Multi-source data fusion converter station device performance degradation prediction method

Abstract

The invention discloses a multi-source data fusion converter station device performance degradation prediction method, and relates to the technical field of converter station operation monitoring, and the method comprises the steps: S1, constructing a device operation basic representation set; s2, performing feature dimension expansion by using the basic representation set; s3, performing multi-source feature fusion by using the multi-dimensional representation vector; s4, performing degradation state index extraction by using the fusion representation set; and S5, performing performance degradation prediction by using the degradation state vector to obtain prediction output. According to the method, the fusion interface alignment of the multi-dimensional representation vectors and the multi-source feature fusion operator are set, and the representation vectors of different sources are combined according to a unified structure rule to form the fusion representation set, so that the sources of various representations can be distinguished, and the operation state of a converter station device can be integrally described; therefore, standard and quantifiable input is provided for subsequent performance degradation prediction, and the prediction precision and the information utilization efficiency are effectively improved.

Description

Technical Field

[0001] This invention relates to the field of converter station operation monitoring technology, specifically a method for predicting the performance degradation of converter station components through multi-source data fusion. Background Technology

[0002] During the operation of a converter station, device performance gradually degrades with long-term operation and changes in external operating conditions. Existing performance monitoring and degradation prediction methods mainly rely on single-type characteristics or simple statistical indicators, which are insufficient to fully reflect the comprehensive impact of various characteristics of converter station devices in steady-state, dynamic, and internal behavior. Under complex operating conditions, traditional methods are prone to insufficient prediction accuracy and inadequate information utilization, failing to provide a reliable data foundation for subsequent performance evolution analysis and health management.

[0003] In existing technologies, multi-source feature fusion usually adopts simple splicing or numerical aggregation methods, which leads to semantic confusion between features from different sources or damages the original representation category boundaries. It is impossible to maintain the traceability of the source of various representations and the unified analysis input requirements, which affects the accuracy and reliability of degradation prediction. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method for predicting the performance degradation of converter station devices using multi-source data fusion, thereby solving the problems mentioned in the background section.

[0005] To achieve the above objectives, the present invention provides the following technical solution: In a first aspect, embodiments of the present invention provide a method for predicting the performance degradation of converter station devices through multi-source data fusion, comprising the following steps: S1. Construct a set of basic characterization features for device operation; S2. Expand the feature dimensions using the basic representation set to obtain a multidimensional representation vector; S3. Use multi-dimensional representation vectors to fuse multi-source features and obtain a fused representation set; S4. Use the fusion representation set to extract degradation state indicators and obtain the degradation state vector; S5. Use the degraded state vector to predict performance degradation and obtain the prediction output.

[0006] To further optimize this technical solution, step S1 serves to provide a set of basic operational representations that can be uniformly parsed and compared for subsequent steps, so that multi-source data have consistency and correspondence when entering subsequent association extraction, fusion modeling and degradation state identification. Step S1 ultimately outputs the basic representation set. , which includes: steady-state characterization set ; Dynamic representation set ; Internal behavioral representation set .

[0007] To further optimize this technical solution, step S2 involves adjusting the basic representation set. By performing structured and vectorized dimensional expansion processing, the representation that originally existed in the form of a set of parameters is transformed into a multi-dimensional representation vector representation that can be uniformly calculated, uniformly transmitted, and uniformly integrated. Step S2, when expanding the feature dimensions, includes the following steps: Structure mapping of collection-level representation objects; The formation of multidimensional representation vectors; Step S2 ultimately forms a set of multidimensional representation vectors: ; in, Steady-state multidimensional representation vector; Dynamic multidimensional representation direction; : Multidimensional representation vector of internal behavior.

[0008] To further optimize this technical solution, step S3, based on step S2, further organizes multi-dimensional representation vectors with different sources and semantics but consistent structural forms into the same fusion representation system, so that it can reflect the comprehensive operating status of converter station devices as a whole. Step S3, when performing multi-source feature fusion, includes the following process: Alignment of the fusion interface for multidimensional representation vectors; Multi-source feature fusion and fusion representation generation.

[0009] To further optimize this technical solution, in step S3, when aligning the fusion interface of the multidimensional representation vectors, the multidimensional representation vectors are subjected to fusion interface alignment processing. The multidimensional representation vectors after interface alignment are represented as follows: .

[0010] To further optimize this technical solution, in step S3, when performing multi-source feature fusion and fusion representation generation, the following will be implemented: , , As a fusion input, it is uniformly organized through a multi-source feature fusion operator; The fusion operator is represented as The fusion result is expressed as: .

[0011] To further optimize this technical solution, step S3 ultimately outputs a fused representation set: ; in: This represents the corresponding steady-state features in the fusion representation space. This represents the corresponding dynamic features in the fusion representation space; This represents the corresponding internal behavioral characteristics in the fusion representation space.

[0012] To further optimize this technical solution, step S4, based on the formation of the fused characterization set and the unified organization of multi-source characterizations into the same fused structure system in step S3, further addresses the needs of device performance degradation analysis by extracting state indicators that can directly reflect the device health and degradation degree from the fused characterization set, and organizing these indicators into a well-defined and quantifiable degradation state vector. .

[0013] To further optimize this technical solution, step S5 has already generated the degenerate state vector in step S4. Based on this, by analyzing the evolution trend of each degradation state index in the vector over time, the future performance state of the converter station devices is quantitatively predicted. Step S5, in predicting performance degradation, includes the following steps: Model input construction; Degradation performance prediction calculation; The predicted output is structured.

[0014] To further optimize this technical solution, the final output obtained in step S5 is a predicted output vector; ; Each dimension This corresponds to the predicted values ​​or degree of degradation of future device performance indicators.

[0015] In a second aspect, embodiments of the present invention provide a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, they implement the steps of a converter station device performance degradation prediction method based on multi-source data fusion as described in the first aspect of the present invention.

[0016] Thirdly, embodiments of the present invention provide a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, they implement the steps of a converter station device performance degradation prediction method based on multi-source data fusion as described in the first aspect of the present invention.

[0017] Compared with existing technologies, this invention provides a method for predicting the performance degradation of converter station devices through multi-source data fusion, which has the following beneficial effects: This multi-source data fusion method for predicting the performance degradation of converter station devices combines representation vectors from different sources according to a unified structural rule by setting a fusion interface alignment for multi-dimensional representation vectors and a multi-source feature fusion operator. This results in a fused representation set, which maintains the source distinguishability of various representations and can comprehensively depict the operating status of converter station devices. This provides standardized and quantifiable input for subsequent performance degradation prediction, effectively improving prediction accuracy and information utilization efficiency. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating a method for predicting the performance degradation of converter station devices based on multi-source data fusion proposed in this invention. Figure 2 This is a schematic diagram of the feature dimension expansion process of a converter station device performance degradation prediction method proposed in this invention, which involves multi-source data fusion. Figure 3 This is a schematic diagram of the multi-source feature fusion process of a converter station device performance degradation prediction method proposed in this invention; Figure 4 This is a schematic diagram of the performance degradation prediction process of a converter station device performance degradation prediction method based on multi-source data fusion proposed in this invention. Detailed Implementation

[0020] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0021] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0022] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places throughout this specification does not necessarily refer to the same embodiment, nor is it a single embodiment or an embodiment selectively excluded from other embodiments.

[0023] Example 1: Reference Figures 1-4 This is the first embodiment of the present invention, which provides a method for predicting the performance degradation of converter station devices based on multi-source data fusion, including the following steps: S1. Construct a set of basic characterization features for device operation; The purpose of step S1 is to provide a set of basic operational representations that can be uniformly parsed and compared for subsequent steps, so that multi-source data have consistency and correspondence when entering subsequent association extraction, fusion modeling and degradation state identification.

[0024] Step S1, in the process of constructing the basic representation set, includes the following steps: Data parsing and channel normalization calibration: Mature data parsing technology is used to format the raw records from electrical quantity monitoring, temperature measurement points, and mechanical vibration sensing channels. Mature equipment channel calibration methods are used to unify the dimensions and synchronize the sampling of various raw quantities, so that data from different sources can participate in subsequent analysis under the same time reference.

[0025] Formation of steady-state characteristic set and dynamic characteristic set: Using mature steady-state quantity extraction technology, average value indicators, effective value indicators, and phase-type characterization quantities are extracted from the formatted electrical quantity, heat, and vibration data within the operating cycle to obtain steady-state characterization content.

[0026] Using mature dynamic response identification technology, dynamic behavior identification is performed on the records of the rising segment, decay segment and short-term disturbance interval in the running curve to obtain response speed, short-term offset and fluctuation amplitude characterization quantities.

[0027] Formation of internal behavioral representation groups: By utilizing mature sequence behavior analysis techniques, the time series of steady-state and dynamic feature groups are structurally described, forming sequential change representations, offset behavior representations, and correlation representations, thus fully constituting the basic operational behavior expression content.

[0028] Step S1 ultimately outputs the basic representation set. , which includes: steady-state characterization set It consists of electrical steady-state quantities, thermal steady-state quantities, and vibration steady-state quantities determined by steady-state quantity extraction technology; Dynamic representation set It consists of the rising segment response, the decay segment response, the disturbance offset, and the short-time fluctuation obtained by dynamic response identification technology; Internal behavioral representation set It consists of the sequence change quantity, offset behavior quantity and internal correlation quantity identified by sequence behavior analysis technology.

[0029] S2. Expand the feature dimensions using the basic representation set to obtain a multidimensional representation vector; Step S2 on the basic representation set By performing structured and vectorized dimensional expansion processing, the representation that originally existed in the form of a set of parameters is transformed into a multi-dimensional representation vector representation that can be uniformly calculated, uniformly transmitted, and uniformly integrated.

[0030] Step S2, when expanding the feature dimensions, includes the following steps: Structure mapping of set-level representation objects: Using the three basic representation sets output in step S1 as input objects, the following processing logic is executed for each type of basic representation set: The representation results that originally existed as "unordered sets" are mapped to vectorized representation objects with fixed-dimensional semantics. This mapping process does not depend on the specific composition of the parameters inside the set, but treats the entire set as a holistic representation unit.

[0031] Formation of multidimensional representation vectors: After completing the set-level structure mapping, corresponding multidimensional representation vector objects are generated for the three types of basic representation sets: From the set of steady-state characteristics The resulting steady-state multidimensional representation vector is denoted as... ; From dynamic representation set The resulting dynamic multidimensional representation vector is denoted as... ; From the set of internal behavioral representations The transformed multidimensional representation vector of internal behavior is denoted as .

[0032] Step S2 ultimately forms a set of multidimensional representation vectors: .

[0033] In existing technologies, the basic parameter set is often used directly as input, lacking a unified constraint on its data structure. In contrast, this step introduces an explicit vectorized structure through set-level representation transformation without expanding the internal content of the set, enabling the basic representation results to have higher consistency and transitivity at the data organization level. This is the main difference between the two in terms of technical implementation path.

[0034] S3. Use multi-dimensional representation vectors to fuse multi-source features and obtain a fused representation set; Step S3, based on step S2, further organizes multidimensional representation vectors with different sources and semantics but consistent structural forms into the same fusion representation system, so that it can reflect the overall operating status of converter station devices as a whole.

[0035] Step S3, when performing multi-source feature fusion, includes the following process: Alignment of the fusion interface for multidimensional representation vectors: The multidimensional representation vector output in step S2 middle, , , While all are multi-dimensional vectors in form, they differ in their source categories and representational semantics. The core feature of this multi-dimensional representation vector fusion interface alignment process is as follows: Do not disassemble or reconstruct the internal dimensions of the vector; No new derived features are introduced; Only at the structural level are a unified fusion identifier and organization index added to each type of multidimensional representation vector; The interface-aligned multidimensional representation vectors are represented as follows: ; The above results still correspond one-to-one with the output of step S2, and only meet the fusion input requirements at the data organization level.

[0036] Multi-source feature fusion and fusion representation generation: After completing the interface alignment, , , As the fusion input, it is uniformly organized through a multi-source feature fusion operator; the fusion operator is represented as follows: This operator is used to combine representation vectors from different sources according to a predetermined structural rule, and the fusion result is expressed as: ; The fusion operator is determined based on existing mature multi-source feature structured fusion technology. Its determination relies on the requirement that features from different sources possess a consistent data organization form before entering the unified analysis process. In this method, the fusion operator adopts a structure-level fusion approach, uniformly encapsulating multi-dimensional representation vectors from steady-state, dynamic, and internal behaviors as independent representation units without transforming their internal dimensions and values. Specifically, the fusion rules are as follows: multi-dimensional representation vectors are organized in parallel according to a pre-defined representation category order, and a clear category identifier and index relationship are retained for each type of representation in the fusion structure. This ensures that the fused representation set simultaneously contains multi-source feature information within the same representation space, and that the features from different sources maintain a distinguishable and traceable structural relationship, thereby meeting the requirement for a unified input format in subsequent device performance degradation analysis.

[0037] Step S3 ultimately outputs a fused representation set: ; in: This represents the corresponding steady-state features in the fusion representation space. This represents the corresponding dynamic features in the fusion representation space; This represents the corresponding internal behavioral characteristics in the fusion representation space.

[0038] In existing mature technologies, multi-source feature fusion usually adopts simple splicing or numerical aggregation methods, which can easily lead to semantic confusion between features from different sources or destroy the original representation category boundaries.

[0039] In contrast, step S3 achieves the fusion of multi-source features by first unifying the representation morphology and then performing structural-level fusion organization, without introducing cross-physical quantity comparisons or changing the original representation semantics. This ensures that the fusion result maintains the traceability of the source and meets the input requirements for subsequent unified analysis.

[0040] S4. Use the fusion representation set to extract degradation state indicators and obtain the degradation state vector; Step S4, based on the fusion characterization set already formed in Step S3 and the multi-source characterizations being organized into the same fusion structure system, further addresses the needs of device performance degradation analysis by extracting state indicators that can directly reflect the device health and degradation level from the fusion characterization set, and organizing these indicators into a well-defined and quantifiable degradation state vector.

[0041] Step S4, when extracting degradation state indicators, includes the following steps: Determination and selection of degradation status indicators: Using the fusion representation set output in step S3 as input, the fusion sub-representations corresponding to steady state, dynamic state, and internal behavior are analyzed according to the source category of each sub-representation in the fusion representation; By using mature equipment condition assessment and health monitoring technologies, the existing characterization dimensions in each fusion sub-characterization are judged, and the characterization that is sensitive to changes in device performance and has stable evolution characteristics over time is selected as the degradation state index.

[0042] Numerical normalization of degradation status indicators: Since the selected degradation state indices may come from different types of fusion component representations, their numerical ranges and dimensions may differ. To ensure the comparability and stability of subsequent analyses, the degradation state indices are numerically normalized. This process employs existing mature data preprocessing techniques to standardize the scale of indicators within the same source category, ensuring that the numerical variations of each indicator remain within a controllable range.

[0043] Construction of the degenerate state vector: After selecting and normalizing the degradation state indicators, the processed degradation state indicators are arranged in a pre-defined order to construct a degradation state vector. During the construction process, the source category and index order information of each dimension of degradation state index are retained, so that the degradation state vector can clearly reflect the composition of different categories of degradation information in terms of structure, thereby providing a stable and unified input object for subsequent degradation evolution analysis.

[0044] Step S4 outputs a degradation state vector, which consists of multiple normalized degradation state indices. This vector comprehensively characterizes the current performance degradation state of the converter station devices within a single vector structure. Its expression is: ; in: This represents the set of degradation state indices derived from steady-state characterization, expressed as: ; This represents the set of degradation state indices derived from dynamic representation, expressed as: ; This represents the set of degradation state indicators derived from internal behavioral representations, expressed as: ; , , Each sub-item in All are degradation state indicators that have been normalized in step S4; vector The structure retains the index category and order information. This degradation state vector serves as the direct input to step S5, providing a basic state description for subsequent device performance degradation trend analysis and prediction.

[0045] S5. Use the degraded state vector to predict performance degradation and obtain the prediction output; Step S5: The degenerate state vector has been generated in step S4. Based on this, by analyzing the evolution trend of each degradation state index in the vector over time, the future performance state of the converter station devices is quantitatively predicted.

[0046] Step S5, in predicting performance degradation, includes the following steps: Model input construction: Degenerate state vector Each sub-vector in ( , , Arrange them in a uniform order to construct standardized prediction input vectors. : ; Preserve the source category and order information for each indicator to ensure that the input vector structure is consistent with the S4 output; add time index or historical sequence information for model prediction.

[0047] Degradation performance prediction calculation: Predictive models trained using mature degradation prediction techniques (such as regression analysis, time series forecasting, or machine learning models) Predict the input vector: ; in: For the prediction period or prediction step size; This indicates the predicted value or degradation level of various performance indicators in the future.

[0048] The model comprehensively considers the effects of steady-state, dynamic, and internal behavior indicators during the calculation process to ensure that the predicted output can reflect the overall effect of multi-source features on performance degradation.

[0049] Predicted output structuring: Prediction results The data is organized into a structured vector, preserving indicator category and order information, and compared with the input vector. correspond; Ensure that each predicted value corresponds one-to-one with the degradation status index in S4, so that it can be directly used for subsequent performance evaluation or operation and maintenance analysis.

[0050] The final output obtained in step S5 is the predicted output vector; ; Each dimension The predicted values ​​or degree of degradation of the corresponding future device performance indicators; The output vector has a clear structure, retaining the category and order information of the input degenerate state vector, and can be directly used for subsequent analysis and decision-making.

[0051] Compared with existing mature technologies, this method uses multi-source, structured degradation state vectors as input in performance degradation prediction, retaining the source and sequence information of steady-state, dynamic and internal behavior indicators. This enables the model to predict degradation trends by integrating multiple features, rather than relying on empirical curves or statistical regressions of a single indicator. At the same time, it ensures that the input-output structure is strictly aligned with the preceding steps, thereby achieving logical closure, feasibility and consistency in prediction across physical quantities.

[0052] Example 2: This embodiment provides a practical application scenario for a multi-source data fusion method for predicting the performance degradation of converter station devices: In the actual operating environment of a converter station, the converter and control unit are subjected to various operating conditions such as current, voltage, and temperature over a long period of time, posing a potential risk of performance degradation. To achieve online prediction of the performance of key components, the following steps can be implemented: Construct a basic characterization set: Obtain the steady-state operating parameters (such as average current, voltage, and temperature rise), dynamic response parameters (such as overload response time and current fluctuation amplitude), and internal behavioral indicators (such as control sequence consistency and action deviation) of the device through field sensors and monitoring systems to form a structured basic characterization set.

[0053] Feature Dimension Expansion: The basic representation set is expanded in dimension, and the representations of steady state, dynamic and internal behavior are organized into multi-dimensional representation vectors. The source information of various representations is preserved and the unified structured processing is carried out to provide standardized input for subsequent fusion analysis.

[0054] Multi-source feature fusion: Using a multi-source feature fusion operator, representation vectors from different sources are combined into a fused representation set according to a unified structural rule, ensuring that the sources of various representations are traceable and forming a comprehensive representation that fully characterizes the operating status of the device.

[0055] Degradation state index extraction: Indicators highly correlated with device performance degradation are selected from the fused characterization set, normalized, and formed into a degradation state vector, which facilitates subsequent performance prediction.

[0056] Performance degradation prediction: Input the degradation state vector into the prediction model to quantitatively predict the future performance state of the device, obtain the degradation level of various performance indicators, and generate a health assessment report to provide a basis for operation and maintenance decisions.

[0057] Example 3: This embodiment also provides a computer device applicable to a converter station device performance degradation prediction method based on multi-source data fusion, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the converter station device performance degradation prediction method based on multi-source data fusion as proposed in the above embodiment.

[0058] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements a multi-source data fusion converter station device performance degradation prediction method as proposed in the above embodiment.

[0059] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0060] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0061] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0062] More specific examples (a non-exhaustive list) of computer-readable media include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which programs can be printed, because programs can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0063] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0064] 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 it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for predicting the performance degradation of converter station devices through multi-source data fusion, characterized in that, Includes the following steps: S1. Construct a set of basic characterizations for device operation; S2. Expand the feature dimensions using the basic representation set to obtain a multidimensional representation vector; S3. Use multi-dimensional representation vectors to fuse multi-source features and obtain a fused representation set; S4. Use the fusion representation set to extract degradation state indicators and obtain the degradation state vector; S5. Use the degraded state vector to predict performance degradation and obtain the prediction output.

2. The method for predicting the performance degradation of converter station devices based on multi-source data fusion according to claim 1, characterized in that, Step S1 provides a set of basic operational representations that can be uniformly parsed and compared for subsequent steps, so that multi-source data have consistency and correspondence when entering subsequent association extraction, fusion modeling and degradation state identification. Step S1 ultimately outputs the basic representation set. This includes: steady-state characterization set ; Dynamic representation set ; Internal behavioral representation set .

3. The method for predicting the performance degradation of converter station devices based on multi-source data fusion according to claim 1, characterized in that, Step S2 involves the basic representation set. By performing structured and vectorized dimensional expansion processing, the representation that originally existed in the form of a set of parameters is transformed into a multi-dimensional representation vector representation that can be uniformly calculated, uniformly transmitted, and uniformly integrated. Step S2, when expanding the feature dimensions, includes the following steps: Structure mapping of collection-level representation objects; The formation of multidimensional representation vectors; Step S2 ultimately forms a set of multidimensional representation vectors: ; in, Steady-state multidimensional representation vector; Dynamic multidimensional representation direction; : Multidimensional representation vector of internal behavior.

4. The method for predicting the performance degradation of converter station devices based on multi-source data fusion according to claim 1, characterized in that, Based on step S2, step S3 further organizes multidimensional representation vectors with different sources and semantics but consistent structural forms into the same fusion representation system, so that it can reflect the comprehensive operating status of converter station devices as a whole. Step S3, when performing multi-source feature fusion, includes the following process: Alignment of the fusion interface for multidimensional representation vectors; Multi-source feature fusion and fusion representation generation.

5. The method for predicting the performance degradation of converter station devices based on multi-source data fusion according to claim 4, characterized in that, In step S3, when performing interface alignment for the fusion of multidimensional representation vectors, the multidimensional representation vectors are subjected to interface alignment processing. The interface-aligned multidimensional representation vectors are represented as follows: 。 6. The method for predicting the performance degradation of converter station devices based on multi-source data fusion according to claim 4, characterized in that, In step S3, during the multi-source feature fusion and fusion representation generation, , , As a fusion input, it is uniformly organized through a multi-source feature fusion operator; The fusion operator is represented as The fusion result is expressed as: 。 7. The method for predicting the performance degradation of converter station devices based on multi-source data fusion according to claim 4, characterized in that, The final output of step S3 is a fused representation set: ； in: This represents the corresponding steady-state features in the fusion representation space. This represents the corresponding dynamic features in the fusion representation space; This represents the corresponding internal behavioral characteristics in the fusion representation space.

8. The method for predicting the performance degradation of converter station devices based on multi-source data fusion according to claim 1, characterized in that, Step S4, based on the fusion characterization set already formed in Step S3 and the multi-source characterizations being uniformly organized into the same fusion structure system, extracts state indicators that directly reflect the health and degradation degree of the device from the fusion characterization set to meet the needs of device performance degradation analysis. These indicators are then organized into well-defined and quantifiable degradation state vectors. .

9. The method for predicting the performance degradation of converter station devices based on multi-source data fusion according to claim 1, characterized in that, In step S5, the degenerate state vector has already been generated in step S4. Based on this, by analyzing the evolution trend of each degradation state index in the vector over time, the future performance state of the converter station devices is quantitatively predicted. Step S5, in predicting performance degradation, includes the following steps: Model input construction; Degradation performance prediction calculation; The predicted output is structured.

10. The method for predicting the performance degradation of converter station devices based on multi-source data fusion according to claim 1, characterized in that, The final output obtained in step S5 is the predicted output vector. ； Each dimension This corresponds to the predicted values ​​or degree of degradation of future device performance indicators.

Images