Method and system for monitoring running state of diesel generator

Abstract

The invention provides a diesel generator operation state monitoring method and system. The method comprises the following steps: extracting time-frequency domain feature information used for representing the stability of the operation state of a diesel generator from multi-source monitoring parameters of the diesel generator under an operation condition; constructing an anomaly identification layer and a fault mechanism-data hybrid drive reasoning layer for the diesel generator; determining a suspected abnormal region of the diesel generator under an operation condition through the abnormal recognition layer and the time-frequency domain feature information, and determining a fault reasoning result of an abnormal type, a fault cause and a fault severity degree of the diesel generator through the fault mechanism-data hybrid drive reasoning layer and the suspected abnormal region; and determining an operation health degree score of the diesel generator, and performing graded diagnosis and early warning on the operation state of the diesel generator according to the operation health degree score and a fault reasoning result. By adopting the scheme of the invention, the abnormal associated area can be accurately identified based on the complex operation data of the diesel generator so as to monitor the operation state of the diesel generator.

Description

Technical Field

[0001] This application relates to the field of operational status monitoring technology, and more specifically, to a method and system for monitoring the operational status of a diesel generator. Background Technology

[0002] Operational status monitoring is a technical means that collects, processes, and analyzes various key parameters generated by equipment during operation in real time or periodically to accurately determine its current working status, assess performance degradation trends, and identify potential anomalies, providing objective basis for early fault warning, maintenance decision optimization, and equipment lifecycle management.

[0003] As an important backup power device, the operating status of diesel generators directly affects the stable operation of supporting systems. During long-term operation, various electrical and mechanical operating data will exhibit complex changing characteristics. At the same time, the operation of the equipment is susceptible to various factors such as load fluctuations, environmental conditions, and component aging. Traditional monitoring methods often rely on single parameter judgment or human experience, which makes it difficult to comprehensively capture potential anomalies in the operation of the equipment, accurately locate the location of the anomaly, analyze the cause of the anomaly, and quantitatively assess the overall health status of the equipment. This can easily lead to missed or misjudged anomalies, and the inability to predict failure risks in advance, affecting the reliable operation and maintenance efficiency of the equipment. Therefore, how to accurately identify anomaly-related areas based on the complex operating data of diesel generators to monitor the operating status of diesel generators has become a problem faced by the industry. Summary of the Invention

[0004] This application provides a method and system for monitoring the operating status of a diesel generator, which can accurately identify abnormal correlation areas based on the complex operating data of the diesel generator to monitor the operating status of the diesel generator.

[0005] In a first aspect, this application provides a method for monitoring the operating status of a diesel generator, comprising the following steps: The multi-source monitoring parameters of the diesel generator under operating conditions are obtained, including the electrical parameters and mechanical operating parameters of the diesel generator. Time-frequency domain analysis is performed on the multi-source monitoring parameters to obtain time-frequency domain feature information that characterizes the stability of the diesel generator's operating state; Based on the mapping relationship between the fault mechanism knowledge base, historical multi-source monitoring parameters and historical abnormal data corresponding to the diesel generator, an anomaly identification layer for pre-diagnosis of abnormal areas of the diesel generator and a fault mechanism-data hybrid driving reasoning layer for reasoning about anomaly types and fault causes are constructed. The anomaly identification layer and the time-frequency domain feature information are used to perform anomaly location and pre-diagnosis of the operating status of the diesel generator, thereby obtaining the suspected anomaly area of ​​the diesel generator under the operating conditions. The fault mechanism-data hybrid driven inference layer performs fault inference on the distortion feature set corresponding to the suspected anomaly area, thereby obtaining the fault inference results of the anomaly type, fault cause and fault severity of the diesel generator. The operating health score of the diesel generator is generated based on the temporal degradation pattern of the multi-source monitoring parameters and the anomaly diagnosis results. The operating health score and the fault inference results are used to perform graded diagnosis and early warning of the operating status of the diesel generator.

[0006] In some embodiments, performing time-frequency domain analysis on the multi-source monitoring parameters to obtain time-frequency domain feature information characterizing the stability of the diesel generator's operating state specifically includes: The multi-source monitoring parameters are preprocessed to obtain preprocessed multi-source monitoring parameters; Time-frequency domain features are extracted from the preprocessed multi-source monitoring parameters to obtain time-frequency domain feature information characterizing the operating state of the diesel generator.

[0007] In some embodiments, based on the mapping relationship between the fault mechanism knowledge base corresponding to the diesel generator, historical multi-source monitoring parameters, and historical abnormal data, an anomaly identification layer for pre-diagnosis of abnormal areas of the diesel generator and a fault mechanism-data hybrid driven inference layer for inferring anomaly types and fault causes are constructed, specifically including: Construct a knowledge base for the fault mechanisms of diesel generators; Obtain the historical multi-source monitoring parameters and historical anomaly data corresponding to the diesel generator; Determine the mapping relationship between the fault mechanism knowledge base, the historical multi-source monitoring parameters, and the historical abnormal data; Based on the mapping relationship, an anomaly identification layer is constructed to perform pre-diagnosis of abnormal areas of the diesel generator; Obtain historical anomaly type data for the diesel generator; A fault mechanism-data hybrid inference layer is constructed by using the mapping relationship and the historical anomaly type data to infer the anomaly type and fault cause of the diesel generator.

[0008] In some embodiments, the anomaly identification layer and the time-frequency domain feature information are used to perform anomaly localization and pre-diagnosis of the operating state of the diesel generator, and the suspected anomaly areas of the diesel generator under operating conditions are obtained, specifically including: Based on the anomaly identification layer, anomaly diagnosis is performed on the time-frequency domain feature information to obtain anomaly diagnosis information in the time-frequency domain feature information; Determine the abnormal coupling judgment rules between the time-domain stability characteristics and frequency-domain energy distribution characteristics of the diesel generator under operating conditions; The abnormality diagnosis information and the abnormality coupling judgment rules are used to locate the abnormality in the operating state of the diesel generator, thereby obtaining the suspected abnormal area of ​​the diesel generator under the operating conditions.

[0009] In some embodiments, the fault inference layer, driven by the fault mechanism-data hybrid approach, performs fault inference on the distortion feature set corresponding to the suspected abnormal region to obtain the fault inference results of the diesel generator anomaly type, fault cause, and fault severity. Specifically, this includes: Determine the distortion feature set corresponding to the suspected abnormal region in the time-frequency domain feature information; The distortion feature set is input into the fault mechanism-data hybrid driven inference layer, and fault inference of the abnormal type or abnormal cause is performed on the distortion feature set to obtain the fault inference results of the abnormal type, fault cause and fault severity of the diesel generator.

[0010] In some embodiments, generating the operational health score of the diesel generator based on the temporal degradation pattern and anomaly diagnosis results of the multi-source monitoring parameters over time specifically includes: Determine the temporal degradation pattern and anomaly diagnosis results of the multi-source monitoring parameters in the time dimension; Determine the baseline timing degradation pattern of the diesel generator; The operational health score of the diesel generator is determined based on the time-series degradation pattern, the anomaly diagnosis results, and the baseline time-series degradation pattern.

[0011] In some embodiments, the graded diagnosis and early warning of the operating status of the diesel generator based on the operating health score and the fault reasoning result specifically includes: The operational health score and the fault reasoning result are deeply fused to obtain diagnostic fusion information of the diesel generator's operating status; Determine the hierarchical diagnostic and early warning decision-making process for the operating status of the diesel generator; The diagnostic fusion information is diagnosed based on the hierarchical diagnostic early warning decision, and the operating status of the diesel generator is diagnosed and warned based on the diagnostic results.

[0012] Secondly, this application provides a diesel generator operating status monitoring system, comprising: The acquisition module is used to acquire multi-source monitoring parameters of the diesel generator under operating conditions. The multi-source monitoring parameters include the electrical parameters and mechanical operating parameters of the diesel generator. The processing module is used to perform time-frequency domain analysis on the multi-source monitoring parameters to obtain time-frequency domain feature information that characterizes the stability of the diesel generator's operating state. The processing module is also used to construct an anomaly identification layer for pre-diagnosis of anomaly areas of the diesel generator and a fault mechanism-data hybrid driving reasoning layer for reasoning about anomaly types and fault causes, based on the mapping relationship between the fault mechanism knowledge base corresponding to the diesel generator, historical multi-source monitoring parameters and historical abnormal data. The processing module is also used to perform anomaly location pre-diagnosis on the operating status of the diesel generator through the anomaly identification layer and the time-frequency domain feature information, to obtain the suspected anomaly area of ​​the diesel generator under the operating conditions, and the fault mechanism-data hybrid driven inference layer performs fault inference on the distortion feature set corresponding to the suspected anomaly area to obtain the fault inference results of the anomaly type, fault cause and fault severity of the diesel generator. The execution module is used to generate an operational health score for the diesel generator based on the temporal degradation pattern and anomaly diagnosis results of the multi-source monitoring parameters in the time dimension, and to perform graded diagnosis and early warning of the operational status of the diesel generator based on the operational health score and the fault inference results.

[0013] Thirdly, this application provides a computer device, the computer device including a memory and a processor, the memory storing code, and the processor being configured to acquire the code and execute the above-described diesel generator operating status monitoring method.

[0014] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for monitoring the operating status of a diesel generator.

[0015] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects: The diesel generator operation status monitoring method and system provided in this application first acquires multi-source monitoring parameters of the diesel generator under operating conditions, including electrical parameters and mechanical operating parameters of the diesel generator; performs time-frequency domain analysis on the multi-source monitoring parameters to obtain time-frequency domain feature information characterizing the stability of the diesel generator's operating status; based on the mapping relationship between the fault mechanism knowledge base corresponding to the diesel generator, historical multi-source monitoring parameters, and historical abnormal data, constructs an anomaly identification layer for pre-diagnosis of abnormal areas of the diesel generator and a fault mechanism-data hybrid driven reasoning layer for inferring anomaly types and fault causes; through the... The anomaly identification layer and the time-frequency domain feature information are used to perform anomaly localization and pre-diagnosis of the diesel generator's operating status, obtaining suspected anomaly areas of the diesel generator under operating conditions. The fault mechanism-data hybrid driven inference layer then performs fault inference on the distortion feature set corresponding to the suspected anomaly areas, obtaining fault inference results of the diesel generator's anomaly type, fault cause, and fault severity. Based on the temporal degradation law of the multi-source monitoring parameters in the time dimension and the anomaly diagnosis results, an operating health score of the diesel generator is generated. The operating health score and the fault inference results are used to perform graded diagnosis and early warning of the diesel generator's operating status.

[0016] Therefore, in the process of monitoring the operating status of diesel generators, this application firstly collects multi-source parameters of two core data types: electrical parameters and mechanical operating parameters of the diesel generator. Time-frequency domain analysis is then performed on these multi-source parameters to extract characteristic information representing the stability of the operating status. This transforms the original time-domain and frequency-domain data into feature data with clear physical meaning, effectively extracting key state information from the data and filtering out noise interference, providing accurate feature support for anomaly identification and health assessment. Secondly, based on historical data mapping relationships, an anomaly identification layer and a fault mechanism-data hybrid driven inference layer are constructed respectively. This achieves layered pre-diagnosis of anomaly location and cause analysis, decomposing the complex anomaly diagnosis task into two independent and complementary modules, improving the clarity and professionalism of the diagnostic logic, and providing a basis for accurate anomaly location. The system provides a standardized analytical framework for identifying common anomalies and their root causes. Then, an anomaly identification layer locates suspected anomaly areas, followed by a fault mechanism-data hybrid-driven inference layer that performs fault reasoning on these areas. This achieves a progressive diagnosis from area location to type / cause identification, accurately pinpointing the anomaly-related areas and specific anomaly types of diesel generators. This solves the core problems of fuzzy anomaly identification and inaccurate location in complex operating data. Finally, a health score is generated based on the temporal degradation patterns of multi-source parameters. The score is then integrated with the fault reasoning results for comprehensive hierarchical diagnosis and early warning. This quantifies the equipment's operational health status and combines anomaly qualitative results to achieve multi-dimensional early warning, improving the comprehensiveness, accuracy, and practicality of the warnings. This enables full-cycle, intelligent monitoring and early warning of equipment operating status. Using this solution, anomaly-related areas can be accurately identified based on complex diesel generator operating data to monitor the generator's operating status. Attached Figure Description

[0017] Figure 1 This is an exemplary flowchart of a diesel generator operating status monitoring method according to some embodiments of this application; Figure 2 This is an exemplary flowchart illustrating the determination of time-frequency domain feature information according to some embodiments of this application; Figure 3 This is an exemplary flowchart illustrating the determination of suspected abnormal regions according to some embodiments of this application; Figure 4 This is a schematic diagram of the structure of a diesel generator operating status monitoring system according to some embodiments of this application; Figure 5 This is a schematic diagram of the structure of a computer device for implementing a diesel generator operating status monitoring method according to some embodiments of this application. Detailed Implementation

[0018] To better understand the technical solution of this application, the technical solution of this application will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0019] refer to Figure 1 The figure is an exemplary flowchart of a diesel generator operating status monitoring method according to some embodiments of this application. The diesel generator operating status monitoring method mainly includes the following steps: In step 101, multi-source monitoring parameters of the diesel generator under operating conditions are obtained, including the electrical parameters and mechanical operating parameters of the diesel generator.

[0020] In practice, when the diesel generator is in normal load operation, multiple types of sensors deployed on the generator body, power output terminal, cooling system, fuel supply system and lubrication system are used to synchronously and in real time collect multi-source monitoring parameters during the generator operation. Among them, electrical parameters include output voltage, output current, active power, reactive power, grid frequency and harmonic content, while mechanical operating parameters include engine speed, oil pressure, oil temperature, coolant temperature, cylinder vibration amplitude, exhaust temperature and fuel pressure.

[0021] In step 102, time-frequency domain analysis is performed on the multi-source monitoring parameters to obtain time-frequency domain feature information that characterizes the stability of the diesel generator's operating state.

[0022] In some embodiments, reference Figure 2 The figure is an exemplary flowchart for determining time-frequency domain feature information in some embodiments of this application. In this embodiment, the time-frequency domain analysis of the multi-source monitoring parameters to obtain time-frequency domain feature information for characterizing the stability of the diesel generator's operating state can be achieved by the following steps: In step 1021, the multi-source monitoring parameters are preprocessed to obtain preprocessed multi-source monitoring parameters; In step 1022, time-frequency domain features are extracted from the preprocessed multi-source monitoring parameters to obtain time-frequency domain feature information characterizing the operating state of the diesel generator.

[0023] In specific implementation, the multi-source monitoring parameters are preprocessed to obtain the preprocessed multi-source monitoring parameters, which can be achieved in the following ways: First, outlier removal is performed, using the 3σ Laida criterion to identify and replace abnormal jump values ​​and outliers in the parameter sequence, while removing invalid data caused by sensor failure; Second, data filtering is performed, using Butterworth low-pass filtering to remove high-frequency noise for non-stationary signals such as mechanical vibration and cylinder temperature, using moving average filtering to smooth data fluctuations for electrical parameters such as voltage and current, and using median filtering to remove pulse interference for periodic interference signals; Third, data normalization is performed, using the minimum-maximum normalization algorithm to uniformly map parameters with different dimensions and value ranges to the [0,1] interval, eliminating the influence of differences in dimensions between parameters on feature extraction; Fourth, time sequence alignment is completed, using linear interpolation to resample multi-source parameters with inconsistent sampling frequencies according to a unified sampling time benchmark, ensuring that the time sequences of all monitoring parameters are completely aligned, and finally obtaining the preprocessed multi-source monitoring parameters. Other methods can also be used in other embodiments, which are not limited here.

[0024] In addition, in specific implementation, the time-frequency domain feature extraction of the preprocessed multi-source monitoring parameters to obtain time-frequency domain feature information characterizing the operating state of the diesel generator can be achieved in the following way: it is divided into two core steps: time-domain feature extraction and frequency-domain feature extraction. In the time-domain feature extraction stage, the statistical features of mean, variance, root mean square, peak value, peak-to-peak value, kurtosis, and skewness of the preprocessed multi-source monitoring parameters are calculated, and waveform features of waveform factor, impulse factor, and margin factor are extracted to cover the steady-state characteristics of electrical parameters and the time-domain fluctuation characteristics of mechanical parameters. In the frequency-domain feature extraction stage, the fast Fourier transform is used to perform spectral decomposition for electrical parameters to extract the features of harmonic content, main frequency amplitude, and spectral energy distribution. The fast Fourier transform is used to complete the frequency domain conversion for mechanical vibration signals to extract the features of main frequency offset, secondary main frequency energy, and spectral entropy. Finally, the time-domain features and frequency-domain features are integrated to form time-frequency domain feature information of the diesel generator operating state covering multiple dimensions and multiple physical quantities of electrical and mechanical components. Other methods can also be used in other embodiments, which are not limited here.

[0025] It should be noted that the time-frequency domain feature information in this application represents the numerical characteristics, temporal degradation law and frequency distribution characteristics of each monitoring parameter of the generator, reflecting the stability of the generator's operating status, the degree of parameter distortion, vibration response characteristics and electrical performance fluctuations. It can be used for refined monitoring of the diesel generator's operating status, fault tracing, health assessment and intelligent early warning analysis.

[0026] In step 103, based on the mapping relationship between the fault mechanism knowledge base corresponding to the diesel generator, historical multi-source monitoring parameters and historical abnormal data, an anomaly identification layer for pre-diagnosis of abnormal areas of the diesel generator and a fault mechanism-data hybrid driving reasoning layer for reasoning about anomaly types and fault causes are constructed.

[0027] In some embodiments, based on the mapping relationship between the fault mechanism knowledge base corresponding to the diesel generator, historical multi-source monitoring parameters, and historical abnormal data, the construction of an anomaly identification layer for abnormal area pre-diagnosis of the diesel generator and a fault mechanism-data hybrid driven inference layer for anomaly type and fault cause reasoning can be achieved by the following steps: Construct a knowledge base for the fault mechanisms of diesel generators; Obtain the historical multi-source monitoring parameters and historical anomaly data corresponding to the diesel generator; Determine the mapping relationship between the fault mechanism knowledge base, the historical multi-source monitoring parameters, and the historical abnormal data; Based on the mapping relationship, an anomaly identification layer is constructed to perform pre-diagnosis of abnormal areas of the diesel generator; Obtain historical anomaly type data for the diesel generator; A fault mechanism-data hybrid inference layer is constructed by using the mapping relationship and the historical anomaly type data to infer the anomaly type and fault cause of the diesel generator.

[0028] In practical implementation, the fault mechanism knowledge base corresponding to the diesel generator can be constructed in the following way: Based on the diesel generator design manual, industry fault standards, and the analysis of a large number of fault cases, a fault mechanism knowledge base covering six major subsystems is formed, including the fuel supply system, lubrication system, cooling system, engine crankshaft and connecting rod mechanism, valve train mechanism, and power generation system. This knowledge base includes various fault modes, fault occurrence mechanisms, fault-feature correspondences, spatial mapping relationships between measurement points and subsystems, fault culpability weights, and industry-standardized fault terminology for each subsystem. Specifically, the fault mechanism knowledge base represents the typical fault modes of each subsystem, the inherent mechanisms of fault occurrence, the correspondence between faults and monitoring features, the spatial mapping relationships between measurement points and subsystems, fault culpability weights, and standardized fault terminology. It reflects the objective laws, characteristic manifestations, and hazard levels of diesel generator faults and is the core theoretical basis for achieving accurate identification of equipment anomalies, fault tracing, and inferential diagnosis.

[0029] In addition, in specific implementation, historical multi-source monitoring parameters and historical anomaly data corresponding to the equipment are comprehensively collected from data sources such as the long-term operation monitoring system of the diesel generator, equipment operation and maintenance ledger, fault repair archives, and historical operating condition records. The historical multi-source monitoring parameters cover historical electrical parameters and historical mechanical operating parameters during generator operation, while the historical anomaly data includes the time of anomaly occurrence, operating conditions, anomaly manifestation, fault location, fault level, maintenance measures, and handling results. The collected raw data is standardized and preprocessed to remove invalid and missing data caused by sensor failure, data transmission errors, and sampling anomalies. Duplicate records are cleaned and the data format is unified to finally obtain complete, valid, and uniformly formatted historical multi-source monitoring parameters and historical anomaly data. Other methods can also be used in other embodiments, which are not limited here.

[0030] In addition, in specific implementation, the mapping relationship between the fault mechanism knowledge base, the historical multi-source monitoring parameters, and the historical abnormal data can be determined in the following way: using conventional correlation analysis methods, combined with the operating principle and fault mechanism of the diesel generator, various fault modes, fault occurrence mechanisms, and fault-feature correspondences in the fault mechanism knowledge base are associated with the temporal degradation law of the historical multi-source monitoring parameters. This clarifies the fault mode and fault mechanism in the fault mechanism knowledge base corresponding to a specific change in a certain type of monitoring parameter. At the same time, abnormal events and abnormal parts in the historical abnormal data are bound with the corresponding changes in historical multi-source monitoring parameters and fault information in the fault mechanism knowledge base, establishing a stable correspondence, i.e., a mapping relationship, among the three. This clarifies the inherent logic of monitoring parameter changes, abnormal events, and fault mechanisms, ensuring that the mapping relationship can accurately reflect the association between data characteristics, abnormal phenomena, and fault mechanisms. Other methods can also be used in other embodiments, which are not limited here.

[0031] In addition, in specific implementation, the anomaly identification layer for pre-diagnosing abnormal areas of the diesel generator based on the mapping relationship can be implemented in the following way: Using each subsystem of the diesel generator as the classification basis, a corresponding feature benchmark library is established based on the mapping relationship between historical multi-source monitoring parameter characteristics and abnormal locations. Specifically, basic data such as parameter characteristic values, fluctuation ranges, and change trends of each subsystem under long-term normal operating conditions are extracted from the mapping relationship, and a feature benchmark data set for normal operation of each subsystem is compiled. Then, combined with the parameter characteristic deviation data and anomaly level information recorded in the mapping relationship under abnormal operating conditions of each subsystem, feature thresholds, fluctuation ranges, and time-series degradation patterns corresponding to different anomaly degrees are added to complete the construction of the feature benchmark library. The benchmark library stores core data such as parameter characteristic thresholds, fluctuation ranges, and time-series degradation patterns of different subsystems under normal and abnormal operating conditions. Simultaneously, feature matching and anomaly judgment rules are formulated, relying on the correspondence between parameter characteristic changes and abnormal locations in the mapping relationship. The algorithm compares real-time feature data with the feature standards of each subsystem in the benchmark library, calculates the degree of feature deviation and matching similarity, and sets reasonable judgment thresholds. Specifically, based on information such as the deviation magnitude of parameter features, frequency of anomalies, and severity level corresponding to different abnormal parts in the mapping relationship, the algorithm performs hierarchical statistics on the degree of feature deviation of each subsystem, distinguishing between minor, moderate, and severe anomalies. Combining generator operation and maintenance experience, judgment thresholds for each feature dimension are defined. Simultaneously, referencing the verification results of historical anomaly data in the mapping relationship, the thresholds are repeatedly calibrated and adjusted to ultimately determine reasonable judgment thresholds suitable for feature matching of each subsystem. When the feature matching degree reaches the threshold requirement, the corresponding subsystem can be determined as a suspected anomaly area. To ensure recognition accuracy, historical anomaly data is used to repeatedly verify, adjust, and optimize the benchmark library and judgment thresholds, ultimately forming an anomaly recognition layer capable of locating and pre-diagnosing anomaly areas. Other methods can also be used in other embodiments, which are not limited here.

[0032] In addition, in specific implementation, historical anomaly type data of diesel generators is collected from sources such as equipment fault repair reports, operation and maintenance summaries, and industry-standard fault terminology databases. This historical anomaly type data includes standardized definitions, anomaly type classifications, typical manifestations, triggering causes, associated anomaly locations, and corresponding parameter characteristic changes for various anomalies. The collected anomaly type data is standardized and organized, fault terminology is unified, and anomaly types are classified and archived. At the same time, a correspondence between anomaly types and anomaly locations and parameter characteristics is established to form a complete and standardized anomaly type database. Other methods can also be used in other embodiments, which are not limited here.

[0033] In addition, in specific implementation, the fault mechanism-data hybrid driving reasoning layer for inferring the anomaly type and fault cause of the diesel generator by constructing the mapping relationship and the historical anomaly type data can be implemented in the following way: establish a fault mechanism association library covering anomaly location, parameter feature changes, anomaly type, and anomaly cause; formulate standardized fault reasoning rules according to the logical chain of anomaly location-parameter feature-anomaly type-core cause; clarify the anomaly type and specific cause corresponding to different suspected anomaly areas; supplement the fault mechanism description of typical faults based on equipment operation and maintenance experience to ensure that the analysis results are easy to understand and the expression is standardized; then verify and optimize the fault reasoning rules through historical fault cases, eliminate rules with ambiguous logic and inaccurate associations, and finally form a fault mechanism-data hybrid driving reasoning layer that can interpret the fault mechanism of suspected anomaly areas and output standardized anomaly types and causes. Other methods can also be used in other embodiments, which are not limited here.

[0034] It should be noted that the historical anomaly data in this application represents a complete record of various abnormal events that have occurred in the past operation of the diesel generator, reflecting the fault performance, location, development process, and handling results of the equipment under different operating conditions and different usage stages; the mapping relationship represents the stable correspondence established between historical multi-source monitoring parameter characteristics and abnormal events, abnormal locations, and abnormal types, reflecting the inherent logic between parameter characteristic changes and abnormal equipment states and fault locations; the anomaly identification layer reflects the characteristic boundary identification layer of normal and abnormal features of each subsystem of the diesel generator equipment, which can be used to compare real-time monitoring features to complete the accurate location of abnormal areas and realize the rapid pre-diagnosis of suspected abnormal parts; the historical anomaly type data reflects the standardized classification system of equipment anomalies and the core characteristics of various anomalies; the fault mechanism-data hybrid driven reasoning layer reflects the transformation logic parsing layer of the diesel generator from numerical anomaly features to standardized fault semantics, which can be used to perform rapid pre-diagnosis reasoning driven by fault mechanism-data hybrid for suspected abnormal areas, and output clear and standardized anomaly type and anomaly cause identification results.

[0035] In step 104, the abnormality identification layer and the time-frequency domain feature information are used to perform abnormality location and pre-diagnosis of the operating status of the diesel generator to obtain the suspected abnormal area of ​​the diesel generator under the operating conditions. The fault mechanism-data hybrid driven inference layer performs fault inference on the distortion feature set corresponding to the suspected abnormal area to obtain the fault inference results of the abnormality type, fault cause and fault severity of the diesel generator.

[0036] In some embodiments, reference Figure 3The figure is an exemplary flowchart for determining suspected abnormal regions in some embodiments of this application. In this embodiment, the abnormality identification layer and the time-frequency domain feature information are used to perform anomaly localization and pre-diagnosis of the operating state of the diesel generator. The suspected abnormal regions of the diesel generator under operating conditions can be obtained by the following steps: In step 1041, anomaly diagnosis is performed on the time-frequency domain feature information based on the anomaly identification layer to obtain anomaly diagnosis information in the time-frequency domain feature information; In step 1042, the abnormal coupling judgment rule of the time domain stability characteristics and frequency domain energy distribution characteristics of the diesel generator under the operating conditions is determined; In step 1043, the abnormal operating status of the diesel generator is located by the abnormal diagnosis information and the abnormal coupling judgment rule to obtain the suspected abnormal area of ​​the diesel generator under the operating conditions.

[0037] In specific implementation, the anomaly diagnosis of the time-frequency domain feature information based on the anomaly identification layer can be achieved in the following way: the time-frequency domain feature information is input into the anomaly identification layer, which has a built-in feature benchmark library corresponding to each equipment subsystem of the diesel generator. The benchmark library stores in detail the standard range of time-domain stability characteristics and frequency-domain energy distribution characteristics of each subsystem under normal operating conditions, as well as the feature deviation standards of each subsystem under different degrees of anomaly. A conventional feature comparison method is used to compare the input real-time time-frequency domain feature information with the benchmark library. The feature standards of the corresponding subsystem are compared item by item and dimension by dimension. It is determined whether the real-time feature value is within the normal standard range and whether the feature time series degradation pattern is consistent with the pattern under normal operating conditions. If the feature value of a certain dimension exceeds the normal range or the feature time series degradation pattern deviates from the normal pattern, the feature is determined to be an abnormal feature. At the same time, key information such as the subsystem to which the abnormal feature belongs, the specific value of the feature deviation, the deviation magnitude, the deviation type, and the duration of the abnormality are recorded. After integrating this information, complete abnormal diagnosis information is obtained. Other methods can be used in other embodiments, which are not limited here.

[0038] It should be noted that the time-domain stability characteristics in this application are indicators that quantify the changes of multi-source monitoring parameters of diesel generators over time. They reflect the stability, fluctuation amplitude, steady-state maintenance ability, and trend of the parameters over time. They include various parameters reflecting the time-series fluctuations and steady-state characteristics of the parameters, such as mean, variance, peak value, peak-to-peak value, root mean square, fluctuation amplitude, time-series change slope, steady-state deviation rate, skewness, and kurtosis. The frequency-domain energy distribution characteristics are indicators that characterize the energy distribution pattern of the monitoring parameter signal in different frequency intervals. They reflect the concentration of signal energy in each frequency component, the dominant frequency offset, the proportion of harmonic energy, the proportion of energy in different frequency intervals, and the uniformity of spectral energy distribution. They include the dominant frequency amplitude, the proportion of energy in the secondary dominant frequency, the proportion of energy in each frequency interval, the harmonic energy content, and the spectral entropy, which reflect the energy distribution state of the signal frequency dimension.

[0039] Furthermore, in specific implementation, the abnormal coupling judgment rule for determining the time-domain stability characteristics and frequency-domain energy distribution characteristics of the diesel generator under operating conditions can be implemented in the following way: Based on the actual operating characteristics of the diesel generator, the statistical results of historical abnormal data, and equipment operation and maintenance experience, the inherent correlation between the time-domain stability characteristics and frequency-domain energy distribution characteristics is sorted out, and an abnormal coupling judgment rule is formulated. Specifically, it is determined that when the same subsystem simultaneously exhibits abnormal time-domain stability characteristics (e.g., parameter fluctuation amplitude exceeds the normal range, or time-series changes are drastic) and abnormal frequency-domain energy distribution characteristics (e.g., main frequency energy shift, or abnormal increase in the energy proportion of a certain frequency range), the subsystem is judged to have a real abnormal risk; if only a single time-domain characteristic abnormality occurs... If a single frequency domain characteristic is abnormal, and the abnormal amplitude is small and the duration is short, it is judged as a pseudo-anomaly affected by non-fault factors such as environmental interference and load fluctuations, and is not included in the scope of real anomaly judgment. At the same time, the coupling judgment criteria are refined according to the operating characteristics of different subsystems. For example, in a mechanical transmission system, if the time domain vibration amplitude is abnormal and the energy ratio of the corresponding transmission frequency in the frequency domain is abnormal, it is judged as a coupling anomaly. This ensures that the rule fits the fault occurrence logic of different subsystems of the diesel generator and can effectively distinguish between real anomalies and pseudo-anomalies. Finally, the above judgment rule is used as the abnormal coupling judgment rule of the time domain stability characteristics and frequency domain energy distribution characteristics of the diesel generator under operating conditions. Other methods can also be used in other embodiments, which are not limited here.

[0040] In addition, in specific implementation, the abnormal location of the diesel generator's operating state through the abnormal diagnosis information and the abnormal coupling judgment rule can be achieved by the following method: combining the abnormal diagnosis information output by the abnormal identification layer with the preset abnormal coupling judgment rule, matching the abnormal features recorded in the abnormal diagnosis information with the correspondence of each subsystem according to the division standard of each equipment subsystem of the diesel generator, filtering out the abnormal features in the abnormal diagnosis information that meet the abnormal coupling judgment rule, that is, features that simultaneously exhibit time-domain and frequency-domain coupling anomalies, and then accurately locating the equipment subsystems with real abnormal risks based on the subsystems to which these coupling abnormal features belong, and finally determining the suspected abnormal area of ​​the diesel generator under the current operating condition, thus completing the abnormal location and pre-diagnosis of the diesel generator's operating state. Other methods can also be used in other embodiments, which are not limited here.

[0041] It should be noted that the anomaly diagnosis information in this application reflects the core content of whether the characteristic values ​​of the diesel generator exceed the normal range, whether the time-series degradation pattern deviates from the baseline, the dimension to which the anomaly belongs and the corresponding equipment subsystem, the magnitude and duration of the anomaly, and can be used to identify the characteristic anomaly points and attributes existing in the equipment operation; the anomaly coupling judgment rule represents the judgment criteria formulated based on the operating characteristics of the diesel generator, reflecting the linkage logic between time-domain stability characteristic anomalies and frequency-domain energy distribution characteristic anomalies, clarifying the distinction between a single characteristic anomaly being a non-fault interference and two types of characteristic anomalies being judged as a real fault, and also refining the judgment conditions of coupled anomalies for different subsystems, which can be used to screen out false anomaly interference and lock in the real equipment fault risk characteristics; the suspected anomaly area reflects the physical parts or functional modules where there is a real anomaly risk under the current operating conditions of the diesel generator, which can be used to provide a precise target range for subsequent anomaly cause fault reasoning, health assessment and graded diagnosis and early warning, and guide the direction of fault tracing and operation and maintenance decision-making.

[0042] In some embodiments, the fault reasoning result of the fault mechanism-data hybrid driven inference layer performing fault reasoning on the distortion feature set corresponding to the suspected abnormal region to obtain the fault reasoning result of the diesel generator abnormality type, fault cause, and fault severity can be achieved by the following steps: Determine the distortion feature set corresponding to the suspected abnormal region in the time-frequency domain feature information; The distortion feature set is input into the fault mechanism-data hybrid driven inference layer, and fault inference of the abnormal type or abnormal cause is performed on the distortion feature set to obtain the fault inference results of the abnormal type, fault cause and fault severity of the diesel generator.

[0043] In specific implementation, determining the distortion feature set corresponding to the suspected abnormal region in the time-frequency domain feature information can be achieved in the following way: Identify the diesel generator equipment subsystem corresponding to the suspected abnormal region; combine the functional characteristics of the subsystem; match and extract all time-frequency domain feature data corresponding to the subsystem from the time-frequency domain feature information characterizing the stability of the equipment's operating state; then compare the extracted feature data with the time-frequency domain feature benchmark of the subsystem under normal operating conditions built into the anomaly identification layer; and filter out feature data that exceeds the normal feature range and deviates from the normal time-series degradation law, which is the distortion feature set corresponding to the suspected abnormal region. This information specifically includes distortion conditions such as excessive parameter fluctuation amplitude, abnormal time-series change slope, excessive steady-state deviation, and signal abrupt changes in the time domain, as well as distortion details such as main frequency shift, abnormal increase or decrease in the energy proportion of a certain frequency range, excessive harmonic content, and uneven spectral energy distribution in the frequency domain. Other methods can also be used in other embodiments, which are not limited here.

[0044] In addition, in specific implementation, the distortion feature set is input into the fault mechanism-data hybrid driving inference layer, and fault inference of the anomaly type or anomaly cause is performed on the distortion feature set to obtain the fault inference results of the diesel generator anomaly type, fault cause, and fault severity. This can be achieved in the following way: the extracted distortion feature set is input into the fault mechanism-data hybrid driving inference layer, which has a complete fault mechanism association library and standardized parsing rules built in. The fault mechanism association library is constructed based on the mapping relationship between historical multi-source monitoring parameters and historical anomaly data, and historical anomaly type data. It contains clear correspondences between each equipment subsystem, time-frequency domain distortion features, anomaly type, and anomaly cause. The parsing rules are formulated in combination with equipment operation and maintenance experience and industry fault handling standards. This process clarifies the anomaly types and causes corresponding to different combinations of distortion features. Using conventional feature comparison and logical reasoning methods, the input distortion feature set is first matched item by item with distortion features in the fault mechanism association library to find the preliminary range of the corresponding anomaly types. Then, combined with parsing rules and the subsystem characteristics of the suspected anomaly area, interference from pseudo-distortion features caused by non-fault factors such as environmental disturbances and load fluctuations is eliminated. This further refines the screening, clarifying the specific anomaly type corresponding to the distortion feature, and inferring the core cause of the anomaly. Finally, a standardized and logically clear fault reasoning result is compiled. This result explicitly includes the current anomaly type definition of the diesel generator, the suspected anomaly area corresponding to the anomaly, and the specific cause of the anomaly. Other methods can be used in other embodiments, which are not limited here.

[0045] It should be noted that the distortion feature set in this application represents abnormal feature data related to the suspected abnormal area of ​​the diesel generator. It reflects the time-domain features of the region, such as excessive fluctuation amplitude, sudden trend change, and excessive steady-state deviation, as well as the frequency dimension anomalies of the frequency domain features, such as main frequency shift, uneven energy distribution, and excessive harmonic content. The fault reasoning results reflect the specific abnormality type, abnormal cause, and related component information of the diesel generator. It can be used to provide clear, standardized, and interpretable abnormal information support for the hierarchical diagnosis and early warning of operating status, health assessment, and operation and maintenance decision-making.

[0046] In step 105, the operating health score of the diesel generator is generated based on the temporal degradation pattern of the multi-source monitoring parameters in the time dimension and the abnormal diagnosis results. The operating health score and the fault reasoning results are used to perform graded diagnosis and early warning of the operating status of the diesel generator.

[0047] In some embodiments, generating the operational health score of the diesel generator based on the temporal degradation pattern and anomaly diagnosis results of the multi-source monitoring parameters in the time dimension can be achieved through the following steps: Determine the temporal degradation pattern and anomaly diagnosis results of the multi-source monitoring parameters in the time dimension; Determine the baseline timing degradation pattern of the diesel generator; The operational health score of the diesel generator is determined based on the time-series degradation pattern, the anomaly diagnosis results, and the baseline time-series degradation pattern.

[0048] In specific implementation, determining the temporal degradation pattern and anomaly diagnosis results of the multi-source monitoring parameters in the time dimension can be achieved in the following way: Comprehensively collect time-series data of multi-source monitoring parameters during the operation of the diesel generator, continuously collecting data at a fixed sampling time step to ensure data continuity and integrity; preprocess the collected time-series data, removing invalid data, missing data, and abnormal jump values ​​caused by sensor failures and data transmission errors; use filtering methods to remove random noise caused by environmental interference and load fluctuations; then align the parameter data of different sampling frequencies according to a unified time reference to ensure that the time-series data of all parameters are consistent in the time dimension; subsequently, use the sliding window method to segment the preprocessed time-series data, setting a reasonable window length based on the diesel generator's operating cycle, for example, 10 minutes per window, dividing the continuous time-series data into several data segments; for each monitoring parameter within each data segment, analyze its temporal degradation pattern in the time dimension, specifically including: observing the upward, downward, or stable trend of parameter values, and calculating the fluctuation amplitude of parameters within the window. The method involves measuring the rate of change of parameters to identify sudden jumps, rapid increases, or sudden decreases, while also recording the periodicity of parameter changes. Finally, by integrating these characteristics of all parameters, the temporal degradation pattern of multi-source monitoring parameters is clarified. Based on the time-series characteristics of the diesel generator's multi-source monitoring parameters, electrical, mechanical, and environmental monitoring data are collected at fixed time steps. After cleaning, denoising, and normalizing the data, the characteristics of each parameter's change trend, fluctuation amplitude, abrupt change nodes, and periodic patterns in the time dimension are extracted. Then, combined with the equipment's operating mechanism and industry maintenance standards, normal fluctuation thresholds, change rate thresholds, and abnormal deviation thresholds for each parameter are set. By comparing the deviation between the monitoring data and the thresholds, it is identified whether the parameters exhibit fluctuations, abrupt changes, or abnormal trends exceeding the normal range. Combining the results of multi-parameter correlation analysis, the abnormality type, abnormality level, abnormal occurrence period, and impact range of the monitoring parameters in the time dimension are comprehensively judged, ultimately determining the abnormality diagnosis result of the multi-source monitoring parameters in the time dimension. Other methods can be used in other embodiments, which are not limited here.

[0049] In addition, in specific implementation, the reference time-series degradation law of the diesel generator can be determined in the following way: the reference time-series degradation law is constructed based on historical monitoring data and industry standard operating condition data under normal equipment operation. First, the historical time-series data of the diesel generator under long-term normal operation is screened out, and segmented according to the same sliding window method. The mean, normal fluctuation range, reasonable rate of change and trend of each parameter in each window are statistically analyzed. Then, combined with the rated operating parameters of the diesel generator, the factory standard and industry operation and maintenance specifications, the normal change range and reasonable change pattern of each monitoring parameter under different operating conditions are clarified. After integration, a complete reference time-series degradation law is formed to ensure that the reference can comprehensively and accurately reflect the parameter change characteristics during normal equipment operation, that is, the reference time-series degradation law of the diesel generator. Other methods can also be used in other embodiments, which are not limited here.

[0050] Furthermore, in specific implementation, determining the operational health score of the diesel generator based on the aforementioned timing degradation pattern, the anomaly diagnosis results, and the baseline timing degradation pattern can be achieved in the following manner: using the baseline timing degradation pattern as the core reference standard, the timing degradation patterns of each subsystem of the diesel generator are compared with the baseline timing degradation pattern one by one, and the degree of deviation between the two in terms of parameter change trends, fluctuation amplitudes, degradation rates, and abrupt change nodes is accurately calculated to quantify the deviation value of the timing degradation of each subsystem; at the same time, combined with the anomaly diagnosis results, the anomaly type, anomaly level, and anomaly impact of each subsystem are clarified. The impact range is determined by setting differentiated quantitative deduction standards based on the anomaly level. The higher the anomaly level, the greater the deduction weight. The deduction range is adjusted in conjunction with the anomaly's impact range to obtain the quantitative deduction value corresponding to the anomaly of each subsystem. Subsequently, a reasonable weighting system is preset, namely the importance of maintenance of core and auxiliary components of the diesel generator, the severity of failure consequences, the downtime risk level, and industry maintenance experience. First, the six major subsystems, including the fuel supply system and lubrication system, are classified as core and auxiliary. Among them, the core subsystems such as the generator system and the engine crankshaft connecting rod mechanism are allocated 60%-80% of the basic weight, while the cooling system and fuel supply system are allocated 60%-80%. Assign a base weight of 20%-40% to the main system and other auxiliary subsystems. Then, dynamically adjust the base weight coefficient by 0.8-1.2 times based on the failure frequency and maintenance cost ratio of each subsystem in historical failure data. Simultaneously, calculate the weight coefficient by multiplying it with the time-series degradation deviation quantification value and anomaly quantification deduction value of each subsystem, clarifying the specific quantitative indicators, coefficient range, and calculation logic of the weight allocation. Multiply the time-series degradation deviation quantification value and anomaly quantification deduction value of each subsystem by their corresponding weights to obtain the weighted score of each subsystem. Finally, calculate the weighted score of all subsystems... The initial comprehensive score is obtained by summing the results. Then, the initial comprehensive score is normalized and converted into a quantitative range of 0-100 points to eliminate the dimensional influence caused by the differences in parameters of different subsystems. Finally, the scoring threshold is calibrated by combining diesel generator operation and maintenance experience, industry standards and equipment factory requirements. The normalized comprehensive score is compared with the calibrated threshold to further fine-tune the scoring accuracy. Finally, an operation health score that can accurately and quantitatively reflect the current operating health status, potential fault risks and deterioration trends of the diesel generator is generated. Other methods can be used in other embodiments, which are not limited here.

[0051] It should be noted that the time-series degradation law in this application represents the numerical variation characteristics of multi-source monitoring parameters of diesel generators in a continuous time series, reflecting the upward / downward trend, fluctuation amplitude, rate of change, periodicity, and sudden changes of parameters over time, and can be used to intuitively present the time-series dynamic performance of equipment operation status; the anomaly diagnosis result represents the quantitative judgment result of the monitoring data deviating from the normal operation benchmark, reflecting the degree of deviation between the operating status of each system of the diesel generator and the preset normal standard, and indicating the type, amplitude, occurrence time, and duration of abnormal fluctuations in the monitoring parameters; the benchmark time-series degradation law represents the data based on the long-term normal operation history data of the equipment, industry standards, and factory specifications. The ideal parameter variation pattern constructed by Fan reflects the normal variation range, reasonable trend, and fluctuation range of the parameters of the diesel generator under fault-free and compliant operating conditions. It can serve as a core reference standard for determining whether the equipment's operating status deviates from the normal. The operating health score reflects the degree of deviation of the overall operating status of the diesel generator equipment from the normal benchmark and the degree of deterioration. It can be used to quickly assess the equipment's health level, identify potential fault risks, and provide a quantitative basis for operation and maintenance decisions and early warning management. The higher the score, the closer the equipment's operating status is to the normal benchmark and the better its health status. The lower the score, the more serious the deviation of the equipment's operating status from the normal benchmark and the worse its health status. This score can intuitively reflect the overall degree of equipment deterioration.

[0052] In some embodiments, the classification and early warning of the operating status of the diesel generator based on the operating health score and the fault reasoning result can be achieved by the following steps: The operational health score and the fault reasoning result are deeply fused to obtain diagnostic fusion information of the diesel generator's operating status; Determine the hierarchical diagnostic and early warning decision-making process for the operating status of the diesel generator; The diagnostic fusion information is diagnosed based on the hierarchical diagnostic early warning decision, and the operating status of the diesel generator is diagnosed and warned based on the diagnostic results.

[0053] In specific implementation, the deep fusion of the operational health score and the fault inference result to obtain the diagnostic fusion information of the diesel generator's operating status can be achieved in the following way: First, clarify the core correlation logic between the operational health score and the fault inference result. The operational health score is a quantitative indicator of 0-100 points, which intuitively reflects the overall health level of the equipment. The fault inference result contains textual information such as specific anomaly types, anomaly causes, suspected anomaly areas, and anomaly severity. During the fusion process, conventional information association and integration methods are used. First, the operational health score is matched with the anomaly severity in the fault inference result. For example, a health score below 50 points corresponds to semantic recognition. The system identifies severe anomalies, moderate anomalies (50-70 points), minor anomalies (70-80 points), and no anomalies (above 80 points). Redundant information is then removed, such as cases where the health score is normal but the fault inference result is a false anomaly. Simultaneously, relevant details are added, binding the deviation of the health score to the cause and area of ​​the anomaly. For example, if the health score is low and the semantic identification indicates abnormal lubrication system oil pressure, then this anomaly is identified as the core cause of the health decline. Finally, this information is integrated to form diagnostic fusion information covering the overall equipment health quantification level, specific anomaly details, anomaly causes, and anomaly areas. Other implementation methods can also be used in other embodiments, which are not limited here.

[0054] In addition, in specific implementation, the graded diagnostic and early warning decision for determining the operating status of the diesel generator can be implemented in the following way: Based on the diesel generator's operation and maintenance specifications, historical fault cases, industry safety standards, and equipment factory requirements, the graded diagnostic and early warning decision for the operating status is determined, specifically divided into four levels: normal, attention, warning, and alarm. The diagnostic fusion information judgment conditions corresponding to each level are clarified: the normal level corresponds to a health score of 80-100 points, the fault inference result is no abnormality, and the equipment operating status meets the benchmark requirements; the attention level corresponds to a health score of 70-80 points, the fault inference result is a minor abnormality, and there is no obvious fault risk; the warning level corresponds to a health score of 50-70 points, the fault inference result is a moderate abnormality, and timely investigation is required; the alarm level corresponds to a health score below 50 points, the fault inference result is a serious abnormality, and emergency handling is required. At the same time, the graded standards are refined for different abnormality types and abnormal areas. For example, the warning threshold is appropriately lowered for electrical system abnormalities compared to other system abnormalities, ensuring that the graded decision is consistent with the actual operating characteristics of the equipment and the degree of fault hazard. Other methods can also be used in other embodiments, which are not limited here.

[0055] It should be noted that the diagnostic fusion information in this application reflects the quantitative results of the overall health level of the equipment, as well as the specific anomaly types, regions, causes, and severity. It can be used to characterize the equipment's operating status, the correspondence between associated anomalies and health, and provide complete feature support for status determination. The hierarchical diagnostic early warning decision reflects the hierarchical classification and severity of equipment operating risks. It can be used to achieve hierarchical determination of abnormal risks, trigger corresponding early warning signals, guide maintenance personnel to carry out inspections and emergency handling according to the level, and provide quantitative basis for fault tracing, rule optimization, and maintenance decision-making.

[0056] In addition, in specific implementation, the diagnosis of the diagnostic fusion information based on the hierarchical diagnostic early warning decision, and the hierarchical diagnostic early warning of the diesel generator's operating status based on the diagnostic results, can be implemented in the following way: using the established hierarchical diagnostic early warning decision as the basis for judgment, the integrated diagnostic fusion information is compared and matched item by item with the judgment conditions of each early warning level. First, the score range to which the operating health score belongs is determined, and then the abnormal details in the fault inference results are combined to verify whether they meet the judgment requirements of the corresponding level. If both the health score and the fault inference results meet a certain early warning level, the equipment is determined to be in that level of operating status; if there is a slight deviation between the two, a comprehensive judgment is made in combination with historical operation and maintenance experience. For example, if the health score is 75 (attention level) but the semantic recognition indicates moderate abnormality, it is adjusted to the warning level to ensure accurate diagnosis. The warning signal for the corresponding level is triggered based on the final diagnosis result. The normal level does not trigger a warning; only the operating status is recorded. The attention level triggers a suggestive warning, reminding maintenance personnel to conduct regular inspections and pay attention to abnormal changes. The warning level triggers a cautionary warning, notifying maintenance personnel to promptly investigate the cause of the abnormality and take preventative measures. The alarm level triggers an emergency warning, immediately issuing an audible and visual alarm signal and reminding maintenance personnel to urgently stop the system for inspection and troubleshooting. Simultaneously, the graded diagnostic warning result and abnormal details are recorded. Other methods can be used in other embodiments, which are not limited here.

[0057] In another aspect, in some embodiments, this application provides a diesel generator operating status monitoring system, with reference to... Figure 4 The figure is a schematic diagram of the structure of a diesel generator operating status monitoring system according to some embodiments of this application. The diesel generator operating status monitoring system 400 includes: an acquisition module 401, a processing module 402, and an execution module 403, which are described below: The acquisition module 401 in this application is mainly used to acquire the multi-source monitoring parameters of the diesel generator under operating conditions. The multi-source monitoring parameters include the electrical parameters and mechanical operating parameters of the diesel generator. Processing module 402, in this application, is used to perform time-frequency domain analysis on the multi-source monitoring parameters to obtain time-frequency domain feature information characterizing the stability of the diesel generator's operating state; It should be noted that the processing module 402 in this application is also used to construct an anomaly identification layer for pre-diagnosis of anomaly areas of the diesel generator and a fault mechanism-data hybrid driving reasoning layer for reasoning about anomaly types and fault causes based on the mapping relationship between the fault mechanism knowledge base, historical multi-source monitoring parameters and historical anomaly data corresponding to the diesel generator. Additionally, it should be noted that the processing module 402 in this application is also used to perform anomaly location pre-diagnosis on the operating state of the diesel generator through the anomaly identification layer and the time-frequency domain feature information, to obtain the suspected abnormal area of ​​the diesel generator under the operating conditions, and the fault mechanism-data hybrid driven inference layer performs fault inference on the distortion feature set corresponding to the suspected abnormal area to obtain the fault inference results of the abnormal type, fault cause and fault severity of the diesel generator. The execution module 403 in this application is mainly used to generate the operating health score of the diesel generator based on the time-series degradation law and abnormal diagnosis results of the multi-source monitoring parameters in the time dimension, and to perform graded diagnosis and early warning of the operating status of the diesel generator based on the operating health score and the fault reasoning results.

[0058] In addition, this application also provides a computer device, which includes a memory and a processor. The memory stores code, and the processor is configured to acquire the code and execute the above-described diesel generator operating status monitoring method.

[0059] In some embodiments, reference Figure 5 The figure is a schematic diagram of the structure of a computer device for implementing a diesel generator operating status monitoring method according to some embodiments of this application. The diesel generator operating status monitoring method in the above embodiments can... Figure 5 The computer device shown is used to implement this, and the computer device 500 includes at least one processor 501, a communication bus 502, a memory 503, and at least one communication interface 504.

[0060] Processor 501 can be a general-purpose central processing unit (CPU) or an application-specific integrated circuit (ASIC).

[0061] The communication bus 502 can be used to transmit information between the aforementioned components.

[0062] Memory 503 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital versatile optical discs, Blu-ray discs, etc.), magnetic disks or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. Memory 503 may exist independently and be connected to processor 501 via communication bus 502. Memory 503 may also be integrated with processor 501.

[0063] The memory 503 stores program code for executing the scheme of this application, and its execution is controlled by the processor 501. The processor 501 executes the program code stored in the memory 503. The program code may include one or more software modules. The method used in the above embodiments can be implemented by the processor 501 and one or more software modules in the program code in the memory 503.

[0064] Communication interface 504 uses any transceiver-like device to communicate with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0065] In a specific implementation, as one example, a computer device may include multiple processors, each of which may be a single-core (single-CPU) processor or a multi-core (multi-CPU) processor. Here, a processor may refer to one or more devices, circuits, and / or processing cores used to process data (e.g., computer program instructions).

[0066] The aforementioned computer device can be a general-purpose computer device or a special-purpose computer device. In specific implementations, the computer device can be a desktop computer, a portable computer, a network server, a handheld digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. This application does not limit the type of computer device.

[0067] In addition, this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for monitoring the operating status of a diesel generator.

[0068] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0069] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for monitoring the operating status of a diesel generator, characterized in that, Includes the following steps: The multi-source monitoring parameters of the diesel generator under operating conditions are obtained, including the electrical parameters and mechanical operating parameters of the diesel generator. Time-frequency domain analysis is performed on the multi-source monitoring parameters to obtain time-frequency domain feature information that characterizes the stability of the diesel generator's operating state; Based on the mapping relationship between the fault mechanism knowledge base, historical multi-source monitoring parameters and historical abnormal data corresponding to the diesel generator, an anomaly identification layer for pre-diagnosis of abnormal areas of the diesel generator and a fault mechanism-data hybrid driving reasoning layer for reasoning about anomaly type and fault cause are constructed. The anomaly identification layer and the time-frequency domain feature information are used to perform anomaly location and pre-diagnosis of the operating status of the diesel generator, thereby obtaining the suspected anomaly area of ​​the diesel generator under the operating conditions. The fault mechanism-data hybrid driven inference layer performs fault inference on the distortion feature set corresponding to the suspected anomaly area, thereby obtaining the fault inference results of the anomaly type, fault cause and fault severity of the diesel generator. The operating health score of the diesel generator is generated based on the temporal degradation pattern of the multi-source monitoring parameters and the anomaly diagnosis results. The operating health score and the fault inference results are used to perform graded diagnosis and early warning of the operating status of the diesel generator.

2. The method as described in claim 1, characterized in that, The time-frequency domain analysis of the multi-source monitoring parameters yields time-frequency domain feature information characterizing the stability of the diesel generator's operating state, specifically including: The multi-source monitoring parameters are preprocessed to obtain preprocessed multi-source monitoring parameters; Time-frequency domain features are extracted from the preprocessed multi-source monitoring parameters to obtain time-frequency domain feature information characterizing the operating state of the diesel generator.

3. The method as described in claim 1, characterized in that, Based on the mapping relationship between the fault mechanism knowledge base corresponding to the diesel generator, historical multi-source monitoring parameters, and historical abnormal data, an anomaly identification layer for pre-diagnosis of abnormal areas of the diesel generator and a fault mechanism-data hybrid driven reasoning layer for inferring anomaly types and fault causes are constructed, specifically including: Construct a knowledge base for the fault mechanisms of diesel generators; Obtain the historical multi-source monitoring parameters and historical abnormal data corresponding to the diesel generator; Determine the mapping relationship between the fault mechanism knowledge base, the historical multi-source monitoring parameters, and the historical abnormal data; Based on the mapping relationship, an anomaly identification layer is constructed to perform pre-diagnosis of abnormal areas of the diesel generator; Obtain historical anomaly type data for the diesel generator; A fault mechanism-data hybrid inference layer is constructed by using the mapping relationship and the historical anomaly type data to infer the anomaly type and fault cause of the diesel generator.

4. The method as described in claim 1, characterized in that, By using the anomaly identification layer and the time-frequency domain feature information to perform anomaly localization and pre-diagnosis of the diesel generator's operating status, the suspected anomaly areas of the diesel generator under operating conditions are obtained, specifically including: Based on the anomaly identification layer, anomaly diagnosis is performed on the time-frequency domain feature information to obtain anomaly diagnosis information in the time-frequency domain feature information; Determine the abnormal coupling judgment rules between the time-domain stability characteristics and frequency-domain energy distribution characteristics of the diesel generator under operating conditions; The abnormality diagnosis information and the abnormality coupling judgment rules are used to locate the abnormality in the operating state of the diesel generator, thereby obtaining the suspected abnormal area of ​​the diesel generator under the operating conditions.

5. The method as described in claim 1, characterized in that, The fault inference layer, driven by the fault mechanism-data hybrid approach, performs fault inference on the distortion feature set corresponding to the suspected abnormal region, and obtains the fault inference results of the diesel generator anomaly type, fault cause, and fault severity. Specifically, these results include: Determine the distortion feature set corresponding to the suspected abnormal region in the time-frequency domain feature information; The distortion feature set is input into the fault mechanism-data hybrid driven inference layer, and fault inference of the abnormal type or abnormal cause is performed on the distortion feature set to obtain the fault inference results of the abnormal type, fault cause and fault severity of the diesel generator.

6. The method as described in claim 1, characterized in that, The operational health score of the diesel generator is generated based on the temporal degradation pattern and anomaly diagnosis results of the multi-source monitoring parameters over time, specifically including: Determine the temporal degradation pattern and anomaly diagnosis results of the multi-source monitoring parameters in the time dimension; Determine the baseline timing degradation pattern of the diesel generator; The operational health score of the diesel generator is determined based on the time-series degradation pattern, the anomaly diagnosis results, and the baseline time-series degradation pattern.

7. The method as described in claim 1, characterized in that, The classification, diagnosis, and early warning of the diesel generator's operating status based on the operational health score and the fault reasoning results specifically include: The operational health score and the fault reasoning result are deeply fused to obtain diagnostic fusion information of the diesel generator's operating status; Determine the hierarchical diagnostic and early warning decision-making process for the operating status of the diesel generator; The diagnostic fusion information is diagnosed based on the hierarchical diagnostic early warning decision, and the operating status of the diesel generator is diagnosed and warned based on the diagnostic results.

8. A diesel generator operating status monitoring system, characterized in that, include: The acquisition module is used to acquire multi-source monitoring parameters of the diesel generator under operating conditions. The multi-source monitoring parameters include the electrical parameters and mechanical operating parameters of the diesel generator. The processing module is used to perform time-frequency domain analysis on the multi-source monitoring parameters to obtain time-frequency domain feature information that characterizes the stability of the diesel generator's operating state. The processing module is also used to construct an anomaly identification layer for pre-diagnosis of anomaly areas of the diesel generator and a fault mechanism-data hybrid driving reasoning layer for reasoning about anomaly types and fault causes, based on the mapping relationship between the fault mechanism knowledge base corresponding to the diesel generator, historical multi-source monitoring parameters and historical abnormal data. The processing module is also used to perform anomaly location pre-diagnosis on the operating status of the diesel generator through the anomaly identification layer and the time-frequency domain feature information, to obtain the suspected anomaly area of ​​the diesel generator under the operating conditions, and the fault mechanism-data hybrid driven inference layer performs fault inference on the distortion feature set corresponding to the suspected anomaly area to obtain the fault inference results of the anomaly type, fault cause and fault severity of the diesel generator. The execution module is used to generate an operational health score for the diesel generator based on the temporal degradation pattern and anomaly diagnosis results of the multi-source monitoring parameters in the time dimension, and to perform graded diagnosis and early warning of the operational status of the diesel generator based on the operational health score and the fault inference results.

9. A computer device, characterized in that, The computer device includes a memory and a processor, the memory storing code, and the processor being configured to acquire the code and execute the diesel generator operating status monitoring method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the diesel generator operating status monitoring method as described in any one of claims 1 to 7.

Images