management method for a data acquisition system and a simulation software according to claim 3, characterized in that The performing spectral analysis on the periodic data set, extracting frequency feature data and classifying it into a frequency domain feature set, includes: Perform fast Fourier transform on the periodic data set to convert the time domain data into frequency domain data; Extract the main frequency components in the spectrum, and calculate the corresponding amplitude, phase and power density; According to the amplitude size of the main frequency components, select the main frequency components exceeding the preset amplitude threshold as the main frequency features; Classify the amplitude, phase and power density of the main frequency features and generate a frequency domain feature set.

5. A functional mapping management method for a data acquisition system and a simulation software according to claim 4, characterized in that, ​ ​ ​ , where is the parameter value at the th row and th column in the adjusted mapping parameter matrix, is the adjustment coefficient, is the real-time running state change amount of the th feature data, , is the actual running state of the th feature data, is the expected running state of the th feature data, is the total number of state change amounts, is the real-time running state change amount of the th feature data, is the adjustment coefficient; ​ 6. The functional mapping management method of a data acquisition system and a simulation software according to claim 5, characterized in that, ​ ​ ​ ​ ​ 7. A functional mapping management method for a data acquisition system and a simulation software according to claim 6, characterized in that, ​ ​ ​ If a parameter conflict is detected, a corrected data set is generated according to the type of the conflicting parameter, where: For a parameter that exceeds the input range of the simulation software model, adjust its normalized value to the input range of the simulation software model; For a missing parameter, find the closest alternative value from the adapted mapping parameter set to replace it; For duplicate or redundant parameters, delete the conflicting items and maintain the uniqueness of the key parameters; Update the adapted mapping parameter set according to the corrected data set and regenerate the simulation input data set.

8. A functional mapping management method for a data acquisition system and a simulation software according to claim 7, characterized in that Comparing the key index differences between the simulation result data and the target device operation data to generate a difference factor set, including: Extract the key index values of the target device operation data, where the key indexes include the fluctuation range, statistical characteristics, and frequency characteristics of time series parameters; Extract the key index values of the simulation result data, and the key index values of the simulation result data correspond one by one to the key index values in the target device operation data; Compare the differences between each key index value in the simulation result data and the target device operation data, including calculating the mean deviation, extreme value deviation, and deviation degree of the change trend of the corresponding key index values; Based on the differences of the key index values, generate a difference factor set according to the preset weight rule, and the preset weight rule is set according to the influence degree of the key index on the simulation result, where the influence degree of the key index is determined by its sensitivity analysis in the simulation software model.

9. A functional mapping management method for a data acquisition system and a simulation software according to claim 8, characterized in that, The calculation formula for the feedback adjustment weight is: , where is the th feedback weight after adjusting the th parameter value in the th simulation result data, is the th feedback weight before adjusting the th parameter value in the th simulation result data, inherited from the adaptation mapping parameter set of the previous round, is the adjustment coefficient, is the th key index value difference between the th simulation result data and the th actual operation data at the th parameter value, is the th parameter value in the th actual operation data, is the th parameter value in the th simulation result data, is the th weight coefficient of the th parameter value among the th difference factors, is the th key index value difference between the th simulation result data and the th actual operation data at the is the total number of difference factors.

Images