New energy data correlation evaluation method and system and storage medium

Abstract

The invention relates to the field of new energy power generation technology and data intelligent analysis, and particularly discloses a new energy data correlation evaluation method and system and a storage medium. A digital space reference model of a new energy station is constructed. And screening out the parameter pairs with time-space co-occurrence and numerical values mutually falling in the dynamic admissible domain of the opposite side to form a candidate association set. And calculating three factors of change trend consistency, fluctuation frequency matching and abnormal event synchronism of the candidate parameter pair, setting a dynamic gate based on an error transfer method for checking, and judging the candidate pair as a strong correlation candidate pair by a passing person. And aiming at the strong correlation candidate pair, constructing a minimum energy control model, and taking the minimum control energy as a correlation intensity quantitative index. Setting a basic threshold value and a high-order threshold value, and classifying the parameter pair association relationship according to threshold value comparison. And high-precision, high-robustness and interpretability strength judgment and type identification of the association relationship among the equipment parameters in the station are realized.

Description

Technical Field

[0001] This invention relates to the field of new energy power generation technology and intelligent data analysis, specifically to a new energy data correlation assessment method, system, and storage medium. Background Technology

[0002] In the process of intelligent operation of new energy power plants, a large number of sensors are deployed inside the plants, continuously collecting massive amounts of operating parameters and environmental monitoring data from various main-level equipment such as wind turbines, photovoltaic inverters, and booster stations. This multi-source, heterogeneous data contains complex interactions between equipment and system operating patterns. Accurately identifying and quantifying the intrinsic relationships between these data parameters is a key foundation for achieving intelligent operation and maintenance, condition assessment, fault early warning, and collaborative optimization of the power plant, and is of great significance for improving the overall efficiency and safety stability of the power plant.

[0003] Currently, the data correlation analysis methods commonly used in the industry mainly rely on traditional statistical theories, such as Pearson correlation coefficient and Spearman rank correlation. However, these classic methods reveal many limitations when dealing with the complex physical system of new energy power plants. First, they are usually static and isolated analyses, focusing only on the statistical characteristics of the data sequence itself, while completely ignoring the actual layout, geometry, and inherent physical limits of the equipment in three-dimensional physical space. This disconnect from physical reality may lead to the analyzed correlations being physically invalid or missing real correlations influenced by spatial location. Second, these methods are sensitive to inherent noise fluctuations, intermittent characteristics, and outliers in the data, lacking robustness and prone to producing false correlations or misjudgments. Furthermore, most of them can only provide a statistically significant correlation value, failing to quantify the cost or energy required for one parameter to have a substantial impact on another, thus making it difficult to distinguish the strength of the correlation and the ease of control.

[0004] Therefore, existing technologies struggle to provide a method that can deeply integrate physical space information, dynamically adapt to operational states, and quantify correlations from a system energy perspective. This results in analysis results that are often inaccurate and poorly interpretable, making it difficult to directly support in-depth applications based on correlation knowledge, thus hindering further improvements in the intelligence level of new energy power plants. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for evaluating the correlation of multi-source parameters in new energy power plants. By constructing a digital spatial benchmark model and combining dynamic allowable domain screening, multi-dimensional consistency factor testing, minimum energy control quantification, Monte Carlo significance verification, and semantic fusion classification, this method achieves high-precision, high-robustness, strong interpretability determination and type identification of the correlation between equipment parameters within the power plant, providing a data foundation for intelligent operation and maintenance and collaborative control of new energy power plants.

[0006] The specific technical solution of this application is as follows:

[0007] According to one aspect of this application, a method for assessing the relevance of new energy data is provided, comprising:

[0008] By integrating the three-dimensional geometric position, shape parameters and surrounding environmental elements of each main-level equipment in the power station, a digital spatial benchmark model of the new energy power station is constructed.

[0009] Based on the physical limits of the equipment and historical operating data, dynamic permissible domains are constructed for each pair of parameters in the digital space benchmark model. Parameter pairs whose values ​​co-occur in time and space and fall within each other's dynamic permissible domains are selected to form a candidate association set.

[0010] For each parameter pair in the candidate association set, calculate its trend consistency factor, fluctuation frequency matching factor, and abnormal event synchronicity factor, and adaptively calculate the variance of each factor based on the error propagation method to set the dynamic gate. Parameter pairs that pass all dynamic gate tests are determined as strong association candidate pairs.

[0011] For strongly correlated candidate pairs, a minimum energy control model is constructed. The minimum control energy required to drive one parameter sequence into another parameter sequence is calculated iteratively. The minimum control energy is defined as a quantitative index of the correlation strength of the parameter pair.

[0012] The distribution of correlation strength quantification indexes for unrelated data sequences is generated by Monte Carlo simulation, and the basic threshold for determining the significance of correlation and the higher-order threshold considering the system's adjustment capability are set accordingly.

[0013] Based on the comparison of basic thresholds, higher-order thresholds, and correlation strength quantification indicators, the correlation between parameter pairs is classified into inherent strong correlation, regulatory correlation, or weak / no correlation.

[0014] As a further option of the method of the present invention, the constructed digital spatial reference model is , which includes Complete information for each device, for each device Record its three-dimensional spatial coordinates Equipment external parameters and environmental factors Digital Space Benchmark Model Represented as: ;

[0015] The construction of the digital spatial benchmark model for new energy power stations includes:

[0016] Multi-source fusion mapping technology is used to collect three-dimensional geometric position data, including the establishment of a global coordinate system, measurement of equipment center coordinates, measurement of equipment spatial orientation, and acquisition of elevation data;

[0017] Combine equipment design drawings and on-site measured data to collect equipment external parameters, including main dimension measurements, structural feature descriptions, auxiliary component positioning, and dynamic component movement range;

[0018] It integrates surrounding environmental elements, including topographic data, meteorological data, building and obstacle data, and electromagnetic environment data, and uses geographic information system technology to unify them under the same spatial reference system;

[0019] The data is preprocessed, including coordinate system normalization, data format standardization, and spatial index establishment. The coordinate system normalization adopts the seven-parameter coordinate transformation method, the data format standardization adopts the IndustryFoundationClasses format, and the spatial index establishment adopts the octree index structure.

[0020] As a further option of the method of the present invention, the construction of the dynamic tolerance domain includes:

[0021] Determine the physical limit boundary of the parameters ,in, and These are the minimum and maximum physical limits of the equipment, respectively.

[0022] Determine the safe operating threshold of the parameters ,in, and These represent the minimum and maximum safe operating thresholds, respectively.

[0023] Calculate the statistical boundary of the parameters under the current spatiotemporal conditions. ,in, and These are the minimum and maximum values ​​within the historical statistical range, respectively.

[0024] The dynamic allowable domain is determined by combining the three factors. :

[0025] .

[0026] As a further option of the method of the present invention, the candidate association set is generated by: initializing the candidate association set. For each parameter pair ,in For each time point ,if and , put the parameters Included in candidate association set Break out of the inner loop and return. ,in, For the first Each device parameter For the first Each device parameter and The first The and the first Each device parameter is in Device parameters at any given time, and The first The and the first Each device parameter is in The dynamic permissive domain at any given moment.

[0027] As a further option of the method of the present invention, the trend consistency factor Measuring distance through improved dynamic time warping:

[0028] ;

[0029] in, for Parameters correspond to the consistency factor of change trend , For parameters Detrending sequence For parameters Detrending sequence The normalized dynamic time-warped distance;

[0030] The fluctuation frequency matching factor The center frequency matching factor of the main modal components is calculated by performing empirical mode decomposition on the sequence.

[0031] ;

[0032] in, for Parameters corresponding to the fluctuation frequency matching factor , For parameters The The center frequencies of the eigenmode functions For parameters The The center frequencies of the eigenmode functions The number of primary modes selected;

[0033] The abnormal event synchronization factor Through mutation point detection and event synchronization rate calculation:

[0034] ;

[0035] in, for Parameters correspond to the synchronization factors of abnormal events , and Parameters and A collection of abnormal events.

[0036] As a further option of the method of the present invention, the setting of the dynamic gate includes:

[0037] Calculate the variance of the consistency factor of the trend of change Variance of fluctuation frequency matching factor Variance of the synchronization factor of abnormal events ;

[0038] Dynamic gate setting ,in, This represents the expected value of the factor under the assumption of no correlation. The constant is 2 to 3. For variance;

[0039] A parameter pair is considered a strong correlation candidate pair only when the trend consistency factor, fluctuation frequency matching factor, and abnormal event synchronicity factor all pass their respective dynamic gate tests.

[0040] As a further option of the method of the present invention, the minimum control energy calculation includes:

[0041] Define the control objective: to make the system state... Tracking reference sequence ;

[0042] Define the performance metric: Minimize control energy;

[0043] Define constraints: physical constraints that control inputs and system states.

[0044] The mathematical formulation of the optimal control problem is:

[0045] ;

[0046] ;

[0047] ;

[0048] Where A is the system matrix and B is the control matrix. For a moment The system status, For a moment The system status, For a moment The control input, To control energy levels, To control the time period, For a moment Control input.

[0049] Solving the above formula yields the minimum control energy value. .

[0050] As a further option of the method of the present invention, the generation of uncorrelated data sequences through Monte Carlo simulation includes:

[0051] Disrupts the temporal order of the original sequence, destroying temporal correlation;

[0052] Parameter sequences are extracted from different devices or different time periods to form sequence pairs;

[0053] The phase is randomized by Fourier transform, which preserves the amplitude spectrum of the sequence but destroys the temporal structure.

[0054] Generate 1000 to 10000 pairs of unrelated sequence pairs, calculate the quantitative index of association strength, and establish a null distribution;

[0055] Base threshold Set as:

[0056] ;

[0057] in, For the distribution of association strength of unrelated sequences Quantiles It is the set of association strengths for unrelated sequence pairs;

[0058] Higher-order threshold Set as:

[0059] ;

[0060] in The adjustment capability coefficient is set to 0.1-0.2.

[0061] As a further option of the method of the present invention, the classification of the correlation between parameter pairs includes:

[0062] Inherent strong association: Furthermore, it exhibits a stable lead-lag relationship, and its business semantics conform to physical laws.

[0063] Regulatory associations: There is a clear leading-lagging relationship, and there is a possibility of adjustment in terms of business semantics;

[0064] Weak / No correlation: They may lack stable temporal relationships or have no semantic basis in business.

[0065] Another aspect of this application provides a new energy data correlation assessment system, the system comprising:

[0066] The digital spatial benchmark modeling module is used to integrate the three-dimensional geometric position, shape parameters and surrounding environmental elements of each main-level equipment in the power station to construct a digital spatial benchmark model of the new energy power station.

[0067] The candidate association set filtering module is used to construct dynamic permissible domains for each pair of parameters in the digital space benchmark model based on the physical limits of the equipment and historical operating data. It filters out parameter pairs whose values ​​co-occur in time and space and fall within each other's dynamic permissible domains to form a candidate association set.

[0068] The strong correlation candidate pair determination module is used to calculate the trend consistency factor, fluctuation frequency matching factor and abnormal event synchronization factor for each parameter pair in the candidate correlation set, and adaptively calculate the variance of each factor based on the error propagation method to set the dynamic gate. The parameter pairs that pass all dynamic gate tests are determined as strong correlation candidate pairs.

[0069] The correlation strength quantification module is used to construct a minimum energy control model for strongly correlated candidate pairs. It iteratively calculates the minimum control energy required to drive one parameter sequence into another parameter sequence, and defines the minimum control energy as the correlation strength quantification index of the parameter pair.

[0070] The association threshold setting module is used to generate the association strength quantification index distribution of unrelated data sequences through Monte Carlo simulation, and to set the basic threshold for determining the significance of association and the higher-order threshold considering the system's adjustment capability.

[0071] The association classification module is used to classify the association between parameter pairs into inherent strong association, regulatory association, or weak / no association based on the comparison between the basic threshold, higher-order threshold and the association strength quantification index.

[0072] A third aspect of this application provides a computer read storage medium storing at least one instruction, at least one program, a code set, or an instruction set, wherein the at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by a processor to implement a new energy data correlation assessment method as described above.

[0073] The beneficial effects of this application are as follows:

[0074] 1) By constructing a digital spatial benchmark model that integrates the three-dimensional geometry, shape, and environmental elements of the equipment, and by building a dynamic allowable domain for the parameters that combines physical limits and historical data, data analysis is grounded in physical reality, which fundamentally avoids false correlations and significantly improves the accuracy and credibility of the results.

[0075] 2) A multi-dimensional factor is used for comprehensive judgment, and the dynamic gate is adaptively set in combination with the error propagation theory, so that the association recognition can fully capture the interactive features of different levels between parameters, while effectively suppressing noise interference and making it more robust.

[0076] 3) It innovatively introduces a minimum energy control model, quantifying the correlation strength as the minimum control energy required to achieve parameter-driven control. This transforms the strength of the correlation from an abstract statistical value into a physically interpretable regulatory cost, providing a new and profound perspective for distinguishing between inherent correlations and potential controllable correlations.

[0077] 4) By using Monte Carlo simulation to set a basic threshold for statistical significance and a higher-order threshold considering the system's adaptability, the correlations are refined and categorized accordingly. This ensures that the analysis results possess both statistical rigor and engineering guidance value, directly supporting accurate decision-making and collaborative optimization strategy formulation in intelligent operation and maintenance.

[0078] In summary, this invention, through a series of innovations including physical modeling, multi-factor dynamic testing, energy quantification, and dual-threshold classification, constructs an evaluation framework that is more accurate, robust, profound, and practical than traditional methods. Attached Figure Description

[0079] Figure 1 A schematic diagram of the overall process of a new energy data correlation assessment method;

[0080] Figure 2 A detailed flowchart of steps S100 for a new energy data correlation assessment method;

[0081] Figure 3 A detailed flowchart of steps S200 for a new energy data correlation assessment method;

[0082] Figure 4 A detailed flowchart of the steps in the S300 method for evaluating the correlation of new energy data.

[0083] Figure 5 A detailed flowchart of the S400 steps in a new energy data correlation assessment method;

[0084] Figure 6 This is a detailed flowchart of the S500 steps for a new energy data correlation assessment method. Detailed Implementation

[0085] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0086] Example 1

[0087] As complex energy systems, renewable energy power plants exhibit intricate coupling relationships among their main equipment. Accurately identifying and quantifying the correlations between these equipment parameters is crucial for improving the plant's operational efficiency and reliability. This invention is based on three pillars: dynamic system energy control theory, spatiotemporal co-occurrence analysis, and multi-factor error propagation theory. It achieves physical information fusion by constructing a digital spatial benchmark model, uses dynamic allowable domain screening for initial selection of spatiotemporal co-occurrence parameter pairs, combines multi-factor dynamic gate testing for robust correlation identification, and quantifies the correlation strength using a minimum energy control model.

[0088] The core theoretical derivation is as follows:

[0089] The digital spatial benchmark model of new energy power stations is represented by a set of equipment parameters. ,in, This represents the total number of parameters. Each parameter... Associate its three-dimensional spatial coordinates Equipment external parameters and environmental factors Digital Spatial Benchmark Model Formal representation: This model provides a physical space constraint basis for subsequent correlation analysis.

[0090] Based on the physical limits of the equipment and historical operating data, for each pair of parameters Construct a dynamic allowance domain. Parameters At any moment Dynamic allowable domain Defined as: The lower boundary and the Upper Realm Due to the physical limits of the equipment Safe operation threshold and historical statistical range Joint decision:

[0091] ;

[0092] ;

[0093] Historical statistics range The following was obtained through kernel density estimation:

[0094] ;

[0095] ;

[0096] in, For parameters At any moment of Quantile estimates It is usually taken as 0.05.

[0097] Parameter pair Included in candidate association set The conditions are:

[0098] ;

[0099] This condition ensures that the parameters co-occur in time and space and that their values ​​are within the dynamic tolerance range of the other.

[0100] For parameter pairs in the candidate association set, three key factors are calculated: Trend Consistency Factor. Measuring distance through improved dynamic time warping:

[0101] ;

[0102] in, For parameters Detrending sequence For parameters Detrending sequence The normalized dynamic time-warped distance.

[0103] Fluctuation frequency matching factor The center frequency matching degree of the main modal components is calculated after performing empirical mode decomposition on the sequence.

[0104] ;

[0105] in, For parameters The The center frequencies of the eigenmode functions The number of primary modes selected.

[0106] Synchronization factor of abnormal events Through mutation point detection and event synchronization rate calculation:

[0107] ;

[0108] in, and Parameters and A collection of abnormal events.

[0109] Based on the error propagation theory, the variance of each factor is calculated in the following way:

[0110] ;

[0111] in, The variance of the consistency factor of the changing trend. For the first input data One source of error, Its variance. Similarly, the variance of the fluctuation frequency matching factor can be calculated. Variance of the synchronization factor of abnormal events .

[0112] The dynamic gate is set as follows: ,in, This represents the expected value of the factor under the assumption of no correlation. It is a constant, usually taken as 2 to 3. Only when all three factor values ​​pass their respective dynamic gate tests are the parameter pairs determined to be strongly correlated candidate pairs.

[0113] For strongly correlated candidate pairs, a minimum energy control model is constructed. The parameter pairs are... Consider it as a dynamic system:

[0114] ;

[0115] The control objective is to find the control sequence. This makes the system state Tracking reference sequence At the same time, minimize control energy: The solution to this optimal control problem gives the minimum control energy. It is defined as a quantitative index of the correlation strength between parameter pairs.

[0116] The correlation strength distribution of unrelated data sequences is generated through Monte Carlo simulation, with a basic threshold. Set as:

[0117] ;

[0118] in, For the distribution of association strength of unrelated sequences Quantiles. Higher-order threshold After considering the system's adaptability, the setting is as follows: ;in, Take a value of 0.1 to 0.2.

[0119] The above theoretical framework provides a solid mathematical foundation for this invention, ensuring the accuracy, robustness, and interpretability of the correlation evaluation. The specific implementation methods of this invention will be described in detail below.

[0120] Please see Figure 1 This illustrates a new energy data correlation assessment method provided by an embodiment of the present invention, the method comprising:

[0121] S100: Integrates equipment 3D data with environmental elements to construct a digital benchmark model of the site.

[0122] S200: Parameter pairs are selected based on physical limits and historical data to form a candidate association set.

[0123] S300: Calculate multiple factors and screen out strongly correlated candidate pairs through dynamic testing.

[0124] S400: Calculate the correlation strength quantification index using the minimum energy control model.

[0125] S500: Generates an unrelated data distribution through simulation and sets a significance threshold.

[0126] S600: Classify the correlation relationships of parameters based on the quantitative index of correlation strength.

[0127] The specific plan is as follows:

[0128] In the data correlation assessment method for new energy sources, S100 provides a spatial constraint basis for correlation analysis by digitally modeling the physical spatial information of new energy power stations. The digital spatial benchmark model integrates the three-dimensional geometric information, shape parameters, and environmental factors of the equipment to form a unified physical information representation framework.

[0129] Please refer to Figure 2 It illustrates a flowchart of an exemplary new energy data correlation assessment method S100 of this application, the contents of which include:

[0130] S110: Collects three-dimensional geometric position data, equipment shape parameters, and surrounding environmental elements.

[0131] In terms of 3D geometric location data acquisition, multi-source fusion mapping technology is used to obtain precise spatial location information of each main-level device within the site. Specifically, 3D geometric location data acquisition includes: establishing a global coordinate system, measuring the center coordinates of the equipment, measuring the spatial orientation of the equipment, and acquiring elevation data.

[0132] In one alternative implementation, the three-dimensional geometric position data acquisition employs UAV LiDAR scanning technology, which extracts the three-dimensional coordinates of the device's feature points through point cloud data processing, achieving centimeter-level accuracy.

[0133] Equipment external parameters reflect the physical dimensions and structural characteristics of the equipment, and are of great significance for analyzing the spatial relationships and physical interactions between equipment. Specifically, the acquisition of equipment external parameters includes: measurement of main dimensions, description of structural features, positioning of auxiliary components, and range of motion of dynamic components.

[0134] In one alternative implementation, the acquisition of equipment external parameters is combined with equipment design drawings and on-site measured data to ensure the accuracy and completeness of the parameters.

[0135] Surrounding environmental factors affect the operating status and performance of equipment and are an important component of digital spatial benchmark models. Specifically, the integration of surrounding environmental factors includes: topographic data, meteorological data, building and obstacle data, and electromagnetic environment data.

[0136] In one alternative implementation, the integration of surrounding environmental elements employs geographic information system (GIS) technology to unify environmental data from different sources under the same spatial reference system.

[0137] S120: Construct a digital space benchmark model.

[0138] The digital spatial benchmark model integrates the equipment's three-dimensional geometric position, shape parameters, and surrounding environmental elements to form a complete digital twin foundation for the site.

[0139] First, the data is preprocessed, including coordinate system normalization, data format standardization, and spatial index creation.

[0140] In one possible implementation, coordinate system normalization uses a seven-parameter coordinate transformation method to convert all data to a unified station coordinate system.

[0141] In one possible implementation, the data format standardization uses the IndustryFoundationClasses format to represent device geometry and attribute information.

[0142] In one possible implementation, the spatial index is built using an octree index structure, which supports efficient spatial queries and neighbor discovery.

[0143] The constructed digital space benchmark model is , which includes Complete information for each device:

[0144] For each device Record its three-dimensional spatial coordinates Equipment external parameters and environmental factors .

[0145] Digital Spatial Benchmark Model Formal representation: ;

[0146] In a new energy data correlation assessment method, S200 constructs a dynamic permissible domain and filters candidate correlation sets based on a digital spatial benchmark model and equipment operation data. This step, by introducing physical constraints and spatiotemporal co-occurrence conditions, initially filters out parameter combinations that are physically impossible to generate strong correlations, significantly improving the efficiency of subsequent analysis.

[0147] Please refer to Figure 3 The diagram illustrates a flowchart of an exemplary new energy data correlation assessment method S200 of this application, the contents of which include:

[0148] S210: Construct a dynamic allowable domain.

[0149] The physical limits of equipment are hard constraints on the allowable range of parameter variation, determined based on the equipment design specifications and the manufacturer's technical parameters.

[0150] Based on historical operational data over a long time scale, the statistical fluctuation characteristics of parameters under different spatiotemporal conditions are learned, providing a data-driven boundary for the dynamic allowable domain.

[0151] Based on the physical limits of the equipment and historical statistical results, a dynamic tolerance domain is constructed for each parameter that varies with time, space and operating status.

[0152] In one possible implementation, the dynamic permission domain construction process includes:

[0153] 1) Determine the physical limit boundary of the parameters ,in, and These are the minimum and maximum physical limits of the equipment, respectively.

[0154] 2) Determine the safe operating threshold of the parameters. ,in, and These represent the minimum and maximum safe operating thresholds, respectively.

[0155] 3) Calculate the statistical boundary of the parameters under the current spatiotemporal conditions. ,in, and These are the minimum and maximum values ​​within the historical statistical range, respectively.

[0156] 4) Determine the dynamic tolerance domain by combining the three factors:

[0157] ;

[0158] S220: Generate a candidate association set.

[0159] Iterate through all parameter pairs, filter out parameter pairs that satisfy spatiotemporal co-occurrence and whose values ​​are within the dynamic tolerance domain of the other, and form a candidate association set.

[0160] In one possible implementation, the candidate association set generation algorithm is as follows:

[0161] Initialize candidate association set For each parameter pair ,in For each time point ,if and , put the parameters Included in candidate association set Break out of the inner loop and return. .

[0162] In a new energy data correlation assessment method, S300 calculates multi-factor features for candidate correlation sets and screens strong correlation candidate pairs through dynamic gate test.

[0163] Please refer to Figure 4 It illustrates a flowchart of an exemplary new energy data correlation assessment method S300 of this application, the contents of which include:

[0164] S310: Calculation of the consistency factor of changing trends.

[0165] The trend consistency factor measures the similarity between two parameters in terms of their trends, reflecting their consistency in macro-dynamic behavior.

[0166] In one possible implementation, the calculation of the trend consistency factor includes the following steps:

[0167] 1) Perform detrending processing on the parameter sequence to eliminate the influence of long-term trends;

[0168] 2) Calculate the improved dynamic time-normalized distance of the detrended series;

[0169] 3) Convert the distance value into a consistency factor value.

[0170] Specifically, the trend consistency factor Measuring distance through improved dynamic time warping:

[0171] ;

[0172] in, For parameters Detrending sequence For parameters Detrending sequence The normalized dynamic time-warped distance.

[0173] S320: Calculation of fluctuation frequency matching factor.

[0174] The fluctuation frequency matching factor measures the similarity between two parameters in terms of fluctuation frequency characteristics, reflecting their consistency in micro-dynamic properties.

[0175] In one possible implementation, the fluctuation frequency matching factor calculation includes:

[0176] 1) Perform empirical mode decomposition on the parameter sequence to obtain the intrinsic mode functions;

[0177] 2) Calculate the center frequency of each intrinsic mode function;

[0178] 3) Match the main frequency components and calculate the matching factor.

[0179] Specifically, the fluctuation frequency matching factor The center frequency matching factor of the main modal components is calculated by performing empirical mode decomposition on the sequence.

[0180] ;

[0181] in, For parameters The The center frequencies of the eigenmode functions For parameters The The center frequencies of the eigenmode functions The number of primary modes selected.

[0182] S330: Calculation of the synchronization factor for abnormal events.

[0183] The abnormal event synchronicity factor measures the degree of synchronization between two parameters at the time of an abnormal event, reflecting the behavioral correlation between the two under special working conditions.

[0184] In one possible implementation, the calculation of the abnormal event synchronization factor includes:

[0185] 1) Identify anomalous events in parameter sequences based on mutation point detection algorithms;

[0186] 2) Establish the time correspondence between abnormal events;

[0187] 3) Calculate the synchronization rate of abnormal events.

[0188] Specifically, the anomaly event synchronization factor Through mutation point detection and event synchronization rate calculation:

[0189] ;

[0190] in, and Parameters and A collection of abnormal events.

[0191] S340: Error propagation analysis and dynamic gate setting.

[0192] Based on the error propagation theory, the impact of input data uncertainty on the calculation results of each factor is analyzed, and the dynamic gate threshold of each factor is adaptively set.

[0193] In one possible implementation, based on error propagation theory, the variance of each factor is calculated as follows:

[0194] ;

[0195] in, The variance of the consistency factor of the changing trend. As a consistency factor for changing trends, For the first input data One source of error, For the first The variance of each error source can be calculated. Similarly, the variance of the fluctuation frequency matching factor can be calculated. Variance of the synchronization factor of abnormal events

[0196] Dynamic gate setting ,in, This represents the expected value of the factor under the assumption of no correlation. This is a constant, typically taken as 2 to 3. A parameter pair is considered a strong correlation candidate only when all three factor values ​​pass their respective dynamic gate tests.

[0197] S350: Determination of strongly correlated candidate pairs.

[0198] Based on the combined test results of the three factors, determine whether the parameter pair is a candidate pair with a strong association.

[0199] In one possible implementation, the rule for determining strongly correlated candidate pairs is as follows:

[0200] 1) The consistency factor of the changing trend must be greater than its dynamic lower gate bound;

[0201] 2) The fluctuation frequency matching factor must be greater than its dynamic gate lower bound;

[0202] 3) The synchronization factor of abnormal events must be greater than the lower bound of its dynamic gate;

[0203] Parameter pairs that simultaneously meet all three conditions are considered strong correlation candidate pairs.

[0204] In a new energy data correlation assessment method, S400 constructs a minimum energy control model for strongly correlated candidate pairs to quantify the correlation strength between parameters.

[0205] Please refer to Figure 5 It illustrates a flowchart of an exemplary new energy data correlation assessment method S400 of this application, the contents of which include:

[0206] S410: Construct a state-space model.

[0207] Strongly correlated candidate pairs The system is modeled as a dynamic control system, with one parameter serving as the system state and the other as the control input.

[0208] In one possible implementation, the system state-space model is represented as:

[0209] ;

[0210] Where A is the system matrix and B is the control matrix. For a moment The system status, For a moment The system status, For a moment The control inputs are estimated from historical data using system identification methods.

[0211] S420: Construct the optimal control problem and calculate the minimum control energy.

[0212] Construct a minimum energy control problem that drives the system state to a desired trajectory.

[0213] In one possible implementation, the optimal control problem is constructed by:

[0214] 1) Define the control objective: to make the system state... Tracking reference sequence ;

[0215] 2) Define the performance metric: Minimize control energy;

[0216] 3) Determine the constraints: physical constraints that control the input and system state.

[0217] The mathematical formulation of the optimal control problem is:

[0218] ;

[0219] ;

[0220] ;

[0221] in, To control energy levels, To control the time period, For a moment Control input.

[0222] Solving the above formula yields the minimum control energy value. .

[0223] The minimum control energy value is defined as a quantitative index of the correlation strength of parameter pairs.

[0224] In a new energy data correlation assessment method, S500 generates the correlation strength distribution of unrelated data sequences through Monte Carlo simulation, and sets the threshold for determining the significance of correlation accordingly.

[0225] Please refer to Figure 6 It illustrates a flowchart of an exemplary new energy data correlation assessment method S500 of this application, the contents of which include:

[0226] S510: Generation of unrelated sequences and establishment of zero distribution.

[0227] A large number of known unrelated data sequence pairs are generated using the Monte Carlo method to establish a null distribution of correlation strength.

[0228] In one possible implementation, the generation of unrelated data sequences includes:

[0229] 1) Disrupt the temporal order of the original sequence, destroy the temporal correlation, and make the temporal order randomized;

[0230] 2) Extract parameter sequences from different devices and time periods to form sequence pairs;

[0231] 3) By randomizing the phase through Fourier transform, the amplitude spectrum of the sequence is preserved but the temporal structure is destroyed, resulting in unrelated sequence pairs.

[0232] The number of unrelated sequence pairs generated is typically 1,000 to 10,000 to ensure the stability of the null distribution estimation. A quantitative index of association strength is calculated for the generated unrelated sequence pairs to establish a null distribution of association strength.

[0233] S520: Basic threshold setting.

[0234] A basic threshold for the significance of associations is set based on the zero distribution, which is used to identify statistically significant strong associations.

[0235] In one possible implementation, the base threshold Set as:

[0236] ;

[0237] in, For the distribution of association strength of unrelated sequences Quantiles This is the set of association strengths for unrelated sequence pairs. Typically... When the correlation strength of a parameter pair is below this threshold, its probability of occurrence under the assumption of no correlation is less than 5%, and it is considered statistically significant.

[0238] S530: High-level threshold setting.

[0239] Based on the basic threshold, and considering the actual adjustment capability of the system, a more lenient higher-order threshold is set.

[0240] In one possible implementation, a higher-order threshold Set as:

[0241] ;

[0242] in The moderating capacity coefficient, set to 0.1-0.2, reflects the level of correlation strength that the system can achieve through regulation. Parameter pairs with correlation strength between the basic threshold and higher-order thresholds are considered to have regulatory correlation potential.

[0243] In a new energy data correlation assessment method, S600 integrates multi-dimensional information to classify the correlation between parameter pairs.

[0244] Based on the strength of the association and a threshold, the association relationships are divided into three categories.

[0245] In one possible implementation, the association classification rules are as follows:

[0246] 1) Inherent strong correlation: Furthermore, it exhibits a stable lead-lag relationship, and its business semantics conform to physical laws.

[0247] 2) Regulatory association: There is a clear leading-lagging relationship, and there is a possibility of adjustment in terms of business semantics;

[0248] 3) Weak / No correlation: They may lack stable temporal relationships or have no semantic basis in business.

[0249] The technical solution of this invention has been practically applied in multiple new energy power plants, all demonstrating significant technical advantages and economic benefits:

[0250] In terms of fault diagnosis, the fault tracing time based on global correlation has been reduced from an average of 4 hours using traditional methods to less than 15 minutes.

[0251] In terms of operational optimization, by exploring regulatory correlations, annual power generation has increased by approximately 1.5% to 2.0%.

[0252] In terms of operation and maintenance costs, based on accurate correlation analysis, the targeting and effectiveness of preventive maintenance are greatly improved, and operation and maintenance costs are reduced by about 20%.

[0253] In terms of personnel training, the global connectivity provides novice operations and maintenance personnel with an intuitive system cognition tool, shortening the training cycle by about 40%.

[0254] The innovation of this invention lies in the deep integration of physical information and data-driven methods. Through multi-level and multi-angle correlation analysis, a knowledge graph that can truly reflect the operation mechanism of new energy power stations is constructed, providing solid technical support for the intelligent operation and maintenance and optimized operation of the power stations.

[0255] Example 2

[0256] According to Embodiment 1 of this application, a new energy data correlation assessment system is provided, the system comprising:

[0257] The digital spatial benchmark modeling module is used to integrate the three-dimensional geometric position, shape parameters and surrounding environmental elements of each main-level equipment in the power station to construct a digital spatial benchmark model of the new energy power station.

[0258] The candidate association set filtering module is used to construct dynamic permissible domains for each pair of parameters in the digital space benchmark model based on the physical limits of the equipment and historical operating data. It filters out parameter pairs whose values ​​co-occur in time and space and fall within each other's dynamic permissible domains to form a candidate association set.

[0259] The strong correlation candidate pair determination module is used to calculate the trend consistency factor, fluctuation frequency matching factor and abnormal event synchronization factor for each parameter pair in the candidate correlation set, and adaptively calculate the variance of each factor based on the error propagation method to set the dynamic gate. The parameter pairs that pass all dynamic gate tests are determined as strong correlation candidate pairs.

[0260] The correlation strength quantification module is used to construct a minimum energy control model for strongly correlated candidate pairs. It iteratively calculates the minimum control energy required to drive one parameter sequence into another parameter sequence, and defines the minimum control energy as the correlation strength quantification index of the parameter pair.

[0261] The association threshold setting module is used to generate the association strength quantification index distribution of unrelated data sequences through Monte Carlo simulation, and to set the basic threshold for determining the significance of association and the higher-order threshold considering the system's adjustment capability.

[0262] The association classification module is used to classify the association between parameter pairs into inherent strong association, regulatory association, or weak / no association based on the comparison between the basic threshold, higher-order threshold and the association strength quantification index.

[0263] Example 4

[0264] According to Embodiment 1 of this application, a computer read storage medium is provided, wherein the storage medium stores at least one instruction, at least one program, code set or instruction set, wherein the at least one instruction, the at least one program, the code set or instruction set is loaded and executed by a processor to implement a new energy data correlation assessment method as described above.

[0265] Those skilled in the art will understand that the embodiments of this application are provided as methods, systems, or computer program products. Therefore, this application takes the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application takes the form of a computer program product implemented on one or more computer storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer program code. The solutions in the embodiments of this application are implemented using various computer languages, exemplified by the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0266] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, are implemented by computer program instructions. These computer program instructions are provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0267] These computer program instructions are also stored in a computer read-memory memory (CROM) that can direct a computer or other programmed data processing device to operate in a specific manner, such that the instructions stored in the CROM produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0268] These computer program instructions are also loaded onto a computer or other programmed data processing device, causing a series of operational steps to be performed on the computer or other programmed device to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmed device for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0269] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0270] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for evaluating the correlation of new energy data, characterized in that, include: By integrating the three-dimensional geometric position, shape parameters and surrounding environmental elements of each main-level equipment in the power station, a digital spatial benchmark model of the new energy power station is constructed. Based on the physical limits of the equipment and historical operating data, dynamic permissible domains are constructed for each pair of parameters in the digital space benchmark model. Parameter pairs whose values ​​co-occur in time and space and fall within each other's dynamic permissible domains are selected to form a candidate association set. For each parameter pair in the candidate association set, calculate its trend consistency factor, fluctuation frequency matching factor, and abnormal event synchronicity factor, and adaptively calculate the variance of each factor based on the error propagation method to set the dynamic gate. Parameter pairs that pass all dynamic gate tests are determined as strong association candidate pairs. For strongly correlated candidate pairs, a minimum energy control model is constructed. The minimum control energy required to drive one parameter sequence into another parameter sequence is calculated iteratively. The minimum control energy is defined as a quantitative index of the correlation strength of the parameter pair. The distribution of correlation strength quantification index for unrelated data sequences is generated by Monte Carlo simulation, and the basic threshold for determining the significance of correlation and the higher-order threshold considering the system's adjustment capability are set accordingly. Based on the comparison of basic thresholds, higher-order thresholds, and correlation strength quantification indicators, the correlation between parameter pairs is classified into inherent strong correlation, regulatory correlation, or weak / no correlation.

2. The new energy data correlation assessment method according to claim 1, characterized in that, The constructed digital space benchmark model is as follows: , which includes Complete information for each device, for each device Record its three-dimensional spatial coordinates Equipment external parameters and environmental factors Digital Space Benchmark Model Represented as: ; The construction of the digital spatial benchmark model for new energy power stations includes: Multi-source fusion mapping technology is used to collect three-dimensional geometric position data, including the establishment of a global coordinate system, measurement of equipment center coordinates, measurement of equipment spatial orientation, and acquisition of elevation data; Combine equipment design drawings and on-site measured data to collect equipment external parameters, including main dimension measurements, structural feature descriptions, auxiliary component positioning, and dynamic component movement range; It integrates surrounding environmental elements, including topographic data, meteorological data, building and obstacle data, and electromagnetic environment data, and uses geographic information system technology to unify them under the same spatial reference system; The data is preprocessed, including coordinate system normalization, data format standardization, and spatial index establishment. The coordinate system normalization adopts the seven-parameter coordinate transformation method, the data format standardization adopts the IndustryFoundationClasses format, and the spatial index establishment adopts the octree index structure.

3. The new energy data correlation assessment method according to claim 1, characterized in that, The construction of the dynamic permission domain includes: Determine the physical limit boundary of the parameters ,in, and These are the minimum and maximum physical limits of the equipment, respectively. Determine the safe operating threshold of the parameters ,in, and These represent the minimum and maximum safe operating thresholds, respectively. Calculate the statistical boundary of the parameters under the current spatiotemporal conditions. ,in, and These are the minimum and maximum values ​​within the historical statistical range, respectively. The dynamic allowable domain is determined by combining the three factors. : 。 4. The new energy data correlation assessment method according to claim 3, characterized in that, The candidate association set is generated as follows: Initialize the candidate association set. For each parameter pair ,in For each time point ,if and , put the parameters Included in candidate association set Return after exiting the inner loop. ,in, For the first Each device parameter For the first Each device parameter and The first The and the first Each device parameter is in Device parameters at any given time, and The first The and the first Each device parameter is in The dynamic permissive domain at any given moment.

5. The new energy data correlation assessment method according to claim 1, characterized in that, The consistency factor of the changing trend Measuring distance through improved dynamic time warping: ； in, for Parameters correspond to the consistency factor of change trend , For parameters Detrending sequence For parameters Detrending sequence The normalized dynamic time-warped distance; The fluctuation frequency matching factor The center frequency matching factor of the main modal components is calculated by performing empirical mode decomposition on the sequence. ； in, for Parameters corresponding to the fluctuation frequency matching factor , For parameters The The center frequencies of the eigenmode functions For parameters The The center frequencies of the eigenmode functions The number of primary modes selected; The abnormal event synchronization factor Through mutation point detection and event synchronization rate calculation: ； in, for Parameters correspond to the synchronization factors of abnormal events , and Parameters and A collection of abnormal events.

6. The new energy data correlation assessment method according to claim 5, characterized in that, The set dynamic gate includes: Calculate the variance of the consistency factor of the trend of change Variance of fluctuation frequency matching factor Variance of the synchronization factor of abnormal events ; Dynamic gate setting ,in, This represents the expected value of the factor under the assumption of no correlation. The constant is 2 to 3. For variance; A parameter pair is considered a strong correlation candidate pair only when the trend consistency factor, fluctuation frequency matching factor, and abnormal event synchronicity factor all pass their respective dynamic gate tests.

7. The new energy data correlation assessment method according to claim 1, characterized in that, The minimum control energy calculation includes: Define the control objective: to make the system state... Tracking reference sequence ; Define the performance metric: Minimize control energy; Define constraints: physical constraints that control inputs and system states; The mathematical formulation of the optimal control problem is: ； ； ； Where A is the system matrix and B is the control matrix. For a moment The system status, For a moment The system status, For a moment The control input, To control energy levels, To control the time period, For a moment Control input; Solving the above formula yields the minimum control energy value. .

8. The new energy data correlation assessment method according to claim 1, characterized in that, The generation of unrelated data sequences through Monte Carlo simulation includes: Disrupts the temporal order of the original sequence, destroying temporal correlation; Parameter sequences are extracted from different devices or different time periods to form sequence pairs; The phase is randomized by Fourier transform, which preserves the amplitude spectrum of the sequence but destroys the temporal structure. Generate 1000 to 10000 pairs of unrelated sequence pairs, calculate the quantitative index of association strength, and establish a null distribution; Base threshold Set as: ； in, For the distribution of association strength of unrelated sequences Quantiles It is the set of association strengths for unrelated sequence pairs; Higher-order threshold Set as: ； in The adjustment capability coefficient is set to 0.1-0.

2.

9. The new energy data correlation assessment method according to claim 8, characterized in that, The classification of the correlation relationships between parameter pairs includes: Inherent strong association: Furthermore, it exhibits a stable lead-lag relationship, and its business semantics conform to physical laws. Regulatory associations: There is a clear leading-lagging relationship, and there is a possibility of adjustment in terms of business semantics; Weak / No correlation: They may lack stable temporal relationships or have no semantic basis in business.

10. A new energy data correlation assessment system, characterized in that, The system for implementing a new energy data correlation assessment method as described in any one of claims 1 to 9 includes: The digital spatial benchmark modeling module is used to integrate the three-dimensional geometric position, shape parameters and surrounding environmental elements of each main-level equipment in the power station to construct a digital spatial benchmark model of the new energy power station. The candidate association set filtering module is used to construct dynamic permissible domains for each pair of parameters in the digital space benchmark model based on the physical limits of the equipment and historical operating data. It filters out parameter pairs whose values ​​co-occur in time and space and fall within each other's dynamic permissible domains to form a candidate association set. The strong correlation candidate pair determination module is used to calculate the trend consistency factor, fluctuation frequency matching factor and abnormal event synchronization factor for each parameter pair in the candidate correlation set, and adaptively calculate the variance of each factor based on the error propagation method to set the dynamic gate. The parameter pairs that pass all dynamic gate tests are determined as strong correlation candidate pairs. The correlation strength quantification module is used to construct a minimum energy control model for strongly correlated candidate pairs. It iteratively calculates the minimum control energy required to drive one parameter sequence into another parameter sequence, and defines the minimum control energy as the correlation strength quantification index of the parameter pair. The association threshold setting module is used to generate the association strength quantification index distribution of unrelated data sequences through Monte Carlo simulation, and to set the basic threshold for determining the significance of association and the higher-order threshold considering the system's adjustment capability. The association classification module is used to classify the association between parameter pairs into inherent strong association, regulatory association, or weak / no association based on the comparison between the basic threshold, higher-order threshold and the association strength quantification index.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps of a new energy data correlation assessment method as described in any one of claims 1 to 9.

Images