Method and system for judging state of new energy power equipment based on digital twinning

Abstract

The invention discloses a new energy electric power equipment state judgment method and system based on digital twinning, and discloses a method for classifying normal and abnormal state data of electric power equipment based on the equivalence of control parameters and external influence parameters, comparing the same kind of data, extracting an initial abnormal factor data section, and judging the state of the electric power equipment based on the initial abnormal factor data section. Forming an initial abnormal factor data segment group; performing secondary classification based on the proximity of the control parameters, the external influence parameters and the initial abnormal factor data segment group, performing effective interception on the initial abnormal factor data in combination with repeated features to obtain effective abnormal factor data segments, and combining the effective abnormal factor data segments into an effective abnormal factor data segment group; analyzing real-time operation data by using the effective abnormal factor data segment group, constructing a virtual model through a digital twinning technology, and comparing dynamic performance to realize state judgment; according to the invention, problems of abnormal factor extraction and real-time monitoring in a complex environment are solved; the method improves the accuracy of state judgment, and is suitable for the safe operation monitoring of new energy power equipment.

Description

Technical Field

[0001] The present invention relates to the technical field of power equipment status monitoring, and in particular to a method and system for judging the status of new energy power equipment based on digital twin. Background Art

[0002] With the rapid development of new energy technologies, the proportion of new energy power equipment such as wind energy and solar energy in the power system has been continuously increasing. During the operation of these equipment, they are affected by various factors, including not only natural environmental conditions (such as temperature, humidity, wind speed), but also changes in electrical parameters such as fluctuations in input current and voltage from the outside. Due to the complexity of the operating environment and the diversity of operating conditions (such as power regulation, load changes) of new energy power equipment, its status monitoring faces significant technical challenges. Traditional status monitoring methods mostly rely on static threshold setting or simple statistical analysis, and it is difficult to adapt to the superimposed effects of dynamic factors such as fluctuations in input current and voltage from the outside on the equipment status, resulting in insufficient accuracy in identifying abnormal states, and misjudgment or missed judgment phenomena are relatively common.

[0003] In addition, the operation data of new energy power equipment has high-dimensionality and time-variability. How to effectively process these data and extract key factors reflecting abnormal states from them is one of the current technical difficulties. Existing methods often lack a systematic process in data classification and abnormal factor extraction. Especially when facing interference from electrical parameters such as fluctuations in input current and voltage from the outside, it is difficult to accurately distinguish the boundary between normal and abnormal states. At the same time, traditional monitoring means have deficiencies in real-time performance and visualization, and cannot directly associate abnormal states with the dynamic characteristics of equipment operation, which limits the efficiency of maintenance decisions. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and system for judging the status of new energy power equipment with more scientific and accurate status judgment.

[0005] The present invention discloses a method for judging the status of new energy power equipment based on digital twin, including: Step S100, classify the normal state data and abnormal state data of the power equipment based on the equivalence of control parameters and external influence parameters, and compare the normal state data and abnormal state data of the same category to determine a number of initial abnormal factor data segments, and obtain a group of initial abnormal factor data segments; Step S200: Classify the initial abnormal factor data segment groups based on the proximity among the control parameters, external influence parameters, and the initial abnormal factor data segment groups, and perform validity truncation on the initial abnormal factor data based on the repeated characteristics of each initial abnormal factor data in the initial abnormal factor data segment groups of the same category, so as to obtain valid abnormal factor data segments. Denote the combination of the valid abnormal factor data segments as the valid abnormal factor data segment group; Step S300: Analyze the real-time operation data of the power equipment collected in real time by using the valid abnormal factor data segment group to realize the judgment of the state of the power equipment.

[0006] In some embodiments disclosed by the present invention, the method for analyzing the real-time operation data of the power equipment by using the valid abnormal factor data segment group includes: Step S301: Construct a virtual three-dimensional model for the power equipment, and construct virtual three-dimensional modules at the corresponding positions in the virtual three-dimensional model based on the positions of different modules relative to the power equipment. A state representation component is set for each virtual three-dimensional module, and the state representation component is used to represent the state of the corresponding virtual three-dimensional module; Step S302: Determine the abnormal factor data types corresponding to the virtual three-dimensional modules, and construct factor data variable intervals for each abnormal factor data type. Each factor data variable interval corresponds to the representation of a state representation component; Step S303: Recognize the valid abnormal factor data segment group as the input data of the virtual three-dimensional model, determine the valid abnormal dynamic representations of different state representation components, and denote the set of several valid abnormal dynamic representations as the valid abnormal dynamic representation group. Recognize the real-time operation data as the input data of the virtual three-dimensional model, determine the real-time dynamic representations of different state representation components, and denote the combination of several real-time dynamic representations as the real-time dynamic representation group; Step S304: Continuously compare the dynamic representation group with different valid abnormal dynamic representation groups in real time, and recognize the abnormal condition corresponding to the matching valid abnormal dynamic representation group as the real-time state of the power equipment.

[0007] In some embodiments disclosed by the present invention, the method for classifying the state data of the power equipment includes: Step S101: Perform curve fitting on the control parameters and external influence parameters in the time dimension to obtain several control parameter curves and several external influence parameter curves; Step S102, dynamically generate sliding windows with several window time widths, and respectively let each sliding window perform variation analysis on the control parameter curve and the external influence parameter curve. If the reference parameter curve and the external influence parameter curve simultaneously show equivalent variations and have the same period parameter segments, then intercept the parameter segments at this time, and record the combination of several parameter segments as the classified parameter segment group; Step S103, classify the corresponding state data with the classified parameter segment group as the classification condition.

[0008] In some embodiments disclosed by the present invention, the method for determining several initial abnormal factor data segments includes: Step S101, analyze the state data to determine several factor data corresponding to each state data, and record it as the factor data group; Step S102, randomly combine the normal state data and the abnormal state data within the same category to obtain a state data comparison group. The method for comparing the state data comparison group includes, based on the corresponding relationship between the control parameters and the external influence parameters among the state data, align the factor data in the factor data group, determine the factor data difference amounts at different time nodes, and based on the continuous manifestation of the factor data difference amounts, determine the initial abnormal factor data segments on the factor data.

[0009] In some embodiments disclosed by the present invention, the method for determining the initial abnormal factor data segments on the factor data based on the continuous manifestation of the factor data difference amounts includes: Step S1021, determine the factor data difference amount at each time node, judge the preset difference amount interval to which each factor data difference amount belongs, and record the set difference influence index corresponding to the preset difference amount interval at the corresponding time node; Step S1022, if the difference influence index is greater than or equal to the preset value, mark the corresponding time node. If it is less than the preset value, integrate the difference influence indexes corresponding to the continuous time nodes before the time node, and based on the integration result, correct the difference influence index. If the corrected difference influence index is greater than or equal to the preset value, mark the corresponding time node; Step S1023, recognize the factor data segments corresponding to the marked and continuous time nodes as the initial abnormal factor data segments; Among them, the expression for calculating the difference influence index is: ; Among them, is the difference influence index after correction, is the difference influence index before correction, is the correction adjustment coefficient, is the integral value of the differential impact index corresponding to the continuous time nodes before the time node, is the integral impact adjustment coefficient, is the integral impact adjustment constant.

[0010] In some embodiments disclosed by the present invention, the method for classifying the initial abnormal factor data section groups includes: Step S201: Extract discrete nodes from the initial abnormal factor data section groups to obtain a number of initial abnormal factor data parameters, construct the initial abnormal factor data parameters into an initial abnormal factor data parameter sequence, and serialize the control parameters and external influence parameters corresponding to the initial abnormal factor data parameter sequence to obtain a control parameter sequence and an external influence parameter sequence; Step S202: Construct a first similarity operator for the comparison between control parameter sequences, construct a second similarity operator for the comparison between external influence parameter sequences, construct a third similarity operator for the comparison between initial abnormal factor data parameter sequences, determine the similarity degree between the initial abnormal factor data section groups based on the first similarity operator, the second similarity operator, and the third similarity operator, and classify the initial abnormal factor data section groups based on the similarity degree.

[0011] In some embodiments disclosed by the present invention, the method for determining the similarity degree between the initial abnormal factor data section groups includes: Step S2021: Set a similarity influence weight coefficient for each initial abnormal factor data parameter sequence, compare the factor data difference amount of each initial abnormal factor data parameter between the initial abnormal factor data parameter sequences, and count the total sum of the factor data difference amounts, compare the control parameter difference amount of each control parameter between the control parameter sequences, and count the total sum of the control parameter difference amounts, compare each external influence parameter difference amount between the external influence parameter sequences, and count the total sum of the external influence parameter difference amounts; Step S2022: Use the similarity influence weight coefficient, the total sum of the control parameter difference amounts, and the total sum of the external influence parameter difference amounts to correct the total sum of the factor data difference amounts to obtain the similarity degree between the initial abnormal factor data section groups; Among them, the expression for calculating the similarity degree is: ; Among them, is the similarity degree, is the preset maximum similarity degree, is the similarity degree conversion coefficient, is the total sum of the factor data difference amounts of the i-th initial abnormal factor data parameter sequence, the similarity influence weight coefficient corresponding to the i-th initial abnormal factor data parameter sequence, is the adjustment constant for the total amount of factor data differences, is the preset first correction coefficient output function, is the total amount of differences in control parameters, is the preset second correction coefficient output function, is the total amount of differences in external influence parameters.

[0012] In some embodiments disclosed by the present invention, the method for performing validity truncation on initial abnormal factor data includes: Step S203: Randomly combine the initial abnormal factor data section groups, compare the corresponding initial abnormal factor data sections between the initial abnormal factor data section groups, and determine the data sections with repeated manifestations among them, which are recorded as valid abnormal factor data sections.

[0013] In some embodiments disclosed by the present invention, a new energy power equipment status judgment system based on digital twin is also disclosed, including: The first module is used to classify the normal state data and abnormal state data of the power equipment based on the equivalence of the control parameters and external influence parameters, compare the normal state data and abnormal state data of the same category, and determine a number of initial abnormal factor data sections to obtain an initial abnormal factor data section group; The second module is used to classify the initial abnormal factor data section group based on the proximity between the control parameters, external influence parameters, and the initial abnormal factor data section group, and perform validity truncation on the initial abnormal factor data based on the repeated characteristics of each initial abnormal factor data in the initial abnormal factor data section group of the same category to obtain valid abnormal factor data sections, and record the combination of the valid abnormal factor data sections as a valid abnormal factor data section group; The third module is used to analyze the real-time operation data of the power equipment collected in real time by using the valid abnormal factor data section group to realize the judgment of the power equipment status.

[0014] The present invention discloses a method and system for judging the state of new energy power equipment based on digital twins. By making use of the equivalence of control parameters and external influence parameters, the normal and abnormal state data of power equipment are classified, the same-category data are compared, the initial abnormal factor data sections are extracted, and an initial abnormal factor data section group is formed. Based on the proximity of the control parameters, external influence parameters and the initial abnormal factor data section group, secondary classification is carried out, and the initial abnormal factor data are effectively intercepted in combination with repeated features to obtain effective abnormal factor data sections, which are combined into an effective abnormal factor data section group. The real-time operation data are analyzed by using the effective abnormal factor data section group, and a virtual model is constructed through digital twin technology and compared with the dynamic performance to realize state judgment. The present invention solves the problems of abnormal factor extraction and real-time monitoring in complex environments, improves the accuracy of state judgment, and is applicable to the safety operation monitoring of new energy power equipment.

[0015] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Description of the Drawings

[0016] Figure 1 It is a method step diagram of the method for judging the state of new energy power equipment based on digital twins disclosed in the embodiment of the present invention. Detailed Embodiments

[0017] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0018] The technical solution of the present invention will be clearly and completely described below in combination with the accompanying drawings and specific embodiments. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and should not be construed as limiting the protection scope of the present invention. Those skilled in the art can make some non-essential improvements and adjustments according to the content of the present invention described below. In the present invention, unless otherwise clearly specified and limited, the technical terms used in the present invention should have the general meaning understood by those skilled in the art of the present invention.

[0019] Embodiment: The present invention discloses a method for judging the state of new energy power equipment based on digital twins. Refer to Figure 1 , including: Step S100, based on the equivalence of control parameters and external influence parameters, classify the normal state data and abnormal state data of power equipment, compare the normal state data and abnormal state data of the same category, determine a number of initial abnormal factor data sections, and obtain an initial abnormal factor data section group.

[0020] The principle of step S100 lies in leveraging the equivalence of control parameters (such as voltage, current, power) and external influence parameters (such as temperature, humidity, fluctuations in external input current and voltage) to preliminarily screen out characteristic sections in the operation data of new energy power equipment that may be related to abnormal states, laying a data foundation for subsequent refined analysis. This process first requires the scientific classification of a large amount of status data, and the classification basis is the similar change patterns of control parameters and external influence parameters in the time dimension, that is, "equivalence". According to (steps S101 - S103), the specific implementation is divided into three sub-steps: First (S101), perform curve fitting on control parameters and external influence parameters in the time dimension to generate several parameter curves, which can intuitively reflect the trend of parameter changes over time, such as voltage fluctuations or temperature rises and falls; then (S102), by dynamically generating sliding windows of different time widths, perform change analysis on these parameter curves. When the control parameter curve and the external influence parameter curve show equivalent changes (such as similar amplitudes and consistent periods) within the same window, intercept the corresponding parameter curve segments and combine these curve segments into a classified parameter curve segment group; finally (S103), using the classified parameter curve segment group as a condition, classify the corresponding status data (including normal and abnormal data) into different categories to ensure that the data within the same category has a similar operation background.

[0021] After completing the data classification, step S100 further focuses on the extraction of abnormal characteristics. According to (steps S101 - S102), within the same category, compare the normal status data and the abnormal status data, and analyze the differences in their factor data groups (such as status indicators like temperature values, vibration values, etc.). Specifically, first (S101), decompose the status data into several factor data to form a factor data group; then (S102), form a status data comparison group by randomly combining normal and abnormal status data. Based on the corresponding relationship between control parameters and external influence parameters, align the factor data and calculate the difference amounts of factor data at different time nodes. (Steps S1021 - S1023) further refine the difference analysis process: First (S1021), determine the difference amount at each time node and classify it into a preset difference amount interval, and assign the corresponding difference influence index; then (S1022), if the difference influence index is greater than or equal to the preset value, mark that time node. If it is less than the preset value, integrate and correct the difference influence index of the previous consecutive time nodes. Finally, if the corrected index meets the standard, mark it; finally (S1023), identify the factor data sections corresponding to the continuously marked time nodes as the initial abnormal factor data sections. The set of these sections is the initial abnormal factor data section group.

[0022] The core of this step lies in achieving the structured classification of data through parameter equivalence and initially locking the abnormal sections by leveraging the continuous characteristics of factor data differences, solving the problems of complex and variable operation data of new energy power equipment and difficult-to-locate abnormal characteristics, and providing a reliable initial data basis for subsequent steps.

[0023] Step S200: Classify the initial abnormal factor data section groups based on the proximity among the control parameters, external influence parameters, and the initial abnormal factor data section groups, and perform valid interception on the initial abnormal factor data based on the repeated characteristics of each initial abnormal factor data in the initial abnormal factor data section groups of the same category to obtain valid abnormal factor data sections. Denote the combination of the valid abnormal factor data sections as the valid abnormal factor data section group.

[0024] The principle of step S200 is to perform secondary classification and validity verification on the initial abnormal factor data extracted in step S100 through the proximity among the control parameters, external influence parameters, and the initial abnormal factor data section groups, as well as the repeated characteristics of the data within the same category, so as to refine more representative and reliable abnormal factor data and provide high-quality input for real-time status judgment. This process is divided into two main stages: proximity classification and repeated feature interception.

[0025] First, the goal of proximity classification is to further subdivide the initial abnormal factor data section groups to ensure higher similarity of the data within the same category. According to (steps S201 - S202), the implementation process is as follows: First (S201), extract the scattered nodes (such as abnormal peaks, inflection points) from the initial abnormal factor data section groups to generate the initial abnormal factor data parameter sequence, and at the same time serialize the corresponding control parameters and external influence parameters to form the control parameter sequence and the external influence parameter sequence; then (S202), construct three groups of proximity operators - the first proximity operator is used to compare the control parameter sequence, the second proximity operator is used to compare the external influence parameter sequence, and the third proximity operator is used to compare the initial abnormal factor data parameter sequence. Through the comprehensive analysis of these three, calculate the proximity degree among the initial abnormal factor data section groups. (Steps S2021 - S2022) provide the calculation details of the proximity degree: First (S2021), set the proximity influence weight coefficient for the initial abnormal factor data parameter sequence, and respectively count the total sum of the factor data difference amount, the control parameter difference amount, and the external influence parameter difference amount; then (S2022), use the correction formula (J expression) of the weight coefficient and the total difference amount to quantify the proximity degree, and classify the initial abnormal factor data section groups into different categories based on this. This proximity analysis comprehensively considers the characteristics of the control parameters, external influence parameters, and the abnormal factors themselves to ensure that the classification results reflect the real abnormal patterns.

[0026] Secondly, after the classification is completed, step S200 uses repeated features to intercept the validity of the abnormal factor data. According to (step S203), within the same category, the initial abnormal factor data segment groups are randomly combined and compared to identify data segments with repeated manifestations among each other. These repeated segments are considered to be more representative of stable abnormal features and are thus intercepted as valid abnormal factor data segments, and finally combined into a group of valid abnormal factor data segments. This process eliminates abnormal factors affected by chance or noise interference through repeatability verification, enhancing the reliability of the data.

[0027] This step optimizes the initial abnormal factor data through proximity classification and repeated feature interception, solves the problem that the initial data may contain redundancy or misjudgment, and provides an accurate abnormal factor benchmark for subsequent real-time analysis based on digital twins.

[0028] In step S300, the real-time operation data of the power equipment collected in real time is analyzed using the group of valid abnormal factor data segments to realize the judgment of the state of the power equipment.

[0029] The principle of step S300 is to use digital twin technology. The group of valid abnormal factor data segments refined in step S200 is used as a reference benchmark for abnormal states and dynamically compared with the real-time operation data of the power equipment collected in real time, so as to achieve accurate state judgment and visually present the results through a virtual model. This process combines virtual modeling, data mapping, and dynamic analysis, and is particularly suitable for complex scenarios such as fluctuations in input current and voltage from the outside world. (Steps S301 - S304), and the specific implementation is divided into four sub-steps.

[0030] First (S301), construct a virtual 3D model for the power equipment, divide it according to different functional modules (such as power generation units, transmission systems), and set virtual 3D modules at corresponding positions. Each module is equipped with state representation components (such as color change, blinking frequency) to express the operating state of the module. Secondly (S302), determine the data types of abnormal factors corresponding to each virtual 3D module (such as temperature anomaly, vibration anomaly), and construct a factor data variable range for each type. Each range corresponds to a form of expression of the state representation component. For example, high temperature corresponds to red blinking, and low voltage corresponds to blue stillness, thus establishing a mapping relationship between abnormal factors and visual representations. Then (S303), use the group of valid abnormal factor data sections as the input of the virtual model to generate several valid abnormal dynamic representations (such as the red blinking mode during high temperature anomaly). The set of these representations is denoted as the group of valid abnormal dynamic representations; at the same time, input the real-time operation data into the model to generate real-time dynamic representations, and combine them into a group of real-time dynamic representations. Finally (S304), compare the group of real-time dynamic representations with the group of valid abnormal dynamic representations one by one. If a real-time dynamic representation matches a valid abnormal dynamic representation, then the corresponding abnormal condition is determined as the current state. For example, when the high temperature anomaly mode is matched, it is judged that the equipment is in an overheated state.

[0031] The core of this step lies in the application of digital twin technology. Through the virtual 3D model, the abstract abnormal factor data is transformed into intuitive state representations, and through the comparison of real-time data with preset abnormal modes, dynamic monitoring is achieved. The design of the state representation components not only improves the accuracy of judgment but also enhances the visualization effect of the results, enabling maintenance personnel to quickly identify problems and take measures.

[0032] In summary, step S300 applies the valid abnormal factor data to real-time state judgment through digital twin virtual modeling and dynamic comparison, solves the deficiencies of traditional methods in terms of dynamics and visualization, and provides efficient support for the safe operation of new energy power equipment.

[0033] In some embodiments disclosed by the present invention, the method for analyzing the real-time operation data of power equipment collected in real time by using the group of valid abnormal factor data sections includes: Step S301, construct a virtual 3D model for the power equipment, and based on the positions of different modules relative to the power equipment, construct virtual 3D modules at corresponding positions in the virtual 3D model. For each virtual 3D module, a state representation component is set, and the state representation component is used to express the state of the corresponding virtual 3D module.

[0034] The principle of step S301 is to construct a digital twin virtual 3D model for new energy power equipment. Through modular design and the introduction of state representation components with the magnitude of deformation as the core, the operating state of the physical equipment is visually mapped into the virtual space, providing a visualization basis for subsequent anomaly analysis and real-time monitoring. Specifically, first, a virtual 3D model is constructed for power equipment (such as wind turbines, photovoltaic inverters). This model not only replicates the overall structure of the equipment but also constructs virtual 3D modules at corresponding positions in the virtual model according to the relative positions of functional modules (such as power generation units, transmission systems, control modules) in the physical equipment to ensure the spatial consistency between the physical and virtual. For each virtual 3D module, a state representation component is set. Here, the magnitude of deformation is used as the manifestation form. For example, the geometric shape of the module (such as a sphere, cube) maintains a standard size under normal conditions and enlarges or shrinks according to the severity of the state during anomalies. For example, the module volume enlarges during high-temperature anomalies and shrinks during voltage deficiencies. This manifestation method of the magnitude of deformation intuitively reflects the module state through geometric changes, has a stronger sense of space compared to traditional color or flashing methods, and facilitates users to quickly perceive the abnormal area from a 3D perspective. This step combines the requirements of step S300 in the overall technical solution, uses digital twin technology to convert the complex operating data of the physical equipment into a visual expression of the virtual model, and the introduction of the magnitude of deformation enhances the recognizability of the state, laying a foundation for subsequent anomaly factor mapping and dynamic analysis.

[0035] Step S302: Determine the data types of anomaly factors corresponding to the virtual 3D modules, and construct a factor data variable interval for each data type of anomaly factor. Each factor data variable interval corresponds to the manifestation of a state representation component.

[0036] The principle of step S302 is to identify the abnormal factor data types related to the virtual 3D module, quantify them into factor data variable intervals, and establish a mapping relationship with the deformation size representation of the state performance component, so as to convert the abnormal features into intuitive geometric changes and provide a quantitative basis for real-time state judgment. Specifically, first analyze the abnormal factor data types that each virtual 3D module may involve. For example, the power generation module may involve temperature abnormalities and voltage abnormalities, and the transmission module may involve vibration abnormalities and current fluctuations. These types are derived from the abnormal factor data extracted in steps S100 and S200. For each abnormal factor data type, construct a factor data variable interval. For example, the temperature interval can be set as "less than 50°C (normal)", "50 - 70°C (warning)", "greater than 70°C (abnormal)", and the voltage interval is "less than 200V (low voltage)", "200 - 240V (normal)", "greater than 240V (overvoltage)". Each factor data variable interval corresponds to a deformation size representation of the state performance component. For example, when the temperature is less than 50°C, the module maintains the standard size; when it is 50 - 70°C, it is enlarged by 1.2 times; when it is greater than 70°C, it is enlarged by 1.5 times; when the voltage is less than 200V, it is reduced to 0.8 times; when it is greater than 240V, it is enlarged to 1.3 times. This mapping method of deformation size directly correlates data changes with geometric features, and users can intuitively judge the abnormal type and severity through the increase or decrease of the module volume. This step combines the technical solutions of the full text, deeply integrates the abnormal factor data with the deformation performance of the digital twin model, solves the problem that abnormal features are difficult to intuitively express, and provides a standardized visualization rule for the dynamic performance generation of step S303.

[0037] Step S303: Recognize the effective abnormal factor data section group as the input data of the virtual 3D model, determine the effective abnormal dynamic performances of different state performance components, record the set of several effective abnormal dynamic performances as the effective abnormal dynamic performance group, recognize the real-time operation data as the input data of the virtual 3D model, determine the real-time dynamic performances of different state performance components, and record the combination of several real-time dynamic performances as the real-time dynamic performance group.

[0038] The principle of step S303 is to input the valid abnormal factor data segment group generated in step S200 and the real-time operation data into the virtual 3D model respectively, generating a valid abnormal dynamic performance group and a real-time dynamic performance group based on the deformation magnitude, providing a visual data basis for subsequent state comparison. This process uses digital twin technology to dynamicize static data into geometric deformation performances. Specifically, first, the valid abnormal factor data segment group (including classified and intercepted abnormal features) is used as the input of the virtual model. According to the mapping relationship between the factor data variable interval defined in step S302 and the deformation magnitude, the performance of each virtual 3D module in the abnormal state is determined. For example, after the input of the temperature abnormal segment (greater than 70 °C), the volume of the power generation module is enlarged to 1.5 times, and after the input of the vibration abnormal segment, the transmission module is enlarged to 1.4 times. These dynamic performances of the deformation magnitude are combined into the valid abnormal dynamic performance, and several performances are aggregated into the valid abnormal dynamic performance group, constituting the known abnormal state pattern library. Then, the real-time collected operation data (including the current values of the control parameters and the external influence parameters) is input into the model to generate the real-time dynamic performance. For example, when the current temperature is 75 °C, the power generation module is enlarged to 1.5 times, and when the voltage is 190 V, the control module is reduced to 0.8 times. These real-time deformations are combined into the real-time dynamic performance group, reflecting the current state of the equipment. The dynamic change of the deformation magnitude makes the abnormal features more easily identifiable in the three-dimensional space. For example, the significant enlargement of the module volume can quickly indicate a serious abnormality. This step, combined with the requirements of step S300 in the overall technical solution, realizes the visual comparison between the abnormal benchmark and the real-time data through the dynamic expression of the deformation magnitude, providing an intuitive basis for the state judgment in step S304.

[0039] Step S304, compare the dynamic performance group with different valid abnormal dynamic performance groups in real time, and identify the abnormal condition corresponding to the matching valid abnormal dynamic performance group as the real-time state of the power equipment.

[0040] The principle of step S304 is to determine whether the current operating data matches the known abnormal patterns by comparing the real-time dynamic performance group based on the deformation magnitude in the virtual 3D model with the effective abnormal dynamic performance group in real time, so as to determine the real-time state of the power equipment, and improve the decision-making efficiency by using the intuitiveness of the deformation magnitude. This process is the core application of digital twin technology in state judgment. Specifically, after generating the real-time dynamic performance group (the deformation performance of the current state) and the effective abnormal dynamic performance group (the deformation benchmark of the abnormal state) in step S303, step S304 compares these two groups of performances in real time. The basis for comparison is the deformation magnitude of the state performance components. For example, in the real-time dynamic performance group, the power generation module is enlarged to 1.5 times, and the transmission module maintains the standard size. If there is the same deformation combination in the effective abnormal dynamic performance group (such as the 1.5-fold magnification mode corresponding to the high-temperature abnormality), the matching is successful. The abnormal condition corresponding to the matched effective abnormal dynamic performance group (such as "high-temperature abnormality") is determined as the current state. If there is no complete match, it may be judged as normal or a new unrecognized abnormality. The way of expressing the deformation magnitude makes the comparison result intuitively visible. For example, the significantly enlarged module directly points to the abnormal area, and users can identify the problem without complex interpretation. This step combines the goal of step S300 in the overall technical solution. Through the dynamic comparison of the deformation magnitude, the effective abnormal factor data is associated with the real-time data, realizing the accurate judgment of the state, and improving the dynamics and response speed of monitoring by virtue of the visualization characteristics of geometric changes, which is particularly applicable to the complex operation scenarios of new energy power equipment.

[0041] In some embodiments disclosed by the present invention, the method for classifying the state data of power equipment includes: Step S101, performing curve fitting on the control parameters and external influence parameters in the time dimension to obtain a number of control parameter curves and a number of external influence parameter curves.

[0042] Step S102, dynamically generating a number of sliding windows with a window time width, and respectively allowing each sliding window to perform change analysis on the control parameter curves and the external influence parameter curves. If the reference parameter curves and the external influence parameter curves simultaneously show the same changes and have the same periodic parameter segments, then intercept the parameter segments at this time, and record the combination of a number of parameter segments as the classified parameter segment group.

[0043] Step S103, using the classified parameter segment group as the classification condition to classify the corresponding state data.

[0044] In some embodiments disclosed by the present invention, the method for determining a number of initial abnormal factor data segments includes: Step S101, analyzing the state data to determine a number of factor data corresponding to each state data, and recording it as the factor data group.

[0045] Step S102: Randomly combine the normal state data and abnormal state data within the same category to obtain a state data comparison group. The method for comparing the state data comparison group includes: Based on the corresponding relationship between the control parameters and external influence parameters among the state data, align the factor data in the factor data group, determine the factor data difference amount at different time nodes, and based on the continuous manifestation of the factor data difference amount, determine the initial abnormal factor data section on the factor data.

[0046] In some embodiments disclosed by the present invention, the method for determining the initial abnormal factor data section on the factor data based on the continuous manifestation of the factor data difference amount includes: Step S1021: Determine the factor data difference amount at each time node, judge the preset difference amount interval to which each factor data difference amount belongs, and configure and record the set difference influence index corresponding to the preset difference amount interval at the corresponding time node.

[0047] Step S1022: If the difference influence index is greater than or equal to the preset value, mark the corresponding time node; if it is less than the preset value, integrate the difference influence indexes corresponding to the continuous time nodes before the time node, and based on the integration result, correct the difference influence index. If the corrected difference influence index is greater than or equal to the preset value, mark the corresponding time node.

[0048] Step S1023: Recognize the factor data section corresponding to the marked and continuous time nodes as the initial abnormal factor data section.

[0049] Among them, the expression for calculating the difference influence index is: 。

[0050] Among them, is the corrected difference influence index, is the difference influence index before correction, is the correction adjustment coefficient, is the integral value of the difference influence indexes corresponding to the continuous time nodes before the time node, is the integral influence adjustment coefficient, is the integral influence adjustment constant.

[0051] In some embodiments disclosed by the present invention, the method for classifying the initial abnormal factor data section group includes: Step S201: Extract the discrete nodes from the initial abnormal factor data section group to obtain a number of initial abnormal factor data parameters, construct the initial abnormal factor data parameters into an initial abnormal factor data parameter sequence, and serialize the control parameters and external influence parameters corresponding to the initial abnormal factor data parameter sequence to obtain a control parameter sequence and an external influence parameter sequence.

[0052] Step S202: Construct a first similarity operator for the comparison between manipulation parameter sequences, construct a second similarity operator for the comparison between external influence parameter sequences, construct a third similarity operator for the comparison between initial abnormal factor data parameter sequences, determine the similarity degree between initial abnormal factor data section groups based on the first similarity operator, the second similarity operator, and the third similarity operator, and classify the initial abnormal factor data section groups based on the similarity degree.

[0053] In some embodiments disclosed by the present invention, the method for determining the similarity degree between initial abnormal factor data section groups includes: Step S2021: Set a similarity influence weight coefficient for each initial abnormal factor data parameter sequence, compare the factor data difference amount of each initial abnormal factor data parameter between the initial abnormal factor data parameter sequences, and count the total sum of the factor data difference amounts, compare the manipulation parameter difference amount of each manipulation parameter between the manipulation parameter sequences, and count the total sum of the manipulation parameter difference amounts, compare each external influence parameter difference amount between the external influence parameter sequences, and count the total sum of the external influence parameter difference amounts.

[0054] Step S2022: Use the similarity influence weight coefficient, the total sum of the manipulation parameter difference amounts, and the total sum of the external influence parameter difference amounts to correct the total sum of the factor data difference amounts, and obtain the similarity degree between the initial abnormal factor data section groups.

[0055] Among them, the expression for calculating the similarity degree is: .

[0056] Among them, is the similarity degree, is the preset maximum similarity degree, is the similarity degree conversion coefficient, is the total sum of the factor data difference amounts of the i-th initial abnormal factor data parameter sequence, the similarity influence weight coefficient corresponding to the i-th initial abnormal factor data parameter sequence, is the adjustment constant of the total sum of the factor data difference amounts, is the preset first correction coefficient output function, is the total sum of the manipulation parameter difference amounts, is the preset second correction coefficient output function, is the total sum of the external influence parameter difference amounts.

[0057] In some embodiments disclosed by the present invention, the method for validly intercepting the initial abnormal factor data includes: Step S203: Randomly combine the initial abnormal factor data section groups, compare the corresponding initial abnormal factor data sections between the initial abnormal factor data section groups, and determine the data sections with repeated manifestations among them, which are recorded as valid abnormal factor data sections.

[0058] In some embodiments disclosed by the present invention, a new energy power equipment status judgment system based on digital twin is also disclosed, including: A first module, configured to classify the normal state data and abnormal state data of the power equipment based on the equivalence of the control parameters and external influence parameters, compare the normal state data and abnormal state data of the same category, and determine a number of initial abnormal factor data sections to obtain an initial abnormal factor data section group; A second module, configured to classify the initial abnormal factor data section groups based on the proximity among the control parameters, external influence parameters, and the initial abnormal factor data section groups, and perform effective truncation on the initial abnormal factor data based on the repeated characteristics of each initial abnormal factor data in the initial abnormal factor data section groups of the same category to obtain valid abnormal factor data sections, and record the combination of the valid abnormal factor data sections as a valid abnormal factor data section group; A third module, configured to analyze the real-time operation data of the power equipment collected in real time by using the valid abnormal factor data section group to realize the judgment of the status of the power equipment.

[0059] The present invention discloses a method and system for judging the status of new energy power equipment based on digital twin. It discloses classifying the normal and abnormal state data of power equipment based on the equivalence of control parameters and external influence parameters, comparing the data of the same category, extracting the initial abnormal factor data sections, and forming an initial abnormal factor data section group; performing secondary classification based on the proximity of the control parameters, external influence parameters, and the initial abnormal factor data section group, and combining the repeated characteristics to perform effective truncation on the initial abnormal factor data to obtain valid abnormal factor data sections, and combining them into a valid abnormal factor data section group; analyzing the real-time operation data by using the valid abnormal factor data section group, constructing a virtual model through digital twin technology and comparing the dynamic performance to realize status judgment; the present invention solves the problems of abnormal factor extraction and real-time monitoring in a complex environment; improves the accuracy of status judgment, and is applicable to the safety operation monitoring of new energy power equipment.

[0060] Through the description of the above embodiments, those skilled in the art can clearly understand that the present invention can be implemented by hardware or by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), including several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various implementation scenarios of the present invention.

[0061] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that they can still modify or equivalently replace the technical solutions of the present invention, and these modifications or equivalent replacements cannot make the modified technical solutions deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for judging the state of new energy power equipment based on digital twins, characterized in that: include: Step S100, based on the equality of the control parameters and the external influence parameters, the normal state data and the abnormal state data of the electric power equipment are classified, and the normal state data and the abnormal state data of the same category are compared to determine a number of initial abnormal factor data segments to obtain an initial abnormal factor data segment group; Step S200, based on the proximity between the control parameters, the external influence parameters and the initial abnormal factor data segment groups, the initial abnormal factor data segment groups are classified, and based on the repetitive characteristics of each initial abnormal factor data in the initial abnormal factor data segment groups of the same category, the initial abnormal factor data are intercepted for effectiveness to obtain effective abnormal factor data segments, and the combination of effective abnormal factor data segments is recorded as an effective abnormal factor data segment group; Step S300: Analyze the real-time operation data of the electric power equipment collected in real time by using the effective abnormal factor data segment group to determine the state of the electric power equipment.

2. The method for determining the state of new energy power equipment based on digital twin according to claim 1 is characterized in that: The method for analyzing the real-time operation data of the electric power equipment collected in real time by using the effective abnormal factor data segment group includes: Step S301, constructing a virtual three-dimensional model for the power equipment, and constructing virtual three-dimensional modules at corresponding positions of the virtual three-dimensional model based on the positions of different modules relative to the power equipment, and setting a state expression component for each point of the virtual three-dimensional module, the state expression component is used to express the state of the corresponding virtual three-dimensional module; Step S302, determining the abnormal factor data type corresponding to the virtual three-dimensional module, and constructing a factor data variable interval for each abnormal factor data type, each factor data variable interval corresponding to a performance of a state performance component; Step S303, identifying the valid abnormal factor data segment group as input data of the virtual three-dimensional model, determining the valid abnormal dynamic performance of different state performance components, recording a set of several valid abnormal dynamic performances as an effective abnormal dynamic performance group, identifying the real-time operation data as input data of the virtual three-dimensional model, determining the real-time dynamic performance of different state performance components, and recording a combination of several real-time dynamic performances as a real-time dynamic performance group; Step S304 , comparing the dynamic performance group with different valid abnormal dynamic performance groups in real time, and identifying the abnormal conditions corresponding to the matching valid abnormal dynamic performance groups as the real-time status of the electric power equipment.

3. The method for determining the state of new energy power equipment based on digital twin according to claim 1 is characterized in that: Methods for classifying the status data of power equipment include: Step S101, performing curve fitting on the control parameters and external influence parameters in the time dimension to obtain a plurality of control parameter curves and a plurality of external influence parameter curves; Step S102, dynamically generate a plurality of sliding windows with window time widths, and respectively allow each sliding window to perform change analysis on the control parameter curve and the external influence parameter curve. If the reference parameter curve and the external influence parameter curve simultaneously show identical changes and have the same parameter curve segment in the same period, the parameter curve segment at this time is intercepted, and the combination of the plurality of parameter curve segments is recorded as a classified parameter curve segment group; Step S103, classifying the corresponding state data based on the classification parameter curve segment group as a classification condition.

4. The method for determining the state of new energy power equipment based on digital twin according to claim 3 is characterized in that: The method of determining a plurality of initial abnormal factor data segments includes: Step S101, analyzing the state data to determine a number of factor data corresponding to each state data, recorded as a factor data group; Step S102, randomly combining normal state data and abnormal state data in the same category to obtain a state data comparison group. The method for comparing the state data comparison group includes aligning the factor data in the factor data group based on the correspondence between the control parameters and the external influence parameters between the state data, determining the difference in factor data at different time nodes, and determining the initial abnormal factor data segment on the factor data based on the continuous expression of the difference in factor data.

5. The method for determining the state of new energy power equipment based on digital twins according to claim 4 is characterized in that: Based on the continuous expression of the difference amount of the factor data, the method for determining the initial abnormal factor data segment on the factor data includes: Step S1021, determining the factor data difference amount at each time node, and determining the preset difference amount interval to which each factor data difference amount belongs, and recording the set difference impact indicator configuration corresponding to the preset difference amount interval at the corresponding time node; Step S1022: if the difference impact index is greater than or equal to the preset value, the corresponding time node is marked; if it is less than the preset value, the difference impact index corresponding to the consecutive time nodes before the time node is integrated, and based on the integration result, the difference impact index is corrected; if the corrected difference impact index is greater than or equal to the preset value, the corresponding time node is marked; Step S1023, identifying the factor data segments corresponding to the marked and continuous time nodes as initial abnormal factor data segments; Among them, the expression for calculating the difference impact index is: ; in, is the corrected difference impact indicator, is the difference impact index before correction, To correct the adjustment factor, is the integral value of the difference impact index corresponding to the continuous time nodes before the time node, is the integral impact adjustment coefficient, Adjust the constant for the integral effect.

6. The method for determining the state of new energy power equipment based on digital twin according to claim 1 is characterized in that: Methods for classifying the initial anomaly factor data segment groups include: Step S201, extracting dispersed nodes from the initial abnormal factor data segment group to obtain a number of initial abnormal factor data parameters, constructing the initial abnormal factor data parameters into an initial abnormal factor data parameter sequence, and serializing the control parameters and external influence parameters corresponding to the initial abnormal factor data parameter sequence to obtain a control parameter sequence and an external influence parameter sequence; Step S202, constructing a first similarity operator for comparison between control parameter sequences, constructing a second similarity operator for comparison between external influence parameter sequences, constructing a third similarity operator for comparison between initial abnormal factor data parameter sequences, determining the degree of similarity between the initial abnormal factor data segment groups based on the first similarity operator, the second similarity operator and the third similarity operator, and classifying the initial abnormal factor data segment groups based on the degree of similarity.

7. The method for determining the state of new energy power equipment based on digital twins according to claim 6 is characterized in that: Methods for determining the similarity between groups of initial anomaly factor data segments include: Step S2021, setting a similar influence weight coefficient for each initial abnormal factor data parameter sequence, comparing the factor data difference of each initial abnormal factor data parameter between the initial abnormal factor data parameter sequences, and counting the sum of the factor data difference, comparing the control parameter difference of each control parameter between the control parameter sequences, and counting the sum of the control parameter difference, comparing the difference of each external influence parameter between the external influence parameter sequences, and counting the sum of the external influence parameter difference; Step S2022, using the similar influence weight coefficient, the sum of the difference of the control parameters and the sum of the difference of the external influence parameters to correct the sum of the difference of the factor data, and obtain the similarity between the initial abnormal factor data segment groups; The expression for calculating the similarity is: ; in, To a similar degree, To preset the maximum similarity, is the conversion coefficient of similarity, is the sum of the factor data differences of the i-th initial abnormal factor data parameter sequence, The similar influence weight coefficient corresponding to the i-th initial abnormal factor data parameter sequence, is the sum adjustment constant of factor data differences, is the preset first correction coefficient output function, is the sum of the differences in the control parameters, is the preset second correction coefficient output function, It is the sum of the differences of external influencing parameters.

8. The method for determining the state of new energy power equipment based on digital twin according to claim 1 is characterized in that: Methods for validating initial abnormal factor data include: Step S203, randomly combining the initial abnormal factor data segment groups, and comparing the corresponding initial abnormal factor data segments between the initial abnormal factor data segment groups, determining the data segments that are repeated with each other, and recording them as valid abnormal factor data segments.

9. The new energy power equipment status judgment system based on digital twins is characterized by: A method for determining the state of new energy power equipment according to any one of claims 1 to 8, comprising: The first module is used to classify the normal state data and abnormal state data of the power equipment based on the equality of the control parameters and the external influence parameters, and compare the normal state data and abnormal state data of the same category to determine a number of initial abnormal factor data segments to obtain an initial abnormal factor data segment group; The second module is used to classify the initial abnormal factor data segment groups based on the proximity between the control parameters, the external influence parameters and the initial abnormal factor data segment groups, and based on the repeated characteristics of each initial abnormal factor data in the initial abnormal factor data segment groups of the same category, perform validity interception on the initial abnormal factor data to obtain valid abnormal factor data segments, and record the combination of valid abnormal factor data segments as a valid abnormal factor data segment group; The third module is used to analyze the real-time operation data of the power equipment collected in real time by using the effective abnormal factor data segment group to determine the status of the power equipment.

Images