Power parameter abnormality early warning monitoring device and method based on embedded system

Abstract

The application relates to the technical field of power detection, and discloses a power parameter abnormality early warning monitoring device and method based on an embedded system. The device comprises a data acquisition and analysis module, a feature fusion analysis module, an abnormality diagnosis and classification module, a baseline self-adaptive correction module and a hierarchical response execution module. The device acquires original power data of various power monitoring devices connected to the embedded system, performs multi-protocol self-adaptive analysis on the original power data, and obtains a standardized power parameter sequence. The device performs multi-dimensional feature fusion analysis on the standardized power parameter sequence, and obtains a multi-dimensional electrical feature vector. The multi-dimensional electrical feature vector is mapped to a preset hierarchical response strategy, and an abnormality type judgment result is generated. The reference baseline is periodically self-adaptively corrected, and an updated reference baseline is obtained. Based on the abnormality type judgment result and the updated reference baseline, hierarchical response operations are performed on the embedded system. The application can improve the accuracy of power parameter abnormality early warning monitoring.

Description

Technical Field

[0001] This invention relates to the field of power detection technology, and in particular to a power parameter anomaly early warning monitoring device and method based on an embedded system. Background Technology

[0002] Existing technologies have significant shortcomings in the data acquisition and parsing stages of power parameter anomaly early warning monitoring in embedded systems. They fail to perform multi-protocol adaptive parsing of raw power data from various power monitoring devices, supporting only a single protocol or a simple parsing method. This results in ineffective compatibility between raw data of different protocol types, making it difficult to form standardized power parameter sequences. Furthermore, they lack rigorous integrity verification and normalization of initial power parameters, directly using only the parsed data, leading to missing data, biases, and insufficient reliability of the foundational data provided for subsequent feature analysis and anomaly detection.

[0003] Existing technologies have significant shortcomings in the feature analysis and early warning response stages of power parameter anomaly early warning monitoring. They fail to perform multi-dimensional fusion analysis of time-domain and frequency-domain features on standardized power parameter sequences, relying solely on single-dimensional features for anomaly judgment. This makes it difficult to comprehensively characterize the operating characteristics of power parameters, leading to one-sided anomaly identification. Furthermore, they lack periodic adaptive correction of the reference baseline, using only a fixed baseline as the judgment standard, which cannot adapt to the dynamic changes in the operating status of embedded systems, causing anomaly judgment thresholds to be out of touch with reality. Finally, they fail to implement tiered responses based on anomaly type and real-time baseline, employing only a uniform early warning or handling method, making it difficult to accurately respond to different anomaly scenarios. This results in low accuracy and practicality of early warning monitoring, failing to meet the monitoring requirements for stable power system operation. Summary of the Invention

[0004] To achieve the above objectives, the present invention provides a power parameter anomaly early warning and monitoring device based on an embedded system, characterized in that the system includes a data acquisition and parsing module, a feature fusion and analysis module, an anomaly diagnosis and classification module, a baseline adaptive correction module, and a hierarchical response execution module, wherein: The data acquisition and parsing module is used to acquire raw power data from various power monitoring devices connected to the embedded system, perform multi-protocol adaptive parsing on the raw power data, and obtain a standardized power parameter sequence of the raw power data. The feature fusion analysis module is used to perform multi-dimensional feature fusion analysis on the standardized power parameter sequence to obtain a multi-dimensional electrical feature vector of the standardized power parameter sequence. The anomaly diagnosis and classification module is used to map the multidimensional electrical feature vector to a preset hierarchical response strategy based on the reference baseline of the embedded system, and generate an anomaly type judgment result for the current power operation status. The baseline adaptive correction module is used to periodically adaptively correct the reference baseline based on the historical operating data of the embedded system to obtain the updated reference baseline of the embedded system. The hierarchical response execution module is used to perform hierarchical response operations on the embedded system based on the anomaly type judgment result and the updated reference baseline.

[0005] In a preferred embodiment, when the data acquisition and parsing module acquires raw power data from various power monitoring devices connected to the embedded system, performs multi-protocol adaptive parsing on the raw power data, and obtains a standardized power parameter sequence from the raw power data, it is specifically used for: Acquire raw power data from the embedded system; The protocol feature identification of the raw power data is performed to obtain the protocol type identification result of the embedded system. The protocol parsing rules corresponding to the protocol type identification result are invoked to parse the original power data and obtain the initial power parameters of the original power data; The initial power parameters are verified for integrity, and the corrected power data is normalized to obtain a standardized power parameter sequence of the original power data.

[0006] In a preferred embodiment, when the feature fusion analysis module performs multidimensional feature fusion analysis on the standardized power parameter sequence to obtain a multidimensional electrical feature vector of the standardized power parameter sequence, it is specifically used for: Extract the time-domain and frequency-domain electrical features from the standardized power parameter sequence; A feature correlation analysis is performed on the time-domain electrical features and the frequency-domain electrical features to obtain the feature correlation matrix of the standardized power parameter sequence; Based on the feature correlation matrix, the time-domain electrical features and the frequency-domain electrical features are fused to obtain the multidimensional electrical feature vector of the standardized power parameter sequence.

[0007] In a preferred embodiment, when the feature fusion analysis module performs feature correlation feature fusion based on the feature correlation matrix to obtain a multidimensional electrical feature vector of the standardized power parameter sequence, it is specifically used for: Based on the preset fusion weights and the feature correlation matrix, the time-domain electrical features and the frequency-domain electrical features are weighted and fused to obtain the initial fused feature vector of the standardized power parameter sequence. The weighted fusion calculation formula is as follows: In the formula, This is the initial fused feature vector. The preset time-domain fusion weighting coefficients, The preset frequency domain fusion weighting coefficients, These are the dominant correlation factors between the time-domain and frequency-domain features extracted from the feature correlation matrix. The time-domain feature vector formed by the aforementioned time-domain electrical characteristics. The frequency domain feature vector is formed by the frequency domain electrical features; The initial fused feature vector is subjected to dimensionality reduction and smoothing processing to obtain the multidimensional electrical feature vector of the standardized power parameter sequence.

[0008] In a preferred embodiment, when the anomaly diagnosis and classification module performs the mapping of the multi-dimensional electrical feature vector to a preset hierarchical response strategy based on the reference baseline of the embedded system to generate an anomaly type judgment result for the current power operation state, it is specifically used for: Based on the reference baseline of the embedded system, a baseline comparison analysis is performed on the multidimensional electrical feature vector; Based on the feature difference analysis results and the preset hierarchical response strategy, the multidimensional electrical feature vectors are mapped to the corresponding response levels to obtain the feature mapping relationship of the multidimensional electrical feature vectors; Based on the feature mapping relationship, the anomaly type and response level corresponding to the multidimensional electrical feature vector are determined to obtain the anomaly type judgment result of the current power operation state.

[0009] In a preferred embodiment, when the baseline adaptive correction module performs periodic adaptive correction of the reference baseline based on the historical operating data of the embedded system to obtain the updated reference baseline of the embedded system, it is specifically used for: Based on the historical operating data of the embedded system, extract the historical electrical feature set of the embedded system; An adaptive assessment of the reference baseline is performed based on the aforementioned historical electrical feature set. Based on the results of the adaptation assessment, the reference baseline is periodically revised to obtain the updated reference baseline.

[0010] In a preferred embodiment, when the baseline adaptive correction module performs periodic correction of the reference baseline based on the adaptive evaluation results to obtain the updated reference baseline, it is specifically used for: Based on the adaptive assessment results, determine the correction status of the reference baseline; When the reference baseline is in a state that needs correction, the correction direction and correction magnitude of the reference baseline are determined based on the historical electrical feature set. Based on the correction direction and the correction magnitude, the reference baseline is iteratively adjusted to obtain the corrected baseline data of the reference baseline; The stationarity of the corrected baseline data is verified to obtain the updated reference baseline.

[0011] In a preferred embodiment, when the hierarchical response execution module performs hierarchical response operations on the embedded system based on the exception type judgment result and the updated reference baseline, it is specifically used for: The impact dimension of the anomaly type judgment result is analyzed to generate the anomaly impact feature vector of the embedded system; The abnormal impact feature vector is combined with the updated reference baseline to perform a comprehensive situation assessment, generating the response situation assessment result of the embedded system. Based on the response situation assessment results, an adaptive response decision instruction set for the embedded system is generated; Based on the adaptive response decision instruction set and the real-time operating parameters of the embedded system, a response action sequence is performed on the embedded system to obtain a response execution status report of the embedded system.

[0012] In a preferred embodiment, the adaptive response decision instruction set includes: When the anomaly type determination result is an overvoltage anomaly and the deviation level is at the warning level, the voltage limit adjustment instruction and the warning log recording instruction of the embedded system are generated. When the abnormality type judgment result is undervoltage abnormality and the deviation degree is at the general abnormality level, the voltage compensation adjustment instruction and abnormal status reporting instruction of the embedded system are generated. When the anomaly type determination result is an overcurrent anomaly and the deviation level is at the general anomaly level, the embedded system generates a load grading cutoff instruction and a protective alarm instruction. When the anomaly type judgment result is that the harmonic content is abnormal and the deviation is at the warning level, the harmonic suppression reference instruction and trend analysis record instruction of the embedded system are generated. When the anomaly type determination result is a frequency deviation anomaly and the deviation degree is at the emergency anomaly level, the embedded system generates an emergency frequency adjustment instruction and a system protection linkage instruction.

[0013] To address the above problems, this invention also provides a power parameter anomaly early warning and monitoring method based on an embedded system, the method comprising: S1. Collect raw power data from various power monitoring devices connected to the embedded system, perform multi-protocol adaptive parsing on the raw power data, and obtain a standardized power parameter sequence of the raw power data; S2. Perform multi-dimensional feature fusion analysis on the standardized power parameter sequence to obtain the multi-dimensional electrical feature vector of the standardized power parameter sequence; S3. Based on the reference baseline of the embedded system, the multidimensional electrical feature vector is mapped to a preset hierarchical response strategy to generate an anomaly type judgment result of the current power operation state; S4. Based on the historical operating data of the embedded system, the reference baseline is periodically and adaptively corrected to obtain the updated reference baseline of the embedded system; S5. Based on the anomaly type judgment result and the updated reference baseline, perform a graded response operation on the embedded system.

[0014] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention lays a precise data foundation for power parameter anomaly early warning monitoring through multi-protocol adaptation and standardized processing. It collects raw power data from various power monitoring devices, extracts initial power parameters through protocol feature identification and adaptive parsing, and generates standardized power parameter sequences through integrity verification and normalization. The time-domain and frequency-domain electrical features of the standardized sequences are fused, combined with correlation matrix weighted fusion and dimensionality reduction smoothing, to obtain a multi-dimensional electrical feature vector, comprehensively characterizing the operating characteristics of power parameters and improving the comprehensiveness of anomaly identification.

[0015] 2. This invention significantly improves the accuracy and practicality of power parameter anomaly early warning monitoring by leveraging dynamic baselines and tiered responses. Based on historical operating data, the reference baseline is periodically revised to generate an updated baseline adapted to dynamic system changes. Multi-dimensional electrical feature vectors are mapped to a tiered response strategy to accurately determine the anomaly type and level. Combined with the updated baseline, a comprehensive situation assessment is performed to generate a targeted adaptive response instruction set, enabling differentiated and precise response operations and ensuring stable power system operation. Attached Figure Description

[0016] Figure 1 This is a system architecture diagram of a power parameter anomaly early warning and monitoring device based on an embedded system provided in an embodiment of the present invention; Figure 2 This is a flowchart illustrating a power parameter anomaly early warning and monitoring method based on an embedded system, as provided in an embodiment of the present invention.

[0017] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments belong to some, but not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “said” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms, and “multiple” generally includes at least two unless the context clearly indicates otherwise.

[0020] Depending on the context, the word "if" or "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0021] Furthermore, the timing of the steps in the following method embodiments is merely an example and not a strict limitation.

[0022] In practice, the server-side equipment deployed by the embedded system-based power parameter anomaly early warning and monitoring device may consist of one or more devices. This embedded system-based power parameter anomaly early warning and monitoring device can be implemented as a business instance, a virtual machine, or a hardware device. For example, it can be implemented as a business instance deployed on one or more devices in a cloud node. Simply put, it can be understood as software deployed on a cloud node, providing power parameter anomaly early warning and monitoring services to various user terminals. Alternatively, it can be implemented as a virtual machine deployed on one or more devices in a cloud node, with application software installed to manage various user terminals. Or, it can also be implemented as a server composed of numerous identical or different types of hardware devices, with one or more hardware devices configured to provide power parameter anomaly early warning and monitoring services to various user terminals.

[0023] In terms of implementation, the power parameter anomaly early warning and monitoring device based on the embedded system and the user terminal are mutually compatible. That is, if the power parameter anomaly early warning and monitoring device based on the embedded system is implemented as an application installed on a cloud service platform, then the user terminal is implemented as a client that establishes a communication connection with the application; or if the power parameter anomaly early warning and monitoring device based on the embedded system is implemented as a website, then the user terminal is implemented as a webpage; or if the power parameter anomaly early warning and monitoring device based on the embedded system is implemented as a cloud service platform, then the user terminal is implemented as a mini-program in an instant messaging application.

[0024] like Figure 1 The figure shown is a system architecture diagram of a power parameter anomaly early warning and monitoring device based on an embedded system provided in an embodiment of the present invention.

[0025] The power parameter anomaly early warning and monitoring device 100 based on an embedded system described in this invention can be installed in a cloud server. In terms of implementation, it can be used as one or more service devices, or as an application installed in the cloud (e.g., a mobile service operator's server, server cluster, etc.), or it can be developed into a website. Depending on the functions implemented, the power parameter anomaly early warning and monitoring device 100 based on an embedded system may include a data acquisition and parsing module 101, a feature fusion and analysis module 102, an anomaly diagnosis and classification module 103, a baseline adaptive correction module 104, and a hierarchical response execution module 105. The module described in this invention can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, stored in the memory of the electronic device.

[0026] In this embodiment of the invention, in the power parameter anomaly early warning and monitoring device based on an embedded system, each of the above-mentioned modules can be implemented independently and can call other modules. Here, "calling" can be understood as one module connecting to multiple modules of another type and providing corresponding services to those connected modules. In the power parameter anomaly early warning and monitoring device based on an embedded system provided by this embodiment of the invention, the applicability of the device architecture can be adjusted by adding modules and directly calling them without modifying the program code, achieving cluster-based horizontal expansion to quickly and flexibly expand the power parameter anomaly early warning and monitoring device based on the embedded system. In practical applications, the above modules can be set in the same device or different devices, or they can be set in a virtual device, such as a service instance in a cloud server.

[0027] The following describes, with reference to specific embodiments, each component and its specific workflow of the power parameter anomaly early warning and monitoring device based on an embedded system: The data acquisition and parsing module 101 is used to acquire raw power data from various power monitoring devices connected to the embedded system, perform multi-protocol adaptive parsing on the raw power data, and obtain a standardized power parameter sequence of the raw power data. In this embodiment of the invention, when the data acquisition and parsing module acquires raw power data from various power monitoring devices connected to the embedded system, performs multi-protocol adaptive parsing on the raw power data, and obtains a standardized power parameter sequence from the raw power data, it is specifically used for: Acquire raw power data from the embedded system; The protocol feature identification of the raw power data is performed to obtain the protocol type identification result of the embedded system. The protocol parsing rules corresponding to the protocol type identification result are invoked to parse the original power data and obtain the initial power parameters of the original power data; The initial power parameters are verified for integrity, and the corrected power data is normalized to obtain a standardized power parameter sequence of the original power data.

[0028] A stable data transmission link is established between the data acquisition and parsing module and the embedded system. This link is compatible with the communication interface specifications of the embedded system and receives raw power data transmitted from various power monitoring devices connected to the embedded system in real time. This data covers raw information generated during power operation, such as instantaneous voltage and current values, power data, and cumulative energy consumption values, ensuring that the collected raw power data fully covers various power monitoring dimensions.

[0029] Protocol features are extracted from the collected raw power data. Key features that can distinguish different communication protocols, such as data frame structure, start identifier, field separator, and check bit format, are identified. The extracted features are compared one by one with the preset feature library of various power communication protocols. Based on the feature matching results, the specific protocol type used by the embedded system to transmit raw power data is determined, and the protocol type identification result is obtained.

[0030] Based on the protocol type identification result, the corresponding protocol parsing rules are retrieved from the protocol parsing rule library of the data acquisition and parsing module. These rules contain detailed parsing specifications such as the definition of data fields, data length, data conversion method, and field correspondence under the protocol. The original power data is disassembled frame by frame according to the parsing rules, and the power-related information corresponding to each field is extracted. The disassembled information is then transformed into power data with clear physical meaning to obtain the initial power parameters of the original power data.

[0031] Based on the integrity criteria for power parameters, the initial power parameters are comprehensively verified to check for issues such as missing data, abnormal fields, and logical conflicts. For incomplete or abnormal data found during the verification, they are supplemented or corrected according to the preset power data correction rules to obtain corrected power data. Subsequently, the corrected power data is normalized using a unified data format and unit standard to eliminate data format differences caused by different power monitoring equipment and different transmission protocols, and finally obtains a standardized power parameter sequence of the original power data.

[0032] The beneficial effects are that it comprehensively collects raw power data from various power monitoring devices connected to the embedded system, covering multiple types of raw power operation information such as instantaneous voltage and current values, power data, and cumulative energy consumption values, ensuring that no data is missed and providing comprehensive basic data support for subsequent analysis.

[0033] The system identifies protocol features in raw power data, accurately matches corresponding protocol parsing rules, and enables adaptive parsing of multiple protocols. This effectively ensures compatibility with the communication protocols of different monitoring devices, overcomes the limitations of single-protocol parsing, and guarantees that all types of raw data can be effectively parsed.

[0034] The original power data is disassembled frame by frame according to the corresponding protocol parsing rules to extract initial power parameters with clear physical meaning, ensuring that the parsing process is standardized and accurate, and avoiding subsequent analysis errors caused by data parsing deviations.

[0035] The initial power parameters are verified for integrity to promptly identify and correct issues such as missing data, abnormal fields, and logical conflicts, ensuring the accuracy and integrity of power data. Then, normalization processing is used to eliminate data format differences caused by different devices and protocols, forming a standardized power parameter sequence with a unified standard.

[0036] The entire process realizes end-to-end processing from multi-device data acquisition and multi-protocol adaptive parsing to data verification and normalization, ensuring that the standardized power parameter sequence is comprehensive, accurate and standardized, laying a solid data foundation for subsequent multi-dimensional feature fusion analysis, anomaly diagnosis and classification and other links, and improving the overall reliability of power parameter anomaly early warning monitoring.

[0037] The feature fusion analysis module 102 is used to perform multi-dimensional feature fusion analysis on the standardized power parameter sequence to obtain the multi-dimensional electrical feature vector of the standardized power parameter sequence. In this embodiment of the invention, when the feature fusion analysis module performs multi-dimensional feature fusion analysis on the standardized power parameter sequence to obtain the multi-dimensional electrical feature vector of the standardized power parameter sequence, it is specifically used for: Extract the time-domain and frequency-domain electrical features from the standardized power parameter sequence; A feature correlation analysis is performed on the time-domain electrical features and the frequency-domain electrical features to obtain the feature correlation matrix of the standardized power parameter sequence; Based on the feature correlation matrix, the time-domain electrical features and the frequency-domain electrical features are fused to obtain the multidimensional electrical feature vector of the standardized power parameter sequence.

[0038] When the feature fusion analysis module performs feature correlation feature fusion based on the feature correlation matrix to obtain a multidimensional electrical feature vector of the standardized power parameter sequence, it is specifically used for: Based on the preset fusion weights and the feature correlation matrix, the time-domain electrical features and the frequency-domain electrical features are weighted and fused to obtain the initial fused feature vector of the standardized power parameter sequence. The weighted fusion calculation formula is as follows: In the formula, This is the initial fused feature vector. The preset time-domain fusion weighting coefficients, The preset frequency domain fusion weighting coefficients, These are the dominant correlation factors between the time-domain and frequency-domain features extracted from the feature correlation matrix. The time-domain feature vector formed by the aforementioned time-domain electrical characteristics. The frequency domain feature vector is formed by the frequency domain electrical features; The initial fused feature vector is subjected to dimensionality reduction and smoothing processing to obtain the multidimensional electrical feature vector of the standardized power parameter sequence.

[0039] Time-domain feature extraction is performed on the standardized power parameter sequence, focusing on the pattern of data change over time. Information such as peak value, valley value, average value, effective value, and rate of change, which can reflect the fluctuation characteristics and statistical features of power parameters in the time dimension, is extracted to form time-domain electrical features. At the same time, frequency-domain transformation processing is performed on the standardized power parameter sequence to transform the power data in the time dimension into information in the frequency dimension. Harmonic content, fundamental frequency, spectrum distribution, and main frequency amplitude, which can reflect the frequency composition and energy distribution characteristics of power parameters, are extracted to obtain frequency-domain electrical features.

[0040] The extracted time-domain electrical features and frequency-domain electrical features are combined in pairs, and the correlation between each pair of features is analyzed one by one. By judging the mutual influence, complementary relationship and repetitive information of different features in describing the operating state of the power system, the correlation between various features is quantitatively characterized. The correlation results of all features are sorted according to fixed arrangement rules to form a feature correlation matrix of standardized power parameter sequence that can comprehensively reflect the correlation between time-domain and frequency-domain electrical features.

[0041] Based on the feature correlation matrix, feature combinations with high correlation in the matrix are identified. For such feature combinations, the core features that can more comprehensively reflect the state of the power system are retained, while duplicate or redundant features are eliminated. At the same time, features with low correlation but that can provide unique information are retained, ensuring that the fused features are neither redundant nor can fully cover the operating characteristics of the power system. Through this correlation-based screening and integration method, time-domain electrical features and frequency-domain electrical features are organically integrated, and finally a multi-dimensional electrical feature vector of standardized power parameter sequences is obtained.

[0042] Referring to the preset fusion weights, which are pre-set based on the importance of time-domain electrical features and frequency-domain electrical features in reflecting the operating state of the power system, the higher the importance, the greater the corresponding weight. Combining the correlation between various features reflected in the feature correlation matrix, each time-domain electrical feature and frequency-domain electrical feature is assigned a corresponding fusion weight. By multiplying each feature by its own weight and summarizing the results, all effective feature information in the time and frequency domains is integrated to obtain the initial fusion feature vector of the standardized power parameter sequence.

[0043] The feature data in the initial fused feature vector is comprehensively reviewed to identify redundant features and noisy data. By retaining core features and removing duplicate or highly correlated redundant information, the dimensionality of the feature vector is reduced. At the same time, a smoothing process is adopted to eliminate abnormal fluctuations and interference signals in the feature data, making the feature data more stable and regular. After dimensionality reduction and smoothing, a multi-dimensional electrical feature vector that can accurately and concisely reflect the operating characteristics of the power system is obtained.

[0044] The time-domain fusion weighting coefficient is preset based on the importance of time-domain electrical characteristics in reflecting the operating status of the power system. The specific value is determined by analyzing the contribution of time-domain characteristics to power system anomaly identification and status assessment. The higher the contribution, the larger the corresponding coefficient value.

[0045] The frequency domain fusion weighting coefficient is pre-set based on the criticality of frequency domain electrical features in characterizing the operating characteristics of the power system. It is combined with the effect of frequency domain features in capturing power parameter frequency anomalies, harmonic interference, etc., and the coefficient is assigned as follows: the more critical the effect, the larger the coefficient value.

[0046] The dominant correlation factor is derived from the feature correlation matrix. It is determined by extracting the value with the strongest correlation between the time-domain electrical features and the frequency-domain electrical features in the matrix. This value directly reflects the core correlation between the two types of features when describing the state of the power system.

[0047] The time-domain feature vector is composed of all extracted time-domain electrical features arranged in a preset order. Each time-domain electrical feature is sequentially assigned to the vector according to its corresponding physical meaning and extraction order, forming a complete time-domain feature vector.

[0048] The frequency domain feature vector is formed by combining all extracted frequency domain electrical features according to a fixed arrangement rule. The arrangement order of the frequency domain electrical features is consistent with the arrangement logic of the corresponding time domain electrical features to ensure the standardization of the feature vector.

[0049] The significance of this formula lies in the fact that it combines preset time-domain and frequency-domain fusion weight coefficients with the dominant correlation factor extracted from the feature correlation matrix to perform weighted fusion calculation of time-domain feature vectors and frequency-domain feature vectors.

[0050] During the calculation process, the weights of the time-domain feature vector and the frequency-domain feature vector are first adjusted by the dominant correlation factor and its complementary value. Then, the importance ratio of the two types of features is further enhanced by the preset time-domain and frequency-domain fusion weight coefficients. The effective information of the time-domain and frequency-domain features is fully integrated by weighted accumulation, and finally the initial fusion feature vector is obtained.

[0051] This fusion approach considers the importance of both time-domain and frequency-domain features, while also fully integrating the correlation between the two types of features. It avoids the limitations of a single feature dimension, enabling the initial fused feature vector to more comprehensively and accurately characterize the core features of the standardized power parameter sequence, thus providing reliable feature support for subsequent power system state analysis.

[0052] The beneficial effects are that the time-domain and frequency-domain electrical features of the standardized power parameter sequence are extracted simultaneously. The time-domain features capture the fluctuation pattern of the data over time, and the frequency-domain features mine the frequency composition and energy distribution of the data. The dual dimensions comprehensively cover the operating characteristics of the power parameters, avoiding the one-sidedness of single-dimensional features.

[0053] Feature correlation analysis is performed on the two types of features to systematically sort out the degree of correlation, complementary relationship and redundant information between different features. The generated feature correlation matrix provides accurate correlation basis for subsequent fusion and ensures that the fusion process is highly targeted.

[0054] Feature fusion is performed based on the feature correlation matrix, prioritizing the retention of core features and eliminating redundant information, while integrating complementary features. This results in a multi-dimensional electrical feature vector that is both concise and comprehensive, accurately representing the core operating patterns of power parameters.

[0055] The entire fusion process achieves the organic integration of time-domain and frequency-domain features, breaking the limitations of a single feature dimension and enabling multi-dimensional electrical feature vectors to have richer information dimensions, providing more reliable feature support for subsequent anomaly diagnosis and classification, and improving the accuracy of anomaly identification.

[0056] The correlation-based fusion method ensures the rationality of feature vectors, avoids interference from invalid information, and makes feature data more consistent with the actual operating state of the power system. This provides a high-quality feature foundation for subsequent comparison with reference baselines and hierarchical response decisions, and helps to improve the scientific nature of overall early warning monitoring.

[0057] The preset time-domain and frequency-domain fusion weight coefficients accurately match the importance of the two types of features in reflecting the operating status of the power system. Combined with the dominant correlation factors extracted from the feature correlation matrix, targeted weighting of time-domain and frequency-domain features is achieved, so that the fusion process not only fits the value of the features themselves, but also takes into account the correlation between features.

[0058] By integrating the effective information of time-domain and frequency-domain feature vectors through weighted fusion calculation, the limitations of a single feature dimension are avoided. The generated initial fused feature vector can comprehensively cover the time fluctuation characteristics and frequency distribution patterns of power parameters, providing a rich feature foundation for subsequent processing.

[0059] The initial fused feature vector is dimensionality reduced to remove redundant features and duplicate information, simplifying the feature dimension while retaining the core effective information, thereby reducing the computational complexity of subsequent anomaly diagnosis and improving monitoring efficiency.

[0060] Smoothing effectively eliminates noisy data and abnormal fluctuations in the initial fused feature vector, making the feature data more stable and regular, reducing the impact of interference factors on anomaly detection, and improving the reliability of the feature vector.

[0061] The entire fusion process forms a complete link from weighted calculation to dimensionality reduction and smoothing. The resulting multidimensional electrical feature vector is comprehensive, concise, and stable. It can accurately characterize the core features of standardized power parameter sequences, provide high-quality feature support for anomaly diagnosis and classification, and help improve the accuracy of power parameter anomaly early warning and monitoring.

[0062] The anomaly diagnosis and classification module 103 is used to map the multidimensional electrical feature vector to a preset hierarchical response strategy based on the reference baseline of the embedded system, and generate an anomaly type judgment result for the current power operation status. In this embodiment of the invention, when the anomaly diagnosis and classification module executes the mapping of the multi-dimensional electrical feature vector to a preset hierarchical response strategy based on the reference baseline of the embedded system to generate an anomaly type judgment result for the current power operation state, it is specifically used for: Based on the reference baseline of the embedded system, a baseline comparison analysis is performed on the multidimensional electrical feature vector; Based on the feature difference analysis results and the preset hierarchical response strategy, the multidimensional electrical feature vectors are mapped to the corresponding response levels to obtain the feature mapping relationship of the multidimensional electrical feature vectors; Based on the feature mapping relationship, the anomaly type and response level corresponding to the multidimensional electrical feature vector are determined to obtain the anomaly type judgment result of the current power operation state.

[0063] The reference baseline of the embedded system is retrieved. This baseline is constructed based on the historical power data of the embedded system under normal operating conditions. It covers the standard range and typical performance of various electrical characteristics under normal operating conditions. Each feature in the multi-dimensional electrical feature vector is compared with the corresponding standard feature in the reference baseline one by one. The deviation of each feature from the baseline standard is analyzed to determine whether the feature is within the normal threshold, the direction and degree of deviation, and to form a complete feature difference analysis result.

[0064] The pre-defined graded response strategy is formulated based on the impact of different anomalies in the power system on operational safety. It divides the system into multiple response levels, each with a clear range of characteristic differences and corresponding processing principles. Combining the previously obtained characteristic difference analysis results, it determines which response level in the pre-defined graded response strategy the overall characteristic differences reflected by the multidimensional electrical feature vectors meet. The multidimensional electrical feature vectors are then associated with that response level to obtain the feature mapping relationship of the multidimensional electrical feature vectors.

[0065] Based on the feature mapping relationship, the response level corresponding to the multi-dimensional electrical feature vector is determined. At the same time, the corresponding anomaly type is identified under each response level in the preset graded response strategy. The anomaly type is predefined based on historical anomaly cases, power system operation mechanism and feature difference pattern. Each response level corresponds to one or more specific power operation anomalies. The corresponding anomaly type is locked through the feature mapping relationship. The anomaly type is combined with the response level to finally obtain the anomaly type judgment result of the current power operation status.

[0066] The beneficial effect is that by using the reference baseline of the embedded system as a unified comparison standard, which is constructed based on the historical data of the system's normal operation, the benchmark for anomaly judgment is consistent with the actual operating characteristics of the system, providing a reliable basis for baseline comparison analysis.

[0067] By performing a baseline comparison on each of the multidimensional electrical feature vectors, the deviation direction, degree and range of each feature from the baseline standard are comprehensively analyzed. The generated feature difference analysis results can fully reflect the abnormal performance of the power operation status and avoid the one-sidedness of single feature comparison.

[0068] The pre-defined hierarchical response strategy clarifies the response level according to the degree of anomaly. Combined with the results of feature difference analysis, it achieves a precise mapping between multi-dimensional electrical feature vectors and response levels. The resulting feature mapping relationship can clearly link the power status and response priority.

[0069] Based on the feature mapping relationship, the corresponding anomaly type and response level are identified. The anomaly type is preset according to historical cases and power system operation mechanism to ensure that the judgment results are both targeted and scientific, enabling staff to quickly grasp the core information of the anomaly.

[0070] The entire process forms a complete logical chain of "baseline comparison - level mapping - type determination", ensuring that the anomaly type judgment results are comprehensive, accurate and in line with the actual operation scenario, providing a clear basis for subsequent graded response operations, and improving the accuracy and efficiency of power parameter anomaly early warning monitoring.

[0071] The baseline adaptive correction module 104 is used to periodically adaptively correct the reference baseline based on the historical operating data of the embedded system to obtain the updated reference baseline of the embedded system. In this embodiment of the invention, when the baseline adaptive correction module performs periodic adaptive correction of the reference baseline based on the historical operating data of the embedded system to obtain the updated reference baseline of the embedded system, it is specifically used for: Based on the historical operating data of the embedded system, extract the historical electrical feature set of the embedded system; An adaptive assessment of the reference baseline is performed based on the aforementioned historical electrical feature set. Based on the results of the adaptation assessment, the reference baseline is periodically revised to obtain the updated reference baseline.

[0072] When the baseline adaptive correction module performs periodic corrections to the reference baseline based on the adaptive evaluation results to obtain the updated reference baseline, it is specifically used for: Based on the adaptive assessment results, determine the correction status of the reference baseline; When the reference baseline is in a state that needs correction, the correction direction and correction magnitude of the reference baseline are determined based on the historical electrical feature set. Based on the correction direction and the correction magnitude, the reference baseline is iteratively adjusted to obtain the corrected baseline data of the reference baseline; The stationarity of the corrected baseline data is verified to obtain the updated reference baseline.

[0073] The historical operating data stored in the embedded system is retrieved. This data covers various power parameter records under different time periods and load conditions during the normal operation of the system. Abnormal data and invalid records are removed through data filtering, and valid data that can truly reflect the normal operating status of the system is retained. Various electrical characteristics such as peak value, average value, and harmonic content are extracted from these valid historical operating data and organized into a historical electrical feature set of the embedded system according to time sequence and feature type.

[0074] Based on the historical electrical feature set, a comprehensive comparison is made between the standard features in the reference baseline and the corresponding features in the historical electrical feature set. The standard range of the reference baseline is analyzed to see if it can still adapt to the current distribution pattern of historical electrical features. The degree of fit between the baseline standard and historical features is judged. The reference baseline is assessed to see if it can still accurately distinguish between normal and abnormal electrical features under the current system operating conditions. A comprehensive assessment conclusion on the adaptability of the reference baseline is formed, namely the adaptability assessment result.

[0075] Based on the results of the adaptability assessment, if the reference baseline has a high degree of fit with the historical electrical feature set, only minor adjustments are made to the standard ranges of individual features in the baseline that have small deviations from the historical features; if the fit is low, it indicates that the reference baseline can no longer accurately adapt to the current operating state of the system, and the standard ranges of each electrical feature are redefined according to the distribution pattern and statistical characteristics presented by the historical electrical feature set, and the core parameters of the reference baseline are corrected.

[0076] A fixed correction cycle is set, which is predetermined based on the stability of the embedded system and the frequency of power load changes. The above evaluation and correction process is repeated according to the set cycle to ensure that the reference baseline can continuously adapt to the changes in the system's operating status. After periodic correction, an updated reference baseline that is more in line with the current actual operating conditions of the embedded system is obtained.

[0077] Based on the adaptability assessment results, and in accordance with the preset baseline adaptation standards, it is determined whether the degree of fit between the reference baseline and the historical electrical feature set meets the system operation requirements. If the standard range of the baseline can accurately cover the normal distribution of historical electrical features and effectively distinguish between normal and abnormal features, the reference baseline is determined to be in a state that does not require correction. If there is a significant deviation between the baseline and historical features, and it cannot accurately reflect the electrical feature patterns of the current system's normal operation, the reference baseline is determined to be in a state that requires correction.

[0078] When it is determined that the reference baseline is in a state of needing correction, the statistical regularity and distribution trend of the historical electrical feature set should be analyzed in depth to clarify the overall deviation direction of the historical electrical features from the reference baseline standard, whether it is generally higher or lower than the baseline standard. At the same time, based on the overall degree of deviation and the concentration range of feature distribution, the specific adjustment range of the reference baseline should be determined to ensure that the correction direction and range can accurately match the actual operating state of the system reflected by the historical electrical features.

[0079] According to the determined correction direction and correction range, each standard feature in the reference baseline is adjusted one by one. The first round of correction is completed to obtain the preliminary corrected baseline. Then, the preliminary corrected baseline is compared with the historical electrical feature set again to check whether the corrected baseline fully matches the historical feature pattern. If there is still a deviation, the correction range is finely adjusted according to the deviation. The above comparison and adjustment process is repeated until the degree of fit between the corrected baseline and the historical electrical feature set meets the preset requirements, and the corrected baseline data of the reference baseline is obtained.

[0080] The stationarity verification method is adopted to compare the corrected baseline data with the electrical characteristics of the system during normal operation for several consecutive periods in recent times. The analysis is to see if the corrected baseline can stably cover the normal electrical characteristics under different time periods and load conditions, and to check whether the baseline standard has frequent fluctuations or unreasonable jumps. If the corrected baseline can continuously and stably adapt to the normal operation characteristics of the system without abnormal fluctuations, the stationarity verification is passed, and the corrected baseline data is determined as the updated reference baseline of the reference baseline.

[0081] The beneficial effect is that historical electrical feature sets can be extracted from the historical operating data of embedded systems. This data covers the normal operating power parameters under different time periods and load conditions, ensuring that the feature sets can fully reflect the true characteristics of the system's long-term operation and provide a rich and reliable basis for baseline correction.

[0082] An adaptive assessment of the reference baseline is performed based on historical electrical feature sets. By comprehensively comparing the degree of fit between the baseline standard and historical features, it is possible to accurately determine whether the baseline is suitable for the current system operating status and avoid abnormal judgment deviations caused by the baseline deviating from reality.

[0083] Based on the results of the adaptability assessment, targeted periodic adjustments are made. When the fit is high, the baseline is fine-tuned, and when the fit is low, the range of characteristic standards is redefined to ensure that the adjusted baseline can dynamically adapt to changes in the system's operating status and break the limitations of a fixed baseline.

[0084] The periodic correction mechanism enables the reference baseline to be continuously iterated and optimized, always keeping in line with the actual operating characteristics of the embedded system. This provides an accurate judgment benchmark for subsequent anomaly diagnosis and classification, improving the accuracy and timeliness of anomaly identification.

[0085] The entire process achieves closed-loop management of the reference baseline from data support and adaptation evaluation to dynamic correction, enabling the baseline to adapt to system changes, providing a stable and accurate core basis for power parameter anomaly early warning monitoring, and ensuring the long-term reliability of the monitoring system.

[0086] Based on the results of the adaptation assessment, the correction status of the reference baseline is accurately determined by comparing it with the preset baseline adaptation standard. The situation of no need for correction and need for correction is clearly distinguished to avoid blind correction or omission of necessary correction, and to ensure that the baseline correction is targeted and reasonable.

[0087] When the reference baseline is in a state that needs correction, based on the historical electrical characteristic set covering different operating conditions of the system, the distribution pattern and deviation trend of the characteristics are analyzed in depth to accurately determine the correction direction and magnitude, so that the correction action fits the actual operating characteristics of the system and avoids baseline inaccuracy caused by adjustments without basis.

[0088] The reference baseline is iteratively adjusted according to the determined correction direction and magnitude. Through multiple rounds of comparison and fine-tuning, the deviation between the baseline and the historical electrical feature set is gradually reduced, ensuring that the corrected baseline data can fully adapt to the feature range of normal system operation and improve the accuracy of the baseline.

[0089] The stability of the corrected baseline data is verified by comparing it with the electrical characteristics of recent multi-cycle normal operation. Fluctuations or unreasonable jumps in the baseline standard are investigated to ensure the stability of the updated reference baseline and provide a continuous and reliable judgment benchmark for subsequent anomaly diagnosis.

[0090] The entire process forms a closed-loop correction mechanism of "state judgment - direction and magnitude determination - iterative adjustment - stability verification", which realizes the refined and dynamic optimization of the reference baseline, so that it always keeps in sync with the operating state of the embedded system. This fundamentally solves the problem of fixed baselines being out of touch with reality, provides precise support for power parameter anomaly early warning monitoring, and ensures the accuracy and timeliness of monitoring results.

[0091] The hierarchical response execution module 105 is used to perform hierarchical response operations on the embedded system based on the anomaly type judgment result and the updated reference baseline.

[0092] In this embodiment of the invention, when the hierarchical response execution module performs hierarchical response operations on the embedded system based on the exception type judgment result and the updated reference baseline, it is specifically used for: The impact dimension of the anomaly type judgment result is analyzed to generate the anomaly impact feature vector of the embedded system; The abnormal impact feature vector is combined with the updated reference baseline to perform a comprehensive situation assessment, generating the response situation assessment result of the embedded system. Based on the response situation assessment results, an adaptive response decision instruction set for the embedded system is generated; Based on the adaptive response decision instruction set and the real-time operating parameters of the embedded system, a response action sequence is performed on the embedded system to obtain a response execution status report of the embedded system.

[0093] The adaptive response decision instruction set includes: When the anomaly type determination result is an overvoltage anomaly and the deviation level is at the warning level, the voltage limit adjustment instruction and the warning log recording instruction of the embedded system are generated. When the abnormality type judgment result is undervoltage abnormality and the deviation degree is at the general abnormality level, the voltage compensation adjustment instruction and abnormal status reporting instruction of the embedded system are generated. When the anomaly type determination result is an overcurrent anomaly and the deviation level is at the general anomaly level, the embedded system generates a load grading cutoff instruction and a protective alarm instruction. When the anomaly type judgment result is that the harmonic content is abnormal and the deviation is at the warning level, the harmonic suppression reference instruction and trend analysis record instruction of the embedded system are generated. When the anomaly type determination result is a frequency deviation anomaly and the deviation degree is at the emergency anomaly level, the embedded system generates an emergency frequency adjustment instruction and a system protection linkage instruction.

[0094] The anomaly type and response level identified in the anomaly type judgment results are deeply decomposed to analyze the scope of the anomaly's impact on the power system operation, including whether it involves core power supply circuits, key electrical equipment, etc. The degree of interference of the anomaly on key operating dimensions such as voltage stability, power balance, and energy efficiency is assessed. At the same time, the chain reactions that the anomaly may trigger and the duration of its impact are identified. These key information related to the impact are integrated and sorted out to generate an anomaly impact feature vector of the embedded system that can comprehensively characterize the impact of the anomaly.

[0095] Retrieve the electrical characteristic standards of the normal operating state in the updated reference baseline, compare each impact index in the abnormal impact characteristic vector with the corresponding standard of the updated reference baseline one by one, analyze the degree of deviation between the abnormal impact and the normal standard, and combine the scope, degree and potential risks of the abnormal impact to comprehensively judge the current overall operating status of the embedded system, clarify whether the system is in a safe and controllable state, requires vigilance and intervention or emergency response state, and form the response status assessment result of the embedded system.

[0096] Based on the response situation assessment results, and in accordance with the preset hierarchical response rule base, which sets corresponding response measures, execution logic, and operation priorities for different response situations, the system situation corresponding to the current assessment results is selected to identify suitable response strategies and clarify the specific operations to be performed, including parameter adjustment, device start-up and shutdown, alarm notification, etc. These operations are integrated according to the execution order and logical relationship to generate an adaptive response decision instruction set for the embedded system containing a series of specific instructions.

[0097] The system collects real-time operating parameters of the embedded system, including key data such as real-time voltage, current, power, and load rate. Based on these real-time operating parameters, it executes corresponding response actions step by step according to the instruction sequence and operation requirements of the adaptive response decision instruction set. During the execution process, it monitors the execution effect of each response action and the changes in system operating parameters in real time, and records key information such as the execution time, execution result, and system state changes of the response actions. These records are integrated and summarized to form a response execution status report of the embedded system that can comprehensively reflect the entire response execution process and the final state of the system.

[0098] When the anomaly type is clearly determined to be an overvoltage anomaly, and the deviation is confirmed to be at the warning level by comparison with the updated reference baseline, the graded response execution module initiates the corresponding response logic and generates a voltage limit adjustment instruction for the embedded system. This instruction explicitly requires the system to control the voltage within the normal range specified by the updated reference baseline by adjusting the voltage regulating component in the power supply circuit. At the same time, a warning log recording instruction is generated to record key information such as the time of the overvoltage anomaly, the current voltage value, the degree of deviation, and the triggered adjustment action, providing complete data support for subsequent traceability analysis.

[0099] When the anomaly type is determined to be undervoltage anomaly, and the deviation is found to be at a general anomaly level, the graded response execution module generates a voltage compensation adjustment command for the embedded system according to preset response rules. The command instructs the system to activate the voltage compensation mechanism, which raises the voltage to the normal standard by supplementing power supply or adjusting circuit parameters. At the same time, an anomaly status reporting command is generated, which accurately reports relevant information about the undervoltage anomaly, including the time of occurrence, the degree of voltage drop, and the compensation measures already implemented, to the system management platform to ensure that management personnel can promptly grasp the anomaly situation.

[0100] When the anomaly is determined to be an overcurrent anomaly and the deviation is classified as a general anomaly, the graded response execution module generates a load graded cut-off instruction for the embedded system. This instruction prioritizes the cut-off according to the importance of the load, prioritizing the cut-off of non-core and non-critical load devices to quickly reduce the current load in the circuit and prevent overcurrent from damaging the lines and core equipment. At the same time, a protective alarm instruction is generated, which issues an audible and visual alarm signal through the system's built-in alarm device to remind on-site personnel to promptly investigate the cause of the overcurrent.

[0101] When the anomaly type judgment result indicates abnormal harmonic content and the deviation level is at the warning level, the graded response execution module generates a harmonic suppression reference instruction for the embedded system. The instruction system starts the harmonic filtering device and filters and suppresses the excessive harmonics according to the preset harmonic suppression strategy, reducing the interference of harmonics on the operation of the power system. At the same time, a trend analysis recording instruction is generated to continuously record the changing trend of harmonic content, the execution effect of suppression measures, and other data, providing a basis for subsequent optimization of harmonic suppression schemes.

[0102] When the anomaly type is determined to be a frequency deviation anomaly, and the deviation reaches the emergency anomaly level, the graded response execution module immediately generates an emergency frequency adjustment command for the embedded system. The core control components of the command system quickly intervene and, by adjusting the output frequency at the generator end or adjusting the load distribution, urgently correct the operating frequency of the power system to return it to the normal range specified by the updated reference baseline. At the same time, a system protection linkage command is generated to activate the system's multiple protection mechanisms, including overload protection and short circuit protection, to prevent the frequency anomaly from further expanding and causing equipment damage or system collapse.

[0103] The beneficial effects are that the results of anomaly type judgment are analyzed from multiple dimensions, comprehensively covering key dimensions such as the scope of anomaly impact, degree of interference, potential chain reactions and duration. The generated anomaly impact feature vector can completely and accurately characterize the effect of anomalies on embedded systems, providing a comprehensive basis for subsequent response decisions.

[0104] By combining the normal operation standards of the updated reference baseline, a comprehensive situation assessment is conducted on the anomaly impact feature vector to accurately determine the current security status level of the system. This avoids response bias caused by judging solely based on the anomaly type and ensures that the response situation assessment results are consistent with the actual operation of the system.

[0105] An adaptive response decision instruction set is generated based on the response situation assessment results. The instruction set matches the corresponding response measures, execution logic and operation priorities for different system situations, so as to realize the precision and differentiation of response decisions and avoid the limitations of a uniform response method.

[0106] Based on real-time operating parameters, the system executes response actions step by step according to the instruction set sequence, while simultaneously monitoring the execution effect of actions and changes in system parameters in real time. This ensures that the response operation dynamically adapts to the real-time state of the system, thereby improving the effectiveness and security of response execution.

[0107] The system records key information throughout the entire response process, and the resulting response execution status report comprehensively reflects the response actions, execution results, and the final state of the system. This provides complete data support for subsequent anomaly tracing and response strategy optimization, and helps the monitoring system to continuously iterate and upgrade.

[0108] The entire process constructs a closed-loop response mechanism of "impact analysis - situation assessment - decision generation - action execution - report feedback", which deeply integrates the anomaly type judgment results with the dynamically updated reference baseline to achieve accurate and efficient hierarchical response to embedded systems, ensure the stable operation of the power system, and improve the practicality and reliability of anomaly early warning monitoring.

[0109] For overvoltage anomaly warning scenarios, the voltage limit adjustment command can quickly control the voltage within the normal range of the updated reference baseline, avoiding potential damage to the equipment caused by overvoltage. The warning log recording command completely retains key information about the anomaly, providing data support for subsequent traceability analysis.

[0110] In the event of a general undervoltage situation, the voltage compensation adjustment command can efficiently restore the voltage to the standard level by providing targeted power replenishment or adjusting parameters. The abnormal status reporting command ensures that management personnel can promptly grasp the details of the abnormality, facilitating subsequent follow-up and handling.

[0111] For general overcurrent anomalies, the load grading and disconnection command precisely disconnects non-core loads according to priority, quickly reducing the circuit load and preventing overcurrent from causing damage to lines or core equipment. The protective alarm command provides audible and visual prompts to help on-site personnel promptly troubleshoot the root cause of the fault.

[0112] For abnormal harmonic content warning levels, the harmonic suppression reference command activates the filtering device to specifically suppress excessive harmonics, reducing interference to the operation of the power system. The trend analysis and recording command continuously tracks harmonic changes and suppression effects, providing a basis for optimizing suppression schemes.

[0113] For frequency deviation anomalies at an emergency level, an emergency frequency adjustment command is executed to quickly correct the operating frequency and bring it back to the normal range. The system protection linkage command activates multiple protection mechanisms to effectively prevent equipment damage or system crashes caused by the expansion of frequency anomalies.

[0114] The instruction set accurately matches response instructions according to the type and degree of deviation, enabling differentiated and scenario-based hierarchical responses. This ensures that various anomalies are handled in a targeted manner while avoiding over-response or under-response, significantly improving the accuracy and safety of power system anomaly handling and guaranteeing stable system operation.

[0115] Reference Figure 2 The diagram shown is a flowchart illustrating a power parameter anomaly early warning and monitoring method based on an embedded system according to an embodiment of the present invention. In this embodiment, the power parameter anomaly early warning and monitoring method based on an embedded system includes: S1. Collect raw power data from various power monitoring devices connected to the embedded system, perform multi-protocol adaptive parsing on the raw power data, and obtain a standardized power parameter sequence of the raw power data; S2. Perform multi-dimensional feature fusion analysis on the standardized power parameter sequence to obtain the multi-dimensional electrical feature vector of the standardized power parameter sequence; S3. Based on the reference baseline of the embedded system, the multidimensional electrical feature vector is mapped to a preset hierarchical response strategy to generate an anomaly type judgment result of the current power operation state; S4. Based on the historical operating data of the embedded system, the reference baseline is periodically and adaptively corrected to obtain the updated reference baseline of the embedded system; S5. Based on the anomaly type judgment result and the updated reference baseline, perform a graded response operation on the embedded system.

[0116] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0117] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0118] 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 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.

Claims

1. A power parameter anomaly early warning and monitoring device based on an embedded system, characterized in that, The system includes a data acquisition and parsing module, a feature fusion and analysis module, an anomaly diagnosis and classification module, a baseline adaptive correction module, and a hierarchical response execution module, wherein: The data acquisition and parsing module is used to acquire raw power data from various power monitoring devices connected to the embedded system, perform multi-protocol adaptive parsing on the raw power data, and obtain a standardized power parameter sequence of the raw power data. The feature fusion analysis module is used to perform multi-dimensional feature fusion analysis on the standardized power parameter sequence to obtain a multi-dimensional electrical feature vector of the standardized power parameter sequence. The anomaly diagnosis and classification module is used to map the multidimensional electrical feature vector to a preset hierarchical response strategy based on the reference baseline of the embedded system, and generate an anomaly type judgment result for the current power operation status. The baseline adaptive correction module is used to periodically adaptively correct the reference baseline based on the historical operating data of the embedded system to obtain the updated reference baseline of the embedded system. The hierarchical response execution module is used to perform hierarchical response operations on the embedded system based on the anomaly type judgment result and the updated reference baseline.

2. The power parameter anomaly early warning and monitoring device based on an embedded system as described in claim 1, characterized in that, When the data acquisition and parsing module acquires raw power data from various power monitoring devices connected to the embedded system, performs multi-protocol adaptive parsing on the raw power data, and obtains a standardized power parameter sequence from the raw power data, it is specifically used for: Acquire raw power data from the embedded system; The protocol feature identification of the raw power data is performed to obtain the protocol type identification result of the embedded system. The protocol parsing rules corresponding to the protocol type identification result are invoked to parse the original power data and obtain the initial power parameters of the original power data; The initial power parameters are verified for integrity, and the corrected power data is normalized to obtain a standardized power parameter sequence of the original power data.

3. The power parameter anomaly early warning and monitoring device based on an embedded system as described in claim 1, characterized in that, When the feature fusion analysis module performs multi-dimensional feature fusion analysis on the standardized power parameter sequence to obtain the multi-dimensional electrical feature vector of the standardized power parameter sequence, it is specifically used for: Extract the time-domain and frequency-domain electrical features from the standardized power parameter sequence; A feature correlation analysis is performed on the time-domain electrical features and the frequency-domain electrical features to obtain the feature correlation matrix of the standardized power parameter sequence; Based on the aforementioned feature correlation matrix, the time-domain electrical features and the frequency-domain electrical features are fused using feature correlation analysis. The multidimensional electrical feature vector of the standardized power parameter sequence is obtained.

4. The power parameter anomaly early warning and monitoring device based on an embedded system as described in claim 3, characterized in that, The feature fusion analysis module performs feature correlation fusion between the time-domain electrical features and the frequency-domain electrical features based on the feature correlation matrix. When obtaining the multidimensional electrical feature vector of the standardized power parameter sequence, it is specifically used for: Based on the preset fusion weights and the feature correlation matrix, the time-domain electrical features and the frequency-domain electrical features are weighted and fused to obtain the initial fused feature vector of the standardized power parameter sequence. The weighted fusion calculation formula is as follows: In the formula, This is the initial fused feature vector. The preset time-domain fusion weighting coefficients, The preset frequency domain fusion weighting coefficients, These are the dominant correlation factors between the time-domain and frequency-domain features extracted from the feature correlation matrix. The time-domain feature vector formed by the time-domain electrical features, The frequency domain feature vector is formed by the frequency domain electrical features; The initial fused feature vector is subjected to dimensionality reduction and smoothing processing to obtain the multidimensional electrical feature vector of the standardized power parameter sequence.

5. The power parameter anomaly early warning and monitoring device based on an embedded system as described in claim 1, characterized in that, When the anomaly diagnosis and classification module executes the mapping of the multi-dimensional electrical feature vector to a preset hierarchical response strategy based on the reference baseline of the embedded system to generate an anomaly type judgment result for the current power operation state, it is specifically used for: Based on the reference baseline of the embedded system, a baseline comparison analysis is performed on the multidimensional electrical feature vector; Based on the feature difference analysis results and the preset hierarchical response strategy, the multidimensional electrical feature vectors are mapped to the corresponding response levels to obtain the feature mapping relationship of the multidimensional electrical feature vectors; Based on the feature mapping relationship, the anomaly type and response level corresponding to the multidimensional electrical feature vector are determined to obtain the anomaly type judgment result of the current power operation state.

6. The power parameter anomaly early warning and monitoring device based on an embedded system as described in claim 1, characterized in that, When the baseline adaptive correction module performs periodic adaptive correction of the reference baseline based on the historical operating data of the embedded system to obtain the updated reference baseline of the embedded system, it is specifically used for: Based on the historical operating data of the embedded system, extract the historical electrical feature set of the embedded system; An adaptive assessment of the reference baseline is performed based on the aforementioned historical electrical feature set. Based on the results of the adaptation assessment, the reference baseline is periodically revised to obtain the updated reference baseline.

7. The power parameter anomaly early warning and monitoring device based on an embedded system as described in claim 6, characterized in that, When the baseline adaptive correction module performs periodic corrections to the reference baseline based on the adaptive evaluation results to obtain the updated reference baseline, it is specifically used for: Based on the adaptive assessment results, determine the correction status of the reference baseline; When the reference baseline is in a state that needs correction, the correction direction and correction magnitude of the reference baseline are determined based on the historical electrical feature set. Based on the correction direction and the correction magnitude, the reference baseline is iteratively adjusted to obtain the corrected baseline data of the reference baseline; The stationarity of the corrected baseline data is verified to obtain the updated reference baseline.

8. The power parameter anomaly early warning and monitoring device based on an embedded system as described in claim 5, characterized in that, When the hierarchical response execution module performs hierarchical response operations on the embedded system based on the exception type judgment result and the updated reference baseline, it is specifically used for: The impact dimension of the anomaly type judgment result is analyzed to generate the anomaly impact feature vector of the embedded system; The abnormal impact feature vector is combined with the updated reference baseline to perform a comprehensive situation assessment, generating the response situation assessment result of the embedded system. Based on the response situation assessment results, an adaptive response decision instruction set for the embedded system is generated; Based on the adaptive response decision instruction set and the real-time operating parameters of the embedded system, a response action sequence is performed on the embedded system to obtain a response execution status report of the embedded system.

9. The power parameter anomaly early warning and monitoring device based on an embedded system as described in claim 8, characterized in that, The adaptive response decision instruction set includes: When the anomaly type determination result is an overvoltage anomaly and the deviation level is at the warning level, the voltage limit adjustment instruction and the warning log recording instruction of the embedded system are generated. When the abnormality type judgment result is undervoltage abnormality and the deviation degree is at the general abnormality level, the voltage compensation adjustment instruction and abnormal status reporting instruction of the embedded system are generated. When the anomaly type determination result is an overcurrent anomaly and the deviation level is at the general anomaly level, the embedded system generates a load grading cutoff instruction and a protective alarm instruction. When the anomaly type judgment result is that the harmonic content is abnormal and the deviation is at the warning level, the harmonic suppression reference instruction and trend analysis record instruction of the embedded system are generated. When the anomaly type determination result is a frequency deviation anomaly and the deviation degree is at the emergency anomaly level, the embedded system generates an emergency frequency adjustment instruction and a system protection linkage instruction.

10. A power parameter anomaly early warning and monitoring method based on an embedded system, characterized in that, The method for using the power parameter anomaly early warning and monitoring device based on an embedded system as described in claim 1: S1. Collect raw power data from various power monitoring devices connected to the embedded system, perform multi-protocol adaptive parsing on the raw power data, and obtain a standardized power parameter sequence of the raw power data; S2. Perform multi-dimensional feature fusion analysis on the standardized power parameter sequence to obtain the multi-dimensional electrical feature vector of the standardized power parameter sequence; S3. Based on the reference baseline of the embedded system, the multidimensional electrical feature vector is mapped to a preset hierarchical response strategy to generate an anomaly type judgment result of the current power operation state; S4. Based on the historical operating data of the embedded system, the reference baseline is periodically and adaptively corrected to obtain the updated reference baseline of the embedded system; S5. Based on the anomaly type judgment result and the updated reference baseline, perform a graded response operation on the embedded system.

Images