A method for generating a remote intelligent operation and maintenance model of a diesel generator set

Abstract

The present application relates to the diesel generating set operation and maintenance technical field, especially to a kind of diesel generating set remote intelligent operation and maintenance model generation method.The method includes the following steps: when diesel generating set is in running state, injection oil pressure parameter and generator end output voltage parameter are collected as operating parameter, and load change process and corresponding time stamp are recorded, to generate operating data sequence;Trend analysis is carried out based on operating data sequence, and the characteristic quantity of generator set stability is extracted, the deviation value of each characteristic quantity is calculated with historical operating data, when the deviation value exceeds preset range, parameter recall operation is executed;The stability index is recalculated for the operating result after continuous multiple recalls;The present application realizes the intelligent monitoring and accurate state prediction of diesel generating set remote operation and maintenance by real-time analysis of trend deviation and stage response law, improves operating stability and fault early warning capability.

Description

Technical Field

[0001] This invention relates to the field of diesel generator set operation and maintenance technology, and in particular to a method for generating a remote intelligent operation and maintenance model for diesel generator sets. Background Technology

[0002] Diesel generator set operation and maintenance largely rely on manual inspections and periodic maintenance, lacking continuous monitoring and refined analysis of operating status. This makes it difficult to detect minor anomalies or potential faults in a timely manner, easily leading to problems such as frequency deviations, voltage fluctuations, and abnormal fuel supply under sudden load changes or prolonged operation, thus affecting power quality and equipment lifespan. Existing intelligent operation and maintenance methods mainly rely on single-parameter monitoring or experience-based threshold judgments for fault early warning. However, these methods often ignore the temporal correlations and response patterns between parameters, making it difficult to accurately assess the dynamic stability of the unit. Furthermore, they have shortcomings in remote monitoring and predictive capabilities. Under the coupled influence of multiple parameters, there are complex nonlinear relationships between fuel injection pressure, output voltage, and load changes. Traditional methods struggle to achieve real-time callbacks and comprehensive stability assessments, resulting in low operation and maintenance efficiency and low reliability. Summary of the Invention

[0003] Therefore, it is necessary to provide a method for generating a remote intelligent operation and maintenance model for diesel generator sets to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, a method for generating a remote intelligent operation and maintenance model for diesel generator sets includes the following steps: Step S1: When the diesel generator set is in operation, collect the injection oil pressure parameters and the output voltage parameters at the generator terminals as operating parameters, and record the load change process and the corresponding timestamps to generate an operating data sequence; Step S2: Perform trend analysis based on the operating data sequence, extract the characteristic quantities of generator set stability, calculate the deviation value of each characteristic quantity with the historical operating data, and execute the parameter callback operation when the deviation value exceeds the preset range. Step S3: Recalculate the stability index based on the running results after multiple consecutive callbacks. Measure the response delay under load using the generator set's frequency recovery time and voltage regulation overshoot. If the response delay is greater than the preset delay threshold, generate a maintenance prompt signal and transmit the maintenance prompt signal and the corresponding running data in the running data sequence to the monitoring terminal. Step S4: When the monitoring terminal detects a long-term fluctuation trend in the running data sequence that lasts for more than a preset period, it divides the running phases and extracts the parameter response pattern data of each phase to form a set of operation and maintenance data relationships, and establishes an operation and maintenance model for remote monitoring and status prediction.

[0005] The beneficial effects of this invention are as follows: During the operation of the diesel generator set, injection oil pressure parameters and generator terminal output voltage parameters are collected in real time, and the load change process and corresponding timestamps are recorded simultaneously, forming a complete operating data sequence, enabling refined acquisition of the actual operating status of the unit. Based on this, trend analysis is performed using the operating data sequence to extract characteristic quantities of the generator set's operational stability. These are then compared item by item with historical operating data to calculate characteristic deviations. This allows for timely identification of abnormal parameters and parameter callback operations, automatically correcting unit parameters that deviate from operating specifications, ensuring stable operation of the unit under load changes or external disturbances. Furthermore, the stability index is recalculated based on the operating results after multiple callbacks, and the unit's response delay is determined based on the frequency recovery time and voltage regulation overshoot. When the response delay exceeds a preset threshold, a maintenance prompt signal is generated and transmitted to the monitoring terminal, enabling remote real-time alarms and maintenance guidance.

[0006] By identifying long-term fluctuation trends, the operation process is divided into several stages. Statistical analysis is performed on the direction of parameter changes and response speed in each stage, extracting parameter response pattern data characteristic of each stage to form an operation and maintenance data relationship set. Combined with the remote monitoring and status prediction model established by the prediction unit, the unit status can be accurately assessed at different operation stages, identifying potential faults or instability trends in advance. Attached Figure Description

[0007] Figure 1 A flowchart illustrating the steps involved in generating a remote intelligent operation and maintenance model for a diesel generator set. Figure 2 Flowchart for intelligent monitoring of parameter status; Figure 3 This is a histogram of oil pressure parameters and voltage fluctuation amplitude. The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0008] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0009] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0010] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0011] To achieve the above objectives, please refer to Figures 1 to 3 A method for generating a remote intelligent operation and maintenance model for diesel generator sets includes the following steps: Step S1: When the diesel generator set is in operation, collect the injection oil pressure parameters and the output voltage parameters at the generator terminals as operating parameters, and record the load change process and the corresponding timestamps to generate an operating data sequence; Step S2: Perform trend analysis based on the operating data sequence, extract the characteristic quantities of generator set stability, calculate the deviation value of each characteristic quantity with the historical operating data, and execute the parameter callback operation when the deviation value exceeds the preset range. Step S3: Recalculate the stability index based on the running results after multiple consecutive callbacks. Measure the response delay under load using the generator set's frequency recovery time and voltage regulation overshoot. If the response delay is greater than the preset delay threshold, generate a maintenance prompt signal and transmit the maintenance prompt signal and the corresponding running data in the running data sequence to the monitoring terminal. Step S4: When the monitoring terminal detects a long-term fluctuation trend in the running data sequence that lasts for more than a preset period, it divides the running phases and extracts the parameter response pattern data of each phase to form a set of operation and maintenance data relationships, and establishes an operation and maintenance model for remote monitoring and status prediction.

[0012] In one embodiment, when the diesel generator set is in rated operating condition, real-time data of injection pressure and generator output voltage are collected by an oil pressure sensor and a voltage acquisition module, respectively, and the load change process and corresponding timestamps are recorded. Data is continuously recorded for 5 minutes with a sampling period of 0.1s to obtain the operating data sequence. During the acquisition process, the oil pressure sensor is installed on the high-pressure oil pipe branch with a range of 0–60MPa, and the voltage acquisition module has a sampling accuracy of ±0.5V. The load change is controlled by an automatic loading device, switching back and forth between 0.6 and 1.0 times the rated load to simulate on-site power supply fluctuations. After the acquisition is completed, the data sequence is automatically time-synchronized and noise-reduced to provide basic data for subsequent trend analysis.

[0013] Trend analysis is performed on the collected oil pressure and voltage data. Short-term fluctuations are smoothed using a moving average method, and then the direction of the average slope change within each 10-second interval is calculated. If the direction of change is consistent across three consecutive intervals, the parameter is considered to have a continuous upward or downward trend. Subsequently, the ratio of oil pressure fluctuation amplitude to voltage fluctuation amplitude is calculated to reflect the fuel response sensitivity under load changes. When this ratio remains stable between 1.2 and 1.4 over a long period, the unit is considered to be in a normal regulation state. The above results are used to form a stability characteristic quantity, which is compared with the standard characteristic quantity under the same load in the historical operating records. If the deviation exceeds 10%, a parameter callback command is generated to notify the control unit to perform corresponding adjustments.

[0014] The control unit fine-tunes the injection advance angle and voltage excitation current according to the parameter callback command, with each adjustment not exceeding 3%. After adjustment, it re-monitors the unit's operating frequency and terminal voltage fluctuations, calculating the frequency recovery time and voltage overshoot. When the frequency recovery time exceeds 3 seconds and the voltage overshoot is greater than 5%, it is determined that the unit's response delay is too high, automatically generating a maintenance prompt signal and uploading the corresponding operating data segment (including oil pressure, voltage, and load information) to the monitoring terminal for remote analysis and judgment of whether the unit has problems such as injection lag or speed regulation wear.

[0015] The monitoring terminal receives and records all operational data and maintenance prompts. Operational data across different time periods is divided into periods, distinguishing between steady-state, load switching, and transitional intervals, and the oil pressure and voltage response curves for each stage are statistically analyzed. By analyzing the response patterns at different stages, a correspondence between operating parameters and unit stability is established, forming a set of operation and maintenance data relationships. Subsequently, a remote intelligent operation and maintenance model is trained based on this relationship set, enabling the model to determine the unit's operating status in real time upon the arrival of new data, predict performance degradation trends, and achieve remote monitoring and health assessment of the diesel generator set.

[0016] Of particular importance, step S1 includes: When the diesel generator set enters a stable operating state, the data acquisition unit is activated to collect the oil pressure signal of the injection system and the output voltage signal of the generator in real time. The collected injection oil pressure parameters and output voltage parameters are paired according to the sampling time sequence to form a parameter correspondence and the sampling time is recorded. During the parameter acquisition process, load changes are monitored, the start and end times of load changes are marked, and corresponding timestamp data is generated. The injection oil pressure parameters, output voltage parameters, and load change timestamps are integrated into a time series to generate a complete operating data sequence.

[0017] In one embodiment, when the diesel generator set is at its rated speed (1500 r / min) and the output voltage is stable at 400V±5V, the data acquisition unit is activated to collect the oil pressure signal at the high-pressure oil pipe port of the injection system and the voltage signal at the generator output terminal in real time. The oil pressure signal is acquired by a pressure sensor with a sampling frequency of 100Hz; the voltage signal is acquired by a voltage sampling module with a sampling frequency of 50Hz. The two sets of data are synchronously paired according to the sampling time axis to form a one-to-one corresponding parameter record. During the acquisition period, the monitoring load control cabinet automatically switches the load state and records the change process of the load from 0.7 times the rated power to full load. The start time and end time of each load change are marked and timestamps are generated. The injection oil pressure parameters, output voltage parameters, and load change timestamps are integrated into a complete operating data sequence according to the sampling time order for subsequent trend analysis and stability judgment.

[0018] In another embodiment, 10 minutes after a cold start of the diesel generator set, when the operating speed gradually stabilizes, the data acquisition unit is activated to collect the oil pressure signal at the injection pump outlet and the voltage signal at the generator bus terminal, respectively. The sampling period is set to 0.05s, and the sampling duration is 300s. During the acquisition process, the monitoring unit detects current changes in real time. When a load switch from light load to heavy load is detected, the instant of the switch is automatically recorded. To ensure data integrity, acquisition continues for 30s after each load change to cover the transition response phase of the generator set. After all sampling is completed, the oil pressure, voltage, and load change time information are integrated in timestamp order to form an operating data sequence with load tags, providing basic data for subsequent trend identification and feature extraction.

[0019] Preferably, step S2 includes: The running data sequence is divided according to parameter categories, and the data correspondence between each parameter is determined; The changes in the values ​​of voltage and oil pressure parameters at continuous sampling points are calculated, and the trend of the parameters is determined. The statistical trend data's persistence and fluctuation range are analyzed to extract features of operational stability and generate a sequence of feature records. The feature record sequence is compared item by item with the pre-stored historical running data to calculate the feature deviation data; When the feature deviation data exceeds the preset range, a parameter callback instruction is triggered, and the corresponding parameter callback operation is executed. When the same parameter triggers a callback instruction in two consecutive trend analysis cycles, the parameter will be marked as a key monitoring state.

[0020] In one embodiment, after the operating data sequence is generated, the data is divided into two groups based on parameter type: voltage and hydraulic pressure. A time correspondence is established to ensure alignment of different parameters at the same sampling time. Subsequently, change calculations are performed on continuous sampling points of voltage and hydraulic parameters, and the rate of change of difference between adjacent sampling points is extracted to determine the upward or downward trend of the parameters. Continuity statistics are performed on trend changes. When changes in the same direction continue for more than five sampling cycles, the trend is considered persistent. Simultaneously, the fluctuation amplitude within the trend interval is calculated to reflect the stability of the generator set within that interval. Based on the combination of persistence and fluctuation amplitude, characteristic quantities of operating stability are extracted, and a characteristic record sequence is generated in chronological order. This characteristic record sequence is then compared item by item with the corresponding parameter records stored in the historical database to calculate the characteristic deviation value. When any deviation exceeds the set stability interval, a parameter callback instruction is automatically triggered. The instruction includes the correction direction and adjustment magnitude of the corresponding parameter. After the callback is executed, the adjustment result is monitored in real time, and the deviation is recalculated in the next trend analysis cycle. When the same parameter is found to trigger the callback instruction twice in a row, the parameter is automatically marked as a key monitoring status so that it can be prioritized for tracking and analysis in subsequent operation cycles.

[0021] In another embodiment, after receiving the operational data sequence, the sampled data is categorized into electrical signals and hydraulic signals based on parameter attributes, and a time index table is established for synchronous invocation in subsequent trend calculations. For each type of parameter, the rate of change sequence of sampling points is calculated using a sliding window method to determine the direction and amplitude of parameter change within adjacent sampling segments. When the direction of change remains consistent and the amplitude of change is stable within a preset range, it is recorded as a stable trend segment; if there is a sudden change in the direction of change or the fluctuation exceeds the limit, it is recorded as an abnormal trend segment. The duration and fluctuation range of each trend segment are statistically analyzed to extract characteristic indicators reflecting stability, forming a feature record sequence arranged by time. Subsequently, this sequence is compared with historical operational samples to calculate the deviation ratio. When the deviation ratio exceeds a set threshold, a parameter callback command is automatically issued, and the control unit performs minor adjustments to the corresponding operational parameters. If the same parameter triggers a callback operation in two consecutive trend analyses, the parameter is automatically added to the key monitoring list, and the sampling frequency and analysis priority of the parameter are increased in subsequent acquisition cycles to promptly detect potential stability hazards.

[0022] Preferably, the calculation of the changes in the values ​​of the voltage parameter at continuous sampling points and the determination of the parameter trend data in step S2 specifically involves: The voltage continuous sampling point sequence was grouped into analysis units of 8 sampling points each; Within each analysis unit, a difference operation is performed on adjacent sampling points and divided by the sampling interval to obtain the voltage instantaneous rate of change sequence; Calculate the proportion of sampling points in the sequence whose absolute value of the rate of change exceeds 0.5V / ms, and record the number of consecutive occurrences of adjacent points with the same direction of change; When the number of consecutive rate of change points in the same direction exceeds 4 and the cumulative voltage offset within the cell exceeds 2V, the voltage is determined to be in a continuous rising or falling phase, which serves as the trend data for the voltage parameter.

[0023] In one embodiment, continuous sampling points of voltage parameters in the running data sequence are analyzed, and the sampling points are paired in chronological order to form continuous sampling segments. Within each sampling segment, the rate of change between adjacent points is calculated to obtain a sequence of instantaneous voltage rate of change, and the direction of change is recorded. When the proportion of sampling points with an absolute value of rate of change exceeding 0.5V / ms exceeds a preset proportion, the number of consecutive occurrences of adjacent points changing in the same direction is counted. If the number of consecutive points changing in the same direction exceeds 4 and the cumulative voltage offset within the sampling segment exceeds 2V, the voltage is determined to be in a continuous rising or falling phase, and this result is taken as the trend of the voltage parameter.

[0024] In another embodiment, a sliding difference operation is performed on the continuous sampling points of the voltage parameter to calculate the instantaneous rate of change between each pair of adjacent sampling points and establish a sequence of change directions. The number of sampling points with a rate of change exceeding 0.5V / ms and the length of consecutive points in the same direction are counted. When the number of consecutive points in the same direction reaches or exceeds 4, and the total voltage offset within the sampling segment is greater than 2V, it is determined that the voltage in that segment shows a continuous upward or downward trend. This trend determination result is used to form the trend information of the voltage parameter and provides a basis for subsequent operational stability analysis and anomaly detection.

[0025] Please refer to [link / reference needed] for further information. Figure 2 The process begins with inputting operational data. First, the data is categorized by parameter type, and the changing trends of key parameters (such as voltage and oil pressure) are calculated. Characteristic quantities representing operational stability are extracted from these trends, and this real-time feature is compared with historical normal data to calculate the feature deviation. Then, the core judgment stage begins: if the deviation is within limits, the process returns to the starting point for continuous monitoring; if it exceeds limits, a parameter callback operation is automatically triggered to attempt correction. A second judgment is then performed to check if the same parameter has shown anomalies in two consecutive analysis cycles. If not, it reverts to regular monitoring; if so, the parameter is marked as a key monitoring state, thus achieving early warning of faults.

[0026] Preferably, the calculation of the changes in the values ​​of the continuously sampled oil pressure parameters and the determination of the parameter trend data in step S2 are specifically as follows: The continuous hydraulic pressure sampling point sequence was grouped into analysis units of 10 sampling points each; Within each analysis unit, the pressure pulse waveform synchronized with the fuel injection action is extracted, and the direction of peak pressure change and the rate of change of response time are measured. By statistically analyzing the direction of peak pressure change and the rate of change of response time over multiple consecutive fuel injection cycles, the trend of oil pressure pulsation is obtained. When the trend of change is continuously rising or falling and the corresponding response time remains unidirectional, it is determined that the fuel pressure trend is upward or downward, and the fuel pressure trend data of the current analysis unit is determined.

[0027] In one embodiment, when processing the acquired continuous oil pressure sampling point sequence, the sampling point sequence is divided into an analysis unit of 10 sampling points each. Within each analysis unit, the pressure pulse waveform synchronized with the fuel injection action is identified, and the direction of change of the peak pressure of the pulse and the rate of change of the response time of each pulse are measured. Subsequently, the direction of change of the peak pressure and the rate of change of the response time in multiple consecutive fuel injection cycles are statistically analyzed to form pulsation trend data. According to the statistical results, when the direction of change of the peak pressure is detected to be continuously rising or falling, and the corresponding response time remains unidirectional, it is determined that the oil pressure in the analysis unit shows a continuous rising or falling trend, and this trend is recorded as the oil pressure trend direction of the current analysis unit. By performing the above operations sequentially on all analysis units, a complete oil pressure trend sequence can be formed, providing basic data for subsequent stability feature extraction.

[0028] In another embodiment, the processing steps for the continuous oil pressure sampling point sequence are as follows: the continuously collected oil pressure data is divided into analysis units of 10 sampling points each, and the pressure pulse waveform related to the injection action is extracted within each unit. Subsequently, the peak pressure direction and response time change rate of the pulse waveform are statistically analyzed over multiple cycles, and the variation patterns within each cycle are recorded. If, during the statistical process, the peak pressure of multiple consecutive cycles shows a unidirectional increase or decrease, and the response time change rate does not exhibit reverse fluctuations, then the oil pressure trend of that analysis unit is determined to be either upward or downward, and the trend type of that unit is marked. By performing this operation on all analysis units, continuous change information of the oil pressure trend throughout the entire operation can be obtained, providing data support for subsequent parameter stability assessment and callback strategies.

[0029] Preferably, the persistence and fluctuation range of statistical trend data, the extraction of operational stability features, and the generation of feature record sequences include: In the trend data, the trend segments are divided, and the number of consecutive cycles in which the oil pressure parameters and voltage change in the same direction within the segments is counted to generate a continuous record. Extract the peak-to-valley difference of oil pressure parameters and the corresponding voltage fluctuation difference in each segment, and calculate the fluctuation amplitude. By pairing continuous records with fluctuation amplitudes in chronological order, a set of correlation parameters for operational trends is formed. Based on the correlation set of operating trend parameters, the distribution of fluctuation amplitude and the concentrated interval of duration of each segment are statistically analyzed, and the stability characteristics of fuel load state are extracted. The features are organized into feature recording units according to the sampling time sequence.

[0030] In one embodiment, when processing the trend data of oil pressure and voltage parameters, trend segments are divided based on trend points of continuous unidirectional change. Within each segment, the number of consecutive cycles in which oil pressure and voltage parameters change in the same direction is counted and recorded as the persistence record of that segment. Subsequently, the difference between the peak and trough values ​​of the oil pressure parameter is extracted within each segment, and the fluctuation difference of the corresponding voltage parameter is extracted simultaneously to calculate the fluctuation amplitude of each segment. The persistence records and fluctuation amplitudes of each segment are paired in chronological order to form a complete set of operational trend parameters. Based on the set of parameters, the distribution range of the fluctuation amplitude and the concentration interval of the persistence cycle of each segment are statistically analyzed, and these data are used to calculate the stability characteristics of the fuel load state. All characteristics are organized in chronological order of sampling time to generate feature record units, providing basic data for subsequent anomaly detection and maintenance analysis.

[0031] In another embodiment, the method for processing trend data is as follows: the entire oil pressure and voltage trend sequence is sequentially divided into continuous segments, and the number of consecutive cycles in which the oil pressure and voltage parameters change in the same direction within each segment is counted as a continuous record. Within each segment, the difference between the peak and trough oil pressure values ​​and the corresponding voltage fluctuation amplitude are measured, and the segment fluctuation amplitude is calculated. The continuous records and fluctuation amplitudes of each segment are paired in chronological order to form an operational trend parameter association set. Subsequently, based on the association set, the distribution characteristics of the fluctuation amplitudes of each segment and the concentrated intervals of the continuous cycles are statistically analyzed, and stability characteristics of the fuel load state are extracted accordingly. The extracted characteristics are organized into feature recording units in chronological order of sampling time, which can be used for fuel supply stability assessment and the formulation of subsequent parameter adjustment strategies.

[0032] Preferably, based on the correlation set of operating trend parameters, the distribution of fluctuation amplitude and the concentrated interval of duration in each segment are statistically analyzed, and the stability characteristics of fuel load state are extracted, including: Histogram statistics were performed on the fluctuation range of the centralized oil pressure parameters and voltage associated with the operating trend parameters, and the quantile interval of the main frequency segment was extracted as the fluctuation range. Perform sliding window counting on the duration sequence to determine the concentrated periodic segment of the oil pressure parameter and voltage response in the same direction; Within a concentrated period, the range of fluctuation amplitude distribution is used to determine the fuel pressure amplitude determination parameters; The response times of the duration period are compared to generate a duration determination parameter; By integrating fuel pressure amplitude determination parameters and continuity determination parameters, a comprehensive score for fuel supply stability is calculated, and this score is used as a stability characteristic quantity of fuel load state.

[0033] In one embodiment, histogram statistics are performed on the fluctuation amplitudes of oil pressure and voltage parameters in the operational trend parameter correlation set to identify the dominant frequency segment and extract its quantile interval as the fluctuation amplitude distribution range. Subsequently, sliding window counting is performed on the continuous period sequence to determine the concentrated period segments where oil pressure and voltage parameters respond in the same direction. Within each concentrated period segment, the fuel pressure amplitude determination parameter is calculated using the fluctuation amplitude distribution range, and the response time of the continuous period is compared to generate a persistence determination parameter. Finally, the fuel pressure amplitude determination parameter and the persistence determination parameter are fused to calculate a comprehensive fuel supply stability score, which is used as a stability characteristic quantity of the fuel load state.

[0034] In another embodiment, for the set of associated operating trend parameters, a histogram is generated for the fluctuation amplitude data of oil pressure and voltage parameters to identify the quantile intervals of the main frequency band and determine the fluctuation amplitude distribution range. Then, a sliding window statistical analysis is performed on the continuous unidirectional response cycles of oil pressure and voltage to filter out concentrated period segments. Within each concentrated period segment, a fuel pressure amplitude determination parameter is determined based on the fluctuation amplitude distribution range, and the response time of the continuous period is compared to generate a persistence determination parameter. The fuel pressure amplitude determination parameter and the persistence determination parameter are combined to calculate a comprehensive fuel supply stability score, which is used as a stability characteristic quantity of the fuel load state.

[0035] Please refer to the following: Figure 3 Histogram analysis was performed on the fluctuation amplitudes of oil pressure and voltage parameters, and the dominant frequency range (oil pressure: 2.0-3.0 MPa, voltage: 2.5-4.0 V) was extracted as the normal fluctuation amplitude distribution range, establishing a benchmark for subsequent judgments. Next, by analyzing the time series, concentrated periodic segments in which the two parameters change in the same direction were identified. Within this key periodic segment, the determined fluctuation range was used to generate fuel pressure amplitude judgment parameters, and a persistence judgment parameter was generated by comparing the number of duration periods. These two judgment parameters were then combined to calculate a comprehensive fuel supply stability score, which serves as a key stability characteristic representing the fuel load state.

[0036] Preferably, the feature record sequence is compared item by item with pre-stored historical operation data, and the feature deviation data is calculated including: In the feature record sequence, corresponding historical running data are selected in order of sampling time to establish parameter comparison units; Within each comparison unit, the difference between the feature record point and the historical running data is calculated to obtain the deviation data; The deviation data is compared with the preset deviation threshold range to determine the deviation status of each feature item; When the deviation of any feature item exceeds the preset threshold continuously, it is recorded as a deviation exceeding the limit point, and the number of times the limit is exceeded within the continuous comparison period is counted. When the same parameter exceeds the limit more than 3 times, it is marked as an abnormal deviation item; The parameter deviation data and the deviation status of each feature item marked as an abnormal deviation item are integrated in the order of sampling time to form feature deviation data.

[0037] In one embodiment, historical operational data corresponding to the sampling time sequence are selected in the feature record sequence to establish parameter comparison units. Within each comparison unit, the difference between the feature record point and the historical operational data is calculated to obtain deviation data. The deviation data is compared with a preset deviation threshold range to determine the deviation status of each feature item. When the deviation of any feature item continuously exceeds the threshold, it is recorded as a deviation exceedance point, and the number of exceedances within a continuous comparison period is counted. When the number of exceedances for the same parameter exceeds 3 times, the parameter is marked as an abnormal deviation item. The deviation data of each parameter marked as an abnormal deviation item and the deviation status of each feature item are integrated according to the sampling time sequence to form complete feature deviation data, which is used for subsequent callback instruction triggering or operation and maintenance analysis.

[0038] In another embodiment, historical operational data corresponding to the feature record sequence are selected in chronological order of sampling time to establish a point-by-point comparison unit. Within each comparison unit, the difference between the feature record value and the historical data is calculated point by point to form a deviation data sequence, which is then compared with a preset deviation threshold to determine the deviation status of each feature item. For deviation points that continuously exceed the threshold, they are recorded as deviation exceedance points, and the number of exceedances within a continuous comparison period is counted. When the number of exceedances of the same parameter within a continuous period exceeds three, the parameter is marked as an abnormal deviation item. The deviation data of all abnormal deviation items and the deviation status of feature items are organized in chronological order of sampling time to generate feature deviation data, which is used for anomaly monitoring and status analysis of the operation and maintenance model.

[0039] Preferably, when the feature deviation data exceeds a preset range, triggering a parameter callback instruction and executing the corresponding parameter callback operation includes: When the characteristic deviation data exceeds the preset range, a parameter callback instruction is generated and sent to the operation control unit; When the control unit receives the parameter callback command, it performs real-time adjustments to the target parameters. The adjusted parameter values ​​are monitored, and the callback action is terminated immediately when the monitoring results fall back to the threshold range.

[0040] In one embodiment, when the characteristic deviation data exceeds a preset range, a parameter callback command is generated and sent to the diesel generator set's operation control unit via a communication interface. Upon receiving the command, the operation control unit adjusts the target parameters in real time according to a pre-set adjustment strategy, while continuously monitoring the adjusted parameter values. If the monitoring results fall back to a preset threshold range, the callback action is immediately stopped, and the callback execution time and adjustment range are recorded for subsequent data analysis and maintenance records.

[0041] In another embodiment, when the detected characteristic deviation data exceeds a preset range, the scheduling unit generates a corresponding parameter callback instruction and sends it to the operation control unit. Upon receiving the callback instruction, the operation control unit dynamically adjusts the specified parameters and collects parameter change data in real time during the adjustment process. It continuously monitors the adjusted parameter values, and when the parameters fall back to within the threshold range, it automatically ends the callback operation and stores the adjustment result along with the deviation data at the time of the callback trigger for establishing operation and maintenance history records and anomaly analysis.

[0042] Of particular importance, step S3 includes: After executing multiple consecutive parameter callback operations, the previous callback records and running data sequences are called to organize the running result data after the callback and generate an evaluation input set. Based on the evaluation input set, the updated stability index is calculated, and the response delay of the generator set under load is determined by comparing the frequency recovery time with the voltage regulation overshoot. The calculated response delay is compared with a preset delay threshold. When the response delay is greater than the delay threshold, a maintenance prompt signal is generated. The maintenance prompt signal and the corresponding operating data are transmitted to the monitoring terminal, and the transmission time and status confirmation information are recorded.

[0043] In one embodiment, after performing multiple consecutive parameter callback operations, the previous callback records and corresponding operating data sequences are invoked to organize the operating results after each callback, including recording the callback time, adjustment range, and the actual response of oil pressure and voltage after the callback, generating an evaluation input set. Based on the evaluation input set, updated stability indicators are calculated, including frequency recovery time and voltage regulation overshoot. By comparing these indicators with preset reference values, the response delay of the generator set under the current load is determined. Subsequently, the calculated response delay is compared with a preset delay threshold. When the response delay exceeds the threshold, a maintenance prompt signal is generated, and this prompt signal is packaged with the corresponding operating data and transmitted to the remote monitoring terminal through the communication interface. At the same time, the signal generation time, data transmission time, and status confirmation information are recorded to ensure that maintenance personnel can obtain operational anomaly prompts and historical data references in a timely manner.

[0044] In another embodiment, after multiple consecutive parameter callback operations, operational data from each callback, including injection oil pressure, output voltage, and callback command execution results, are collected and integrated into an evaluation input set. Using this evaluation input set, the generator set's stability indicators are calculated, with a focus on analyzing frequency recovery time and voltage regulation overshoot under load changes, thereby determining the response delay under the current operating state. If the response delay exceeds a pre-set delay threshold, a maintenance prompt signal is generated and simultaneously sent to the monitoring terminal along with the callback-related operational data. The signal transmission time and the monitoring terminal's confirmation receipt are also recorded to facilitate subsequent operation and maintenance analysis and decision-making.

[0045] Preferably, step S4 includes: In the running data sequence, a preset periodic threshold is set, and the parameter change amplitude in the continuous sampling segment is periodically detected. When the duration of fluctuation exceeds the preset periodic threshold, it is determined to be a long-term fluctuation trend. After detecting a long-term fluctuation trend, the process is divided into several stages according to the time sequence and the range of fluctuation amplitude, and a stage index is established for each stage. In each operational phase, the direction of parameter change and response speed are statistically analyzed, and parameter response pattern data characteristic of each phase are extracted. The stage indexes of each operational phase are mapped to the parameter response pattern data in chronological order to form an operation and maintenance data relationship set; Based on the operation and maintenance data relationship set, the parameter response pattern data is input into the prediction unit to establish an operation and maintenance model that can be used for remote monitoring and status prediction.

[0046] In one embodiment, a preset periodic threshold is set in the operational data sequence to periodically detect the changes in oil pressure and output voltage parameters within continuous sampling segments. When the duration of parameter fluctuations exceeds the preset threshold, the segment is determined to be a long-term fluctuation trend. Subsequently, based on the time sequence and the fluctuation amplitude range, the entire operation process is divided into several operational stages, and a unique stage index is established for each stage. For each operational stage, the data segment corresponding to the stage index is called to statistically analyze the direction and rate of change of oil pressure and voltage, and extract the parameter response pattern data for each stage, including key indicators such as the rate of change, peak-to-valley difference, and fluctuation duration. The stage indexes of each operational stage are mapped to the parameter response pattern data in chronological order to form a complete set of operation and maintenance data relationships. This set of relationships is then input into the prediction unit for analysis to establish an operation and maintenance model that can be used for remote monitoring and status prediction, thereby realizing dynamic monitoring of the diesel generator set's operating status and prediction of future trends.

[0047] In another embodiment, preset periodic thresholds are set for the oil pressure and output voltage parameters in the operating data sequence, and parameter changes in continuous sampling segments are detected to identify fluctuation ranges that consistently exceed the thresholds as long-term fluctuation trends. Based on these long-term fluctuation trends, the operating process is divided into several stages, and a stage index is assigned to each stage. For each stage, the direction of oil pressure and voltage parameter changes is analyzed, the time difference between the parameter change value and adjacent sampling points is calculated to generate response speed data, and the peak and valley changes of each stage are statistically analyzed to form parameter response law data with stage characteristics. All stage indices and their corresponding parameter response law data are integrated in chronological order to generate an operation and maintenance data relationship set, which is then input into the remote prediction unit. Through data fitting and state prediction algorithms, a remote monitoring and state prediction operation and maintenance model for the diesel generator set is established, thereby providing operation and maintenance personnel with continuous state analysis and operation trend prediction.

[0048] Preferably, in each operating phase, the statistical parameters' change direction and response speed are analyzed, and the parameter response pattern data representing the phase characteristics are extracted, including: During the runtime phase, the data segment corresponding to the phase index is called to perform sequential analysis of the parameter change values ​​within the phase and determine the direction of parameter change. Calculate the response time data corresponding to the parameter change values ​​between adjacent sampling points, and generate response speed data based on the ratio of parameter change value to time difference; Statistical data on the direction of parameter change and response speed are used to generate data on parameter response patterns.

[0049] In one embodiment, based on a pre-defined stage index, the data segment corresponding to each operating stage is invoked, and sequential analysis is performed on the continuous sampling points of the oil pressure parameters and output voltage parameters within that stage. By comparing the numerical changes of adjacent sampling points, the direction of change of each parameter within that stage is determined, including rising, falling, or remaining stable. Simultaneously, the parameter change value and time difference between adjacent sampling points are calculated to obtain the response speed data for each sampling point, and the distribution of parameter change direction and response speed within the stage is statistically analyzed. Based on these statistical results, parameter response pattern data for each operating stage is generated, covering information such as parameter change direction, response speed, fluctuation amplitude, and duration, providing a foundation for data comparison and trend analysis between subsequent operating stages.

[0050] In another embodiment, data segments for each operating stage are extracted sequentially according to the stage index. The changes in oil pressure and output voltage parameters within each stage are analyzed sequentially to determine the direction of parameter change. Furthermore, the numerical change and corresponding time difference for each pair of adjacent sampling points are calculated, and the ratio of these two values ​​is used to generate response speed data, reflecting the instantaneous response characteristics of the parameters to load changes. Simultaneously, the overall distribution of parameter change direction and response speed within each stage is statistically analyzed, including the duration of continuous rise or fall segments, peak change amplitude, and concentrated response speed intervals. By synthesizing these statistical results, complete parameter response pattern data for each stage is generated, providing detailed basis for stage characteristic analysis, anomaly detection, and status prediction in the remote operation and maintenance model.

[0051] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A method for generating a remote intelligent operation and maintenance model for a diesel generator set, characterized in that, Includes the following steps: Step S1: When the diesel generator set is in operation, collect the injection oil pressure parameters and the output voltage parameters at the generator terminals as operating parameters, and record the load change process and the corresponding timestamps to generate an operating data sequence; Step S2: Perform trend analysis based on the operating data sequence, extract the characteristic quantities of generator set stability, calculate the deviation value of each characteristic quantity with the historical operating data, and execute the parameter callback operation when the deviation value exceeds the preset range. Step S3: Recalculate the stability index based on the running results after multiple consecutive callbacks. Measure the response delay under load using the generator set's frequency recovery time and voltage regulation overshoot. If the response delay is greater than the preset delay threshold, generate a maintenance prompt signal and transmit the maintenance prompt signal and the corresponding running data in the running data sequence to the monitoring terminal. Step S4: When the monitoring terminal detects a long-term fluctuation trend in the running data sequence that lasts for more than a preset period, it divides the running phases and extracts the parameter response pattern data of each phase to form a set of operation and maintenance data relationships, and establishes an operation and maintenance model for remote monitoring and status prediction.

2. The method for generating a remote intelligent operation and maintenance model for a diesel generator set according to claim 1, characterized in that, Step S2 includes: The running data sequence is divided according to parameter categories, and the data correspondence between each parameter is determined; The changes in the values ​​of voltage and oil pressure parameters at continuous sampling points are calculated, and the trend of the parameters is determined. The statistical trend data's persistence and fluctuation range are analyzed to extract features of operational stability and generate a sequence of feature records. The feature record sequence is compared item by item with the pre-stored historical running data to calculate the feature deviation data; When the feature deviation data exceeds the preset range, a parameter callback instruction is triggered, and the corresponding parameter callback operation is executed. When the same parameter triggers a callback instruction in two consecutive trend analysis cycles, the parameter will be marked as a key monitoring state.

3. The method for generating a remote intelligent operation and maintenance model for a diesel generator set according to claim 2, characterized in that, In step S2, the calculation of the changes in the values ​​of the voltage parameter at continuous sampling points and the determination of the parameter trend data are as follows: The voltage continuous sampling point sequence was grouped into analysis units of 8 sampling points each; Within each analysis unit, a difference operation is performed on adjacent sampling points and divided by the sampling interval to obtain the voltage instantaneous rate of change sequence; Calculate the proportion of sampling points in the sequence whose absolute value of the rate of change exceeds 0.5V / ms, and record the number of consecutive occurrences of adjacent points with the same direction of change; When the number of consecutive rate of change points in the same direction exceeds 4 and the cumulative voltage offset within the cell exceeds 2V, the voltage is determined to be in a continuous rising or falling phase, which serves as the trend data for the voltage parameter.

4. The method for generating a remote intelligent operation and maintenance model for a diesel generator set according to claim 2, characterized in that, In step S2, the calculation of the changes in the values ​​of the continuously sampled oil pressure parameters and the determination of the parameter trend data are as follows: The continuous hydraulic pressure sampling point sequence was grouped into analysis units of 10 sampling points each; Within each analysis unit, the pressure pulse waveform synchronized with the fuel injection action is extracted, and the direction of peak pressure change and the rate of change of response time are measured. By statistically analyzing the direction of peak pressure change and the rate of change of response time over multiple consecutive fuel injection cycles, the trend of oil pressure pulsation is obtained. When the trend of change is continuously rising or falling and the corresponding response time remains unidirectional, it is determined that the fuel pressure trend is upward or downward, and the fuel pressure trend data of the current analysis unit is determined.

5. The method for generating a remote intelligent operation and maintenance model for a diesel generator set according to claim 2, characterized in that, The statistical trend data's persistence and fluctuation range are analyzed to extract operational stability features and generate feature record sequences, including: In the trend data, the trend segments are divided, and the number of consecutive cycles in which the oil pressure parameters and voltage change in the same direction within the segments is counted to generate a continuous record. Extract the peak-to-valley difference of oil pressure parameters and the corresponding voltage fluctuation difference in each segment, and calculate the fluctuation amplitude. By pairing continuous records with fluctuation amplitudes in chronological order, a set of correlation parameters for operational trends is formed. Based on the correlation set of operating trend parameters, the distribution of fluctuation amplitude and the concentrated interval of duration of each segment are statistically analyzed, and the stability characteristics of fuel load state are extracted. The features are organized into feature recording units according to the sampling time sequence.

6. The method for generating a remote intelligent operation and maintenance model for a diesel generator set according to claim 5, characterized in that, Based on the correlation set of operating trend parameters, the distribution of fluctuation amplitude and the concentrated interval of duration in each segment are statistically analyzed, and the stability characteristics of fuel load state are extracted, including: Histogram statistics were performed on the fluctuation range of the centralized oil pressure parameters and voltage associated with the operating trend parameters, and the quantile interval of the main frequency segment was extracted as the fluctuation range. Perform sliding window counting on the duration sequence to determine the concentrated periodic segment of the oil pressure parameter and voltage response in the same direction; Within a concentrated period, the range of fluctuation amplitude distribution is used to determine the fuel pressure amplitude determination parameters; The response times of the duration period are compared to generate duration determination parameters; By integrating fuel pressure amplitude determination parameters and continuity determination parameters, a comprehensive score for fuel supply stability is calculated, and this score is used as a stability characteristic quantity of fuel load state.

7. The method for generating a remote intelligent operation and maintenance model for a diesel generator set according to claim 2, characterized in that, The feature record sequence is compared item by item with pre-stored historical running data, and the feature deviation data is calculated, including: In the feature record sequence, corresponding historical running data are selected in order of sampling time to establish parameter comparison units; Within each comparison unit, the difference between the feature record point and the historical running data is calculated to obtain the deviation data; The deviation data is compared with the preset deviation threshold range to determine the deviation status of each feature item; When the deviation of any feature item exceeds the preset threshold continuously, it is recorded as a deviation exceeding the limit point, and the number of times the limit is exceeded within the continuous comparison period is counted. When the same parameter exceeds the limit more than 3 times, it is marked as an abnormal deviation item; The parameter deviation data and the deviation status of each feature item marked as an abnormal deviation item are integrated in the order of sampling time to form feature deviation data.

8. The method for generating a remote intelligent operation and maintenance model for a diesel generator set according to claim 2, characterized in that, When the feature deviation data exceeds the preset range, a parameter callback instruction is triggered, and the corresponding parameter callback operation is executed, including: When the characteristic deviation data exceeds the preset range, a parameter callback instruction is generated and sent to the operation control unit; When the control unit receives the parameter callback command, it performs real-time adjustments to the target parameters. The adjusted parameter values ​​are monitored, and the callback action is terminated immediately when the monitoring results fall back to the threshold range.

9. The method for generating a remote intelligent operation and maintenance model for a diesel generator set according to claim 1, characterized in that, Step S4 includes: In the running data sequence, a preset periodic threshold is set, and the parameter change amplitude in the continuous sampling segment is periodically detected. When the duration of fluctuation exceeds the preset periodic threshold, it is determined to be a long-term fluctuation trend. After detecting a long-term fluctuation trend, the process is divided into several stages according to the time sequence and the range of fluctuation amplitude, and a stage index is established for each stage. In each operational phase, the direction of parameter change and response speed are statistically analyzed, and parameter response pattern data characteristic of each phase are extracted. The stage indexes of each operational phase are mapped to the parameter response pattern data in chronological order to form an operation and maintenance data relationship set; Based on the operation and maintenance data relationship set, the parameter response pattern data is input into the prediction unit to establish an operation and maintenance model that can be used for remote monitoring and status prediction.

10. The method for generating a remote intelligent operation and maintenance model for a diesel generator set according to claim 9, characterized in that, In each operational phase, the direction of parameter change and response speed are statistically analyzed, and the parameter response patterns characteristic of each phase are extracted, including: During the runtime phase, the data segment corresponding to the phase index is called to perform sequential analysis of the parameter change values ​​within the phase and determine the direction of parameter change. Calculate the response time data corresponding to the parameter change values ​​between adjacent sampling points, and generate response speed data based on the ratio of parameter change value to time difference; Statistical data on the direction of parameter change and response speed are used to generate data on parameter response patterns.

Images