Power transmission and distribution equipment operation state self-optimization system based on large parameter model

Abstract

The invention discloses a power transmission and distribution equipment operation state self-optimization system based on a large parameter model, relates to the technical field of power transmission and distribution equipment optimization, and is used for collecting operation data of equipment and generating a space-time synchronization signal set. Searching feature pairs appearing across signals on a time axis; feature pairs with the number larger than a preset threshold value are extracted and marked as symbiotic fault pairs, and a symbiotic fault mode is generated; analyzing the energy equilibrium state distortion degree and the energy conversion efficiency deviation degree, and fusing to generate a health index; determining the fault type and the residual service life based on the physical layer evidence and the effect layer evidence; operating parameters most sensitive to equipment health are extracted; according to the sensitivity coefficient of the operation parameter and in combination with the current health index, the suggested fine tuning amount of each parameter is analyzed; and the stability of the system after parameter adjustment is monitored in real time, feedback calibration is formed, and a historical maintenance case library is updated. The accuracy of data analysis is improved; and the security of the control strategy is ensured.

Description

Technical Field

[0001] This invention relates to the field of power transmission and distribution equipment optimization technology, specifically a power transmission and distribution equipment operation state self-optimization system based on a large parameter model. Background Technology

[0002] As an indispensable and crucial link in the power system, the power transmission and distribution system plays a central and pivotal role in the entire process of transmitting electricity from power plants to user terminals. Its operational status not only directly affects the quality and stability of power supply but also profoundly impacts the normal operation of social production and daily life. In today's society, with rapid economic development, various industries are increasingly reliant on electricity, resulting in a continuous increase in electricity demand and increasingly stringent requirements for the reliability of power supply.

[0003] While mainstream diagnostic models and operational strategies can detect anomalies, their decision-making processes are like black boxes, lacking a deep understanding of the physical mechanisms of equipment. This results in poor interpretability of diagnostic results, making them difficult for operations and maintenance personnel to fully trust. More importantly, most existing systems stop at fault alarms or simple diagnoses, failing to automatically translate diagnostic conclusions into effective control commands. Operations and maintenance work still heavily relies on human experience for decision-making and intervention. This passive response model is not only inefficient but also unable to achieve online adaptive optimization of operating parameters to suppress fault development. Furthermore, it lacks the ability to continuously learn and evolve from historical successful intervention cases, making it difficult to achieve a breakthrough in the system's intelligence level.

[0004] Therefore, this invention discloses a self-optimization system for the operating status of power transmission and distribution equipment based on a large parameter model to solve the above problems. Summary of the Invention

[0005] The purpose of this invention is to provide a self-optimization system for the operating status of power transmission and distribution equipment based on a large parameter model, so as to solve the problems raised in the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a self-optimization system for the operating status of power transmission and distribution equipment based on a large parameter model, characterized in that the system includes an initial fault analysis module, a health analysis module, an evidence fusion diagnosis module, and a parameter adjustment module; The initial fault analysis module is used to collect the equipment's operating data and generate a spatiotemporal synchronization signal set; it uses signal processing and feature extraction technology to identify abnormal feature points or feature segments; it searches for feature pairs that appear across signals on the time axis; it extracts feature pairs with a number greater than a preset threshold and records them as co-occurring fault pairs, and generates co-occurring fault modes. The health analysis module is used to centrally analyze the energy information of each device from the spatiotemporal synchronization signal; analyze the energy balance distortion degree and energy conversion efficiency deviation degree, and fuse them to generate a health index; The evidence fusion diagnostic module is used to integrate and splice the health index with the fault location results to form physical layer evidence; guided by the fault location results, it traces back and focuses on specific signal features in the spatiotemporal synchronization signal set to extract effect layer evidence; and confirms the fault type and remaining service life based on the physical layer evidence and effect layer evidence. The parameter adjustment module extracts the operating parameters most sensitive to equipment health based on the diagnostic results; analyzes the suggested fine-tuning amount for each parameter based on the sensitivity coefficient of the operating parameters and in combination with the current health index; monitors the stability of the system after parameter adjustment in real time, forms feedback calibration, and updates the historical maintenance case library.

[0007] Based on the above: the initial fault analysis module includes a data acquisition and alignment unit, a feature extraction unit, and a fault symbiosis analysis unit; The data acquisition and alignment unit is used to collect the operating data of the equipment, including but not limited to current, vibration, sound and temperature data; the administrator presets a sequence of physical events, using physical events as time reference points; the raw signals collected by all sensors are time-shifted and aligned to eliminate inherent timing differences caused by transmission and sampling, and to generate a set of spatiotemporal synchronization signals; the sequence of physical events includes but is not limited to the rising or falling edge of the coil current pulse during opening or closing operations and known mechanical shocks; The feature extraction unit is used to identify abnormal feature points or feature segments in each individual signal using signal processing and feature extraction techniques. These signal processing and feature extraction techniques include, but are not limited to, peak detection, wavelet packet decomposition, spectral analysis, and energy calculation. The abnormal feature points or feature segments include, but are not limited to, detecting partial discharge peak values ​​from ultrasonic signals, extracting vibration energy from specific frequency bands from vibration signals, analyzing sudden increases in harmonic distortion rate from current signals, and capturing abnormal temperature rise rates from temperature signals. The vibration energy extracted from specific frequency bands in vibration signals includes the vibration energy of bearing fault characteristic frequency bands. The fault coexistence analysis unit is used to search for feature pairs that appear across signals on the time axis; extract feature pairs with a number greater than a preset threshold and record them as coexisting fault pairs, and generate coexisting fault patterns; the coexisting fault patterns include precursor symptoms, subsequent symptoms, time intervals, and confidence levels; the patterns are classified and stored according to the physical fault types corresponding to the coexisting fault patterns; the physical fault types include bearing outer ring faults, circuit breaker failure to operate, and insulation degradation.

[0008] This invention fundamentally solves the problem of data timing asynchrony caused by transmission and sampling delays by using time-shift alignment based on physical events, laying a solid foundation for subsequent accurate analysis of cross-signal causal relationships. By extracting specific and physically meaningful microscopic features such as "partial discharge peak value" and "specific frequency band vibration energy" from each individual signal, it can effectively capture early fault signs that are difficult to detect by the human ear or conventional monitoring, improving monitoring sensitivity. By "searching for cross-signal feature pairs" and forming "co-existing fault modes," it no longer views signal anomalies in isolation, but reveals the causal transmission chain within the equipment where one physical phenomenon triggers another (such as discharge followed by vibration). This greatly improves the interpretability and reliability of fault diagnosis and reduces the false alarm rate.

[0009] Based on the above: the health analysis module includes an energy analysis unit, a balance and efficiency analysis unit, and a health index unit; The energy analysis unit is used to centrally analyze the energy information of each device from the spatiotemporal synchronization signal. For rotating machinery or vibrating components with reciprocating motion, the vibration acceleration signal is deducted from the DC component and integrated twice to obtain the vibration mechanical energy. For electrical equipment or motor drive systems, the instantaneous active power is calculated using synchronously acquired current and voltage signals. The electromagnetic energy consumed in the analysis period is obtained by integrating the instantaneous active power within an analysis period. Temperature field information is obtained through deployed temperature sensors, combined with the heat capacity parameters of the equipment, and the heat accumulation or consumption within the analysis period is calculated using the temperature change rate. The heat accumulation or consumption is generated by integrating the product of the temperature change rate and the heat capacity parameters of the equipment over time. The balance and efficiency analysis unit is used to analyze the energy balance distortion degree and the energy conversion efficiency deviation degree. The energy balance distortion degree is equal to the difference between the input electrical energy and the output energy divided by the input electrical energy. The output energy is equal to the sum of the output vibration mechanical energy, the lost heat energy, and the energy dissipated in the form of vibration. The health index unit is used to integrate several energy flow distortion indicators and generate a health index through weighted fusion; the several energy flow distortion indicators include energy balance distortion degree and several energy conversion efficiency deviation degrees; based on the abnormal patterns of energy balance distortion degree and energy conversion efficiency deviation degree, combined with the energy flow path, the fault is located to a specific physical link.

[0010] Based on the above: For the conversion of electromagnetic energy to mechanical energy, the conversion efficiency is equal to the ratio of useful mechanical energy to electromagnetic energy; the useful mechanical energy is equal to the integral of the product of the torque and rotational speed of the output shaft; the difference between the target conversion efficiency and the actual conversion efficiency is calculated and divided by the target conversion efficiency, and recorded as the energy conversion efficiency deviation.

[0011] Based on the above, locating the fault to a specific physical component includes: if the rate of change of energy equilibrium distortion is greater than the threshold, and the mechanical energy of vibration increases simultaneously, while the change in energy conversion efficiency deviation is less than the threshold, the fault is identified as: loose or damaged mechanical structure, resulting in additional vibration energy dissipation, but the core conversion function is not yet severely damaged; if the rate of change of energy conversion efficiency deviation is greater than the threshold, and the heat energy increases simultaneously, the fault is identified as: abnormal electromagnetic conversion, increased winding losses or magnetic circuit fault, leading to decreased efficiency and overheating.

[0012] This invention's "Energy Equilibrium Distortion Degree" and "Energy Conversion Efficiency Deviation Degree" can quantify the severity of faults. By analyzing the combined change patterns of these indicators, it can clearly pinpoint whether the fault is in the "mechanical structure" or the "electromagnetic conversion link," achieving a leap from "problem discovery" to "problem localization," providing a direct basis for subsequent maintenance decisions. By calculating indicators such as "electromagnetic energy to mechanical energy conversion efficiency," it directly reflects whether the equipment's ability to perform its core functions is sound. The health index integrates multiple energy distortions, providing a quantitative reflection of the overall health status of the equipment, making it more comprehensive and reliable than single-parameter alarms.

[0013] Based on the above: the evidence fusion diagnostic module includes a physical evidence integration unit, an effect evidence integration unit, and a similarity analysis unit; The physical evidence integration unit is used to integrate and splice the health index with the fault location result to form physical layer evidence; the fault location result is one or more tags used to identify the location where the energy flow abnormality occurs, including but not limited to electromagnetic conversion links, mechanical transmission links or insulator systems. The effect evidence integration unit is used to extract effect layer evidence by backtracking and focusing on specific signal features in the spatiotemporal synchronization signal set, guided by the fault location results. The similarity analysis unit is used to normalize the quantitative indicators in the physical layer evidence and the effect layer evidence and splice them into the current feature vector; in the pre-built historical maintenance case library, the similarity measurement algorithm is used to find several historical cases that are most similar to the current feature vector and form the nearest neighbor set; each case in the historical case library contains a historical feature vector, the finally confirmed fault type and the actual remaining service life. Based on the fault types identified in the nearest neighbor set, a weighted voting algorithm is used to determine the final fault type; the weight of the weighted vote is equal to the similarity between the historical feature vector of the historical case and the current feature vector; based on the actual remaining service life of the nearest neighbor set, a regression model is used to calculate the predicted remaining service life of the current device and output the prediction confidence interval.

[0014] Based on the above: In the extraction of evidence of the effect layer, for faults located in the electromagnetic conversion link, the analysis includes: analyzing the electrodynamic force by calculating the square of the current signal; the electrodynamic force is equal to the product of the square of the current signal and the proportional coefficient; and analyzing the peak value, average electrodynamic force, and amplitude of the electrodynamic force; and integrating the discharge quantity, discharge frequency, and temperature data of the partial discharge signal, and constructing an insulation degradation assessment index by normalization and weighted fusion.

[0015] This invention integrates evidence at two levels—the physical layer and the effect layer—mimicking the diagnostic thinking of human experts: first, energy flow analysis is used to roughly locate the problem (physical layer), and then detailed feature analysis is used for specific verification (effect layer). This improves the accuracy of data analysis. Through a "similar analysis unit," the current case is matched with historical maintenance cases, enabling case-based reasoning. It provides diagnostic conclusions and similar historical evidence, making the diagnostic results more convincing. Simultaneously, it can directly output remaining service life predictions, achieving an advancement from condition monitoring to service life management.

[0016] Based on the above: the parameter adjustment module includes a feedforward fine-tuning monitoring unit and a feedback update unit; The feedforward fine-tuning monitoring unit, based on the final fault type, finds several operating parameters that are most effective in suppressing the fault of that type from a preset fault-parameter sensitivity table, and obtains the sensitivity coefficient of each operating parameter; based on the sensitivity coefficient of the operating parameters and combined with the current health index, it analyzes the suggested fine-tuning amount for each parameter. The feedback update unit is used to immediately start the stability monitoring program after parameter adjustment, and to track several system stability indicators in real time. The stability indicators include, but are not limited to, the fluctuation rate of output voltage or current, the total vibration of the shaft system, and the power oscillation amplitude. If the system stability indicators are detected to exceed the safety threshold, feedback calibration is triggered. The most recent parameter adjustment is rolled back proportionally according to a predetermined strategy. After the optimization operation has been verified for stability and proven to be effective, the system starts the self-learning process; the successful optimization case, including but not limited to: the initial symbiotic fault mode, the finally confirmed fault type, the parameter adjustment amount used, and the health index after adjustment, are added to the historical maintenance case library as new knowledge samples.

[0017] Based on the above: the suggested fine-tuning amount is equal to the product of the global gain coefficient, the health index deviation rate, the sensitivity coefficient allocation value, and the allowable adjustment range; the health index deviation rate is equal to the difference between the health index baseline value and the current health index divided by the health index baseline value; the sensitivity coefficient allocation value is equal to the ratio of the sensitivity coefficient of the current parameter to the sum of the absolute values ​​of the sensitivity coefficients of all adjusted parameters.

[0018] This invention not only diagnoses faults but also automatically provides feedforward optimization suggestions based on a "fault-parameter sensitivity table," proactively suppressing fault development and improving equipment reliability and efficiency. Through "stability monitoring" and "feedback calibration" mechanisms, system stability is immediately assessed after parameter adjustments, and automatic rollback is possible in case of anomalies, preventing secondary faults caused by improper optimization and ensuring the safety of the control strategy. A "self-learning process" transforms successful optimization cases into new knowledge stored in a historical database, enabling the system to accumulate operational experience and become increasingly intelligent with use. Subsequent encounters with similar situations will result in faster and more accurate diagnosis and optimization, achieving true "self-optimization."

[0019] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention fundamentally solves the problem of data timing asynchrony caused by transmission and sampling delays by performing time-shift alignment based on physical events, laying a solid foundation for subsequent accurate analysis of cross-signal causal relationships; by extracting specific and physically meaningful microscopic features such as "partial discharge peak value" and "specific frequency band vibration energy" from each individual signal, it can effectively capture early fault signs that are difficult to detect by the human ear or conventional monitoring, improving monitoring sensitivity; by "searching for cross-signal feature pairs" and forming "co-existing fault modes," it no longer views signal anomalies in isolation, but reveals the causal transmission chain within the equipment where one physical phenomenon triggers another. This improves the interpretability and reliability of fault diagnosis and reduces the false alarm rate. This invention quantifies the severity of faults using "energy equilibrium distortion degree" and "energy conversion efficiency deviation degree." By analyzing the combined change patterns of these indicators, it clearly pinpoints whether the fault lies in the "mechanical structure" or the "electromagnetic conversion link," achieving a leap from "problem discovery" to "problem localization," providing a direct basis for subsequent maintenance decisions. Calculating indicators such as "electromagnetic energy to mechanical energy conversion efficiency" directly reflects the equipment's ability to perform its core functions. The health index integrates multiple energy distortions, providing a quantitative reflection of the overall health status of the equipment, which is more comprehensive and reliable than single-parameter alarms. This invention integrates evidence at two levels—"physical layer" and "effect layer"—mimicking the diagnostic thinking of human experts: first, a rough location is determined through energy flow analysis, and then specific verification is achieved through refined feature analysis. This improves the accuracy of data analysis. The "similar analysis unit" matches current cases with historical maintenance cases, enabling case-based reasoning. It provides diagnostic conclusions and similar historical evidence, making the diagnostic results more convincing. Simultaneously, it can directly output remaining life predictions, achieving an advancement from condition monitoring to lifespan management. This invention not only diagnoses faults but also automatically provides feedforward optimization suggestions based on a "fault-parameter sensitivity table," proactively suppressing fault development and improving equipment reliability and efficiency. Through "stability monitoring" and "feedback calibration" mechanisms, system stability is immediately assessed after parameter adjustments, and automatic rollback is possible in case of anomalies, preventing secondary faults caused by improper optimization and ensuring the safety of the control strategy. A "self-learning process" transforms successful optimization cases into new knowledge stored in a historical database, enabling the system to accumulate operational experience and become increasingly intelligent with use. Subsequent encounters with similar situations will result in faster and more accurate diagnosis and optimization, achieving true "self-optimization." Attached Figure Description

[0020] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1This is a schematic diagram of the structure of a power transmission and distribution equipment operation state self-optimization system based on a large parameter model according to the present invention. Detailed Implementation

[0021] 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 are only some embodiments of the present invention, and not all embodiments. 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.

[0022] Please see Figure 1 The present invention provides a technical solution: a self-optimization system for the operating status of power transmission and distribution equipment based on a large parameter model, characterized in that the system includes an initial fault analysis module, a health analysis module, an evidence fusion diagnosis module, and a parameter adjustment module; The initial fault analysis module is used to collect equipment operating data and generate a spatiotemporal synchronization signal set; it uses signal processing and feature extraction technology to identify abnormal feature points or feature segments; it searches for feature pairs that appear across signals on the time axis; feature pairs with a number greater than a preset threshold are recorded as co-occurring fault pairs and co-occurring fault modes are generated. The initial fault analysis module includes a data acquisition and alignment unit, a feature extraction unit, and a fault symbiosis analysis unit. The data acquisition and alignment unit is used to collect the operating data of the equipment, including but not limited to current, vibration, sound and temperature data; the administrator presets a physical event sequence, using physical events as time reference points; the raw signals collected by all sensors are time-shifted and aligned to eliminate inherent timing differences caused by transmission and sampling, and generate a set of spatiotemporal synchronization signals; the physical event sequence includes but is not limited to the rising or falling edge of the coil current pulse during opening or closing operations and known mechanical shocks; The feature extraction unit is used to identify abnormal feature points or feature segments in each individual signal using signal processing and feature extraction techniques. These techniques include, but are not limited to, peak detection, wavelet packet decomposition, spectral analysis, and energy calculation. Abnormal feature points or feature segments include, but are not limited to, detecting partial discharge peaks in ultrasonic signals, extracting vibration energy from specific frequency bands in vibration signals, analyzing sudden increases in harmonic distortion rate from current signals, and capturing abnormal temperature rise rates from temperature signals. Extracting vibration energy from specific frequency bands in vibration signals includes the vibration energy of bearing fault characteristic frequency bands. The fault coexistence analysis unit is used to search for feature pairs that appear across signals on the time axis; extract feature pairs with a number greater than a preset threshold and record them as coexisting fault pairs, and generate coexisting fault patterns; coexisting fault patterns include precursor symptoms, subsequent symptoms, time intervals, and confidence levels; the patterns are classified and stored according to the physical fault types corresponding to the coexisting fault patterns; physical fault types include bearing outer ring faults, circuit breaker failure to operate, and insulation degradation.

[0023] The health analysis module is used to centrally analyze the energy information of various devices from spatiotemporal synchronization signals; analyze the energy balance distortion degree and energy conversion efficiency deviation degree, and integrate them to generate a health index; The health analysis module includes an energy analysis unit, a balance and efficiency analysis unit, and a health index unit; The energy analysis unit is used to centrally analyze the energy information of various devices from spatiotemporal synchronization signals. For rotating machinery or reciprocating vibration components, the vibration acceleration signal is de-DC and integrated twice to obtain the vibration mechanical energy. For electrical equipment or motor drive systems, instantaneous active power is calculated using synchronously acquired current and voltage signals. The electromagnetic energy consumed within an analysis cycle is obtained by integrating the instantaneous active power over that cycle. Temperature field information is acquired through deployed temperature sensors, combined with the equipment's thermal capacity parameters, and the temperature change rate is used to calculate the heat accumulation or consumption within the analysis cycle. The heat accumulation or consumption is generated by integrating the product of the temperature change rate and the equipment's thermal capacity parameters over time. The balance and efficiency analysis unit is used to analyze the energy balance distortion degree and the energy conversion efficiency deviation degree. The energy balance distortion degree is equal to the difference between the input electrical energy and the output energy divided by the input electrical energy; the output energy is equal to the sum of the output vibration mechanical energy, the lost heat energy, and the energy dissipated in the form of vibration. The health index unit is used to integrate several energy flow distortion indicators and generate a health index through weighted fusion. The several energy flow distortion indicators include the energy balance distortion degree and several energy conversion efficiency deviation degrees. Based on the abnormal patterns of the energy balance distortion degree and energy conversion efficiency deviation degree, combined with the energy flow path, the fault is located to a specific physical link.

[0024] For the conversion of electromagnetic energy to mechanical energy, the conversion efficiency is equal to the ratio of useful mechanical energy to electromagnetic energy; the useful mechanical energy is equal to the integral of the product of the torque and speed of the output shaft; the difference between the target conversion efficiency and the actual conversion efficiency is calculated and divided by the target conversion efficiency, and is recorded as the energy conversion efficiency deviation.

[0025] Locating the fault to a specific physical component includes: if the rate of change of energy equilibrium distortion is greater than the threshold, and the mechanical energy of vibration increases synchronously, while the change in energy conversion efficiency deviation is less than the threshold, the fault is identified as: loose or damaged mechanical structure, resulting in additional vibration energy dissipation, but the core conversion function has not been severely damaged; if the rate of change of energy conversion efficiency deviation is greater than the threshold, and the heat energy increases, the fault is identified as: abnormal electromagnetic conversion, increased winding loss or magnetic circuit fault, leading to decreased efficiency and overheating.

[0026] The evidence fusion diagnostic module is used to integrate and splice the health index with the fault location results to form physical layer evidence; guided by the fault location results, it traces back and focuses on specific signal features in the spatiotemporal synchronization signal set to extract effect layer evidence; and confirms the fault type and remaining service life based on the physical layer evidence and effect layer evidence. The evidence fusion diagnostic module includes a physical evidence integration unit, an effect evidence integration unit, and a similarity analysis unit; The physical evidence integration unit is used to integrate and splice the health index with the fault location results to form physical layer evidence; the fault location result is one or more tags used to identify the location where the energy flow anomaly occurs, including but not limited to electromagnetic conversion links, mechanical transmission links or insulator systems; The effect evidence integration unit is used to extract effect layer evidence by backtracking and focusing on specific signal features in the spatiotemporal synchronization signal set, guided by the fault location results. The similarity analysis unit is used to normalize the quantitative indicators in the physical layer evidence and the effect layer evidence and concatenate them into the current feature vector; in the pre-built historical maintenance case library, the similarity measurement algorithm is used to find several historical cases that are most similar to the current feature vector and form the nearest neighbor set; each case in the historical case library contains the historical feature vector, the finally confirmed fault type and the actual remaining service life; Based on the fault types identified in the nearest neighbor set, a weighted voting algorithm is used to determine the final fault type. The weight of the weighted vote is equal to the similarity between the historical feature vector of the historical case and the current feature vector. Based on the actual remaining useful life of the nearest neighbor set, a regression model is used to calculate the predicted remaining useful life of the current device and output the prediction confidence interval.

[0027] In extracting evidence of the effect layer, for faults located in the electromagnetic conversion link, the analysis includes: analyzing the electrodynamic force by calculating the square of the current signal; the electrodynamic force is equal to the product of the square of the current signal and the proportional coefficient; and analyzing the peak value, average electrodynamic force, and amplitude of the electrodynamic force; and integrating the discharge quantity, discharge frequency, and temperature data of the partial discharge signal, and constructing an insulation degradation assessment index by normalization and weighted fusion.

[0028] Based on the diagnostic results, the parameter adjustment module extracts the operating parameters that are most sensitive to the health of the equipment; based on the sensitivity coefficient of the operating parameters and combined with the current health index, it analyzes the suggested fine-tuning amount for each parameter; it monitors the stability of the system after parameter adjustment in real time, forms feedback calibration, and updates the historical maintenance case library.

[0029] The parameter adjustment module includes a feedforward fine-tuning monitoring unit and a feedback update unit; Based on the final fault type, the feedforward fine-tuning monitoring unit finds several operating parameters that are most effective in suppressing the fault of that type from a preset fault-parameter sensitivity table and obtains the sensitivity coefficient of each operating parameter; based on the sensitivity coefficient of the operating parameter and combined with the current health index, it analyzes the suggested fine-tuning amount of each parameter. Example 1: In this example, the system diagnosed a motor with a fault type of "early bearing lubrication failure," with a current health index HI=70 (baseline 100) and a downward trend. By querying the "Fault-Parameter Sensitivity Table," the two most effective operating parameters for suppression were identified: P1: Load current limit (sensitivity coefficient S1 = -0.5); P2: Cooling fan speed (sensitivity coefficient S2 = +0.3); In this embodiment, the allowable adjustment range of P1, P1_range, is 50A; the allowable adjustment range of P2, P2_range, is 1000RPM; and the global gain coefficient is 0.1. For P1: Fine-tuning is recommended. ; For P2: Fine-tuning is recommended. ; The feedback update unit is used to immediately start the stability monitoring program after parameter adjustment, and to track several system stability indicators in real time. The stability indicators include, but are not limited to, the fluctuation rate of output voltage or current, the total vibration of the shaft system, and the power oscillation amplitude. If the system stability indicators are detected to exceed the safety threshold, feedback calibration is triggered. The most recent parameter adjustment is rolled back proportionally according to a predetermined strategy. After the optimization operation has been verified for stability and proven to be effective, the system starts the self-learning process; the successful optimization case, including but not limited to: the initial symbiotic fault mode, the finally confirmed fault type, the parameter adjustment amount used, and the health index after adjustment, are added to the historical maintenance case library as new knowledge samples.

[0030] It is recommended that the fine-tuning amount equal to the product of the global gain coefficient, the health index deviation rate, the sensitivity coefficient allocation value, and the allowable adjustment range; the health index deviation rate equals the difference between the health index baseline value and the current health index divided by the health index baseline value; the sensitivity coefficient allocation value equals the ratio of the sensitivity coefficient of the current parameter to the sum of the absolute values ​​of the sensitivity coefficients of all adjusted parameters.

[0031] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0032] 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 invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A self-optimization system for the operating state of power transmission and distribution equipment based on a large parameter model, characterized in that, The system includes an initial fault analysis module, a health analysis module, an evidence fusion diagnostic module, and a parameter adjustment module; The initial fault analysis module is used to collect equipment operating data and generate a spatiotemporal synchronization signal set; it uses signal processing and feature extraction technology to identify abnormal feature points or feature segments; and it searches for feature pairs that appear across signals on the time axis. Feature pairs with a number greater than a preset threshold are identified as co-occurring fault pairs, and co-occurring fault modes are generated. The health analysis module is used to centrally analyze the energy information of each device from the spatiotemporal synchronization signal; Analyze the degree of distortion of energy equilibrium state and the deviation of energy conversion efficiency, and combine them to generate a health index; The evidence fusion diagnostic module is used to integrate and splice the health index with the fault location results to form physical layer evidence; Guided by the results of fault location, we trace back and focus on specific signal characteristics in the spatiotemporal synchronization signal set to extract evidence at the effect layer; based on the physical layer evidence and the effect layer evidence, we confirm the fault type and remaining service life. The parameter adjustment module extracts the operating parameters that are most sensitive to the health of the equipment based on the diagnostic results; and analyzes the suggested fine-tuning amount for each parameter based on the sensitivity coefficient of the operating parameters and the current health index. The system's stability is monitored in real time after parameter adjustments, feedback calibration is generated, and the historical maintenance case library is updated.

2. The self-optimization system for the operating status of power transmission and distribution equipment based on a large parameter model according to claim 1, characterized in that: The initial fault analysis module includes a data acquisition and alignment unit, a feature extraction unit, and a fault co-occurrence analysis unit. The data acquisition and alignment unit is used to acquire the device's operating data, which includes, but is not limited to, current, vibration, sound, and temperature data. The administrator presets a sequence of physical events, using physical events as time reference points; the raw signals collected by all sensors are time-shifted and aligned to eliminate inherent timing differences caused by transmission and sampling, generating a set of spatiotemporal synchronization signals; the sequence of physical events includes, but is not limited to, the rising or falling edge of the coil current pulse during opening or closing operations, as well as known mechanical shocks; The feature extraction unit is used to identify abnormal feature points or feature segments in each individual signal using signal processing and feature extraction techniques. These signal processing and feature extraction techniques include, but are not limited to, peak detection, wavelet packet decomposition, spectral analysis, and energy calculation. The abnormal feature points or feature segments include, but are not limited to, detecting partial discharge peak values ​​from ultrasonic signals, extracting vibration energy from specific frequency bands from vibration signals, analyzing sudden increases in harmonic distortion rate from current signals, and capturing abnormal temperature rise rates from temperature signals. The vibration energy extracted from specific frequency bands in vibration signals includes the vibration energy of bearing fault characteristic frequency bands. The fault co-occurrence analysis unit is used to search for feature pairs that appear across signals on the time axis; Feature pairs with a number greater than a preset threshold are extracted and recorded as co-existing fault pairs, and co-existing fault patterns are generated. The co-existing fault patterns include precursor symptoms, subsequent symptoms, time intervals, and confidence levels. The patterns are classified and stored according to the physical fault types corresponding to the co-existing fault patterns. The physical fault types include bearing outer ring faults, circuit breaker failure to operate, and insulation degradation.

3. The self-optimization system for the operating status of power transmission and distribution equipment based on a large parameter model according to claim 2, characterized in that: The health analysis module includes an energy analysis unit, a balance and efficiency analysis unit, and a health index unit; The energy analysis unit is used to centrally analyze the energy information of each device from the spatiotemporal synchronization signal. For rotating machinery or vibrating components with reciprocating motion, the vibration acceleration signal is deducted from the DC component and integrated twice to obtain the vibration mechanical energy. For electrical equipment or motor drive systems, the instantaneous active power is calculated using synchronously acquired current and voltage signals. The electromagnetic energy consumed in the analysis period is obtained by integrating the instantaneous active power within an analysis period. Temperature field information is obtained through deployed temperature sensors, combined with the heat capacity parameters of the equipment, and the heat accumulation or consumption within the analysis period is calculated using the temperature change rate. The heat accumulation or consumption is generated by integrating the product of the temperature change rate and the heat capacity parameters of the equipment over time. The balance and efficiency analysis unit is used to analyze the energy balance distortion degree and the energy conversion efficiency deviation degree. The energy balance distortion degree is equal to the difference between the input electrical energy and the output energy divided by the input electrical energy. The output energy is equal to the sum of the output vibration mechanical energy, the lost heat energy, and the energy dissipated in the form of vibration. The health index unit is used to integrate several energy flow distortion indicators and generate a health index through weighted fusion; the several energy flow distortion indicators include energy balance distortion degree and several energy conversion efficiency deviation degrees; based on the abnormal patterns of energy balance distortion degree and energy conversion efficiency deviation degree, combined with the energy flow path, the fault is located to a specific physical link.

4. The self-optimization system for the operating status of power transmission and distribution equipment based on a large parameter model according to claim 3, characterized in that: For the conversion of electromagnetic energy to mechanical energy, the conversion efficiency is equal to the ratio of useful mechanical energy to electromagnetic energy; the useful mechanical energy is equal to the integral of the product of the output shaft's torque and rotational speed. The difference between the target conversion efficiency and the actual conversion efficiency is calculated and then divided by the target conversion efficiency. This difference is recorded as the energy conversion efficiency deviation.

5. The self-optimization system for the operating status of power transmission and distribution equipment based on a large parameter model according to claim 3, characterized in that: Locating the fault to a specific physical component includes: if the rate of change of energy equilibrium distortion is greater than the threshold, and the mechanical energy of vibration increases synchronously, while the change in energy conversion efficiency deviation is less than the threshold, the fault is identified as: loose or damaged mechanical structure, resulting in additional vibration energy dissipation, but the core conversion function has not been severely damaged; if the rate of change of energy conversion efficiency deviation is greater than the threshold, and the heat energy increases, the fault is identified as: abnormal electromagnetic conversion, increased winding loss or magnetic circuit fault, leading to decreased efficiency and overheating.

6. The self-optimization system for the operating status of power transmission and distribution equipment based on a large parameter model according to claim 3, characterized in that: The evidence fusion diagnostic module includes a physical evidence integration unit, an effect evidence integration unit, and a similarity analysis unit; The physical evidence integration unit is used to integrate and splice the health index with the fault location results to form physical layer evidence; The fault location result is one or more tags used to identify the location where the energy flow abnormality occurs, including but not limited to electromagnetic conversion links, mechanical transmission links, or insulator systems; The effect evidence integration unit is used to extract effect layer evidence by backtracking and focusing on specific signal features in the spatiotemporal synchronization signal set, guided by the fault location results. The similarity analysis unit is used to normalize the quantitative indicators in the physical layer evidence and the effect layer evidence and then concatenate them into the current feature vector. In the pre-built historical maintenance case library, a similarity measurement algorithm is used to find several historical cases that are most similar to the current feature vector, forming a nearest neighbor set; each case in the historical case library contains a historical feature vector, the finally confirmed fault type, and the actual remaining service life; Based on the fault types identified in the nearest neighbor set, a weighted voting algorithm is used to determine the final fault type; the weight of the weighted vote is equal to the similarity between the historical feature vector of the historical case and the current feature vector. Based on the actual remaining useful life of the nearest neighbor set, the predicted remaining useful life of the current device is calculated using a regression model, and the prediction confidence interval is output.

7. The self-optimization system for the operating state of power transmission and distribution equipment based on a large parameter model according to claim 6, characterized in that: The analysis of the extracted effect layer evidence includes: analyzing electrodynamic force by calculating the square of the current signal; the electrodynamic force is equal to the product of the square of the current signal and the proportionality coefficient; and analyzing the peak value, average electrodynamic force, and amplitude of the electrodynamic force; and constructing an insulation degradation assessment index by fusing the discharge quantity, discharge frequency, and temperature data of the partial discharge signal after normalization and weighted fusion.

8. The self-optimization system for the operating state of power transmission and distribution equipment based on a large parameter model according to claim 6, characterized in that: The parameter adjustment module includes a feedforward fine-tuning monitoring unit and a feedback update unit; The feedforward fine-tuning monitoring unit, based on the final fault type, finds several operating parameters that are most effective in suppressing the fault of that type from a preset fault-parameter sensitivity table, and obtains the sensitivity coefficient of each operating parameter; based on the sensitivity coefficient of the operating parameters and combined with the current health index, it analyzes the suggested fine-tuning amount for each parameter. The feedback update unit is used to immediately start the stability monitoring program after parameter adjustment, and to track several system stability indicators in real time. The stability indicators include, but are not limited to, the fluctuation rate of output voltage or current, the total vibration of the shaft system, and the power oscillation amplitude. If the system stability indicators are detected to exceed the safety threshold, feedback calibration is triggered. The most recent parameter adjustment is rolled back proportionally according to a predetermined strategy. After the optimization operations have been verified for stability and proven effective, the system initiates a self-learning process. The successful optimization cases, including but not limited to: the initial symbiotic failure mode, the finally confirmed failure type, the parameter adjustment amount used, and the health index after adjustment, will be added to the historical maintenance case library as new knowledge samples.

9. A self-optimization system for the operating state of power transmission and distribution equipment based on a large parameter model according to claim 8, characterized in that: The suggested fine-tuning amount is equal to the product of the global gain coefficient, the health index deviation rate, the sensitivity coefficient allocation value, and the allowable adjustment range; the health index deviation rate is equal to the difference between the health index baseline value and the current health index divided by the health index baseline value. The sensitivity coefficient allocation value is equal to the ratio of the sensitivity coefficient of the current parameter to the sum of the absolute values ​​of the sensitivity coefficients of all adjusted parameters.

Images