A multi-time scale early warning control method and an electromechanical controller

Abstract

The application relates to the technical field of electromechanical controllers, and discloses a multi-time-scale early warning control method and an electromechanical controller. By constructing a pulse impact and creep loss coupling cause-effect mapping reference library, the method realizes effective decoupling and recoupling of fault features of different time scales, so that the original nonlinear coupling relationship can be accurately characterized. When the time sequence cause-effect locking grouping matching technology is applied, the information interference problem among fast, medium and slow time scale fault features in the prior art is overcome, the cross-scale information processing capability is improved, the corrected creep loss judgment envelope curve is generated, the invalid pulse signals of electromagnetic clutter are effectively filtered out by combining the pulse adaptive constraint threshold, the purity and accuracy of the monitoring data are ensured, the coupling control instructions are generated based on the credible fault coupling feature cluster, the feedback mechanism is formed, the coupling cause-effect mapping reference library is continuously optimized, and the early warning accuracy and response speed of the equipment are improved.

Description

Technical Field

[0001] This invention relates to the field of electromechanical controller technology, specifically to a multi-timescale early warning control method and an electromechanical controller. Background Technology

[0002] Multi-timescale early warning for electromechanical controllers is a hierarchical early warning mechanism for fault prevention throughout the entire life cycle of electromechanical systems. Its core is to divide multiple timescales according to the speed of fault evolution and the scope of impact, and to design differentiated monitoring frequencies, feature extraction and early warning strategies for different scales, so as to achieve full-coverage early warning capabilities from rapid protection against instantaneous faults to prediction of gradual degradation trends.

[0003] Current technologies for multi-timescale early warning mainly employ hierarchical threshold and trend fusion early warning techniques. The underlying principle is to break down the fault evolution characteristics of electromechanical systems according to time scales, and then apply differentiated algorithms for hard threshold determination, dynamic statistical analysis, and long-term trend prediction to different time dimensions (fast, medium, and slow). This is combined with multi-scale monitoring data to achieve hierarchical early warning decisions. However, due to the nonlinear coupling relationship between fault characteristics at different time scales, the interaction between long-term mechanical wear and instantaneous impact is difficult to accurately represent using existing models. Furthermore, high-frequency noise easily interferes with the degradation trend analysis at low-frequency scales, resulting in insufficient cross-scale information coupling processing capabilities and a prominent problem of false alarms in equipment early warning systems. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a multi-timescale early warning and control method, which solves the problems in the background technology.

[0005] To achieve the above objectives, the present invention provides a multi-timescale early warning control method, comprising: Acquire multiple fast-sequence pulse impact characteristic data, slow-sequence creep loss characteristic data, and a benchmark time-sequence sample dataset under stable and normal operating conditions generated by the electromechanical controller. Construct a benchmark library for coupled causal mapping between pulse impact and creep loss based on the benchmark time-sequence sample dataset. Based on the coupled causal mapping benchmark library, multiple fast-time pulse impact characteristic data and slow-time creep loss characteristic data are subjected to time-series causal locking group matching to obtain time-series locking matching results, and a pulse and creep real-time time-series locking mapping link is generated based on the time-series locking matching results. An initial creep loss timing determination envelope curve is generated based on the coupled causal mapping reference library, and the envelope offset compensation amount is obtained based on the pulse and creep real-time timing locking mapping link to correct the initial creep loss timing determination envelope curve, thus obtaining a corrected creep loss determination envelope curve. Based on the corrected creep loss determination envelope curve, an adaptive pulse constraint threshold is obtained to filter out electromagnetic noise and invalid pulse signals generated by the pulse acquisition channel, resulting in a clean and effective pulse feature sequence. Based on the clean and effective pulse feature sequence and the real-time creep loss time series data corresponding to the corrected creep loss determination envelope curve, a two-way time series mutual verification and false detection is performed to obtain a reliable fault coupling feature cluster. Based on the trusted fault coupling feature cluster matching coupling causal mapping benchmark library, the time-series closed-loop causal verification rules are used to generate coupling control instructions in a hierarchical manner, and the coupling control instructions are sent to the electromechanical controller to collect the actual load coupling parameters and pulse and creep dual time-series feature offsets before and after the electromechanical controller receives the coupling control instructions in real time. Based on the actual load coupling parameters and dual-time-series feature offsets, control feedback timing samples are obtained, and the coupling causal mapping benchmark library is iteratively optimized based on multiple sets of control feedback timing samples summarized in the periodic data to synchronously update the initial creep loss timing judgment envelope curve.

[0006] The present invention also provides a multi-timescale early warning control system, comprising: The module is used to acquire multiple fast-sequence pulse impact characteristic data, slow-sequence creep loss characteristic data and benchmark time-sequence sample dataset under stable and normal operating conditions generated by the operation of the electromechanical controller, and to construct a benchmark library for coupled causal mapping of pulse impact and creep loss based on the benchmark time-sequence sample dataset. The binding module is used to perform time-series causal locking group matching on multiple fast-series pulse impact characteristic data and slow-series creep loss characteristic data according to the coupled causal mapping benchmark library, obtain time-series locking matching results, and generate pulse and creep real-time time-series locking mapping links according to the time-series locking matching results. The correction module is used to generate an initial creep loss timing determination envelope curve based on the coupled causal mapping reference library, and to obtain the envelope offset compensation amount based on the pulse and creep real-time timing locking mapping link to correct the initial creep loss timing determination envelope curve, thereby obtaining a corrected creep loss determination envelope curve. The verification module is used to obtain the pulse adaptive constraint threshold based on the corrected creep loss judgment envelope curve to filter out electromagnetic noise invalid pulse signals generated by the pulse acquisition channel, obtain a clean and effective pulse feature sequence, and perform bidirectional time series mutual verification and false detection based on the clean and effective pulse feature sequence and the real-time creep loss time series data corresponding to the corrected creep loss judgment envelope curve to obtain a reliable fault coupling feature cluster. The generation module is used to generate coupling control instructions in a hierarchical manner according to the temporal closed-loop causal verification rules in the coupled causal mapping benchmark library based on the trusted fault coupling feature cluster, and to send the coupling control instructions to the electromechanical controller to collect in real time the actual load coupling parameters and the pulse and creep dual temporal feature offsets before and after the electromechanical controller receives the coupling control instructions. The optimization module is used to obtain the control feedback timing sample based on the actual load coupling parameters and dual timing feature offset, and to iteratively optimize the coupled causal mapping benchmark library based on multiple sets of control feedback timing samples summarized in the period to synchronously update the initial creep loss timing judgment envelope curve.

[0007] The present invention also provides an electromechanical controller, the electromechanical controller comprising methods for performing the above-described multi-timescale early warning control method.

[0008] Compared with the prior art, the present invention provides a multi-timescale early warning and control method, which has the following beneficial effects: This invention constructs a causal mapping benchmark library for coupled pulse impact and creep loss. This method effectively decouples and recouples fault features at different time scales, enabling accurate characterization of previously existing nonlinear coupling relationships. By applying time-series causal locking grouping matching technology, it overcomes the information interference problem between fault features at fast, medium, and slow time scales in existing technologies, improving the cross-scale capability of information processing. By generating a corrected creep loss judgment envelope curve and combining it with pulse adaptive constraint thresholds, it effectively filters out electromagnetic clutter and invalid pulse signals, ensuring the purity and accuracy of monitoring data. Based on a reliable fault coupling feature cluster, it generates coupled control commands, forming a feedback mechanism to continuously optimize the causal mapping benchmark library. This solves the false alarm problem caused by the lack of effective coupling considerations in traditional multi-time-scale early warning technologies, improving the early warning accuracy and response speed of equipment. It has significant technological advancements and practical value. Attached Figure Description

[0009] Figure 1 This is a flowchart of the multi-timescale early warning control method according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of the multi-timescale early warning control system according to an embodiment of the present invention. Detailed Implementation

[0010] 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.

[0011] Example 1

[0012] Please see Figure 1 This invention provides a multi-timescale early warning control method, comprising: S1. Acquire multiple fast-sequence pulse impact characteristic data, slow-sequence creep loss characteristic data, and a benchmark time-sequence sample dataset under stable and normal operating conditions generated by the operation of the electromechanical controller, and construct a benchmark library for coupled causal mapping of pulse impact and creep loss based on the benchmark time-sequence sample dataset. This step acquires fast and slow time-series characteristic data of the electromechanical controller under different operating conditions, and constructs a benchmark library for the coupling causal mapping of pulse impact and creep loss based on the benchmark time-series sample dataset. This effectively solves the problem of insufficient separation and identification of fault features at multiple time scales in the existing technology. By refining data acquisition, it can comprehensively capture the dynamic changes of the electromechanical system under complex operating conditions, providing an accurate data foundation for subsequent steps, thereby significantly improving the accuracy and reliability of fault early warning. S2. Based on the coupled causal mapping reference library, perform time-series causal locking group matching on multiple fast-series pulse impact characteristic data and slow-series creep loss characteristic data to obtain time-series locking matching results, and generate pulse and creep real-time time-series locking mapping links based on the time-series locking matching results. This step generates a real-time temporal locking mapping link between pulse and creep by performing temporal causal locking grouping and matching on fast-series pulse impact data and slow-series creep loss data. This improves the coupling analysis capability of multi-timescale data. Compared with the traditional hard threshold mode, this technology allows for more flexible capture of nonlinear relationships between different time scales, reduces false alarms caused by multi-timescale coupling, and thus improves the system's response capability and accuracy to sudden events. S3. Generate an initial creep loss timing determination envelope curve based on the coupled causal mapping reference library, and obtain the envelope offset compensation amount based on the pulse and creep real-time timing locking mapping link to correct the initial creep loss timing determination envelope curve, thereby obtaining a corrected creep loss determination envelope curve. By generating an initial creep loss time-series judgment envelope curve and performing offset compensation correction, this step can accurately reflect the creep loss changes caused by impact. This technical feature solves the problem that traditional methods cannot consider the misleading effect of instantaneous impact on long-term wear trends, thereby introducing higher accuracy in modeling, effectively reducing the risk of false alarms and missed alarms caused by misjudged faults, and improving the reliable operation level of the system. S4. Based on the corrected creep loss determination envelope curve, obtain the pulse adaptive constraint threshold to filter out electromagnetic noise and invalid pulse signals generated by the pulse acquisition channel, obtain a clean and effective pulse feature sequence, and perform bidirectional time series mutual verification and false detection based on the clean and effective pulse feature sequence and the real-time creep loss time series data corresponding to the corrected creep loss determination envelope curve to obtain a reliable fault coupling feature cluster. This step removes invalid pulse signals affected by electromagnetic clutter by setting pulse adaptive constraint thresholds, thereby obtaining a clean and effective pulse feature sequence, which improves the reliability of signal processing and enables accurate identification of effective pulse features even in dynamic situations. It also enhances the accuracy of bidirectional time-series mutual verification and counterfeit detection. This level of automation and intelligence in information processing provides an efficient real-time monitoring function and provides important data support for early identification of potential faults. S5. Generate coupling control instructions in a hierarchical manner according to the temporal closed-loop causal verification rules in the coupled causal mapping benchmark library based on the trusted fault coupling feature cluster, and send the coupling control instructions to the electromechanical controller to collect the actual load coupling parameters and the pulse and creep dual temporal feature offsets before and after the electromechanical controller receives the coupling control instructions in real time. This step generates hierarchical coupling control instructions based on the causal verification rules in the reliable fault coupling feature cluster matching coupling causal mapping benchmark library, and sends them to the electromechanical controller. This enables the system to achieve adaptive control capability, dynamically respond to sudden events and operation adjustments, and overcomes the shortcomings of the lack of real-time control and feedback mechanisms in the existing technology. As a result, more accurate load management and fault early warning are achieved in actual operation. S6. Obtain the control feedback timing sample based on the actual load coupling parameters and dual timing feature offset, and iteratively optimize the coupling causal mapping benchmark library based on the multiple sets of control feedback timing samples summarized in the period to synchronously update the initial creep loss timing judgment envelope curve. By acquiring time-series samples of regulatory feedback and iteratively optimizing the coupled causal mapping benchmark library based on these samples, this step effectively overcomes the shortcomings of existing algorithms that are highly dependent on time-series samples, and enhances the self-learning ability of the causal mapping library. This continuous optimization mechanism not only ensures the real-time updating of the envelope curve, but also promotes the long-term stability of the model, enabling the fault warning system to maintain high adaptability and accuracy when facing new operating conditions.

[0013] In one embodiment, step S1, which involves constructing a baseline library for coupled causal mapping between pulse shock and creep loss based on the baseline time-series sample dataset, includes: S11. The reference time series sample dataset is split into a pulse impact time series sample subset and a creep loss time series sample subset, and time series paired coupling sample groups under the same operating sequence are selected according to the pulse impact time series sample subset and the creep loss time series sample subset; pulse impact amplitude and cumulative pulse trigger frequency data are extracted from multiple time series paired coupling sample groups, and the distribution range of full sample parameters is statistically analyzed according to the pulse impact amplitude and cumulative pulse trigger frequency data; the distribution range of full sample parameters is equally divided into multiple threshold boundary values ​​to delineate the pulse and creep coupling interval boundaries of multiple levels under the fault-free operating condition of the electromechanical controller; This step addresses the shortcomings of current technologies in multi-timescale analysis by splitting the benchmark time-series sample dataset into pulse impact time-series sample subsets and creep loss time-series sample subsets, and by screening time-series paired coupling sample groups under the same operating time series. It solves the problem of data classification and coupling relationship processing in multi-timescale analysis by accurately extracting and statistically analyzing the distribution range of parameters in the whole sample and forming multiple gear boundary values ​​under fault-free operating conditions. This can effectively delineate the pulse and creep coupling region, providing a clear classification basis for subsequent analysis, improving the coupling identification accuracy of fault features at multiple timescales, and effectively distinguishing the impact of instantaneous impact and long-term wear on electromechanical systems on the time scale, thereby reducing the risk of misjudgment and missed judgment and improving the overall reliability of the monitoring system. S12. Classify all time-series paired coupling sample groups according to the boundaries of each coupling interval to obtain multiple gear interval classified coupling samples. Extract the measured creep loss increment, cumulative pulse trigger count, and gear average pulse impact amplitude corresponding to each gear interval classified coupling sample to solve for the corresponding creep acceleration coupling coefficient. The calculation formula is as follows: ;in, Indicates the first The creep acceleration coupling coefficient for each gear level Indicates the first The measured creep loss increment for each gear. Indicates the first The cumulative number of pulse triggers for each gear range. Indicates the first The average pulse impact amplitude of each gear; In this step, the time-series paired coupling sample groups are classified according to the coupling interval boundaries. The measured creep loss increment, cumulative pulse trigger count, and average pulse impact amplitude of each gear interval are extracted, thereby providing a data-driven calculation of the creep acceleration coupling coefficient. This technical feature emphasizes the integration of the causal relationship between creep loss and pulse impact amplitude, which helps to clarify the dynamic behavior of mechanical wear under different working conditions. By introducing the calculation formula of the creep acceleration coupling coefficient, the impact of instantaneous impact on long-term creep loss can be accurately quantified, effectively overcoming the problem that existing models fail to accurately characterize nonlinear coupling relationships. This achieves higher scientificity and effectiveness in data analysis and solves the problem of false alarms caused by unclear causal relationships in existing technologies. S13. Based on the multiple coupling interval boundaries and creep acceleration coupling coefficients, a benchmark library calibration parameter set is formed. Based on the coupling interval boundary in the benchmark library calibration parameter set, the creep acceleration coupling coefficient of the same level is matched and bound to obtain a single-level causal mapping unit. Multiple single-level causal mapping units are integrated and encapsulated in an orderly manner to construct and generate a benchmark library for coupling causal mapping between pulse impact and creep loss. This step, by summarizing multiple coupling interval boundaries and creep acceleration coupling coefficients, forms a benchmark library calibration parameter set and binds creep acceleration coupling coefficients of the same level to generate a single-level causal mapping unit. This embodies an efficient and structured knowledge storage and dynamic mapping mechanism, which not only improves the accuracy and operability of coupled causal mapping but also achieves the standardization and modularization of model parameters, facilitating rapid response and adjustment in the future. By integrating multi-stage causal mapping units, the constructed benchmark library provides a systematic solution that effectively addresses the problem of insufficient cross-scale information coupling processing capability in existing technologies for multi-level prediction. It ensures adaptability and responsiveness to new data in the future and further enhances the intelligence and adaptability of the fault early warning system.

[0014] In one embodiment, step S2, which involves performing time-series causal locking grouping and matching on multiple fast-series pulse impact characteristic data and slow-series creep loss characteristic data according to the coupled causal mapping benchmark library to obtain time-series locking matching results, and generating a pulse and creep real-time time-series locking mapping link based on the time-series locking matching results, includes: S21. Extract pulse causal classification criteria and creep loss causal classification criteria from the coupled causal mapping benchmark library, and obtain effective pulse triggering threshold and significant creep loss threshold respectively based on the pulse causal classification criteria and creep loss causal classification criteria; remove small-amplitude electromagnetic disturbance pulses with pulse impact amplitudes less than the effective pulse triggering threshold from multiple fast time-series pulse impact feature data based on the effective pulse triggering threshold to obtain the effective pulse time-series dataset; By extracting pulse causal classification criteria and creep loss causal classification criteria based on the coupled causal mapping benchmark library, and setting effective pulse triggering threshold and significant creep loss threshold respectively, noise and invalid signals in the feature data are effectively eliminated. The threshold setting ensures that only pulse and creep data with significant features are included in subsequent analysis, reducing mismatches and unnecessary computational burden in signal processing, improving data quality and reliability. In multi-timescale monitoring, it can effectively cope with the impact of high-frequency noise on low-frequency trend analysis, thereby enhancing the coupling processing capability of cross-scale information. S22. Based on the significant creep loss threshold, normal temperature drift loss data with real-time creep rates less than the significant creep loss threshold are removed from multiple slow-time-series creep loss feature data to obtain a creep loss response dataset; the acquisition timestamps attached to each feature data in the effective pulse time-series dataset and the creep loss response dataset are extracted, and the corresponding feature entries in the effective pulse time-series dataset and the creep loss response dataset in the same time-series interval are matched according to the acquisition timestamps to complete the time-series locking and pairing, and the time-series locking and pairing result is obtained. That is, when the absolute value of the difference between the acquisition time of a single fast-scale transient feature and the acquisition time of a single slow-scale degradation feature is less than the preset time-series tolerance, it is determined to be a feature entry that can be bound and paired. By removing normal temperature drift loss data whose real-time creep rate is below the significant creep loss threshold, the extracted creep response dataset is ensured to mainly contain effective data related to actual physical state changes. By accurately matching the timestamps within the dataset, effective coupling of transient features (pulse) and slow-changing features (creep) can be achieved. This step reduces information cross-interference, improves the accuracy of feature matching, and solves the interference problem caused by the nonlinear coupling relationship between features of different time scales in the existing technology, thus providing a more reliable basis for equipment condition monitoring and fault early warning. S23. Construct a time-series coupling index linked list based on the time-series locking matching results, and generate a pulse and creep real-time time-series locking mapping link by connecting all time-stamped pairing coupling features in the time-series locking matching results in series according to the time-series coupling index linked list. The step of constructing the time-series coupling index linked list is to split the transient data number, degenerate data number and time-series label of each pairing feature entry in the time-series locking matching results, and fill the form fields according to the transient data number, degenerate data number and time-series label to construct the time-series coupling index linked list. By constructing a time-coupling index linked list and concatenating all time-scale paired coupling features, a real-time time-series locking mapping link for pulses and creep is generated. This step effectively integrates feature data from different time scales. This mechanism allows the real-time monitoring system to accurately track subtle changes in equipment status, enabling it to dynamically adapt to health assessment needs under complex operating conditions. This structure design based on the coupling index linked list overcomes the problem of information isolation between features at different time scales in traditional methods, demonstrating the advanced nature and operability of real-time monitoring technology for multiple time scales. Compared with the traditional hierarchical threshold and trend fusion method, this step provides a new data integration method, significantly improving the dynamic monitoring capability of equipment status evolution and greatly enhancing the accuracy of fault alarms.

[0015] In one embodiment, step S3, which involves generating an initial creep loss timing determination envelope curve based on the coupled causal mapping reference library and obtaining an envelope offset compensation amount based on the pulse and creep real-time timing locking mapping link to correct the initial creep loss timing determination envelope curve, and obtaining a corrected creep loss determination envelope curve, includes: S31. Extract the steady-state creep rate calibration parameters corresponding to each gear position according to the coupled causal mapping benchmark library, and fit the equipment benchmark creep loss evolution curve according to multiple steady-state creep rate calibration parameters; delineate the upper limit degradation threshold and the lower limit degradation threshold according to the benchmark creep loss evolution curve, and form a standard creep loss floating band according to the upper limit degradation threshold and the lower limit degradation threshold; extract the interval center degradation rate of the standard creep loss floating band, and construct a benchmark trend line according to the interval center degradation rate and combine it with the upper limit degradation threshold and the lower limit degradation threshold to form a boundary constraint, and generate the initial creep loss time series judgment envelope curve; By extracting steady-state creep rate calibration parameters from the coupled causal mapping benchmark library and constructing a benchmark creep loss evolution curve, accurate modeling of equipment degradation characteristics is achieved. This modeling process can calibrate precise upper and lower degradation thresholds and form a standard creep loss floating band, providing excellent boundary constraints for equipment condition monitoring. It effectively avoids the simplification of time scales in traditional hierarchical threshold methods, and can better capture nonlinear coupling relationships under multiple time scales, thereby accurately characterizing the dynamic interaction between long-term mechanical wear and instantaneous impact, reducing the probability of false alarms caused by signal noise, and improving the accuracy of early warning. S32. Retrieve the real-time pulse sequence of time-stamped locking according to the pulse and creep real-time timing locking mapping link, and calculate the total number of pulse triggers within the period according to the real-time pulse sequence; retrieve the creep acceleration coupling coefficient of the corresponding gear in the coupling causal mapping reference library according to the total number of pulse triggers, and calculate the envelope offset compensation amount according to the creep acceleration coupling coefficient and the total number of pulse triggers, wherein the calculation formula is: ;in, This indicates the envelope offset compensation amount. Indicates the first The creep acceleration coupling coefficient for each gear level Indicates the total cumulative number of pulse triggers; By analyzing real-time pulse sequences, statistically analyzing the total number of cumulative pulse triggers, and retrieving the coupled causal mapping benchmark library, the system can quickly acquire and calculate the envelope offset compensation amount. The technical feature of this process is that it is closely coupled with the real-time performance of the equipment, and can dynamically reflect the degradation characteristics of the equipment in a rapidly developing state. By introducing a creep acceleration coupling coefficient, this step integrates the relationship between the pulse trigger amount and the equipment degradation, providing a new measurement and adjustment method. It effectively breaks through the bottleneck of traditional monitoring systems when fusing degradation characteristics at different time scales, improves the filtering ability of high-frequency noise, thereby enhancing the ability to process fault characteristics of the equipment at various time scales, and effectively improving the response speed and accuracy of fault warning. S33. The entire segment of the initial creep loss timing judgment envelope curve is longitudinally shifted and adjusted according to the envelope offset compensation amount, and the corrected creep loss judgment envelope curve is generated according to the curve parameters after the shift is completed. By longitudinally shifting the initial creep loss time-series judgment envelope curve, the adjustment strategy based on real-time feedback is reflected. By directly utilizing the calculated envelope offset compensation amount, the time-series curve of equipment condition monitoring is effectively corrected. This not only ensures the accuracy of monitoring data but also enhances the system's sensitivity to fault characteristics at different time scales through the feedback mechanism. This adjustment strategy enables more accurate fault trend prediction, avoids the shortcomings of traditional models in handling long-term degradation trends, overcomes the interference of high-frequency noise on low-frequency information analysis, improves the coupling processing capability of cross-scale information, and significantly reduces the probability of false alarms in equipment warnings.

[0016] In one embodiment, step S4, which involves obtaining a pulse adaptive constraint threshold based on the corrected creep loss determination envelope curve to filter out electromagnetic clutter and invalid pulse signals generated by the pulse acquisition channel, thereby obtaining a clean and effective pulse feature sequence, and performing bidirectional time-series mutual verification and falsification based on the clean and effective pulse feature sequence and the real-time creep loss time-series data corresponding to the corrected creep loss determination envelope curve to obtain a reliable fault coupling feature cluster, includes: S41. Extract the multi-segment degradation rate boundary values ​​of the corrected creep loss judgment envelope curve, and decompose the built-in creep grading calibration parameters according to the degradation rate boundary values. Divide the degradation rate intervals into multiple levels according to the creep grading calibration parameters, and determine the real-time loss health level of the electromechanical controller according to the degradation rate interval into which the current real-time creep rate of the electromechanical controller falls. Retrieve the anti-pulse impact benchmark parameters of the corresponding level from the coupled causal mapping benchmark library according to the real-time loss health level, and calculate the corresponding pulse adaptive constraint threshold according to each anti-pulse impact benchmark parameter. The calculation formula is as follows: ;in, Indicates the first A pulse adaptive constraint threshold for a real-time loss health level. Indicates the first The anti-pulse shock reference parameters for each real-time wear and tear health level. Indicates the operating condition coupling correction coefficient; In this step, by dividing the degradation rate range into multiple levels, the real-time wear and health status of the electromechanical controller under different operating conditions can be accurately reflected. This not only enables real-time monitoring of the controller's status but also effectively identifies fault characteristics, achieving accurate interpretation of nonlinear coupling relationships at different time scales. By retrieving the anti-pulse impact reference parameters and calculating the adaptive constraint threshold, the sensitivity to impact noise can be adaptively adjusted, thereby enhancing the reliability of the fault early warning system and avoiding false alarms caused by external noise in conventional technologies. The establishment of a dynamic mapping relationship between real-time data and historical references helps to enhance the timeliness and accuracy of equipment status assessment, solving the shortcomings of existing technologies in multi-time scale monitoring. S42. Each of the fast-sequence pulse impact feature data is compared one by one with the corresponding pulse adaptive constraint threshold to filter out invalid pulse signals of electromagnetic clutter in the acquisition link that are below the constraint threshold, and the pure and effective pulse feature sequence is retained; according to the pulse and creep real-time timing locking mapping link, the real-time creep loss timing data corresponding to the correction creep loss judgment envelope curve in the timing matching of the pure and effective pulse feature sequence is retrieved, and the abnormal pulse impact entries in the pure and effective pulse feature sequence and the abnormal creep loss entries in the real-time creep loss timing data are extracted respectively. This step effectively filters out invalid electromagnetic clutter pulse signals below the threshold by comparing the fast-sequence pulse impact feature data with the corresponding pulse adaptive constraint threshold one by one, retaining the pure and valid pulse feature sequence. By adopting a precise threshold judgment mechanism, the false alarm and false alarm rates can be significantly reduced, ensuring that a signal is generated only when specific and real fault characteristics appear. Based on this core technical feature, abnormal pulse impact entries and real-time creep loss entries are further extracted, laying a solid foundation for subsequent verification and anomaly detection. This association and integration of heterogeneous data effectively promotes the coupling processing of information across time scales, enhances the accuracy of the overall early warning system, and provides practical technical support for the development of a new generation of intelligent fault diagnosis systems. S43. Perform forward timing verification on the abnormal pulse impact entries, remove pseudo-pulse data of working condition disturbances without matching creep increments, perform reverse timing verification on the abnormal creep loss entries, remove pseudo-loss data of sensor drift lacking pulse impact support, and integrate to obtain a reliable fault coupling feature cluster. In this step, the process of performing forward timing verification on abnormal pulse impact items and reverse timing verification on abnormal creep loss items can efficiently eliminate invalid pseudo-data, thereby achieving high-reliability extraction of fault features and improving the system's resistance to noise and interference. Especially in nonlinear dynamic environments with multiple time scales, it can extract stable fault features, breaking through the limitations of existing technologies in representing the complex interaction between long-term wear and instantaneous impact. Through this deep integration, hysteresis effect and dynamic response are effectively integrated to form a more complete fault coupling feature cluster, providing high-quality data support for scientific decision-making. This not only improves the timeliness and accuracy of equipment health assessment, but also provides a reference for the health management of more complex equipment in the future, thereby promoting the advancement of intelligent maintenance technology.

[0017] In one embodiment, step S5, which involves generating coupling control instructions hierarchically based on the temporal closed-loop causal verification rules in the trusted fault coupling feature cluster matching coupling causal mapping benchmark library, includes: S51. Extract the real-time pulse impact amplitude and synchronous creep loss increment according to the reliable fault coupling feature cluster, and retrieve the preset time-series closed-loop reference calibration parameters according to the coupling causal mapping reference library. The time-series closed-loop reference calibration parameters include the time-series closed-loop reference impact limit, the time-series closed-loop reference creep limit, the pulse coupling weight coefficient, and the creep coupling weight coefficient. In this step, by extracting the real-time pulse impact amplitude and synchronous creep loss increment from the reliable fault coupling feature cluster, and combining them with the time-series closed-loop benchmark calibration parameters in the coupled causal mapping benchmark library for analysis, the ability to identify and process fault features is improved. This process can establish a real-time dynamic feedback mechanism, which not only ensures accurate monitoring under complex working conditions, but also provides a detailed data foundation for the subsequent generation of coupled control commands. By retrieving parameters including the time-series closed-loop benchmark impact limit, creep limit, and coupling weight coefficient, this step effectively reduces the impact of insufficient modeling of the nonlinear characteristics of the system in traditional methods, thereby realizing the accurate expression of the dynamic relationship between mechanical wear and instantaneous impact. This enables the system to be optimized and adjusted in real time for different working conditions, directly improving the response speed and accuracy of traditional multi-timescale early warning technology, and avoiding the risk of equipment false alarms and failures. S52. Calculate the fault coupling matching index based on the real-time pulse impact amplitude, synchronous creep loss increment, and timing closed-loop reference calibration parameters, wherein the calculation formula is: ;in, Indicates the fault coupling matching index. Represents the pulse coupling weighting coefficient. Indicates the real-time pulse impact amplitude. Indicates the timing closed-loop reference impact limit. This represents the creep coupling weighting coefficient. This represents the increment of synchronous creep loss. Indicates the time-series closed-loop reference creep limit; In this step, the fault coupling matching index is calculated using a formula. Combined with real-time pulse impact amplitude, synchronous creep loss increment, and time-series closed-loop benchmark calibration parameters, a quantitative fault matching assessment method is provided. By integrating the two characteristics of pulse impact and creep loss through weighted coefficients, a comprehensive indicator value is formed, which can effectively reflect the overall health status of the system. Compared with previous methods, this calculation method demonstrates innovation in high-level nonlinear coupling data processing. It not only breaks the limitation of static thresholds but also nonlinearly integrates instantaneous and continuous effects, achieving a more profound physical correlation analysis of fault evolution characteristics at different time scales. The introduction of the fault coupling matching index provides a strong basis for subsequent fault warning level classification, improves the overall system's ability to identify complex fault modes, ensures that corresponding countermeasures can be quickly initiated when potential faults occur, and greatly enhances the system's safety and reliability. S53. The fault coupling matching index is compared with the preset three-level matching degree boundary thresholds in sequence, and the fault warning levels are divided into three levels: mild, moderate and severe according to the range of the values. The preset time-series coupling handling strategies are matched according to the three levels of fault warning levels. That is, the real-time online monitoring strategy is matched according to the mild warning level, the load reduction control strategy is matched according to the moderate warning level, and the emergency shutdown protection strategy is matched according to the severe warning level. The three types of graded coupling control commands of online monitoring, load reduction operation and emergency shutdown are generated according to the preset time-series coupling handling strategies of each level. By comparing the fault coupling matching index with a preset matching degree threshold and classifying the fault warning levels, this step effectively achieves precise adjustment of the fault response strategy. This hierarchical mechanism enables the rapid and accurate implementation of response measures for mild, moderate, and severe faults, employing various dynamic control strategies such as online monitoring, load reduction, and emergency shutdown. By establishing a hierarchical warning mechanism, the problem of false alarms and missed alarms caused by cross-timescale feature coupling in traditional methods is effectively solved. The implementation of the hierarchical handling strategy not only ensures the emergency response capability of the equipment under different fault levels but also significantly improves the intelligence and flexibility of fault handling, thereby providing more scientific and comprehensive decision support for equipment management. Through this intelligent control method, equipment operation and maintenance costs can be reduced, safety assurance levels can be improved, and the overall operating efficiency of the electromechanical system can be enhanced.

[0018] In one embodiment, step S6, which involves obtaining control feedback timing samples based on the actual load coupling parameters and dual-time-series feature offsets, and iteratively optimizing the coupling causal mapping benchmark library based on multiple sets of periodically summarized control feedback timing samples to synchronously update the initial creep loss timing determination envelope curve, includes: S61. Extract the load coupling correction coefficient of the electromechanical controller according to the actual load coupling parameters, and extract the pulse timing offset and creep loss timing offset according to the pulse and creep dual timing feature offsets before and after regulation; bind and aggregate the load coupling correction coefficient, pulse timing offset and creep loss timing offset to generate a single regulation feedback timing sample, and summarize multiple regulation feedback timing samples in the whole cycle according to the preset update cycle to obtain the cycle summary dataset; By extracting the load coupling correction coefficient of the electromechanical controller based on the actual load coupling parameters, and then combining it with the impact pulse and creep time-series characteristic offset, accurate acquisition and comprehensive analysis of control feedback can be achieved. By binding and aggregating the load coupling correction coefficient, pulse time-series offset, and creep loss time-series offset, a single control feedback time-series sample is formed. This technical feature effectively solves the problem of insufficient expression of nonlinear coupling relationship of fault characteristics at multiple time scales in the existing technology. By summarizing multiple control feedback samples throughout the entire cycle, the transient and long-term trends in equipment operation can be clearly captured, providing more accurate and reliable basic data for subsequent fault prediction and early warning. This reduces false alarms caused by insufficient cross-scale information coupling processing capabilities and has significant practical significance in improving the level of data integration, thus promoting the development of mechanical system health monitoring technology towards higher precision. S62. Utilize the periodic summary dataset to statistically analyze the data distribution range of transient amplitude and degradation rate at each gear level, and redefine the boundary values ​​based on the data distribution range to correct the boundaries of the multi-gear coupling interval and the creep acceleration coupling coefficient within the coupled causal mapping benchmark library, so as to achieve iterative optimization of the coupled causal mapping benchmark library and obtain the optimized coupled causal mapping benchmark library. By utilizing the statistical analysis of the periodic summary dataset, the data distribution range of transient amplitude and degradation rate at each gear level is clarified, and the boundary values ​​are redefined based on these ranges. This corrects the multi-gear coupling interval boundaries of the coupling causal mapping benchmark library, fully demonstrating the data-driven dynamic optimization capability. It effectively overcomes the problem of inaccurate threshold setting caused by nonlinear coupling relationships in existing technologies. By adjusting the coupling interval boundaries and creep acceleration coupling coefficient in real time, this step not only improves the sensitivity and accuracy of equipment fault judgment, but also takes into account the interference of high-frequency noise on low-frequency trend analysis. It can more accurately reflect the actual health status of the equipment, improve the reliability of the equipment early warning system, and lay a solid foundation for realizing online monitoring and early fault warning. S63. Extract the updated steady-state creep rate calibration parameters from the optimized coupled causal mapping benchmark library, and refit the new standard creep loss floating band based on the updated steady-state creep rate calibration parameters to obtain a new degradation fluctuation range. Update the initial creep loss time series judgment envelope curve synchronously based on the new degradation fluctuation range. Based on the optimized coupled causal mapping benchmark library, the updated steady-state creep rate calibration parameters are extracted and refitted to obtain a new standard creep loss floating band. This achieves real-time dynamic correction and adaptive updates, effectively addressing model inaccuracies caused by the interaction between long-term wear and instantaneous impacts. It overcomes the problems of delayed response and false alarms in current early warning systems. By synchronously updating the initial creep loss time series judgment envelope curve, a more scientific information processing method is provided, significantly improving the intelligence level of fault detection. This facilitates faster fault diagnosis, enhances the timeliness of fault maintenance, and ultimately promotes the safety and reliability of equipment maintenance.

[0019] Example 2

[0020] Please see Figure 2 The present invention also provides a multi-timescale early warning control system, comprising: The module is used to acquire multiple fast-sequence pulse impact characteristic data, slow-sequence creep loss characteristic data and benchmark time-sequence sample dataset under stable and normal operating conditions generated by the operation of the electromechanical controller, and to construct a benchmark library for coupled causal mapping of pulse impact and creep loss based on the benchmark time-sequence sample dataset. The binding module is used to perform time-series causal locking group matching on multiple fast-series pulse impact characteristic data and slow-series creep loss characteristic data according to the coupled causal mapping benchmark library, obtain time-series locking matching results, and generate pulse and creep real-time time-series locking mapping links according to the time-series locking matching results. The correction module is used to generate an initial creep loss timing determination envelope curve based on the coupled causal mapping reference library, and to obtain the envelope offset compensation amount based on the pulse and creep real-time timing locking mapping link to correct the initial creep loss timing determination envelope curve, thereby obtaining a corrected creep loss determination envelope curve. The verification module is used to obtain the pulse adaptive constraint threshold based on the corrected creep loss judgment envelope curve to filter out electromagnetic noise invalid pulse signals generated by the pulse acquisition channel, obtain a clean and effective pulse feature sequence, and perform bidirectional time series mutual verification and false detection based on the clean and effective pulse feature sequence and the real-time creep loss time series data corresponding to the corrected creep loss judgment envelope curve to obtain a reliable fault coupling feature cluster. The generation module is used to generate coupling control instructions in a hierarchical manner according to the temporal closed-loop causal verification rules in the coupled causal mapping benchmark library based on the trusted fault coupling feature cluster, and to send the coupling control instructions to the electromechanical controller to collect in real time the actual load coupling parameters and the pulse and creep dual temporal feature offsets before and after the electromechanical controller receives the coupling control instructions. The optimization module is used to obtain the control feedback timing sample based on the actual load coupling parameters and dual timing feature offset, and to iteratively optimize the coupled causal mapping benchmark library based on multiple sets of control feedback timing samples summarized in the period to synchronously update the initial creep loss timing judgment envelope curve.

[0021] In one embodiment, the building module includes: The filtering unit is used to split the benchmark time series sample dataset into a pulse impact time series sample subset and a creep loss time series sample subset, and to filter time series paired coupling sample groups under the same running time according to the pulse impact time series sample subset and the creep loss time series sample subset. The delineation unit is used to extract pulse impact amplitude and cumulative pulse trigger frequency data based on multiple sets of time-series paired coupling sample groups, and to delineate the pulse and creep coupling interval boundaries of multiple gears under fault-free operating conditions of the electromechanical controller based on the pulse impact amplitude and cumulative pulse trigger frequency data. The solving unit is used to classify all time-series paired coupling sample groups according to the boundary of each coupling interval to obtain multiple gear interval classified coupling samples, and extract the measured creep loss increment, the cumulative pulse triggering number of intervals and the gear average pulse impact amplitude corresponding to each gear interval classified coupling sample to solve the corresponding creep acceleration coupling coefficient. The construction unit is used to form a benchmark library calibration parameter set by summarizing multiple coupling interval boundaries and creep acceleration coupling coefficients, and to construct a benchmark library for coupling causal mapping between pulse shock and creep loss based on the benchmark library calibration parameter set.

[0022] Example 3

[0023] The present invention also provides an electromechanical controller, the electromechanical controller comprising methods for performing the above-described multi-timescale early warning control method.

[0024] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A multi-time scale early warning control method, characterized in that, include: Acquire multiple fast-sequence pulse impact characteristic data, slow-sequence creep loss characteristic data, and a benchmark time-sequence sample dataset under stable and normal operating conditions generated by the electromechanical controller. Construct a benchmark library for coupled causal mapping between pulse impact and creep loss based on the benchmark time-sequence sample dataset. Based on the coupled causal mapping benchmark library, multiple fast-time pulse impact characteristic data and slow-time creep loss characteristic data are subjected to time-series causal locking group matching to obtain time-series locking matching results, and a pulse and creep real-time time-series locking mapping link is generated based on the time-series locking matching results. An initial creep loss timing determination envelope curve is generated based on the coupled causal mapping reference library, and the envelope offset compensation amount is obtained based on the pulse and creep real-time timing locking mapping link to correct the initial creep loss timing determination envelope curve, thus obtaining a corrected creep loss determination envelope curve. Based on the corrected creep loss determination envelope curve, an adaptive pulse constraint threshold is obtained to filter out electromagnetic noise and invalid pulse signals generated by the pulse acquisition channel, resulting in a clean and effective pulse feature sequence. Based on the clean and effective pulse feature sequence and the real-time creep loss time series data corresponding to the corrected creep loss determination envelope curve, a two-way time series mutual verification and false detection is performed to obtain a reliable fault coupling feature cluster. Based on the trusted fault coupling feature cluster matching coupling causal mapping benchmark library, the time-series closed-loop causal verification rules are used to generate coupling control instructions in a hierarchical manner, and the coupling control instructions are sent to the electromechanical controller to collect the actual load coupling parameters and pulse and creep dual time-series feature offsets before and after the electromechanical controller receives the coupling control instructions in real time. Based on the actual load coupling parameters and dual-time-series feature offsets, control feedback timing samples are obtained, and the coupling causal mapping benchmark library is iteratively optimized based on multiple sets of control feedback timing samples summarized in the periodic data to synchronously update the initial creep loss timing judgment envelope curve.

2. The multi-timescale early warning control method according to claim 1, characterized in that, The step of constructing a baseline library for coupled causal mapping between pulse shock and creep loss based on the baseline time-series sample dataset includes: The baseline time series sample dataset is split into a pulse impact time series sample subset and a creep loss time series sample subset, and time series paired coupling sample groups under the same running time are selected based on the pulse impact time series sample subset and the creep loss time series sample subset. Based on multiple sets of time-paired coupling sample groups, pulse impact amplitude and cumulative pulse trigger frequency data are extracted, and the pulse and creep coupling interval boundaries of multiple gears under fault-free operating conditions of the electromechanical controller are delineated based on the pulse impact amplitude and cumulative pulse trigger frequency data. All time-series paired coupling sample groups are classified according to the boundary of each coupling interval to obtain multiple gear interval classified coupling samples. The measured creep loss increment, cumulative pulse triggering number of intervals and gear average pulse impact amplitude corresponding to each gear interval classified coupling sample are extracted to solve the corresponding creep acceleration coupling coefficient. A benchmark library calibration parameter set is formed by summarizing multiple coupling interval boundaries and creep acceleration coupling coefficients, and a benchmark library for coupling causal mapping between pulse shock and creep loss is constructed based on the benchmark library calibration parameter set.

3. The multi-timescale early warning control method according to claim 1, characterized in that, The step of performing time-series causal locking grouping and matching on multiple fast-series pulse impact characteristic data and slow-series creep loss characteristic data according to the coupled causal mapping benchmark library to obtain time-series locking matching results, and generating a pulse and creep real-time time-series locking mapping link based on the time-series locking matching results, includes: Based on the coupled causal mapping benchmark library, pulse causal classification criteria and creep loss causal classification criteria are extracted, and effective pulse triggering threshold and significant creep loss threshold are obtained based on the pulse causal classification criteria and creep loss causal classification criteria, respectively. Based on the effective pulse triggering threshold, small electromagnetic disturbance pulses with pulse impact amplitudes less than the effective pulse triggering threshold are removed from multiple fast-sequence pulse impact feature data to obtain the effective pulse time series dataset. Based on the significant creep loss threshold, normal temperature drift loss data with real-time creep rates lower than the significant creep loss threshold are removed from multiple slow-series creep loss characteristic data to obtain the creep loss response dataset. Extract the acquisition timestamps attached to each feature data in the effective pulse time series dataset and creep loss response dataset, and match the corresponding feature entries in the effective pulse time series dataset and creep loss response dataset within the same time series interval according to the acquisition timestamps to complete the time series locking pairing and obtain the time series locking matching result; Based on the timing lock-in matching results, a timing coupling index linked list is constructed, and all time-scaled pairing coupling features within the timing lock-in matching results are connected in series according to the timing coupling index linked list to generate a pulse and creep real-time timing lock-in mapping link.

4. The multi-timescale early warning control method according to claim 1, characterized in that, The step of generating an initial creep loss timing determination envelope curve based on the coupled causal mapping reference library, and obtaining an envelope offset compensation amount based on the pulse and creep real-time timing locking mapping link to correct the initial creep loss timing determination envelope curve, thereby obtaining a corrected creep loss determination envelope curve, includes: The steady-state creep rate calibration parameters corresponding to each gear position are extracted according to the coupled causal mapping benchmark library, and the equipment benchmark creep loss evolution curve is fitted according to multiple steady-state creep rate calibration parameters. The standard creep loss floating band is divided according to the benchmark creep loss evolution curve, and the initial creep loss timing judgment envelope curve is generated according to the standard creep loss floating band. The real-time pulse sequence of time-stamped locking is retrieved according to the pulse and creep real-time timing locking mapping link, and the total number of pulse triggers within the period is calculated according to the real-time pulse sequence. The creep acceleration coupling coefficient of the corresponding gear in the coupled causal mapping reference library is retrieved according to the total cumulative pulse triggering amount, and the envelope offset compensation amount is obtained according to the creep acceleration coupling coefficient and the total cumulative pulse triggering amount. The initial creep loss timing determination envelope curve is corrected based on the envelope offset compensation amount to obtain the corrected creep loss determination envelope curve.

5. The multi-timescale early warning control method according to claim 1, characterized in that, The steps of obtaining a pulse adaptive constraint threshold based on the corrected creep loss determination envelope curve to filter out electromagnetic clutter and invalid pulse signals generated by the pulse acquisition channel, obtaining a clean and effective pulse feature sequence, and performing bidirectional time series mutual verification and false detection based on the clean and effective pulse feature sequence and the real-time creep loss time series data corresponding to the corrected creep loss determination envelope curve to obtain a reliable fault coupling feature cluster include: The built-in creep grading calibration parameters are analyzed based on the correction creep loss determination envelope curve, and the real-time loss health level of the electromechanical controller is divided according to the creep grading calibration parameters. According to the real-time loss health level, retrieve the anti-pulse impact benchmark parameters of the corresponding level in the coupled causal mapping benchmark library, and obtain the corresponding pulse adaptive constraint threshold according to each anti-pulse impact benchmark parameter. Each fast-sequence pulse impact feature data is compared one by one with the corresponding pulse adaptive constraint threshold to filter out invalid pulse signals of electromagnetic clutter in the acquisition link that are below the constraint threshold, and the pure and effective pulse feature sequence is retained. According to the pulse and creep real-time timing locking mapping link, retrieve the real-time creep loss timing data corresponding to the correction creep loss determination envelope curve in the timing matching of the pure effective pulse feature sequence, and extract the abnormal pulse impact entries in the pure effective pulse feature sequence and the abnormal creep loss entries in the real-time creep loss timing data respectively. Forward timing verification is performed on the abnormal pulse impact entries to remove pseudo-pulse data of operating condition disturbances without corresponding creep increments. Reverse timing verification is performed on the abnormal creep loss entries to remove pseudo-loss data of sensor drift lacking pulse impact support. The results are then integrated to obtain a reliable fault coupling feature cluster.

6. The multi-timescale early warning control method according to claim 1, characterized in that, The step of generating coupling control instructions hierarchically based on the temporal closed-loop causal verification rules in the coupled causal mapping benchmark library according to the trusted fault coupling feature cluster includes: The real-time pulse impact amplitude and synchronous creep loss increment are extracted based on the reliable fault coupling feature cluster, and the preset time-series closed-loop reference calibration parameters are retrieved based on the coupled causal mapping reference library. The fault coupling matching index is obtained based on the real-time pulse impact amplitude, synchronous creep loss increment and time-series closed-loop reference calibration parameters, and three levels of fault warning are divided according to the fault coupling matching index. Based on the three levels of fault warning, a preset timing coupling handling strategy is matched, and three types of hierarchical coupling control commands are generated according to the preset timing coupling handling strategy of each level: online monitoring, load reduction operation, and emergency shutdown.

7. The multi-timescale early warning control method according to claim 1, characterized in that, The step of obtaining the control feedback time series samples based on the actual load coupling parameters and dual-time series feature offsets, and iteratively optimizing the coupling causal mapping benchmark library based on multiple sets of control feedback time series samples summarized periodically to synchronously update the initial creep loss time series determination envelope curve includes: The load coupling correction coefficient of the electromechanical controller is extracted based on the actual load coupling parameters, and the pulse timing offset and creep loss timing offset are extracted based on the pulse and creep dual timing characteristic offsets before and after regulation. The load coupling correction coefficient, pulse timing offset, and creep loss timing offset are bound and aggregated to generate a single control feedback timing sample. Multiple control feedback timing samples within the entire cycle are then aggregated according to a preset update cycle to obtain a cycle summary dataset. By using the periodic summary dataset to correct the boundaries of multi-level coupling intervals and creep acceleration coupling coefficients in the coupled causal mapping benchmark library, iterative optimization of the coupled causal mapping benchmark library is achieved, resulting in an optimized coupled causal mapping benchmark library. A new standard creep loss floating band is obtained by refitting the optimized coupled causal mapping benchmark library, resulting in a new degradation fluctuation range. The initial creep loss time series determination envelope curve is then updated synchronously based on the new degradation fluctuation range.

8. A multi-timescale early warning control system, characterized in that, include: The module is used to acquire multiple fast-sequence pulse impact characteristic data, slow-sequence creep loss characteristic data and benchmark time-sequence sample dataset under stable and normal operating conditions generated by the operation of the electromechanical controller, and to construct a benchmark library for coupled causal mapping of pulse impact and creep loss based on the benchmark time-sequence sample dataset. The binding module is used to perform time-series causal locking group matching on multiple fast-series pulse impact characteristic data and slow-series creep loss characteristic data according to the coupled causal mapping benchmark library, obtain time-series locking matching results, and generate pulse and creep real-time time-series locking mapping links according to the time-series locking matching results. The correction module is used to generate an initial creep loss timing determination envelope curve based on the coupled causal mapping reference library, and to obtain the envelope offset compensation amount based on the pulse and creep real-time timing locking mapping link to correct the initial creep loss timing determination envelope curve, thereby obtaining a corrected creep loss determination envelope curve. The verification module is used to obtain the pulse adaptive constraint threshold based on the corrected creep loss judgment envelope curve to filter out electromagnetic noise invalid pulse signals generated by the pulse acquisition channel, obtain a clean and effective pulse feature sequence, and perform bidirectional time series mutual verification and false detection based on the clean and effective pulse feature sequence and the real-time creep loss time series data corresponding to the corrected creep loss judgment envelope curve to obtain a reliable fault coupling feature cluster. The generation module is used to generate coupling control instructions in a hierarchical manner according to the temporal closed-loop causal verification rules in the coupled causal mapping benchmark library based on the trusted fault coupling feature cluster, and to send the coupling control instructions to the electromechanical controller to collect in real time the actual load coupling parameters and the pulse and creep dual temporal feature offsets before and after the electromechanical controller receives the coupling control instructions. The optimization module is used to obtain the control feedback timing sample based on the actual load coupling parameters and dual timing feature offset, and to iteratively optimize the coupled causal mapping benchmark library based on multiple sets of control feedback timing samples summarized in the period to synchronously update the initial creep loss timing judgment envelope curve.

9. A multi-timescale early warning control system according to claim 8, characterized in that, The building module includes: The filtering unit is used to split the benchmark time series sample dataset into a pulse impact time series sample subset and a creep loss time series sample subset, and to filter time series paired coupling sample groups under the same running time according to the pulse impact time series sample subset and the creep loss time series sample subset. The delineation unit is used to extract pulse impact amplitude and cumulative pulse trigger frequency data based on multiple sets of time-series paired coupling sample groups, and to delineate the pulse and creep coupling interval boundaries of multiple gears under fault-free operating conditions of the electromechanical controller based on the pulse impact amplitude and cumulative pulse trigger frequency data. The solving unit is used to classify all time-series paired coupling sample groups according to the boundary of each coupling interval to obtain multiple gear interval classified coupling samples, and extract the measured creep loss increment, the cumulative pulse triggering number of intervals and the gear average pulse impact amplitude corresponding to each gear interval classified coupling sample to solve the corresponding creep acceleration coupling coefficient. The construction unit is used to form a benchmark library calibration parameter set by summarizing multiple coupling interval boundaries and creep acceleration coupling coefficients, and to construct a benchmark library for coupling causal mapping between pulse shock and creep loss based on the benchmark library calibration parameter set.

10. An electromechanical controller, characterized in that, The electromechanical controller includes a method for performing the multi-timescale early warning control method according to any one of claims 1-7.

Images