Anti-seismic emergency power supply vehicle load data processing method based on time series analysis

Abstract

The application discloses an anti-seismic emergency power supply vehicle load data processing method based on time series analysis and relates to the technical field of machine learning. Position correction is performed based on load electrical parameter data and attitude state data to form an anti-seismic emergency multi-source time series data set reflecting the synchronous relationship between the load change process and the disturbance process; the time sequence correlation between the load electrical parameter change and the auxiliary state change is analyzed to select external disturbance suppression or load switching abnormality; the load state and the key load power supply stability are evaluated to generate power supply adjustment instructions, which are helpful to solve the problem that in the disaster relief scene, especially in the case that the ground is uneven after the earthquake and the vehicle body parking attitude is easy to change, the real load switching event and the external disturbance pseudo-abnormality may be mixed in the time sequence representation and difficult to distinguish, leading to distorted load state recognition and further affecting the stable control of the anti-seismic emergency power supply vehicle on the emergency power supply process.

Description

Technical Field

[0001] This invention relates to the field of machine learning technology, and in particular to a method for processing load data of earthquake-resistant emergency power supply vehicles based on time series analysis. Background Technology

[0002] Earthquake-resistant emergency power supply vehicles are crucial mobile power supply equipment used for emergency power restoration, temporary power supply, and rapid power restoration of critical electrical equipment after an earthquake. They typically need to continuously supply power and monitor externally connected loads under conditions of road damage, complex load types, and significant fluctuations in on-site operating conditions. To ensure the safety, stability, and rational scheduling of the power supply process, existing technologies generally require the collection, recording, transmission, and analysis of various load operation data during the earthquake-resistant emergency power supply vehicle's operation. This forms a corresponding load data processing flow, providing a data foundation for load status assessment, anomaly identification, and subsequent power supply adjustments.

[0003] Existing earthquake-resistant emergency power vehicles typically require continuous data collection on the voltage, current, active power, reactive power, frequency, and load fluctuations of connected loads during post-earthquake power restoration, temporary power supply, and rapid restoration of critical loads. This data is recorded, stored, and analyzed in a time-series manner to understand load change patterns, identify abnormal fluctuations, and provide a basis for power dispatching. Generally, existing technologies rely on monitoring devices or data acquisition terminals mounted on the emergency power vehicle to acquire corresponding load data according to a preset sampling period. This load data includes, but is not limited to, the three-phase voltage, three-phase current, active power, reactive power, and apparent power of each load branch, forming a chronological load data sequence. Data cleaning, noise reduction, missing data filling, time-series alignment, and statistical analysis are then used to obtain results suitable for load assessment. Furthermore, historical operating curves are combined to determine load change trends, peak-valley characteristics, and sudden increases or decreases.

[0004] However, in earthquake relief scenarios, earthquake-resistant emergency power supply vehicles usually need to perform emergency power supply tasks under conditions of uneven ground after an earthquake and easily changing vehicle parking posture. Unlike the relatively stable load-bearing environment faced by general emergency power supply vehicles, loads such as medical equipment, communication equipment, pumping equipment, and temporary lighting equipment after an earthquake often have the characteristics of irregular start-stop, strong impact fluctuations, and frequent short-term changes. At the same time, changes in vehicle tilt and support status can easily disturb the on-board monitoring results, resulting in the collected load time series being mixed with real load change information and false anomaly information caused by posture disturbances and wiring fluctuations.

[0005] Existing technologies mostly rely on standardized sampling, smoothing, and trend judgment of load data under conventional load conditions. This makes it difficult to effectively distinguish between the actual load switching behavior after an earthquake and the non-steady-state anomalies caused by external disturbances. Specifically, when medical equipment, drainage equipment, or communication equipment are connected, disconnected, or switch operating modes in a short period of time, the resulting surges in current, power jumps, and frequency fluctuations can easily be confused with the instantaneous sampling offsets, amplitude jitters, and local distortions caused by vehicle tilting or changes in support status in the time series. For voltage, current, and power data collected from different monitoring channels, existing processing methods typically only perform denoising, smoothing, or threshold judgment within a fixed window. On the one hand, this can easily misjudge actual load mutations as abnormal disturbances; on the other hand, it can misidentify pseudo-anomalies caused by external disturbances as normal load fluctuations. This leads to distorted load status assessment results, delayed or biased anomaly identification, and affects the timeliness and accuracy of power supply regulation and decision-making for critical loads by earthquake-resistant emergency power supply vehicles.

[0006] At earthquake relief sites, when earthquake-resistant emergency power vehicles simultaneously supply power to temporary medical points, communication base stations, and pumping equipment, if a high-power pumping device is suddenly activated due to on-site emergency needs, the corresponding monitoring channel will show changes in current instantaneously increasing, active power rapidly changing, and frequency fluctuating for a short time. At the same time, if the power vehicle is parked on uneven ground or its support status changes, it may also produce instantaneous sampling offsets, waveform jitter, or local distortions similar to the above load switching characteristics within a similar time period, making it difficult for the monitoring system to accurately determine whether the time-series changes originate from actual load access behavior or external disturbance factors.

[0007] In disaster relief scenarios, especially when performing emergency power supply tasks on uneven ground after an earthquake and when the vehicle's parking posture is easily changed, there may be a mixture and difficulty in distinguishing between real load switching events and pseudo-anomalies caused by external disturbances in the time series representation. This leads to a decrease in the accuracy of load status identification and further affects the stable control of the earthquake-resistant emergency power supply vehicle on the emergency power supply process. Summary of the Invention

[0008] This invention provides a method for processing load data of an earthquake-resistant emergency power supply vehicle based on time series analysis. The specific implementation is as follows: Load electrical parameter data of the earthquake-resistant emergency power supply vehicle is extracted, and attitude state data is acquired simultaneously. Position correction is performed based on the load electrical parameter data and attitude state data to form an earthquake-resistant emergency multi-source time-series dataset reflecting the synchronous relationship between load change processes and disturbance processes under a unified time axis. The temporal correlation between load electrical parameter changes and auxiliary state change sequences formed by arranging attitude state data according to a unified time axis is analyzed based on the earthquake-resistant emergency multi-source time-series dataset, generating corresponding anomaly source differentiation information. External disturbance suppression or load switching anomalies are selected based on the anomaly source differentiation information. The current load state of the earthquake-resistant emergency power supply vehicle and the power supply stability of key loads are evaluated, and corresponding power supply adjustment commands and operation control results are generated. The load electrical parameter data includes the average branch voltage of the emergency power supply branch, the total voltage on the output side of the earthquake-resistant emergency power supply vehicle, and the total current on the output side of the earthquake-resistant emergency power supply vehicle. The attitude state data includes grounding pressure, vehicle body lateral tilt angle, and vehicle body longitudinal tilt angle.

[0009] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: 1. By extracting load electrical parameter data from the earthquake-resistant emergency power supply vehicle and simultaneously acquiring attitude status data, position correction is performed based on the load electrical parameter data and attitude status data to form a multi-source time-series dataset for earthquake-resistant emergency power supply that reflects the synchronous relationship between load change process and disturbance process under a unified time axis. This helps improve the time-series alignment and data correlation capabilities of the earthquake-resistant emergency power supply vehicle for multi-source heterogeneous monitoring data under complex post-earthquake conditions, and is conducive to more accurately characterizing the real load change state and external disturbance impact state during the emergency power supply process. Based on the analysis of the time-series correlation between load electrical parameter changes and auxiliary state change sequences formed by arranging attitude status data according to a unified time axis, corresponding anomaly source differentiation information is generated. Based on the anomaly source differentiation information, external disturbance suppression or load switching anomaly is selected, which helps to effectively identify the source of abnormal load fluctuations and accurately distinguish between external disturbance anomalies and real load switching anomalies. The current load state and key load power supply stability of the earthquake-resistant emergency power supply vehicle are evaluated, and corresponding power supply adjustment commands and operation control results are generated, which helps to achieve state adaptive adjustment, priority protection of key loads, and timely early warning of abnormal risks during the earthquake-resistant emergency power supply process.

[0010] 2. Based on the sampling timestamp differences between load electrical parameter data and attitude state data, corresponding time-series compensation parameters are generated. Position correction is performed according to the time-series compensation parameters. After correction, a multi-source time-series dataset for earthquake emergency response is obtained, and the synchronous change relationship and response relationship between load change sequence and auxiliary state change sequence are analyzed to generate anomaly source differentiation information. Compared with the shortcomings of existing methods, such as asynchronous sampling of multi-source monitoring data, low data alignment accuracy, reliance on a single moment or fixed threshold for anomaly judgment, and difficulty in accurately identifying anomaly sources, its advantages are that it can combine time offset characteristics to perform unified compensation and fine alignment of data, and further perform correlation analysis of anomaly sources based on synchronous change relationship and response relationship, thereby improving the accuracy and reliability of anomaly identification in earthquake emergency scenarios.

[0011] 3. By distinguishing the source of anomalies as external disturbances, external disturbance suppression is initiated. After suppression, the current load status of the earthquake emergency power supply vehicle and the power supply stability of critical loads are evaluated. Conversely, after the load switching anomaly suppression is completed, the current load status of the earthquake emergency power supply vehicle and the power supply stability of critical loads are evaluated. Compared with the shortcomings of existing methods, which often easily include false anomalies caused by external disturbances in the load assessment or misjudge real load switching fluctuations as external interference, this method has the advantage of being able to take corresponding suppression measures for anomalies from different sources before entering the subsequent evaluation stage. This reduces the interference of false anomalies or sudden switching on the evaluation results and improves the authenticity and pertinence of load status identification and critical load power supply stability assessment.

[0012] 4. In cases where a short-term connection of a high-power load causes a significant current surge on the output side of the earthquake-resistant emergency power supply vehicle, the load assessment parameters and load stability assessment parameters are input into the emergency abnormal state assessment model when the safe load-bearing conditions are met. This generates corresponding power supply adjustment commands and operation control results. When the safe load-bearing conditions are not met, an overload warning is output. Compared with existing methods, which often have the disadvantage of difficulty in distinguishing between tolerable impacts and dangerous overload states in a timely manner, this method has the advantage of being able to first perform a safe load-bearing check based on the current surge intensity during emergency power supply before deciding whether to enter the adjustment assessment process. This improves the earthquake-resistant emergency power supply vehicle's adaptability to sudden impact loads and its overload risk prevention and control capabilities, ensuring the continuity of power supply to critical loads and the overall safety of power supply operation. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings involved in the description of the embodiments will be briefly introduced below. Obviously, the following drawings are only drawings of some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without departing from the concept of the present invention.

[0014] Figure 1 This is a flowchart of the earthquake-resistant emergency power supply vehicle load data processing method based on time series analysis provided in an embodiment of the present invention; Figure 2 This is the finite element mesh model of the power supply vehicle provided in this embodiment of the invention; Figure 3 This is a horizontal seismic time history waveform provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of position correction provided in an embodiment of the present invention; Figure 5 This is a schematic diagram illustrating the generation of anomaly source differentiation information provided in an embodiment of the present invention; Figure 6 These are comparison charts of the preprocessing effects of load time series data provided in the embodiments of the present invention. In the chart, (a) is the original load time series data, (b) is the data after anomaly removal and interpolation, and (c) is the data after sliding smoothing filtering. Detailed Implementation

[0015] The technical objectives, implementation schemes, and beneficial effects of the embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0016] Before providing a detailed explanation of the embodiments of this application, the application scenarios in the embodiments of this application will be described first.

[0017] Example 1, as Figure 1 The flowchart shown is a method for processing load data of earthquake-resistant emergency power supply vehicles based on time series analysis. The processing flow of this method may include the following steps: The load electrical parameter data of the earthquake-resistant emergency power supply vehicle is extracted, and the attitude state data representing the vehicle body attitude state and support state are acquired simultaneously. Based on the load electrical parameter data and attitude state data, position correction is performed to form an earthquake-resistant emergency multi-source time series dataset that reflects the synchronous relationship between the load change process and the disturbance process under a unified time axis. This is beneficial to the time sequence alignment and correlation representation capabilities of data in earthquake-resistant emergency scenarios, and provides a unified and reliable data foundation for distinguishing between real load changes and the impact of external disturbances.

[0018] Based on the analysis of multi-source time-series datasets for earthquake emergency response, the temporal correlation between load electrical parameter changes and auxiliary state change sequences formed by arranging attitude state data according to a unified time axis is analyzed. Corresponding anomaly source differentiation information is generated, and external disturbance suppression or load switching anomalies are selected based on the anomaly source differentiation information. This helps to improve the accuracy of identifying the source of abnormal fluctuations and reduce the situation where false anomalies caused by changes in vehicle attitude and support status are misjudged as real load switching anomalies.

[0019] The system assesses the current load status and power supply stability of critical loads of the earthquake-resistant emergency power supply vehicle, and generates corresponding power supply adjustment commands and operation control results. This helps improve the accuracy of status assessment and the timeliness of adjustment response during emergency power supply, thereby enhancing the continuous power supply guarantee capability of critical loads and the overall operational stability of the earthquake-resistant emergency power supply vehicle.

[0020] In this embodiment, the present invention focuses on the emergency power supply process of an earthquake-resistant emergency power supply vehicle under complex post-earthquake conditions. First, it synchronously acquires load electrical parameter data and attitude state data characterizing the vehicle's attitude and support status, and combines this with position correction to form a unified time-series dataset for earthquake-resistant emergency power supply. This improves the temporal consistency and correlation accuracy of the multi-source heterogeneous data from the source, providing a reliable data foundation for subsequent anomaly analysis. Based on this, it further generates anomaly source differentiation information based on the temporal correlation between changes in load electrical parameters and changes in auxiliary status. This allows subsequent processing to move beyond a general identification of anomalies and instead select external disturbance suppression or load switching anomaly handling methods based on the anomaly source, thereby enhancing the connection and specificity between preceding data processing and subsequent anomaly determination. Subsequently, based on the specifically processed data, it evaluates the current load status of the earthquake-resistant emergency power supply vehicle and the power supply stability of key loads, generating corresponding power supply adjustment commands and operational control results. This ensures that the anomaly identification results can directly serve subsequent power supply adjustment and operational control.

[0021] It should be noted that the semi-trailer and body of the earthquake-resistant emergency power supply vehicle provided by this invention have been evaluated through earthquake resistance testing. Figure 2 The image shows a finite element mesh model of a power supply vehicle provided in an embodiment of this invention. The power supply vehicle is appropriately simplified by removing semi-trailer accessories such as wheels and transmission mechanisms, while retaining load-bearing components such as longitudinal beams, crossbars, and outriggers. The mass of the removed parts is evenly distributed across the load-bearing components by adjusting material density. Equipment inside the vehicle, such as the diesel generator set, fuel tank, electrical cabinet, and cable winch, is simplified to concentrated mass. A structural finite element model is established using solid elements and a small number of plate and shell elements. The model has a total of 100,595 nodes and 123,111 elements. Figure 3 As shown, this is a horizontal earthquake time history waveform provided in an embodiment of the present invention. The horizontal axis represents time, and the vertical axis represents ground motion acceleration, which is used to show the time history variation of ground motion in the horizontal direction.

[0022] Furthermore, load electrical parameter data of the earthquake-resistant emergency power supply vehicle is extracted, and attitude status data characterizing the vehicle's attitude and support status are acquired simultaneously. The specific process is as follows: At the preset power distribution node of the earthquake-resistant emergency power supply vehicle, load electrical parameter data corresponding to the load operating status are collected according to a preset sampling period. The load electrical parameter data includes the average branch voltage of the emergency power supply branch, the total voltage of the output side of the earthquake-resistant emergency power supply vehicle, and the total current of the output side of the earthquake-resistant emergency power supply vehicle. Among them, the average branch voltage of the emergency power supply branch refers to the voltage value obtained by averaging multiple voltage values ​​collected by voltage sensors at the output end of a certain emergency power supply branch of the earthquake-resistant emergency power supply vehicle within a preset analysis period. The total voltage of the output side of the earthquake-resistant emergency power supply vehicle refers to the overall output voltage value corresponding to the main power distribution bus of the earthquake-resistant emergency power supply vehicle monitored by voltage sensors. The total current of the output side of the earthquake-resistant emergency power supply vehicle refers to the overall output current value corresponding to the main power distribution bus of the earthquake-resistant emergency power supply vehicle monitored by current sensors.

[0023] At a predetermined position on the vehicle body of the earthquake-resistant emergency power supply vehicle, attitude status data characterizing the vehicle's tilt degree are collected. The attitude status data includes ground pressure, lateral tilt angle, and longitudinal tilt angle. Ground pressure refers to the pressure value exerted on the ground by the predetermined ground bearing point of the earthquake-resistant emergency power supply vehicle, monitored by a pressure sensor. Lateral tilt angle refers to the tilt angle of the vehicle body relative to the horizontal reference plane in the left or right lateral direction, monitored by a tilt sensor. Longitudinal tilt angle refers to the tilt angle of the vehicle body relative to the horizontal reference plane in the front or rear longitudinal direction, monitored by a tilt sensor. The timestamps corresponding to the load electrical parameter data and attitude status data are read, and corresponding time-series compensation parameters are generated based on the difference in sampling timestamps between the load electrical parameter data and attitude status data. Position correction is performed according to the time-series compensation parameters, and after correction, a multi-source time-series dataset for earthquake-resistant emergency power supply is obtained.

[0024] Based on the sampling timestamp differences between load electrical parameter data and attitude state data, corresponding time-series compensation parameters are generated. The specific process is as follows: The acquisition timestamps corresponding to each data point in the load electrical parameter data and attitude state data are obtained respectively; for the acquisition timestamps corresponding to the load electrical parameter data and attitude state data within the same preset analysis time window, the time difference between the two is calculated to obtain the corresponding time offset, which is the absolute value of the difference between the acquisition timestamps corresponding to the load electrical parameter data monitored by the timer and the acquisition timestamps corresponding to the attitude state data. This can transform the sampling asynchrony problem between different monitoring channels in seismic emergency scenarios, which was originally difficult to directly unify, into a quantifiable time offset problem; the time offset is then calculated based on the sampling timestamp differences between the load electrical parameter data and attitude state data. The time offset obtained from the data is input into a preset seismic emergency timing compensation model, and the output is the timing compensation parameters of the seismic emergency power supply vehicle. The timing compensation parameters include offset direction parameters and offset amount parameters, which can not only determine the leading or lagging relationship of the corresponding data point relative to the unified time axis reference position, but also determine the length of the time difference that the corresponding data point needs to be corrected. This provides a direct basis for subsequent targeted forward or backward correction of the load electrical parameter data and attitude status data. Among them, the offset direction parameter represents the leading or lagging direction of the load electrical parameter data or attitude status data relative to the unified time axis reference position, and the offset amount parameter is used to characterize the length of the time difference that the corresponding load electrical parameter data and attitude status data need to be corrected under the unified time axis.

[0025] It should be noted that the preset seismic emergency time-series compensation model can be constructed based on the random forest regression algorithm: Historical load electrical parameter data and historical attitude state data of multiple seismic emergency power supply vehicles under different emergency power supply scenarios, as well as the historical acquisition timestamps corresponding to each data point, are obtained; based on the historical acquisition timestamps, the historical time offset between the historical load electrical parameter data and the historical attitude state data is calculated, and the corresponding target offset direction parameter and target offset parameter are marked to form a model training sample set. The sample input information in the model training sample set includes at least the historical time offset and its corresponding historical load electrical parameter data and historical attitude state data, while the sample output information includes at least the target offset direction parameter and target offset parameter; the model training sample set is then input into the random forest regression model for training, so that the random forest regression model learns the mapping relationship between the time offset, the characteristics of load electrical parameter changes, the characteristics of attitude state changes, and the time-series compensation parameters.

[0026] After the model training is completed, validation samples are selected to verify the accuracy of the trained random forest regression model. When the error between the offset direction parameter and offset parameter output by the model and the corresponding labeled result in the validation sample meets the preset error threshold, the trained random forest regression model is determined as the preset earthquake emergency time series compensation model, which is used to output the corresponding time series compensation parameters according to the current time offset and related state characteristics.

[0027] Position correction is performed according to the time-series compensation parameters. After the correction is completed, a multi-source time-series dataset for earthquake-resistant emergency power supply vehicle is obtained. Specifically, based on the obtained time-series compensation parameters of the earthquake-resistant emergency power supply vehicle, forward or backward correction is performed on the load electrical parameter data and attitude state data. After the position correction is completed, the load electrical parameter data and attitude state data are aligned according to a unified time axis to form a multi-source time-series dataset for earthquake-resistant emergency power supply vehicle containing load electrical parameter data and attitude state data.

[0028] like Figure 4 The diagram shown is a position correction schematic provided by an embodiment of the present invention. The specific process of performing forward or backward correction on the load electrical parameter data and attitude state data is as follows: When the offset direction parameter is displayed as a lag direction, that is, when the acquisition time of the load electrical parameter data or attitude state data lags behind the unified time axis reference position, the data point is forward corrected according to the corresponding offset parameter. Specifically, the corresponding position of the data point corresponding to the load electrical parameter data or attitude state data on the time axis is moved along the time advance direction by the time difference length corresponding to the offset parameter.

[0029] When the offset direction parameter is displayed as the leading direction, that is, when the load electrical parameter data or attitude status data acquisition time is ahead of the unified time axis reference position, the data point is corrected by shifting it backward according to the corresponding offset parameter. Specifically, the corresponding position of the data point of the load electrical parameter data or attitude status data on the time axis is moved along the time delay direction by the time difference length corresponding to the offset parameter.

[0030] By using the offset direction parameter to determine the leading or lagging direction of the data point relative to the unified time axis reference position, and using the offset parameter to determine the corresponding time correction length, forward or backward correction can be performed on the load electrical parameter data and attitude state data. This can bring the originally time-misaligned data points closer to the target position on the unified time axis, thereby improving the time alignment between load change data and attitude state data and reducing the correlation analysis deviation caused by inconsistent sampling times.

[0031] Interpolation compensation is performed for vacant positions appearing on the unified time axis after forward shift correction or backward shift correction; interpolation compensation refers to filling compensation data into corresponding vacant positions according to a preset interpolation rule based on valid data points at adjacent moments before and after the vacant position, wherein the preset interpolation rule may be linear interpolation; this helps reduce the problem of discontinuity of the data sequence after position correction, thereby facilitating maintaining the integrity and continuity of the multi-source time series data set for seismic emergency response, and avoiding the impact of local data missing on subsequent anomaly identification and state assessment results; for example, when a vacant position is formed at moment t2 of the unified time axis after a data point of a certain attitude state data is corrected by backward shift, the previous valid data point of the vacant position is located at moment t1 with a corresponding data value a1, and the next valid data point is located at moment t3 with a corresponding data value a3, and t1<t2<t3, the linear interpolation rule can be followed to perform linear interpolation calculation on the data value a1 and the data value a3 according to the relative position of moment t2 between moment t1 and moment t3, so as to obtain compensation data corresponding to moment t2, and fill the compensation data into the corresponding vacant position; for example, if the ground contact pressure corresponding to moment t1 is 12kN and the ground contact pressure corresponding to moment t3 is 10kN, and moment t2 is located at the middle position between moment t1 and moment t3, the compensated ground contact pressure corresponding to moment t2 can be calculated by linear interpolation as 11kN, and 11kN is filled into the vacant position.

[0032] For overlapping positions appearing on the unified time axis after forward shift correction or backward shift correction, fusion processing is performed on overlapping data points, which helps reduce the problem of discontinuity of the data sequence after position correction, thereby facilitating maintaining the integrity and continuity of the multi-source time series data set for seismic emergency response, and avoiding the impact of local data missing on subsequent anomaly identification and state assessment results; performing fusion processing on overlapping data points refers to, for a plurality of overlapping data points located at the same position on the unified time axis, calculating a corresponding target fusion value according to a preset fusion rule, and taking the target fusion value as the final data of the position, wherein the preset fusion rule may be averaging, for example, if the total output side currents corresponding to two overlapping data points at the same moment are 98A and 102A respectively, 100A can be taken as the target fusion value corresponding to the moment, and 100A is taken as the final data at the position on the unified time axis.

[0033] In this embodiment, by collecting the average branch voltage, total output voltage, and total output current of the emergency power supply branch at the preset power distribution nodes of the earthquake-resistant emergency power supply vehicle, and simultaneously acquiring attitude state data such as grounding pressure, vehicle body lateral tilt angle, and vehicle body longitudinal tilt angle, the load electrical parameter data and the attitude state data characterizing the vehicle body attitude and support status can jointly participate in subsequent analysis. This allows the load change process in the earthquake-resistant emergency scenario to no longer be judged as an isolated electrical phenomenon, but to be associated with external factors such as vehicle body attitude disturbance and grounding force changes. Furthermore, by reading the acquisition timestamps corresponding to each data point, a timing compensation parameter is generated based on the sampling timestamp difference, and the load electrical parameter data and attitude state data are corrected by forward or backward shifting. At the same time, interpolation compensation for missing positions and fusion of overlapping positions are combined. This combined processing effectively improves the temporal misalignment problem caused by asynchronous sampling, sampling delay, or local overlap of multi-source monitoring data in earthquake emergency scenarios, thereby forming a unified time-axis multi-source time-series dataset for earthquake emergency, improving the temporal consistency and correspondence accuracy of multi-source heterogeneous data. On this basis, the load change process and attitude disturbance process can establish a more accurate synchronization and response relationship under the same time reference, which is conducive to more reliable identification of the actual load changes and the impact of external disturbances. Therefore, this invention not only helps to improve the multi-source data alignment capability and data availability of earthquake emergency power vehicles under complex post-earthquake conditions, but also helps to provide a more accurate data foundation for subsequent anomaly source identification, load status assessment, and power supply regulation and control, thereby improving the reliability of power supply guarantee for critical loads and the stability of emergency power supply operation of the whole vehicle.

[0034] Furthermore, based on the seismic emergency multi-source time-series dataset, the temporal correlation between load electrical parameter changes and auxiliary state change sequences formed by arranging attitude state data according to a unified time axis is analyzed. Specifically, this involves: acquiring load electrical parameter data and attitude state data within the same analysis time window from the seismic emergency multi-source time-series dataset; within the same analysis time window, forming a load change sequence from the load electrical parameter data arranged according to a unified time axis, and forming an auxiliary state change sequence from the attitude state data arranged according to a unified time axis; extracting the change amplitude and duration corresponding to the load change sequence and auxiliary state change sequence respectively; and based on the change amplitude and duration, analyzing the synchronous change relationship and response relationship between the load change sequence and the auxiliary state change sequence to generate anomaly source differentiation information.

[0035] By acquiring load electrical parameter data and attitude status data within the same analysis time window from a multi-source time-series dataset for earthquake emergency response, and forming load change sequences and auxiliary status change sequences under a unified time axis, the electrical load change process of the earthquake emergency power supply vehicle can be analyzed in correspondence with the auxiliary status change processes such as vehicle attitude and support status under the same time-series reference. Furthermore, by extracting the change amplitude and duration corresponding to the load change sequence and auxiliary status change sequence respectively, and analyzing the synchronous change relationship and response relationship between the two based on the change amplitude and duration, the anomaly judgment originally based solely on fluctuations of a single electrical parameter can be expanded to a correlation judgment combined with the external disturbance state. This allows for a more accurate distinction between abnormal fluctuations caused by actual load switching and pseudo-abnormal fluctuations caused by external factors such as attitude changes and support changes. Therefore, this process helps improve the accuracy and specificity of anomaly source identification in earthquake emergency scenarios, reduces misjudgments and omissions, provides a reliable basis for subsequent selection of external disturbance suppression or load switching anomaly handling methods, and thus improves the credibility of load state assessment and critical load power supply stability analysis.

[0036] like Figure 5 The diagram illustrates the generation of anomaly source differentiation information according to an embodiment of this invention. The specific process for generating the anomaly source differentiation information is as follows: The parameter changes of the load change sequence and the auxiliary state change sequence within the current analysis time window are obtained, and the corresponding parameter changes are used as the load change amplitude and the auxiliary state change amplitude, respectively. Specifically, the voltage difference between the average branch voltage of the emergency power supply branch and the voltage at the previous moment or the beginning of the window, the change in the total voltage of the earthquake-resistant emergency power supply vehicle output side relative to the corresponding preset voltage, and the change in the total current of the earthquake-resistant emergency power supply vehicle output side relative to the corresponding preset current value are used as the parameter changes corresponding to the load change sequence. The change in grounding pressure relative to the corresponding preset grounding pressure and the vehicle body lateral tilt angle can also be considered. The changes in the angle of the horizontal reference position and the changes in the longitudinal tilt angle of the vehicle body relative to the horizontal reference position are used as the parameter changes corresponding to the auxiliary state change sequence. For example, when the average branch voltage of the emergency power supply branch changes from 220V to 212V, the total voltage on the output side of the earthquake-resistant emergency power supply vehicle changes from 400V to 389V, and the total current on the output side changes from 80A to 108A within a certain analysis time window, 8V, 11V, and 28A can be used as the corresponding load change amplitudes, respectively. When the ground pressure changes from 12kN to 9kN, the lateral tilt angle of the vehicle body changes from 1.2° to 2.8°, and the longitudinal tilt angle of the vehicle body changes from 0.8° to 2.1° within the same analysis time window, 3kN, 1.6°, and 1.3° can be used as the corresponding auxiliary state change amplitudes, respectively.

[0037] The system acquires the duration segments in the load change sequence and auxiliary state change sequence monitored by timers where parameters exceed the corresponding preset limit thresholds, and uses the length of these duration segments as the load change duration and auxiliary state change duration, respectively. Specifically, the system uses the time periods when the average branch voltage of the emergency power supply branch is lower than the preset branch voltage lower limit threshold, the time periods when the total voltage on the output side of the earthquake-resistant emergency power supply vehicle deviates from the preset total voltage acceptable range, and the time periods when the total current on the output side is higher than the preset current impulse threshold as the duration segments corresponding to the load change duration. The system also uses the time periods when the ground pressure is lower than the preset ground pressure threshold and the time periods when the vehicle body lateral tilt angle is higher than the preset ground pressure threshold as the duration segments corresponding to the load change duration. The time periods for which the preset lateral tilt angle threshold and the time periods for which the vehicle body longitudinal tilt angle is higher than the preset longitudinal tilt angle threshold are used as the duration periods corresponding to the duration of auxiliary state changes. For example, when the total output current is higher than 100A for 4 consecutive seconds and the total output voltage is lower than 392V for 3 consecutive seconds within a certain analysis time window, 4 seconds and 3 seconds can be used as the corresponding durations of load changes, respectively. When the ground pressure is lower than 10kN for 5 consecutive seconds, the vehicle body lateral tilt angle is higher than 2.5° for 4 consecutive seconds, and the vehicle body longitudinal tilt angle is higher than 2.0° for 3 consecutive seconds within the same analysis time window, 5 seconds, 4 seconds, and 3 seconds can be used as the corresponding durations of auxiliary state changes, respectively.

[0038] When both the load change amplitude and the auxiliary status change amplitude meet the corresponding synchronous change conditions, and both the load change duration and the auxiliary status change duration meet the corresponding response conditions, it indicates that the current load electrical parameter changes and vehicle attitude state changes have a high degree of consistency in terms of change intensity and duration. The corresponding abnormal fluctuations are more likely caused by coupling from external attitude disturbances, grounding status changes, or support status changes. The anomaly source identification information shows an external disturbance anomaly, initiating external disturbance suppression. After suppression, the current load state of the earthquake emergency power supply vehicle and the power supply stability of critical loads are assessed. Conversely, if the load electrical parameter changes and auxiliary status changes do not form a synchronous response relationship that meets the preset conditions, the corresponding abnormal fluctuations are more likely to originate from... The actual load surge caused by the load connection, disconnection, or operation mode switching of the emergency power supply branch is displayed as a load switching anomaly in the anomaly source differentiation information. Load switching anomaly suppression is initiated, and after suppression, the current load status of the earthquake emergency power supply vehicle and the power supply stability of critical loads are evaluated. The synchronous change condition indicates that the load change amplitude and the auxiliary status change amplitude are both within the preset qualified synchronous change range. The qualified synchronous change range includes the preset qualified load change range and the preset qualified auxiliary status change range. The response condition indicates that the duration of the load change and the duration of the auxiliary status change are both within the corresponding qualified response range. The qualified response range includes the preset load change duration response range and the preset auxiliary status change duration response range.

[0039] It should be noted that, in order to facilitate a unified determination of the operating status of earthquake-resistant emergency power supply vehicles under different emergency power supply scenarios, this invention pre-constructs a database, which includes at least qualified synchronous change ranges, qualified response ranges, and preset impact thresholds. Taking the preset qualified load change range as an example, the construction process of the preset range is explained as follows: The historical load change amplitudes of multiple earthquake-resistant emergency power supply vehicles under different emergency power supply scenarios are obtained. The load operation results corresponding to each historical load change amplitude are manually marked. The load operation results include at least one or more of the following: qualified power supply status, short-term fluctuation recoverable status, and abnormal fluctuation status. Historical load change amplitude samples corresponding to the qualified power supply status and short-term fluctuation recoverable status are selected, and the corresponding load change distribution intervals are statistically obtained based on the historical load change amplitude samples. Finally, the load change distribution intervals that meet the preset coverage ratio requirements are determined as the preset qualified load change ranges and stored in the database.

[0040] Specifically, the historical load change amplitude samples after screening can be sorted according to the numerical size, and the value corresponding to the preset lower quantile in the sorted samples can be used as the preset qualified load change range lower limit, and the value corresponding to the preset upper quantile can be used as the preset qualified load change range upper limit; wherein, the preset lower quantile and preset upper quantile can be set by preset personnel according to the stability requirements under different seismic emergency power supply scenarios.

[0041] Specifically, external disturbance suppression means that data points in the multi-source time-series dataset for earthquake emergency response that meet the synchronization and response conditions are smoothed using a moving average smoothing algorithm. This can flexibly weaken short-term fluctuations caused by external disturbances such as ground pressure fluctuations, vehicle lateral tilt, or vehicle longitudinal tilt. This allows data points corresponding to external disturbance anomalies to retain the overall trend while reducing the impact of local spikes and jitters, thus helping to reduce the interference of false anomalies from external disturbances on subsequent load status assessment and critical load power supply stability analysis. For example, a preset number of adjacent data points are obtained from the multi-source time-series dataset for earthquake emergency response, and the average value of these preset number of data points is calculated. The average result is then used as the smoothed result corresponding to the target data point. Load switching anomaly suppression means that data points in the multi-source time-series dataset for earthquake emergency response that do not meet the synchronization or response conditions are removed. This helps to avoid introducing abnormal data without associated synchronization characteristics into the subsequent processing flow.

[0042] In this embodiment, based on the multi-source time-series dataset for earthquake-resistant emergency power supply vehicles, the changes in load electrical parameters and auxiliary states such as grounding pressure, vehicle body lateral tilt angle, and vehicle body longitudinal tilt angle are included in the same analysis time window and unified time axis for correlation analysis. First, by extracting the load change amplitude, auxiliary state change amplitude, and corresponding duration, a synchronous change relationship and response relationship between load-side changes and external disturbance-side changes are established. Then, based on the synchronous change relationship and response relationship, anomaly source differentiation information is generated, so that anomaly judgment is no longer limited to the isolated identification of single voltage, current, or power anomalies, but can be associated and attributed to the causes of anomalies by combining attitude state and support state changes. On this basis, when an external disturbance anomaly is determined, the external disturbance suppression process is entered; when a load switching anomaly is determined, the load switching anomaly suppression process is entered, so that anomalies from different sources can correspond to different subsequent processing paths. Finally, the results after targeted processing are used for the current load state and key load power supply stability assessment of the earthquake-resistant emergency power supply vehicle.

[0043] Furthermore, the current load status and power supply stability of the earthquake-resistant emergency power supply vehicle are evaluated. Specifically, this involves extracting load assessment parameters and load stability assessment parameters that characterize the current operating status of the earthquake-resistant emergency power supply vehicle, and inputting these parameters into the emergency abnormal state assessment model to generate corresponding power supply adjustment commands and operation control results. The power supply adjustment commands represent control command information generated based on the current load assessment parameters and load stability assessment parameters of the earthquake-resistant emergency power supply vehicle to adjust the emergency power supply output status, including one or more of the following: output power adjustment commands, voltage adjustment commands, and frequency adjustment commands.

[0044] It should be noted that the emergency abnormal state assessment model can also be constructed based on the random forest regression algorithm. Its construction process is similar to that of the preset earthquake emergency time series compensation model. The sample input information in the model training sample set corresponding to the emergency abnormal state assessment model includes at least historical load assessment parameters and historical load stability assessment parameters, and the sample output information includes at least the corresponding power supply adjustment commands and operation control results.

[0045] Among them, the output power adjustment command refers to the control command information generated to adjust the active power level allocated from the output side of the earthquake-resistant emergency power supply vehicle to each emergency power supply branch or the total output terminal, specifically including one of the following: increase output power command, decrease output power command, and maintain current output power command; voltage regulation command refers to the control command information generated to adjust the voltage level on the output side of the earthquake-resistant emergency power supply vehicle or the corresponding voltage level of each emergency power supply branch, specifically including one of the following: boost regulation command, buck regulation command, and voltage stabilization regulation command; frequency regulation command refers to the control command information generated to adjust the emergency power supply output frequency to a preset qualified frequency range, specifically including one of the following: frequency boost regulation command, frequency buck regulation command, and frequency stabilization regulation command; the operation control result represents the early warning result formed by the earthquake-resistant emergency power supply vehicle executing the corresponding power supply regulation command based on the output result of the emergency abnormal state assessment model, including power supply abnormality warning. The system should display a power supply qualification indicator; the load assessment parameters should include at least one or more of the following: total current load and load fluctuation range. Total load refers to the total load level corresponding to all emergency loads carried by the output side of the earthquake-resistant emergency power supply vehicle during the current analysis period, which can be characterized by the sum of active power monitored by power sensors in each emergency power supply branch. Load fluctuation range refers to the difference between the maximum and minimum values ​​of the total load on the output side of the earthquake-resistant emergency power supply vehicle during the current analysis period; the load stability assessment parameters should include at least one or more of the following: emergency voltage fluctuation and emergency current fluctuation. Emergency voltage fluctuation refers to the difference between the actual voltage value and the rated voltage value on the output side of the earthquake-resistant emergency power supply vehicle monitored by voltage sensors during the current analysis period. Emergency current fluctuation refers to the difference between the actual current value and the rated current value on the output side of the earthquake-resistant emergency power supply vehicle monitored by current sensors during the current analysis period.

[0046] In this embodiment, by extracting the load assessment parameters and load stability assessment parameters that characterize the current operating status of the earthquake-resistant emergency power supply vehicle, and inputting these parameters into the emergency abnormal state assessment model, the overall load status of the earthquake-resistant emergency power supply vehicle and the power supply stability status of key loads can be jointly assessed, rather than making isolated judgments based on a single electrical parameter. This is beneficial for a more comprehensive reflection of the load carrying capacity and power supply stability level during the current emergency power supply process. Furthermore, by generating corresponding power supply adjustment commands and operation control results based on the emergency abnormal state assessment model, the results of the preceding state assessment can be directly applied to the subsequent emergency power supply output adjustment process, making the output power adjustment, voltage adjustment, and frequency adjustment more targeted and real-time.

[0047] Example 2, based on the method of Example 1, assesses the current load status and power supply stability of the earthquake-resistant emergency power supply vehicle during post-earthquake emergency power supply when a high-power load is briefly connected, causing a significant current surge on the output side of the vehicle. This assessment includes: obtaining the current surge intensity on the output side of the earthquake-resistant emergency power supply vehicle during the current analysis period; determining whether the vehicle meets the safe load-bearing conditions; inputting load assessment parameters and load stability assessment parameters into the emergency abnormal state assessment model when the conditions are met to generate corresponding power supply adjustment commands and operation control results; and outputting an overload warning when the conditions are not met. The safe load-bearing conditions indicate that the current surge intensity on the output side of the earthquake-resistant emergency power supply vehicle does not exceed a preset surge threshold. The current surge intensity can be characterized by the ratio between the peak instantaneous current on the output side and the rated current threshold during the current analysis period. The peak instantaneous current represents the maximum value among the instantaneous current values ​​monitored by the current sensor at each moment during the current analysis period.

[0048] Similarly, it is necessary to explain the preset values ​​in the database of this invention, such as the construction process of the preset impact threshold: The historical current impact intensity of multiple earthquake-resistant emergency power supply vehicles under different emergency power supply scenarios is obtained. The load-bearing operation status corresponding to each historical current impact intensity is manually marked. The load-bearing operation status includes at least one or more of the following: safe load-bearing status, short-term impact recoverable status, and overload abnormal status. Historical current impact intensity samples corresponding to the safe load-bearing status and short-term impact recoverable status are selected, and the corresponding impact intensity distribution range is statistically obtained based on the historical current impact intensity samples. Finally, the upper limit value of the impact intensity distribution that meets the preset coverage ratio requirement is determined as the preset impact threshold, and the preset impact threshold is stored in the database. Specifically, the selected historical current impact intensity samples can be sorted according to the numerical size, and the value corresponding to the preset upper quantile in the sorted samples is used as the preset impact threshold. The preset upper quantile can be set by preset personnel according to the allowable short-term impact tolerance under earthquake-resistant emergency power supply scenarios, so that the preset impact threshold can cover the instantaneous current rise caused by the short-term access of normal emergency loads, and can effectively distinguish abnormal current impacts that may cause continuous overload risks.

[0049] In this embodiment, during post-earthquake emergency power supply, to address the issue that short-term access to high-power loads can easily lead to significant current surges on the output side of the earthquake-resistant emergency power supply vehicle, this invention first obtains the current surge intensity on the output side during the current analysis period and uses whether it exceeds a preset surge threshold as the basis for determining safe load conditions. This adds a pre-emptive safety check step for instantaneous surge risks before load condition assessment. This effectively distinguishes between short-term tolerable normal surges and dangerous surges that may cause overload risks, preventing the continuation of the routine assessment and adjustment process even when significant overload risks have already occurred. Furthermore, when the safe load conditions are met, the load assessment parameters and load stability assessment parameters are input into the emergency abnormal state assessment model to generate corresponding power supply adjustment commands and operation control results. This ensures that subsequent output power, voltage, and frequency adjustments are based on safe load conditions, thereby improving the reliability and targeting of power supply adjustments. Conversely, when the safe load conditions are not met, a load overload warning is directly output, enabling more timely exposure and intervention of potential overload risks.

[0050] like Figure 6 The comparison chart of the preprocessing effect of the load time series data is shown in three sub-charts: Figure (a) is the original load time series data chart, which intuitively shows the large amount of sensor noise, extreme spikes and abnormal fluctuations in the earthquake emergency multi-source time series dataset, and the data quality is poor; Figure (b) is the data chart after anomaly removal and interpolation. Through anomaly removal and interpolation completion, the data regularity is greatly improved; Figure (c) is the data chart after sliding smoothing filter. The sliding window filter effectively suppresses high-frequency noise, and the load trend is clearly distinguishable. Among them, the vertical axis of the sub-chart is the load rate, which specifically refers to the percentage ratio of the real-time output power of the emergency power vehicle to the rated power of the equipment, representing the current load operation degree of the equipment. The higher the value, the heavier the load of the equipment; the horizontal axis is the running time.

[0051] The above-disclosed embodiments are merely some examples of the present invention and should not be construed as limiting the scope of the present invention. Therefore, any equivalent variations made in accordance with the claims of the present invention are still within the scope of the present invention.

Claims

1. A method for processing load data of earthquake-resistant emergency power supply vehicles based on time series analysis, characterized in that, The method includes: The load electrical parameter data of the earthquake-resistant emergency power vehicle is extracted, and the attitude status data is acquired simultaneously. Based on the load electrical parameter data and attitude status data, position correction is performed to form a multi-source time series dataset of earthquake-resistant emergency power vehicles that reflects the synchronous relationship between the load change process and the disturbance process under a unified time axis. Based on the seismic emergency multi-source time series dataset, analyze the temporal correlation between load electrical parameter changes and auxiliary state change sequences formed by arranging attitude state data according to a unified time axis, generate corresponding anomaly source differentiation information, and select external disturbance suppression or load switching anomaly based on the anomaly source differentiation information. The current load status and power supply stability of the earthquake-resistant emergency power supply vehicle are assessed, and corresponding power supply adjustment commands and operation control results are generated. The load electrical parameter data includes the average branch voltage of the emergency power supply branch, the total voltage on the output side of the earthquake-resistant emergency power supply vehicle, and the total current on the output side of the earthquake-resistant emergency power supply vehicle. The attitude status data includes grounding pressure, vehicle body lateral tilt angle, and vehicle body longitudinal tilt angle.

2. The method for processing earthquake-resistant emergency power supply vehicle load data based on time series analysis according to claim 1, characterized in that, The specific process for extracting the load electrical parameter data of the earthquake-resistant emergency power supply vehicle and simultaneously acquiring attitude status data is as follows: At the preset power distribution nodes of the earthquake-resistant emergency power supply vehicle, load electrical parameter data corresponding to the load operating status are collected according to the preset sampling cycle. At predetermined positions on the earthquake-resistant emergency power supply vehicle body, attitude state data characterizing the degree of vehicle body tilt are collected; Read the acquisition timestamps corresponding to the load electrical parameter data and attitude status data, and generate corresponding timing compensation parameters based on the sampling timestamp differences between the load electrical parameter data and attitude status data. Perform position correction according to the timing compensation parameters, and obtain the seismic emergency multi-source timing dataset after the correction is completed.

3. The method for processing earthquake-resistant emergency power supply vehicle load data based on time series analysis according to claim 2, characterized in that, The process of generating corresponding timing compensation parameters based on the sampling timestamp differences between load electrical parameter data and attitude state data is as follows: Obtain the acquisition timestamps corresponding to each data point in the load electrical parameter data and attitude status data respectively; For the acquisition timestamps of load electrical parameter data and attitude status data that are within the same preset analysis time window, calculate the time difference between the two to obtain the corresponding time offset. The time offset obtained based on load electrical parameter data and attitude state data is input into the preset seismic emergency timing compensation model, and the timing compensation parameters of the seismic emergency power supply vehicle are output. The timing compensation parameters include offset direction parameters and offset amount parameters.

4. The method for processing earthquake-resistant emergency power supply vehicle load data based on time series analysis according to claim 3, characterized in that, The position is corrected according to the time-series compensation parameters. After the correction is completed, a multi-source time-series dataset for earthquake emergency response is obtained, specifically: Based on the obtained timing compensation parameters of the earthquake-resistant emergency power supply vehicle, forward or backward corrections are performed on the load electrical parameter data and attitude state data. After the position correction is completed, the load electrical parameter data and attitude state data are aligned according to a unified time axis to obtain an earthquake-resistant emergency multi-source timing dataset.

5. The method for processing earthquake-resistant emergency power supply vehicle load data based on time series analysis according to claim 4, characterized in that, The specific process for performing forward or backward correction on the load electrical parameter data and attitude state data is as follows: When the offset direction parameter is displayed as a lag direction, the data point is shifted forward according to the corresponding offset amount parameter. When the offset direction parameter is displayed as a leading direction, the data point is corrected by shifting it backward according to the corresponding offset amount parameter. For gaps that appear on the same timeline after forward or backward correction, interpolation compensation is performed. For overlapping positions that appear on the same time axis after forward or backward correction, the overlapping data points are fused.

6. The method for processing earthquake-resistant emergency power supply vehicle load data based on time series analysis according to claim 1, characterized in that, The analysis of the temporal correlation between load electrical parameter changes and auxiliary state change sequences formed by arranging attitude state data according to a unified time axis based on the multi-source time-series dataset for earthquake emergency response is specifically expressed as follows: Acquire load electrical parameter data and attitude status data within the same analysis time window in a multi-source time-series dataset for earthquake emergency response; Within the same analysis time window, load electrical parameter data are arranged according to a unified time axis to form a load change sequence, and attitude state data are arranged according to a unified time axis to form an auxiliary state change sequence; Extract the magnitude and duration of the load change sequence and the auxiliary state change sequence respectively. Based on the magnitude and duration, analyze the synchronous change relationship and response relationship between the load change sequence and the auxiliary state change sequence to generate anomaly source differentiation information.

7. The method for processing earthquake-resistant emergency power supply vehicle load data based on time series analysis according to claim 6, characterized in that, The specific process for generating anomaly source differentiation information is as follows: Obtain the parameter changes of the load change sequence and the auxiliary state change sequence within the current analysis time window, and use the corresponding parameter changes as the load change magnitude and the auxiliary state change magnitude. Obtain the duration segment in the load change sequence and auxiliary state change sequence monitored by the timer where the parameter exceeds the corresponding preset limit threshold, and use the time length corresponding to the duration segment as the load change duration and the auxiliary state change duration. When both the load change magnitude and the auxiliary status change magnitude meet the corresponding synchronous change conditions, and both the load change duration and the auxiliary status change duration meet the corresponding response conditions, the anomaly source identification information is displayed as an external disturbance anomaly. External disturbance suppression is initiated, and after suppression, the current load status of the earthquake emergency power supply vehicle and the power supply stability of the critical loads are evaluated. Conversely, when the anomaly source identification information is displayed as a load switching anomaly, load switching anomaly suppression is initiated, and after suppression, the current load status of the earthquake emergency power supply vehicle and the power supply stability of the critical loads are evaluated. The synchronous change condition indicates that both the load change amplitude and the auxiliary status change amplitude are within the pre-set qualified synchronous change range; The response condition indicates that both the duration of the load change and the duration of the auxiliary state change are within the corresponding qualified response range.

8. The method for processing earthquake-resistant emergency power supply vehicle load data based on time series analysis according to claim 7, characterized in that, The external disturbance suppression refers to smoothing the data points in the multi-source time-series dataset of earthquake emergency response that meet the synchronous change conditions and response conditions: The load switching anomaly suppression means removing data points from the earthquake emergency multi-source time series dataset that do not meet the synchronous change conditions or response conditions.

9. The method for processing earthquake-resistant emergency power supply vehicle load data based on time series analysis according to claim 7, characterized in that, The assessment of the current load status and power supply stability of the earthquake-resistant emergency power supply vehicle for critical loads specifically indicates: Extract the load assessment parameters and load stability assessment parameters, and input the load assessment parameters and load stability assessment parameters into the emergency abnormal state assessment model to generate corresponding power supply adjustment commands and operation control results; The power supply adjustment command represents control command information generated to adjust the emergency power supply output state based on the current load assessment parameters and load stability assessment parameters of the earthquake-resistant emergency power supply vehicle. The operation control result represents the early warning result formed by the earthquake-resistant emergency power supply vehicle executing the corresponding power supply adjustment command based on the output result of the emergency abnormal state assessment model; The load assessment parameters include at least one or more of the current total load and load fluctuation range; The load stability assessment parameters include at least one or more of the emergency voltage fluctuation and emergency current fluctuation.

10. The method for processing earthquake-resistant emergency power supply vehicle load data based on time series analysis according to claim 7, characterized in that, The assessment of the current load status and power supply stability of the earthquake-resistant emergency power supply vehicle for critical loads also includes: The current surge intensity on the output side of the earthquake-resistant emergency power supply vehicle during the current analysis period is obtained to determine whether the earthquake-resistant emergency power supply vehicle meets the safe load conditions. If the safe load conditions are met, the load assessment parameters and load stability assessment parameters are input into the emergency abnormal state assessment model to generate corresponding power supply adjustment commands and operation control results. When the safety load conditions are not met, an overload warning message will be output. The safety load condition means that the current surge intensity on the output side of the earthquake-resistant emergency power supply vehicle does not exceed the preset surge threshold.

Images