Intelligent control system for on-grid and off-grid switching of mobile energy storage vehicles

Abstract

The present invention relates to the field of mobile energy storage vehicle control technology, and discloses an intelligent control system for on-grid and off-grid switching of mobile energy storage vehicles, which includes a grid-connected state detection module and a switching control module. The grid-connected state detection module collects grid connection point data in multiple dimensions, extracts characteristics such as voltage instability coefficient and frequency offset, and obtains the grid connection state level through dynamic threshold analysis and trend analysis. Based on this level, the switching control module integrates multi-dimensional parameters such as the historical switching delay time, power output range, and remaining capacity of the energy storage unit, calculates the switching fluctuation index, unit efficiency mean, etc., obtains a priority score through multi-dimensional parameter fusion, selects the optimal execution unit, generates a task execution sequence, and dynamically displays the unit status on a visual interface. The system realizes intelligent evaluation and optimized scheduling of on-grid and off-grid switching, improves switching accuracy, stability, and efficiency, and is suitable for efficient operation and management of mobile energy storage vehicles.

Description

Technical Field

[0001] The present invention relates to the technical field of mobile energy storage vehicle control, and in particular to an intelligent control system for on-grid and off-grid switching of mobile energy storage vehicles. Background Art

[0002] With the rapid development of the new energy industry, mobile energy storage vehicles, as flexible energy storage devices, have been widely used in scenarios such as emergency power supply, grid peak regulation, and renewable energy consumption. During actual operation, mobile energy storage vehicles often need to switch between on-grid and off-grid modes to adapt to varying power demands and grid conditions. However, existing on-grid and off-grid switching control systems have numerous drawbacks, making them difficult to meet the requirements for efficient, stable, and safe operation of mobile energy storage vehicles.

[0003] From the perspective of grid-connected status monitoring, traditional monitoring methods often focus solely on a single parameter, such as grid voltage or frequency, failing to comprehensively and multi-dimensionally reflect the actual grid connection status. For example, focusing solely on voltage amplitude while ignoring factors such as frequency offset, phase synchronization deviation, and load fluctuation can lead to misjudgments of grid stability, thereby affecting the accuracy of switching decisions. Furthermore, traditional monitoring methods typically use fixed thresholds for analysis and are unable to dynamically adjust thresholds based on real-time data. This makes them difficult to adapt to dynamic changes in grid operating conditions and prone to false positives or omissions.

[0004] In terms of switching control strategies, existing systems lack a comprehensive evaluation and optimized scheduling mechanism for energy storage units. When switching between on-grid and off-grid, most systems simply switch according to the access order or fixed priority of the energy storage units, without fully considering the historical operating data, current status, and coordination with other units of each energy storage unit. For example, factors such as the switching delay time, load change type, power output range, response delay time, remaining capacity, and energy transmission distance of the energy storage unit are not taken into account, resulting in problems such as excessive load fluctuations, low switching efficiency, and shortened service life of the energy storage unit during the switching process. At the same time, the task execution sequence of the traditional system lacks a dynamic adjustment mechanism, and it is impossible to optimize the task execution sequence in a timely manner according to the real-time changes in the grid connection status, affecting the overall performance and reliability of the system.

[0005] In terms of data processing and analysis, traditional control systems handle monitoring data in a simplistic manner, typically limited to data collection and basic statistics, lacking in-depth data mining and trend analysis. For example, without constructing dynamic trend curves and performing trend analysis to predict changing trends in grid connection status, switching strategies cannot be formulated in advance, resulting in delayed system response. Furthermore, traditional systems lack scientific and rational methods for multi-parameter fusion, making it impossible to effectively integrate the multiple parameters that influence switching decisions, making it difficult to form comprehensive and accurate priority scores, which in turn affects the selection of switching units and the rationality of task allocation.

[0006] In terms of visualization and monitoring, the existing system's visualization interface has a single function and is unable to display the real-time coordinates, task status, remaining capacity, switching alarm status and other information of the energy storage unit in real time and intuitively. This is not conducive to operators to grasp the system's operation status in a timely manner, and it is difficult to discover and deal with potential problems in a timely manner. Summary of the Invention

[0007] The purpose of the present invention is to provide an intelligent control system for mobile energy storage vehicles and off-grid switching to solve the problems raised in the above background technology.

[0008] To achieve the above objectives, the present invention provides the following technical solution: an intelligent control system for switching between on-grid and off-grid mobile energy storage vehicles, the system comprising:

[0009] Grid-connected status detection module and switching control module;

[0010] The grid-connected state detection module collects multi-dimensional data based on real-time monitoring of grid connection points, including grid voltage amplitude data, frequency offset data, phase synchronization deviation data, and load fluctuation data, and extracts state features to obtain voltage instability coefficient, frequency offset, phase deviation index, and load disturbance value at each sampling moment, and performs dynamic threshold analysis on them to obtain grid-connected stability index corresponding to each sampling moment; constructs a two-dimensional coordinate system with time as the horizontal axis and grid-connected stability index as the vertical axis, marks grid-connected stability index values ​​at each sampling moment as monitoring points in the coordinate system, uses piecewise linear interpolation to connect adjacent monitoring points to generate a dynamic trend curve, performs trend analysis on the dynamic trend curve to obtain the current grid connection state level, and transmits it to the switching control module;

[0011] The switching control module executes on-grid and off-grid switching strategies and energy storage unit scheduling based on the grid connection status level.

[0012] Preferably, the specific operation process of the switching control module is:

[0013] Obtain the grid connection status levels of all energy storage units, sort them from low to high, and select the energy storage unit with the lowest level as the switching reference unit;

[0014] The historical switching delay duration of each energy storage unit and the load change type corresponding to each delay are retrieved. The interval between two adjacent switches is calculated and variance processing is performed to obtain the switching stability period. The load fluctuation amplitude of each switching is extracted and averaged to obtain the disturbance reference amplitude. The switching stability period and the disturbance reference amplitude are correlated and analyzed to obtain the switching fluctuation index of each energy storage unit.

[0015] Retrieve the historical operating data of each energy storage unit, including the power output range, response delay time, and grid connection status level of each operation. Calculate the output efficiency value by the ratio of power range to response delay time. Combined with the grid connection status level, the efficiency weight is assigned to obtain the average unit efficiency of each energy storage unit.

[0016] The current remaining capacity of each energy storage unit and the energy transmission distance to the switching reference unit are obtained. The switching fluctuation index, unit efficiency mean, current remaining capacity, and energy transmission distance are integrated into a multi-dimensional parameter to obtain the priority score of each energy storage unit. The energy storage unit with the highest score is selected as the execution unit and assigned to it the grid-connected and off-grid switching task of the switching reference unit.

[0017] Count all task queues of all execution units and generate a task execution sequence from low to high according to the grid connection status level. The execution units complete the switching tasks in sequence. After each task is completed, the queue status is updated and an operation log is generated. The log contains the execution unit power output range, response delay time and grid connection status level change data;

[0018] Integrate the real-time coordinates and task status of all energy storage units, mark the executing, pending, and completed tasks with different color blocks in the visual interface, and dynamically update the current remaining capacity and switching alarm status of the energy storage units;

[0019] The energy storage unit is the core execution component of the mobile energy storage vehicle to realize the functions of electric energy storage, scheduling and on-grid and off-grid switching.

[0020] Preferably, the calculation process of the output efficiency value is as follows:

[0021] The interval corresponding to the power output range is divided into several equal power segments. The maximum output duration and average response delay of each power segment are calculated. The efficiency benchmark value is obtained through power segment priority allocation and delay weighted calculation. The efficiency benchmark value is nonlinearly fitted with the grid connection status level to obtain the unit efficiency mean.

[0022] Preferably, feature extraction of grid voltage amplitude data includes:

[0023] Set the standard voltage reference range of the grid connection point, perform a sliding window comparison between the real-time collected grid voltage amplitude data and the standard voltage reference range, and count the duration and number of fluctuations outside the range; generate the voltage instability coefficient based on the product of the duration and the number of fluctuations.

[0024] Preferably, feature extraction of frequency offset data includes:

[0025] Multiple frequency sensors are deployed at the grid connection point, and the frequency differences between adjacent sensors are calculated to generate a gradient distribution map. Spatial clustering analysis is performed based on the extreme value positions and mean offsets in the distribution map to obtain the frequency offset.

[0026] Preferably, feature extraction of phase synchronization deviation data includes:

[0027] The phase signal of the grid connection point is collected through a phase detection device to separate the synchronous component and the asynchronous component. The amplitude-frequency changes of the asynchronous component are envelope tracked, and the phase deviation index is obtained by combining the trend fitting of the synchronous component.

[0028] Preferably, the dynamic threshold analysis is implemented as follows:

[0029] The benchmark threshold interval is generated based on historical power grid abnormal event data, and sliding window statistics are performed in combination with the real-time collected grid stability indicators to dynamically adjust the upper and lower limits of the benchmark threshold interval.

[0030] Preferably, the specific steps of trend analysis are:

[0031] The inflection points and plateau segments in the dynamic trend curve are extracted, and the slope change rate between adjacent inflection points and plateau segments is calculated; the grid connection status level is determined based on the matching degree between the slope change rate and the preset status template.

[0032] Preferably, the process of multi-dimensional parameter fusion is as follows:

[0033] The switching fluctuation index, unit efficiency mean, current remaining capacity and energy transmission distance are normalized respectively, and linearly superimposed through the preset priority coefficient to output the priority score.

[0034] Preferably, the feature extraction process of the load fluctuation data includes:

[0035] A standard load fluctuation threshold is set for the grid connection point. The real-time collected load fluctuation data is differentially compared with the standard load fluctuation threshold. The fluctuation amplitude and duration exceeding the standard load fluctuation threshold are counted. The load disturbance value is generated based on the product of the fluctuation amplitude and duration.

[0036] Compared with the prior art, the present invention has the following beneficial effects:

[0037] At the grid-connected status monitoring level, the system uses the grid-connected status detection module to collect and conduct in-depth real-time analysis of multi-dimensional data such as voltage amplitude, frequency offset, phase synchronization deviation, and load fluctuation at the grid connection point. By extracting state features from each dimensional data, such as calculating the voltage instability coefficient, frequency offset, phase deviation index, and load disturbance value, and performing dynamic threshold analysis, the actual stability of the power grid can be more comprehensively and accurately reflected. By constructing a two-dimensional coordinate system to generate a dynamic trend curve and perform trend analysis, the grid connection status level can be determined in real time, providing a scientific and reliable basis for switching decisions. This multi-dimensional, dynamic monitoring method overcomes the shortcomings of traditional single-parameter monitoring and fixed threshold analysis, significantly improving the accuracy and timeliness of grid status judgments and reducing the risk of misjudgment and missed judgments.

[0038] In terms of switching control strategy, the switching control module conducts a comprehensive evaluation and optimized scheduling of energy storage units based on the grid connection status level. By obtaining the grid connection status level of the energy storage units and sorting them, the unit with the lowest level is selected as the switching reference unit to ensure that switching starts from the link with the worst stability, thereby improving the targeted nature of the switching. The historical operating data of each energy storage unit, such as the historical switching delay time, load change type, power output range, and response delay time, are analyzed to calculate the switching fluctuation index and the mean unit efficiency. The multi-dimensional parameter fusion is combined with the current remaining capacity and energy transmission distance to obtain the priority score and select the optimal execution unit, thus achieving a comprehensive and objective evaluation of the energy storage unit and ensuring the scientific and rational allocation of the switching task. The task execution sequence is generated according to the grid connection status level, and the queue status and operation log are dynamically updated, making the switching process more orderly and efficient. The task execution order can be adjusted in time according to the changes in the grid status, thereby improving the system's responsiveness and overall performance.

[0039] In the data processing and analysis phase, the system employs a variety of advanced data processing methods. These include sliding window comparison and product calculation of grid voltage amplitude data to generate voltage instability coefficients, gradient distribution plot analysis and spatial clustering of frequency offset data, envelope tracking and trend fitting of phase synchronization deviation data, sliding window adjustment of dynamic thresholds based on historical data and real-time statistics, and analysis of inflection points and plateaus, as well as slope rate matching, on trend curves. These methods fully tap into the inherent information of monitoring data, enhance its utilization value, and provide strong support for accurately assessing grid status and formulating switching strategies. The normalization and linear superposition in the multidimensional parameter fusion process enable the comprehensive consideration of multiple influencing factors using a unified standard, enhancing the scientific nature and credibility of priority scoring.

[0040] In terms of visualization and monitoring, the system integrates the real-time coordinates and task status of all units. The visual interface displays active, pending, and completed tasks with different color blocks, and dynamically updates the current remaining capacity and switching alarm status of energy storage units. This feature enables operators to intuitively understand system operations in real time, identify potential issues promptly, and take appropriate measures. This improves system monitorability and management efficiency, while reducing the complexity and risk of manual operation errors.

[0041] Furthermore, through optimized design of each link, the system achieves intelligent and automated on-grid and off-grid switching, reducing manual intervention, improving switching efficiency and reliability, extending the service life of the energy storage units, and reducing system operating costs. Furthermore, the system can better adapt to the dynamic changes in grid operation, improving the adaptability and stability of mobile energy storage vehicles in different scenarios, and providing strong technical support for the efficient utilization of new energy and the development of smart grids. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 This is a working principle diagram of the mobile energy storage vehicle on-grid and off-grid switching intelligent control system of the present invention;

[0043] Figure 2 It is the allocation diagram of the switching control module tasks;

[0044] Figure 3 Process diagram for calculating output efficiency value;

[0045] Figure 4 This is the implementation diagram of dynamic threshold analysis;

[0046] Figure 5 A diagram showing the status of the visualization interface. DETAILED DESCRIPTION

[0047] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0048] See also Figure 1-Figure 5 The present invention relates to an intelligent control system for switching between on-grid and off-grid mobile energy storage vehicles, the system comprising:

[0049] The grid-connected status detection module is based on multi-dimensional collection of real-time monitoring data of the grid connection point, including grid voltage amplitude data, frequency offset data, phase synchronization deviation data, and load fluctuation data. State feature extraction is performed on each type of data to obtain the voltage instability coefficient, frequency offset, phase deviation index, and load disturbance value at each sampling moment, and a dynamic threshold analysis is performed on them to obtain the grid stability index corresponding to each sampling moment. Specifically, a two-dimensional coordinate system is constructed with time as the horizontal axis and the grid stability index as the vertical axis. The grid stability index value at each sampling moment is marked as a monitoring point in the coordinate system. Adjacent monitoring points are connected using piecewise linear interpolation to generate a dynamic trend curve. The dynamic trend curve is subjected to trend analysis to obtain the current grid connection status level, which is then transmitted to the switching control module.

[0050] The switching control module executes the on-grid switching strategy and energy storage unit scheduling based on the grid connection status level. The specific process is: obtain the grid connection status level of all energy storage units, sort them from low to high, and select the energy storage unit with the lowest level as the switching reference unit; retrieve the historical switching delay time of each energy storage unit and the load change type corresponding to each delay, calculate the interval between two adjacent switches and perform variance processing to obtain the switching stability period, extract the load fluctuation amplitude of each switch and perform mean processing to obtain the disturbance reference amplitude, perform correlation analysis on the switching stability period and the disturbance reference amplitude to obtain the switching fluctuation index of each energy storage unit; retrieve the historical operation data of each energy storage unit, including the power output range, response delay time and grid connection level of each operation, calculate the output efficiency value by the ratio of power range to response delay time, and perform efficiency weight distribution based on the grid connection status level to obtain the unit efficiency average of each energy storage unit; obtain the load fluctuation amplitude of each switch, and perform mean processing on the load fluctuation amplitude of each switch to obtain the disturbance reference amplitude, and perform correlation analysis on the switching stability period and the disturbance reference amplitude to obtain the switching fluctuation index of each energy storage unit; retrieve the historical operation data of each energy storage unit, including the power output range, response delay time and grid connection level of each operation, calculate the output efficiency value by the ratio of power range to response delay time, and perform efficiency weight distribution based on the grid connection status level to obtain the unit efficiency average of each energy storage unit; obtain the load fluctuation amplitude of each storage unit, and perform mean processing on the load fluctuation amplitude of each switch to obtain the load fluctuation index of each energy storage unit. The current remaining capacity of the energy storage unit and the energy transmission distance with the switching reference unit are calculated. The switching fluctuation index, unit efficiency mean, current remaining capacity and energy transmission distance are multi-dimensionally integrated to obtain the priority score of each energy storage unit. The unit with the highest score is selected as the execution unit, and the grid-connected and off-grid switching task of the switching reference unit is assigned to it. All task queues of the execution unit are counted, and a task execution sequence is generated from low to high according to the grid connection status level. The execution unit completes the switching tasks in sequence. Each time a task is completed, the queue status is updated and an operation log is generated. The log contains the power output range of the execution unit, the response delay time and the grid connection level change data. The real-time coordinates and task status of all energy storage units are integrated, and the executing, pending and completed tasks are marked with different color blocks in the visual interface, and the current remaining capacity and switching alarm status of the energy storage unit are dynamically updated.

[0051] Example 1:

[0052] In this embodiment, the feature extraction process for the grid voltage amplitude data and frequency offset data is as follows:

[0053] To extract features from grid voltage amplitude data, we first need to establish a voltage monitoring system at the grid connection point. Specifically, voltage sensors deployed at the grid connection point collect grid voltage amplitude data in real time at a preset sampling frequency (e.g., 100 times per second), forming a continuous voltage signal sequence. The system sets a standard voltage reference range for the grid connection point. This range is determined based on the grid's rated voltage and national power quality standards. For example, for a three-phase 380V grid, the standard voltage reference range can be set to ±10% of the rated value (i.e., 342V-418V).

[0054] During the feature extraction phase, a sliding window algorithm is used to process the real-time grid voltage amplitude data. The sliding window duration can be set to 1 minute, with each window movement lasting 1 second, ensuring real-time and continuous data processing. Within each sliding window, the system compares the real-time grid voltage amplitude data collected within the window with the standard voltage reference range point by point, and calculates the following parameters:

[0055] Duration of out-of-range conditions: When the voltage amplitude at a certain moment exceeds the standard range, the accumulated duration of the out-of-range period (in seconds) is recorded, starting from that moment until the voltage amplitude returns to within the standard range. If the voltage continues to exceed the range throughout the entire window, the duration is considered the window duration.

[0056] Number of fluctuations: Within a single sliding window, the complete process of the voltage amplitude jumping from within the standard range to outside the range and then returning to the standard range is counted as one fluctuation, and the number of fluctuations within the window is counted.

[0057] Based on the duration and number of fluctuations obtained from the above statistics, the system calculates the voltage instability coefficient. The calculation formula is:

[0058]

[0059] This coefficient reflects the degree and frequency of voltage deviations from the standard range within a certain period of time. A larger value indicates poorer voltage stability. For example, if the voltage exceeds the range for 20 seconds and fluctuates three times within a certain window, the voltage instability coefficient is 60.

[0060] To extract features from frequency offset data, the system deploys multiple frequency sensors (e.g., three) at different locations at the grid connection point, forming a spatially distributed monitoring network. Each sensor synchronously collects grid frequency data, using the same sampling frequency as the voltage sensors. The physical spacing between adjacent sensors is determined by the actual layout of the grid connection point. For example, within a distribution substation, sensors can be deployed at the incoming, mid-point, and outgoing lines, with spacing of approximately 5-10 meters.

[0061] During the feature extraction process, the frequency difference between adjacent sensors is first calculated. Taking sensors A and B as an example, the frequency difference is ,in 、 The real-time frequency values ​​collected by the two sensors are shown. By calculating the frequency differences between all adjacent sensor pairs, a frequency gradient distribution graph is generated. This distribution graph uses sensor location as the horizontal axis and frequency difference as the vertical axis, presenting the spatial frequency trend as a line graph or scatter plot.

[0062] Next, we conduct spatial cluster analysis. The specific steps are as follows:

[0063] Determine the location of extreme values: In the frequency gradient distribution diagram, identify the maximum and minimum points of the frequency difference. These extreme points usually correspond to areas or boundaries with abnormal frequency in the power grid.

[0064] Calculate mean offset: Calculate the average value of the frequency differences of all adjacent sensors as the reference value of frequency offset.

[0065] Cluster analysis: Using the density-based spatial clustering of applications with noise (DBSCAN) algorithm, we cluster the frequency difference data using the extreme value as the core point and a radius that is a multiple of the mean shift (e.g., 1.5). Regions with similar frequency differences are grouped into the same cluster. Each cluster represents a region with a frequency shift characteristic. The frequency shift value for that region is determined by calculating the mean or median of the frequency differences within the cluster.

[0066] For example, if the frequencies of three sensors are 50.1Hz, 49.8Hz, and 50.0Hz, respectively, the frequency differences between adjacent sensor pairs are 0.3Hz (sensor A and B) and 0.2Hz (sensor BC). The frequency gradient distribution plot shows a large frequency difference extreme between sensor A and B. Cluster analysis identifies this region as a region of significant frequency offset, with a frequency offset of 0.3Hz.

[0067] During data processing, the system aligns the voltage instability coefficient and frequency offset in time series to ensure they correspond to the same sampling moment. At each sampling moment, the voltage instability coefficient, frequency offset, and subsequent processed phase deviation index and load disturbance values ​​are combined to form a multidimensional data vector representing the grid stability indicator.

[0068] It should be noted that during the feature extraction process, the sliding window duration, the number and spacing of sensors deployed, and the parameters of the clustering algorithm can all be adjusted based on the actual grid operating environment. For example, in high-frequency fluctuating industrial grid scenarios, the sliding window duration can be shortened to 30 seconds and the number of sensors deployed can be increased to 5 to improve monitoring accuracy. In low-frequency fluctuating scenarios such as rural distribution networks, the sliding window duration can be extended to 2 minutes and the number of sensors can be reduced to 2 to reduce system costs.

[0069] Furthermore, the system utilizes a parallel computing architecture to collect and process grid voltage amplitude and frequency offset data. This enables real-time data processing via a field-programmable gate array (FPGA) or digital signal processor (DSP), ensuring feature extraction and coefficient calculation within milliseconds, meeting the real-time requirements for on-grid and off-grid switching of mobile energy storage vehicles. For data transmission, industrial-grade fieldbuses (such as ModbusTCP) or wireless communication technologies (such as 5G) are used to transmit processed feature data in real time to the central processor of the grid-connected status detection module for subsequent dynamic threshold analysis and trend analysis.

[0070] During abnormal data processing, the system implements a data validity verification mechanism. For grid voltage amplitude data, if the value at a sampling point exceeds the sensor's measurement range (e.g., 0-1000V) or exhibits a significant jump (e.g., the difference between adjacent sampling points exceeds 50% of the rated voltage), the data is deemed invalid and filled in using valid data from previous and subsequent data using interpolation methods (e.g., linear interpolation). For frequency offset data, if the frequency difference between adjacent sensors exceeds the theoretical maximum value for normal grid operation (e.g., 1Hz), this is considered a communication failure or sensor anomaly, triggering a system alarm and activating backup sensors for data collection.

[0071] By extracting the characteristics of grid voltage amplitude and frequency offset data described above, the system can comprehensively characterize the changes in electrical parameters at grid connection points in both temporal and spatial dimensions, providing a solid data foundation for the subsequent generation of grid stability indicators and the assessment of grid connection status. This process does not rely on empirical conclusions or experimental effect assumptions; instead, it derives characteristic parameters solely through real-time monitoring of physical quantities, mathematical statistics, and algorithmic analysis, ensuring the objectivity and accuracy of data processing.

[0072] Example 2:

[0073] In this embodiment, the feature extraction and dynamic threshold analysis of phase synchronization deviation data are implemented as follows:

[0074] The feature extraction of phase synchronization deviation data is based on high-precision phase signal acquisition and signal processing algorithms. The system uses phase detection devices (such as synchronized phasor measurement units (PMUs)) deployed at the grid connection point to collect three-phase voltage and current phase signals in real time. The devices have built-in Beidou satellite clocks to achieve nanosecond-level time synchronization, ensuring the temporal and spatial consistency of phase data. The acquisition frequency is set to 100 times per second, generating a phase sequence with a timestamp. ,in For the Sampling time, is the phase value at the corresponding moment.

[0075] During the signal preprocessing phase, a finite impulse response (FIR) filter is used to eliminate noise interference. The filter has a passband of 45-55Hz (for a 50Hz power grid) and a stopband attenuation of no less than 60dB to ensure the integrity of the fundamental phase signal. The preprocessed signal enters the synchronous and asynchronous component separation phase, where the Kalman filter algorithm is used to construct a state space model:

[0076]

[0077] Among them, the state vector , is the synchronous component (fundamental wave phase), is the asynchronous component (phase fluctuation amount); is the state transfer matrix, which describes the linear relationship between the states at adjacent moments; is the observation matrix, which maps the state vector to the observation space; and are process noise and observation noise respectively, which obey zero-mean Gaussian distribution. The predicted value is calculated by recursion With updated value , to achieve synchronization component With asynchronous component The optimal estimate of .

[0078] For the separated asynchronous components When analyzing the amplitude-frequency variation, Hilbert transform is used to extract the instantaneous amplitude envelope. Assume that the discrete sequence of the asynchronous component is , whose Hilbert transform sequence is , then the instantaneous amplitude envelope for:

[0079]

[0080] To eliminate high frequency glitches, Perform 5-point sliding average filtering to obtain a smoothed envelope curve , reflecting the slow-changing trend of the asynchronous component. The trend fitting of the synchronous component adopts cubic polynomial interpolation, the formula is:

[0081]

[0082] in, 、 、 、 The polynomial coefficients are used to calculate the synchronization components of the first 30 sampling points by the least square method. ( The phase deviation index is calculated by combining the synchronous component fitting residual and the asynchronous component envelope, and is defined as:

[0083]

[0084] Where, is the weight coefficient (the value range is 0.4-0.6, configured according to the characteristics of the power grid), is the absolute error between the actual value and the fitted value of the synchronization component, is the historical maximum phase error value, is the smoothed envelope value of the asynchronous component at the current moment, This is the historical maximum envelope value.

[0085] The implementation of dynamic threshold analysis relies on historical data drive and real-time data adaptive adjustment. The system first calculates the average value of the phase deviation index at each sampling moment based on the normal operation data of the past three years (excluding the fault period). and standard deviation , construct the initial benchmark threshold interval as In real-time monitoring, a sliding window (window length is 30 minutes, including 1800 sampling points) is used to calculate the mean of the current data. and standard deviation , dynamically adjust the baseline threshold boundary according to the following rules:

[0086] Center offset adjustment: If , then the center of the threshold interval moves to ;

[0087] Width scaling adjustment: If , then the threshold interval width is proportional to Extension, upper limit is , the lower limit is ;

[0088] Edge shrinkage adjustment: If 10 consecutive sampling points are located in an area less than 5% of the interval width away from the current threshold boundary, the boundary will be shrunk toward the center by 5% of the current interval width.

[0089] The threshold adjustment process uses exponential smoothing to suppress high-frequency fluctuations, such as the current threshold mean From the historical average and real-time average Update by weight:

[0090]

[0091] Ensure smooth transition of threshold intervals to avoid misjudgment due to short-term data fluctuations. The system automatically updates the historical database every month and recalculates and , adapt to changes in the long-term operating characteristics of the power grid.

[0092] In terms of hardware implementation, the phase detection device utilizes a dual-redundant design, with hot-swap mechanisms between the primary and backup devices ensuring continuous monitoring. Signal processing algorithms are executed in parallel by a field-programmable gate array (FPGA). Convolution operations such as FIR filtering and Hilbert transforms are accelerated by hardware multipliers, while matrix operations in the Kalman filter are optimized through a pipeline architecture, ensuring overall processing latency of less than 10ms. Data transmission utilizes the industrial-grade OPC UA protocol, transmitting the phase deviation index and threshold range in real time to the central control system via a dedicated 5G network, keeping communication latency to less than 50ms.

[0093] Exception handling mechanisms include:

[0094] Data validity check: If the phase signal jump exceeds If the time difference is 2.5ms under the power frequency, it is marked as invalid data and filled by linear interpolation of the previous and next effective values;

[0095] Equipment status monitoring: Real-time monitoring of PMU clock synchronization status, signal-to-noise ratio and other indicators. If the satellite signal is lost for more than 1 minute, it will automatically switch to the built-in rubidium atomic clock to maintain time synchronization and trigger an alarm;

[0096] Threshold conflict detection: When the deviation between the dynamically adjusted benchmark threshold interval and the historical benchmark interval exceeds 30%, the manual review process is automatically initiated to determine the rationality of the threshold adjustment based on changes in grid operation (such as switching on and off of large equipment).

[0097] The entire processing flow does not rely on empirical parameters or hypothetical conclusions; instead, it achieves feature extraction and dynamic threshold adaptation solely through signal processing algorithms and statistical analysis. The calculation of the phase deviation index integrates time-domain residual and frequency-domain envelope features. The dynamic threshold interval, through the dual constraints of historical distribution and real-time trends, ensures accurate characterization of the grid's phase synchronization state, providing a reliable basis for on-grid and off-grid switching decisions. The system ensures stability and robustness in complex grid environments through hardware redundancy, data verification, and adaptive learning mechanisms.

[0098] Example 3:

[0099] In this example, trend analysis and multi-dimensional parameter fusion are implemented based on specific power grid scenarios. Assume that three mobile energy storage vehicles (ESV-01, ESV-02, and ESV-03) are deployed in the distribution network of an industrial park. Their on-grid and off-grid switching control utilizes the intelligent control system of the present invention. One afternoon between 2:00 PM and 3:00 PM, large-scale industrial equipment in the park is scheduled to be switched on and off, causing voltage fluctuations and frequency deviations in the power grid. The system needs to monitor the grid connection status in real time and dispatch the energy storage vehicles.

[0100] In terms of trend analysis, the system first processes the dynamic trend curve. Taking the grid stability index curve of the ESV-01 access point as an example, this curve is generated by weighted fusion of the voltage instability coefficient, frequency offset, phase deviation index, and load disturbance value. The system analyzes it through the following steps:

[0101] Inflection point identification: At 2:15 PM, the slope of the curve changes significantly, from a gentle rise to a rapid increase. The system marks this point as an inflection point. Analysis shows that this moment corresponds to the startup of a large air compressor within the park, which caused a sudden increase in grid load.

[0102] Stable segment classification: During the two time periods of 2:00 PM to 2:15 PM and 2:30 PM to 3:00 PM, the slope of the curve approaches zero, and the system identifies it as a stable segment. The former corresponds to the normal operation of the power grid, while the latter corresponds to the stable operation of the equipment after startup.

[0103] Slope Change Rate Calculation: During the fluctuation period from 2:15 PM to 2:30 PM, the system calculates the slope change rate between adjacent inflection points and stable segments. For example, from 2:15 PM to 2:20 PM, the slope change rate was 0.05 units per minute, indicating a rapid deterioration in grid stability indicators. However, from 2:25 PM to 2:30 PM, the slope change rate dropped to 0.01 units per minute, indicating that the system was recovering stability.

[0104] The system has four preset status templates: green (stable), yellow (caution), orange (warning), and red (emergency). Each template corresponds to a different slope change rate range and indicator absolute value range. By matching the real-time calculated slope change rate and indicator value with the preset templates, the system determined the current grid connection status level to be orange warning at 2:20 PM and triggered the corresponding control strategy.

[0105] In terms of multi-dimensional parameter fusion, the system needs to select the optimal execution unit from the three energy storage vehicles. The parameters of each energy storage unit are as follows:

[0106] ESV-01: Located in the southeast corner of the campus, it has 75% remaining capacity and an energy transmission distance of 800 meters to the switching benchmark unit. Historical data shows that its switching fluctuation index under similar load fluctuation scenarios is 0.25 (a lower index indicates smoother switching), and its average unit efficiency is 0.85 (the maximum value is 1).

[0107] ESV-02: Located in the center of the campus, it has 60% remaining capacity and a transmission distance of 300 meters. Its historical switching fluctuation index is 0.32, and its average unit efficiency is 0.91.

[0108] ESV-03: Located in the northwest corner of the campus, it has 90% remaining capacity and a transmission distance of 1,200 meters. Its historical switching fluctuation index is 0.18, and its average unit efficiency is 0.88.

[0109] The system first normalizes each parameter. For example, the normalization formula for remaining capacity is actual capacity divided by 100%. Therefore, the normalized remaining capacities for ESV-01, ESV-02, and ESV-03 are 0.75, 0.60, and 0.90, respectively. The normalization formula for energy transmission distance is 1 minus (actual distance divided by maximum distance). In this example, the maximum distance is 1200 meters, so the normalized transmission distances for the three devices are 0.33, 0.75, and 0, respectively.

[0110] The system uses the statistical distribution of historical data to normalize the switching fluctuation index and unit efficiency mean. Assuming the historical minimum fluctuation index is 0.15 and the maximum is 0.40, the normalized fluctuation indexes for ESV-01, ESV-02, and ESV-03 are 0.80, 0.52, and 1.00, respectively. Similarly, assuming the historical efficiency mean has a minimum of 0.70 and a maximum of 0.95, the normalized efficiency means for the three units are 0.60, 1.00, and 0.72, respectively.

[0111] Next, we perform linear superposition to calculate the priority score. The system assigns weight coefficients to each parameter as follows: remaining capacity 0.3, energy transmission distance 0.2, switching fluctuation index 0.3, and unit efficiency mean 0.2. The calculation results are as follows:

[0112] ESV-01: 0.75×0.3+0.33×0.2+0.80×0.3+0.60×0.2=0.701

[0113] ESV-02: 0.60×0.3+0.75×0.2+0.52×0.3+1.00×0.2=0.676

[0114] ESV-03: 0.90×0.3+0.0×0.2+1.00×0.3+0.72×0.2=0.714

[0115] Based on the priority score, the system selects ESV-03 as the execution unit and assigns it the task of switching the base unit on and off the grid. At this point, the system detects that ESV-03's current task queue already has a pending task (scheduled for execution at 2:45 PM), while ESV-01's task queue is empty. Considering that ESV-03's priority score is only slightly higher than ESV-01's and that ESV-01 has no task conflicts, the system initiates the task reallocation process.

[0116] During the task redistribution process, the system retrieved real-time status data from ESV-01 and ESV-03. It discovered that while ESV-03 had a high remaining capacity, it was currently in a battery balancing state and was not expected to complete until 2:35 PM. ESV-01, on the other hand, could immediately respond to the task. Furthermore, while ESV-01's energy transmission distance was longer, optimizing the transmission path (utilizing backup lines within the campus) reduced actual transmission losses to a level comparable to that of ESV-03. Taking these factors into consideration, the system ultimately reallocated the task to ESV-01.

[0117] After receiving the task, the ESV-01 executed the on-grid and off-grid switching operation according to the preset switching strategy. First, the system fine-tuned the energy storage vehicle's output power to match the grid's current load, avoiding power surges during the switching process. At 2:22 PM, the ESV-01 began the on-grid to off-grid transition. The entire process lasted 25 seconds, during which the system monitored the grid connection status indicators in real time.

[0118] During the switchover process, the voltage instability coefficient increased from 0.32 to 0.45, and the frequency offset increased from 0.05 Hz to 0.12 Hz, but neither exceeded the system's thresholds. The phase deviation index reached 0.68 (the threshold is 0.8) at the moment of switchover before quickly falling back to 0.35. The system recorded all parameter changes during the switchover and generated a detailed operation log, including the power output range of the actuator unit (gradually decreasing from 300 kW when connected to the grid to 150 kW after disconnection), the response delay (180 ms), and the change in grid connection level (recovery from orange warning to yellow caution).

[0119] After the switchover, the ESV-01 entered off-grid operation, providing continuous power to critical loads within the campus, such as the data center and security systems. The system continued to monitor its remaining capacity and grid status. When the grid stabilized at 2:40 PM, the system re-initiated trend analysis and multi-dimensional parameter fusion to assess whether the ESV-01 needed to be reconnected to the grid.

[0120] At this point, the parameters of each energy storage vehicle have changed: ESV-01's remaining capacity has dropped to 68%, ESV-02 has been recharged to 75%, and ESV-03 has completed battery balancing and its remaining capacity remains at 90%. The system recalculates the priority score, and the results show that ESV-03 is still the best choice. However, considering that ESV-01 is currently off-grid, the operational complexity of reconnecting it to the grid is relatively low, and its remaining capacity can still meet emergency needs for a period of time, the system decides to maintain ESV-01's off-grid status while dispatching ESV-03 to increase its grid-connected output power to alleviate grid pressure.

[0121] Throughout the scheduling process, the system displays the status of each energy storage unit in real time through a visual interface. Task status is color-coded: green indicates completed, yellow indicates in-progress, and red indicates pending. The system also dynamically updates the remaining capacity of the energy storage unit (displayed as a bar graph) and the status of switching alarms (such as abnormal battery temperature and communication interruption). Operators can access historical data and operation logs at any point in time through the interface, facilitating post-analysis and system optimization.

[0122] Example 4:

[0123] This example uses a mobile energy storage vehicle cluster in a city's emergency power supply scenario as an example to explain the calculation process and application of output efficiency values. Assuming the cluster consists of four energy storage vehicles (numbered E-01 through E-04), each with a rated power of 500 kW, the system calculates output efficiency values ​​based on each vehicle's historical operating data to support on-grid and off-grid switching priority decisions.

[0124] Taking the energy storage vehicle E-01 as an example, its power output range is 0-500kW. The system divides this range into five equal power segments: 0-100kW (low power segment), 100-200kW, 200-300kW, 300-400kW, and 400-500kW (high power segment). Based on historical data statistics, the operating parameters of each power segment are as follows:

[0125] 0-100kW: Maximum output duration is 8 hours (corresponding to low-load periods at night), with an average response delay of 200ms (the battery management system needs to be preheated during the startup phase);

[0126] 100-200kW: Maximum duration is 5 hours, with an average response delay of 150ms (the system enters a stable operating state).

[0127] 200-300kW: Maximum duration is 3 hours, with an average response delay of 120ms (power output and cooling system are balanced).

[0128] 300-400kW: Maximum duration is 2 hours, with an average response delay of 180ms (battery internal resistance increases at high power, and response speed decreases slightly);

[0129] 400-500kW: Maximum duration is 1 hour, with an average response delay of 250ms (overload protection is triggered, requiring additional battery temperature monitoring).

[0130] The system assigns priority to each power segment, following the principle of "segments closer to rated power have higher priority". The priority coefficients from low power segment to high power segment are set as 1, 2, 3, 4, and 5 respectively. The calculation of the efficiency benchmark value requires combining priority and response delay. For example, the efficiency benchmark value of the low power segment is:

[0131]

[0132] The efficiency benchmark value in the high power range is:

[0133]

[0134] The same applies to the intermediate power range, with the efficiency benchmark values ​​for each range being calculated as 5, 13.3, 25, 22.2, and 20. Taking the weighted average of the benchmark values ​​for each power range (the weight is the historical operating time percentage of each range), assuming that the operating time percentages for the low, medium, and high power ranges are 30%, 50%, and 20%, respectively, the efficiency benchmark values ​​for E-01 are:

[0135]

[0136] The grid connection status level adjusts the efficiency benchmark value through nonlinear fitting. The system presets four grid connection levels: Level 1 (stable), Level 2 (mild fluctuation), Level 3 (moderate fluctuation), and Level 4 (severe fluctuation). Taking the E-01 operating at Level 2 as an example, historical data shows that its efficiency benchmark value has decreased by an average of 15%. Therefore, the average unit efficiency of each energy storage unit after adjustment is:

[0137]

[0138] If the grid is connected at the third level, and the efficiency benchmark value decreases by 25%, the average unit efficiency of each energy storage unit is:

[0139]

[0140] Compared with the calculation process of other energy storage vehicles:

[0141] E-02: In the high power range (400-500kW), due to battery aging, the average response delay increases to 300ms, causing the efficiency benchmark value in this range to drop to , the overall efficiency average is lower than E-01;

[0142] E-03: Equipped with a new thermal management system, the response delay in the 300-400kW power range is reduced to 100ms, and the efficiency benchmark value in this range is increased to , raising the overall efficiency average;

[0143] E-04: In historical operation, the low-power segment accounted for 60% (mainly used for nighttime peak shaving). Although the response delay was low, the overall efficiency average was at a medium level due to the low priority coefficient.

[0144] In an actual dispatch scenario, assuming that the grid connection status level is level 2 at a certain moment, the system needs to select the unit to execute and switch off the grid from the four energy storage vehicles. Other parameters of each vehicle are as follows:

[0145] E-01: Remaining capacity 80%, energy transmission distance 400 meters, switching fluctuation index 0.28;

[0146] E-02: Remaining capacity 75%, transmission distance 600 meters, switching fluctuation index 0.35;

[0147] E-03: Remaining capacity 65%, transmission distance 200 meters, switching fluctuation index 0.22;

[0148] E-04: Remaining capacity 90%, transmission distance 800 meters, switching fluctuation index 0.30.

[0149] When combining the average unit efficiency of each energy storage unit with other parameters (remaining capacity, transmission distance, and switching fluctuation index) for multi-dimensional integration, the system normalizes each parameter. For example, the average unit efficiency of the E-03 is 18.5 (higher than the E-01's 12.92), which is normalized to 0.9 (assuming a historical maximum of 20); 65% remaining capacity is normalized to 0.65; a transmission distance of 200 meters (the shortest) is normalized to 1; and a switching fluctuation index of 0.22 (lower) is normalized to 0.85. Priority scores are calculated using preset weights (average efficiency 0.4, remaining capacity 0.2, transmission distance 0.2, and switching fluctuation index 0.2):

[0150]

[0151] The priority ratings for E-01 are:

[0152]

[0153] By comparison, we can see that E-03 has the highest priority score, and the system selects it as the execution unit.

[0154] During the switching process, the E-03's power output range is dynamically adjusted based on the average efficiency. Due to its higher efficiency in the 300-400kW power range, the system prioritizes its output power at 350kW, resulting in a response delay of 120ms, which is optimal for this power range. After the switch is complete, the system records the power output range (300-400kW), response delay (120ms), and grid connection level change (from Level 2 to Level 2), and updates the historical database to provide new samples for subsequent efficiency calculations.

[0155] If the grid level is subsequently upgraded to level 3, the unit efficiency average of E-03 will be automatically adjusted based on the historical fitting relationship. For example, its efficiency benchmark value will be reduced from 18.5 to , after normalization it is 0.69, and the priority score drops accordingly to:

[0156]

[0157] At this time, if the remaining capacity of E-01 increases to 85%, its priority score may surpass that of the previous one, and the system will re-evaluate the execution unit, reflecting the dynamic linkage between efficiency value and grid-connected status.

[0158] Example 5:

[0159] This example describes the process of extracting and processing load fluctuation data and operating the switching control module, using a smart microgrid system in an industrial park as an example. The park deploys five mobile energy storage vehicles (ESV-01 to ESV-05), each with a rated power of 1 MW. The intelligent control system of this invention enables on-grid and off-grid switching.

[0160] At 10:00 AM one day, large industrial equipment (such as an electric arc furnace) within the industrial park was scheduled to start up, predicting grid load fluctuations. The system preemptively entered monitoring mode and set a standard load fluctuation threshold of ±15% at the grid connection point (based on a 95% confidence interval of the industrial park's historical load data). This real-time load data was transmitted to the control system via industrial Ethernet with a sampling period of 100ms.

[0161] During the load fluctuation data feature extraction phase, the system compares real-time load data with standard thresholds. For example, at 10:05 a.m., when the arc furnace started up, the system detected a load jump from 4.2 MW to 5.8 MW, exceeding the upper limit of the standard threshold (4.83 MW). The system immediately began counting the amplitude and duration of the fluctuation: the amplitude was 1.6 MW (5.8 MW - 4.2 MW), and the duration began at 10:05:00.

[0162] When the load exceeded the threshold for 5 seconds (i.e., at 10:05:05), the system determined a significant load fluctuation and generated an initial load disturbance value. During this fluctuation, the system continuously monitored load changes and found that after the arc furnace started, the load fluctuated between 5.5 and 6.0 MW, not stabilizing at 5.7 MW until 10:07:30. The system recorded the maximum amplitude (1.6 MW) and duration (150 seconds) of the entire fluctuation process, ultimately calculating the load disturbance value as the product of these two values ​​(240 MW·second).

[0163] After obtaining the load disturbance value, the switching control module synchronously obtains the current status of each energy storage unit. At this time, the status of the five energy storage vehicles is as follows:

[0164] ESV-01: Remaining capacity 85%, energy transmission distance to the switching reference unit (arc furnace access point) is 300 meters, and currently performing normal charging tasks;

[0165] ESV-02: Remaining capacity 60%, transmission distance 500 meters, in hot standby mode;

[0166] ESV-03: Remaining capacity 72%, transmission distance 400 meters, supplying power to another load;

[0167] ESV-04: Remaining capacity 90%, transmission distance 600 meters, in cold standby mode;

[0168] ESV-05: Remaining capacity 55%, transmission distance 200 meters, battery balancing in progress.

[0169] The system integrates this data with parameters such as the switching fluctuation index and unit efficiency average. The historical switching fluctuation indexes for each energy storage vehicle are: ESV-01 (0.25), ESV-02 (0.32), ESV-03 (0.28), ESV-04 (0.20), and ESV-05 (0.35); and the unit efficiency averages are: ESV-01 (0.88), ESV-02 (0.82), ESV-03 (0.85), ESV-04 (0.90), and ESV-05 (0.79).

[0170] During the multi-dimensional parameter fusion process, the system first normalizes each parameter. For example, the remaining capacity is based on 100%, the transmission distance is based on the longest distance (600 meters), the switching fluctuation index is based on the historical minimum value (0.15), and the average unit efficiency is based on the historical maximum value (0.95). The system then calculates the priority score of each energy storage vehicle based on the preset weights (remaining capacity 0.3, transmission distance 0.2, switching fluctuation index 0.2, and average unit efficiency 0.3).

[0171] The calculation results showed that ESV-04 had the highest priority score, making it the optimal execution unit. However, considering that ESV-04 was in cold standby mode and required five minutes to start up, while ESV-01, while having a slightly lower priority, could respond immediately and had a shorter transmission distance, the system ultimately selected ESV-01 to perform the switchover task.

[0172] Upon receiving the task, the ESV-01 immediately terminated regular charging and initiated power regulation. At 10:05:10, the system sent a power boost command to the ESV-01, requiring it to increase its output power from 0 to 800 kW within 30 seconds. The ESV-01's battery management system responded by adjusting the output voltage and frequency through the DC / AC converter, while also monitoring parameters such as battery temperature and SOC to ensure operation within safe limits.

[0173] During the power ramp-up process, the system monitored the status of the grid connection point in real time. When the ESV-01's output power reached 500kW, the grid load fluctuation amplitude dropped from 1.6MW to 1.2MW, and the duration was shortened to 45 seconds, correspondingly reducing the load disturbance value. This demonstrates that the energy storage vehicle's rapid response effectively suppressed load fluctuations and mitigated the impact on the grid.

[0174] Once the ESV-01 was fully connected to the grid (at 10:05:40), its output power stabilized at 800kW, and grid load fluctuations further decreased to ±5%, returning to normal levels. The system recorded key data from the entire switching process: the ESV-01's power output range (0-800kW), response delay (30 seconds), and the change in grid connection level (from level 3 fluctuation to level 2 stability).

[0175] After the switchover, the system continuously monitored ESV-01's remaining capacity and the grid's status. At 10:30, when the electric arc furnace entered stable operation and load fluctuations significantly decreased, the system evaluated whether adjustments to the energy storage vehicle scheduling strategy were necessary. At this point, ESV-01's remaining capacity had dropped to 78%, while ESV-04 had completed its startup preparations and was in hot standby mode.

[0176] The system recalculated the priority scores of each energy storage vehicle and found that ESV-04's score had surpassed ESV-01. Considering ESV-04's higher remaining capacity and superior average unit efficiency, the system decided to perform an energy storage vehicle rotation. At 10:32, the system sent a grid connection command to ESV-04 while gradually reducing ESV-01's output power to achieve a smooth handover. The entire rotation process lasted two minutes, during which grid load fluctuations remained within ±3%.

[0177] Throughout the dispatch process, the system displays the status of each energy storage unit in real time through a visual interface. The left side of the interface displays the location and transmission path of each energy storage vehicle in map format, while the right side displays key parameters (remaining capacity, power output, switching status, etc.) in table format. Operators can view historical load fluctuation curves, energy storage vehicle response times, and other data through the interface, facilitating system performance evaluation and optimizing dispatch strategies.

[0178] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "includes," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0179] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent control system for switching between on-grid and off-grid for mobile energy storage vehicles, characterized in that: include: Grid-connected status detection module and switching control module; The grid-connected state detection module collects multi-dimensional data based on real-time monitoring of grid connection points, including grid voltage amplitude data, frequency offset data, phase synchronization deviation data, and load fluctuation data, and extracts state features to obtain voltage instability coefficient, frequency offset, phase deviation index, and load disturbance value at each sampling moment, and performs dynamic threshold analysis on them to obtain grid-connected stability index corresponding to each sampling moment; constructs a two-dimensional coordinate system with time as the horizontal axis and grid-connected stability index as the vertical axis, marks grid-connected stability index values ​​at each sampling moment as monitoring points in the coordinate system, uses piecewise linear interpolation to connect adjacent monitoring points to generate a dynamic trend curve, performs trend analysis on the dynamic trend curve to obtain the current grid connection state level, and transmits it to the switching control module; The switching control module executes the on-grid and off-grid switching strategy and energy storage unit scheduling based on the grid connection status level; The specific operation process of the switching control module is as follows: Obtain the grid connection status levels of all energy storage units, sort them from low to high, and select the energy storage unit with the lowest level as the switching reference unit; The historical switching delay duration of each energy storage unit and the load change type corresponding to each delay are retrieved. The interval between two adjacent switches is calculated and variance processing is performed to obtain the switching stability period. The load fluctuation amplitude of each switching is extracted and averaged to obtain the disturbance reference amplitude. The switching stability period and the disturbance reference amplitude are correlated and analyzed to obtain the switching fluctuation index of each energy storage unit. Retrieve the historical operating data of each energy storage unit, including the power output range, response delay time, and grid connection status level of each operation. Calculate the output efficiency value by the ratio of power range to response delay time. Combined with the grid connection status level, the efficiency weight is assigned to obtain the average unit efficiency of each energy storage unit. Obtain the current remaining capacity of each energy storage unit and the energy transmission distance to the switching reference unit, and perform multi-dimensional parameter fusion of the switching fluctuation index, unit efficiency mean, current remaining capacity and energy transmission distance to obtain the priority score of each energy storage unit; The energy storage unit with the highest score is selected as the execution unit, and the task of switching the reference unit on and off the grid is assigned to it; All task queues of the execution unit are counted, and a task execution sequence is generated from low to high according to the grid connection status level. The execution unit completes the switching tasks in sequence. After each task is completed, the queue status is updated and an operation log is generated. The log contains the execution unit power output range, response delay time, and grid connection status level change data; Integrate the real-time coordinates and task status of all energy storage units, mark the executing, pending, and completed tasks with different color blocks in the visual interface, and dynamically update the current remaining capacity and switching alarm status of the energy storage units; The energy storage unit is the core execution component of the mobile energy storage vehicle to realize the functions of electric energy storage, scheduling and on-grid and off-grid switching.

2. The intelligent control system for on-grid and off-grid switching of mobile energy storage vehicles according to claim 1 is characterized in that: The calculation process of the output efficiency value is as follows: The interval corresponding to the power output range is divided into several equal power segments. The maximum output duration and average response delay of each power segment are calculated. The efficiency benchmark value is obtained through power segment priority allocation and delay weighted calculation. The efficiency benchmark value is nonlinearly fitted with the grid connection status level to obtain the unit efficiency mean.

3. The intelligent control system for on-grid and off-grid switching of mobile energy storage vehicles according to claim 1 is characterized in that: Feature extraction of grid voltage amplitude data includes: Set the standard voltage reference range of the grid connection point, perform a sliding window comparison between the real-time collected grid voltage amplitude data and the standard voltage reference range, and count the duration and number of fluctuations outside the range; generate the voltage instability coefficient based on the product of the duration and the number of fluctuations.

4. The intelligent control system for switching between on-grid and off-grid of a mobile energy storage vehicle according to claim 3, characterized in that: Feature extraction of frequency offset data includes: Multiple frequency sensors are deployed at the grid connection point, and the frequency differences between adjacent sensors are calculated to generate a gradient distribution map. Spatial clustering analysis is performed based on the extreme value positions and mean offsets in the distribution map to obtain the frequency offset.

5. The intelligent control system for on-grid and off-grid switching of mobile energy storage vehicles according to claim 4 is characterized in that: Feature extraction of phase synchronization deviation data includes: The phase signal of the grid connection point is collected through a phase detection device to separate the synchronous component and the asynchronous component. The amplitude-frequency changes of the asynchronous component are envelope tracked, and the phase deviation index is obtained by combining the trend fitting of the synchronous component.

6. The intelligent control system for on-grid and off-grid switching of mobile energy storage vehicles according to claim 1, characterized in that: Dynamic threshold analysis is implemented as follows: The benchmark threshold interval is generated based on historical power grid abnormal event data, and sliding window statistics are performed in combination with the real-time collected grid stability indicators to dynamically adjust the upper and lower limits of the benchmark threshold interval.

7. The intelligent control system for switching between on-grid and off-grid of a mobile energy storage vehicle according to claim 6, characterized in that: The specific steps of trend analysis are: The inflection points and plateau segments in the dynamic trend curve are extracted, and the slope change rate between adjacent inflection points and plateau segments is calculated; the grid connection status level is determined based on the matching degree between the slope change rate and the preset status template.

8. The intelligent control system for on-grid and off-grid switching of mobile energy storage vehicles according to claim 1, characterized in that: The process of multidimensional parameter fusion is as follows: The switching fluctuation index, unit efficiency mean, current remaining capacity and energy transmission distance are normalized respectively, and linearly superimposed through the preset priority coefficient to output the priority score.

9. The intelligent control system for on-grid and off-grid switching of mobile energy storage vehicles according to claim 1, characterized in that: The feature extraction process of load fluctuation data includes: A standard load fluctuation threshold is set for the grid connection point. The real-time collected load fluctuation data is differentially compared with the standard load fluctuation threshold. The fluctuation amplitude and duration exceeding the standard load fluctuation threshold are counted. The load disturbance value is generated based on the product of the fluctuation amplitude and duration.

Images