Charging stopping integrated management method and system based on digital twinning

Abstract

The application discloses a kind of based on digital twinning's stop fills integrated management method and system, belong to new energy automobile technical field, this method includes: acquisition charging power sequence and battery state parameter, constructs characteristic vector and calculates its amplitude.In the anti-occupying position countdown period receives again charging request, current and historical charging event is mapped as node, to define phase angle with countdown remaining time proportion, the final phase coupling strength between nodes is calculated in combination with amplitude.Obtain the phase angle distribution of historical normal charging event, calculate the standardization of current node Deviation degree and filter out high deviation history node, construct complex vector to determine weighted synchronization degree and high deviation network average phase.Finally, charging intention confidence score is generated from weighted synchronization degree, average phase and deviation degree, and hierarchical countdown control instruction is generated according to this and mapped to digital twinning model.Through the application, the robustness of anti-occupying mechanism can be improved.

Description

Technical Field

[0001] This invention belongs to the field of new energy vehicle technology, specifically relating to a method and system for integrated control of charging and shutdown based on digital twins. Background Technology

[0002] To ensure efficient utilization of charging station resources, when a charging pile transitions from a charging in progress state to a completed state, the system initiates an anti-occupancy countdown. If the vehicle does not leave within the specified time, an occupancy fee is charged. However, if a user re-initiates a charging request just before the countdown expires, the charging pile receives a valid handshake signal from the vehicle's battery management system, the state machine re-enters the charging in progress state, and the countdown is reset. The system cannot determine whether the charging request reflects a genuine user need or is merely a formality to circumvent the occupancy fee based solely on the state transition, causing the anti-occupancy mechanism to fail.

[0003] To address the aforementioned issues, threshold rules can be set to identify suspected evasion behaviors, such as single charging sessions lasting less than 5 minutes or energy increments below 1 kWh. When anomalies are detected, the countdown timer can be refused or the next countdown timer shortened. This method can filter out some obvious short-duration fake charging behaviors to a certain extent. However, once users learn the system's judgment boundaries through repeated trials, they can still continuously evade detection using carefully designed parameter combinations, causing the anti-occupancy mechanism to fail. Summary of the Invention

[0004] To address the aforementioned problems, this invention provides a digital twin-based integrated control method and system for charging and stopping, thereby resolving the issues present in the background art.

[0005] To achieve the aforementioned objectives, this invention proposes a digital twin-based integrated charging / stop management method, comprising: Obtain the charging power sequence output by the charging pile, construct a feature vector based on the charging power sequence and battery state parameters, and calculate the Euclidean norm of the feature vector to obtain the amplitude. When a recharging request is received during the anti-occupancy countdown, the current charging process of the charging pile is defined as a charging event, the current charging event is mapped to the current node, and the historical charging events are mapped to historical nodes. The ratio of the remaining countdown time of the current node to the initial countdown time is mapped to the phase angle of the current node. Based on the amplitude and phase angle of the current node and the historical nodes, the final phase coupling strength between the current node and the historical nodes is calculated. Obtain the phase angle distribution of normal charging events in historical nodes, calculate the standardized deviation based on the phase angle of the current node and the phase angle distribution, filter out high deviation historical nodes from the historical nodes, construct a complex vector based on the amplitude and phase angle of the high deviation historical nodes, and calculate the weighted synchronization degree and the average phase of the high deviation network based on the final phase coupling strength and the complex vector. The charging intention confidence score is calculated based on the weighted synchronization degree, the average phase of the high deviation network, and the standardized deviation degree. The corresponding countdown control command is generated according to the charging intention confidence score and mapped to the digital twin model.

[0006] This invention also provides a digital twin-based integrated charging and shutdown control system, which is used to implement the above-described method. The system includes: The feature acquisition module obtains the charging power sequence output by the charging pile, constructs a feature vector based on the charging power sequence and battery state parameters, and calculates the Euclidean norm of the feature vector to obtain the amplitude. The associated network construction module, when receiving a recharging request during the anti-occupancy countdown, defines the current charging process of the charging pile as a charging event, maps the current charging event to the current node, maps the historical charging events to historical nodes, maps the ratio of the remaining countdown time of the current node to the initial countdown time to the phase angle of the current node, and calculates the final phase coupling strength between the current node and the historical node based on the amplitude and phase angle of the current node and the historical node. The deviation calculation module obtains the phase angle distribution of normal charging events in historical nodes, calculates the standardized deviation degree based on the phase angle of the current node and the phase angle distribution, filters high deviation historical nodes from historical nodes, constructs a complex vector based on the amplitude and phase angle of the high deviation historical nodes, and calculates the weighted synchronization degree and the average phase of the high deviation network based on the final phase coupling strength and the complex vector. The decision mapping module calculates the charging intention confidence score based on weighted synchronization degree, high deviation network average phase and standardized deviation degree, generates corresponding countdown control commands based on the charging intention confidence score, and maps them to the digital twin model.

[0007] The beneficial effects of this invention are as follows: This invention first constructs a feature vector based on charging power sequence and battery state parameters and calculates the amplitude. When a recharging request is received during the anti-occupancy countdown, the current and historical charging events are mapped to nodes, and the remaining countdown timer is mapped to a phase angle. The final phase coupling strength between nodes is calculated based on the amplitude, thereby quantifying the correlation between the current event and historical events in terms of triggering timing and physical characteristics. Then, high-deviation historical nodes are selected, and their amplitude and phase angle are used to construct complex vectors. Based on the final phase coupling strength, a weighted synchronization degree and the average phase of the high-deviation network are calculated to capture the characteristics of historical malicious event clusters. Finally, a charging intent confidence score is generated based on the fusion of weighted synchronization degree, average phase, and deviation degree, and a hierarchical countdown control command is generated accordingly. This invention overcomes the defect that fixed threshold rules are easily circumvented by repeated trial and error, can identify formalized charging operations, and effectively improves the robustness of the anti-occupancy mechanism. Attached Figure Description

[0008] Figure 1 This is a flowchart illustrating the steps of a digital twin-based integrated control method for charging and stopping according to the present invention. Figure 2 This is a comparison diagram showing the robustness of the control method of this invention against evasion in existing technologies; Figure 3 This is a schematic diagram of the integrated control system for charging and stopping based on digital twins according to the present invention. Detailed Implementation

[0009] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0010] like Figure 1 As shown, a digital twin-based integrated control method for charging and stopping includes: Obtain the charging power sequence output by the charging pile, construct a feature vector based on the charging power sequence and battery state parameters, and calculate the Euclidean norm of the feature vector to obtain the amplitude.

[0011] In this embodiment, a feature vector is constructed based on the charging power sequence and battery state parameters, including: The change in battery state of charge during the charging process is obtained from the battery state parameters. The theoretical energy transfer is obtained by multiplying the change in battery state of charge by the preset nominal battery capacity. The actual energy transfer is calculated based on the charging power sequence. The energy transfer index is calculated based on the actual energy transfer and the theoretical energy transfer.

[0012] The charging pile's internal current and voltage sensors measure the actual output voltage and current values ​​in real time and calculate the charging power. Based on this power, a charging power sequence arranged in chronological order is generated. The State of Charge (SOC) value at the start and end of charging is obtained; the difference between these two values ​​represents the change in battery state of charge (SOC). Multiplying this SOC by the battery's nominal capacity yields the theoretical energy transfer. The collected charging power sequence is then integrated over time, from the start to the end of charging; the integration result represents the actual energy transfer. Finally, the energy transfer index is calculated using the following formula. The calculation formula is: ,in, This represents the actual amount of energy transferred. The theoretical energy transfer value represents the deviation between the actual energy transferred into the battery and the theoretical energy calculated based on changes in State of Charge (SOC). When a user adjusts the target SOC from 80% to 81% to avoid charging fees (i.e., only charging 1% of the battery capacity), the theoretical energy transfer value is 1% of the battery capacity. However, the actual charging time is extremely short, and the actual energy transfer value may only be 30% of the theoretical value, resulting in a relatively high energy transfer value. Conversely, if the user actually continues charging, the actual energy transfer value is basically consistent with the theoretical energy transfer value, and the value of this value is close to zero.

[0013] Calculate the sum of squares of power changes in the charging power sequence, and then perform a negative exponential mapping on the sum of squares of power changes to obtain the power index.

[0014] For the power index, the power difference between adjacent points in the charging power sequence is calculated, the differences are squared and summed to obtain the sum of squares of power changes. This sum of squares of power changes is then used as the independent variable of an exponential function, and after negative exponential mapping, the power index is obtained. The specific calculation formula is as follows: ,in In this embodiment, the preset scaling factor is used. The value range is from 0.001 to 0.01. In practical applications, Calibration can be performed based on the rated power and sampling frequency of the charging pile. This represents the sum of squares of the power changes.

[0015] The power index represents the smoothness of the charging power curve. During actual charging, the charging pile output power follows the physical laws of the battery charging curve, and the power changes continuously and steadily over time. The power difference between adjacent sampling points is small, and the sum of squares of power changes is also small. After negative exponential mapping, the power index is close to 1. However, when a user briefly initiates charging to trigger a handshake, the charging process is short and involves human intervention, resulting in abrupt changes in the power curve, a larger sum of squares of power changes, and a smaller power index value.

[0016] Get the remaining time of the anti-occupancy countdown, get the preset initial countdown time, calculate the ratio of the remaining time to the initial countdown time, and if the ratio is less than the preset time boundary ratio, set the time indicator to the first value; otherwise, set it to the second value.

[0017] When a charging station transitions from charging in progress to finished charging, the system initiates an anti-occupancy countdown, with an initial countdown duration set by a preset value. The system records the remaining countdown timer in real time. Upon receiving a new charging request, the system reads the remaining time at that moment. It calculates the ratio of the remaining time to the initial countdown timer as the countdown progress ratio and checks if this ratio is less than 0.2. If it is less than 0.2, it indicates the charging request occurred within the last 20% of the countdown window, and the system sets the time indicator to 1 (the first value). If the ratio is greater than or equal to 0.2, the system sets the time indicator to 0 (the second value). The time indicator represents the proximity between the charging request initiation time and the countdown end time. When users operate to avoid occupancy fees, they typically trigger charging just before the countdown ends, resulting in a time indicator of 1. Normal users' charging needs are unaffected by the countdown; the initiation time of charging requests is randomly distributed throughout the countdown period, making a time indicator of 1 less likely.

[0018] Obtain the battery temperature change rate from the battery status parameters. If the temperature change rate is less than the preset temperature change threshold and the charging duration is less than the preset time threshold, set the thermal balance index to the first value; otherwise, set it to the second value.

[0019] The normalized energy transfer index, power index, time index, and heat balance index are used as feature vectors.

[0020] The system reads the battery temperature at the start and end of charging, calculates the temperature change, and uses the ratio of this temperature change to the charging duration as the temperature change rate. If the temperature change rate is less than a preset threshold and the charging duration is less than 5 minutes, it indicates that the charging process did not cause a significant thermal effect, violating the physical laws of charging, and the system sets the thermal balance index to 1. Otherwise, the system sets the thermal balance index to 0. The thermal balance index represents whether the charging behavior conforms to the battery's thermodynamic characteristics. When the user briefly initiates charging, the charging time is extremely short, and the battery temperature remains almost unchanged; this index is 1. During actual charging, the battery temperature continuously rises as charging progresses; this index is 0.

[0021] The system combines energy transfer indicators, power indicators, time indicators, and thermal balance indicators in sequence into a four-dimensional column vector to obtain a feature vector. This feature vector reflects the characteristics of the current charging event in terms of energy transfer rationality, power curve smoothness, triggering timing specificity, and thermodynamic compliance.

[0022] When a recharging request is received during the anti-occupancy countdown, the current charging process of the charging pile is defined as a charging event, the current charging event is mapped to the current node, and the historical charging events are mapped to historical nodes.

[0023] In the node network, nodes store the phase and amplitude information of each charging event, while the edge weight set stores the phase coupling relationship between nodes. This embodiment only constructs nodes for charging events involving power outages and subsequent recharging. When the charging pile transitions from a charging in progress state to a completed state, the system initiates an anti-occupancy countdown. During the countdown, if the charging pile receives another charging request from the vehicle, causing its state to transition back from completed to charging, the system defines this process as a charging event. For normal single-charge, fully charged, and vehicle-departure processes without state transitions, the system does not construct event nodes.

[0024] The ratio of the remaining countdown time of the current node to the initial countdown time is mapped to the phase angle of the current node. Based on the amplitude and phase angle of the current node and the historical nodes, the final phase coupling strength between the current node and the historical nodes is calculated.

[0025] In this embodiment, the basic phase coupling strength is calculated based on the amplitude of the current node, the amplitude of the historical node, the phase angle of the current node, and the phase angle of the historical node. The user identifiers stored in the current node and the historical node are obtained. If the user identifiers of the current node and the historical node are the same, the indicator function is set to the first weight value; otherwise, it is set to the second weight value. The user home gain coefficient is calculated based on the indicator function and the preset gain constant. The basic phase coupling strength is multiplied by the user home gain coefficient to obtain the final phase coupling strength.

[0026] The countdown progress percentage is mapped to a phase angle using the following formula: ,in Let be the phase angle of the i-th charging event. Let represent the countdown progress percentage for the i-th charging event. This mapping linearly transforms the countdown progress into coordinates in angle space. The Euclidean norm of the eigenvector corresponding to the charging event is calculated to obtain the amplitude. The amplitude calculation formula is: ,in Let be the amplitude of the i-th charging event. The normalized energy transfer index, The normalized power index, The normalized time index This is a normalized thermal equilibrium index. The amplitude represents the overall intensity of the physical quantity characteristics of the event; a larger value indicates a higher degree of anomaly in the event.

[0027] The system combines phase angle and amplitude into a set of state parameters, which identify the position and intensity of the i-th charging event in phase space. The node network is defined as a site-level global network, with historical nodes derived from all charging events that have occurred on vehicles within the charging site. When creating a node, the user identifier is stored as a node attribute. Event nodes created based on the current charging event are defined as current nodes, and previously created nodes are defined as historical nodes. For the current node and any historical node, the system calculates the fundamental phase coupling strength between them based on the following formula: ,in The phase coupling strength between the current node and historical nodes. The amplitude of the current event. For the amplitude of historical events, The phase angle of the current event. The phase angle of a historical event.

[0028] Read the user ID of the current node and the user IDs of historical nodes to determine if the two nodes belong to the same user. Calculate the user affiliation gain coefficient based on the following formula. ,in User home gain coefficient The preset gain constant, For the indicator function, when the current node user identifier With historical node user identifier The value is 1 when they are the same and 0 when they are different. The system multiplies the basic phase coupling strength by the user's home gain coefficient to obtain the final phase coupling strength.

[0029] The final phase coupling strength is stored in the edge weight set. By introducing user affiliation gain, the historical events of the same user receive higher weight in the phase coupling calculation, enabling the system to identify malicious patterns of new users based on the global history of the site, and to be more sensitive to repeated malicious behavior of the same user.

[0030] Repeat the above calculation process for all historical nodes in the network to obtain the phase coupling strength between the current node and each historical node, and store these coupling strength values ​​in the edge weight set.

[0031] Obtain the phase angle distribution of normal charging events in historical nodes, calculate the standardized deviation based on the phase angle of the current node and the phase angle distribution, filter high deviation historical nodes from the historical nodes, construct a complex vector based on the amplitude and phase angle of the high deviation historical nodes, and calculate the weighted synchronization degree and the average phase of the high deviation network based on the final phase coupling strength and the complex vector.

[0032] The normal event sample is obtained from historical charging events where the charging duration is greater than a preset duration and the battery state of charge increment is greater than a preset proportion. The standard deviation of the phase angle of all normal event samples is calculated. Based on the phase angle of the current node, the mean phase angle of normal events, and the standard deviation of the phase angle of normal events, the standardized deviation is calculated.

[0033] The system analyzes historical charging event data for all users within the charging station, selecting events with a charging duration greater than 10 minutes and a SOC increase greater than 5% as normal event samples. The system calculates the phase angle distribution of these normal event samples and performs an arithmetic mean calculation on the phase angles of all normal event samples to obtain the mean phase angle. The system also calculates the standard deviation of the phase angle for each normal event sample.

[0034] Standardized deviation is calculated based on the following formula: ,in For standardized deviation, The phase angle of the current event. The mean phase angle of a normal event. The standard deviation represents the phase angle standard deviation of a normal event. The standardized deviation indicates the degree to which the trigger time of the current charging event deviates from the normal user trigger time. A larger value indicates that the trigger time of the current event is closer to the end of the countdown and deviates more from the statistical regularity of normal charging behavior. When the standardized deviation is less than 1, it indicates that the phase angle of the current event is within one standard deviation of the normal distribution, belonging to a normal trigger time. When the standardized deviation is greater than 2, it indicates that the phase angle of the current event has exceeded two standard deviations of the normal distribution, belonging to an abnormal trigger time.

[0035] Nodes whose phase angle exceeds the screening threshold and whose amplitude is greater than the preset abnormal threshold are selected from all historical nodes to form a high deviation historical node set. Complex numbers are constructed based on the amplitude and phase angle of each node in the high deviation historical node set. The complex numbers are then weighted and summed with the final phase coupling strength of the corresponding nodes to obtain a weighted complex number sum. The weighted complex number sum is then combined with the final phase coupling strength of the high deviation historical node set to obtain the weighted complex order parameter. The weighted synchronization degree and the average phase of the high deviation network are calculated using the weighted complex order parameter.

[0036] Historical nodes that meet preset malicious characteristic criteria are selected from historical nodes to form a high-deviation historical node set. During selection, a preset anomaly threshold is used, and the sum of the mean phase angle of normal events and twice the standard deviation of the phase angle of normal events is also used as the selection threshold. For the j-th node in the historical node set, if its phase angle is greater than the selection threshold and its amplitude is greater than the preset anomaly threshold, the node is added to the high-deviation historical node set. This selection method ensures that the nodes in the set not only correspond to events triggered at the end of the countdown, but also exhibit abnormal physical quantity characteristics such as unreasonable energy transfer or discontinuous power, thereby eliminating normal nodes that, although triggered at the end, are genuine short-term recharges.

[0037] The final phase coupling strength between the current node and all nodes within the high-deviation historical node set is summed to obtain the final phase coupling strength sum. When the final phase coupling strength sum is zero, it indicates that there are no event records in the station's history that meet the malicious characteristic conditions. The system directly sets the weighted synchronization degree to 0 and skips the subsequent weighted synchronization degree calculation steps.

[0038] When the final sum of phase coupling strengths is greater than zero, the complex number of each node in the set of high-deviation historical nodes is determined based on the following formula: ,in, and Let be the amplitude and phase angle of the j-th high-deviation history node, respectively. Multiply the complex number by the final phase coupling strength to obtain a weighted complex number. Accumulate the weighted complex numbers for all high-deviation history nodes to obtain a weighted complex sum. Divide the weighted complex sum by the sum of the final phase coupling strengths of the high-deviation nodes to obtain the reweighted complex order parameter. The reweighted complex order parameter corresponds to a vector in the complex plane, and the direction and magnitude of this vector reflect the overall trend and intensity of the phase distribution of the high-deviation history nodes.

[0039] The magnitude of the weighted complex sequence parameter is calculated to obtain the weighted synchronization degree. The weighted synchronization degree represents the degree of phase consistency between the current event and a group of known charging events within the site that are at the end of their countdown and exhibit abnormal physical quantities. When the phase angles of high deviations from historical nodes are highly concentrated and have large amplitudes, the magnitude of the weighted complex sum is close to the sum of phase coupling strengths, and the weighted synchronization degree is close to 1. When the phase angles of high deviations from historical nodes are dispersed, the complex vectors in different directions cancel each other out, the magnitude of the weighted complex sum is small, and the weighted synchronization degree is close to 0. A value close to 1 indicates that the current event highly overlaps with the group of historical malicious events in the site's phase space, while a value close to 0 indicates that although the current event is in a high deviation region, its distribution is inconsistent with that of historical malicious events.

[0040] In this embodiment, after calculating the weighted synchronization degree, the method further includes: If the weighted synchronization degree is less than the preset effective resonance threshold, the absolute difference between the phase angle of the current node and the phase angle of each historical node in the high deviation historical node set is calculated. Based on the absolute difference, local coherent subclusters are selected from the high deviation historical node set. A complex vector is constructed based on the amplitude and phase angle of each historical node in the local coherent subcluster. Based on the final phase coupling strength and the complex vector of the local coherent subcluster, the local weighted synchronization degree and the local high deviation network average phase are recalculated. The local weighted synchronization degree is used as the weighted synchronization degree, and the local high deviation network average phase is used as the high deviation network average phase.

[0041] In this embodiment, local coherent subclusters are selected from the set of historical nodes with high deviation based on the absolute difference, including: Calculate the absolute difference between the phase angle of each historical node in the high deviation historical node set and the phase angle of the current node. Sort all historical nodes in the high deviation historical node set in ascending order of absolute difference, and select the first preset number of historical nodes to form a local coherent sub-cluster.

[0042] In actual site operation, when multiple malicious user groups with different time preferences exist simultaneously within the site, for example, the first group typically triggers charging when the anti-occupancy countdown has 10% remaining, with their phase angle distribution concentrated around θ≈1.8π, while the second group typically triggers when the countdown has 1% remaining, with their phase angle distribution concentrated around θ≈1.98π. In this case, nodes in the high-deviation historical node set exhibit a bimodal or even multimodal clustered distribution in the phase space. Since the calculation of the complex sequence parameter is essentially a weighted summation of complex vectors, when the phase angles of different malicious clusters exhibit angular separation distributions on the complex plane, their corresponding complex vectors have different directions, leading to vector cancellation during the accumulation process. For example, if the weighted complex sum of the first type of malicious cluster points to the direction of angle φ1 on the complex plane, and the weighted complex sum of the second type of malicious cluster points to the direction of angle φ2, and |φ1-φ2| is close to π, the two vectors are almost opposite, and the magnitude of their vector sum will shrink sharply or even approach zero, resulting in the final calculated weighted synchronization degree value being far lower than the actual consistency strength of malicious behavior.

[0043] To address this issue, after calculating the weighted synchronization degree, it is compared with a preset effective resonance threshold. The effective resonance threshold is determined based on historical data from the field station, and in this embodiment, it is set to 0.35. If the weighted synchronization degree is greater than the effective resonance threshold, it indicates that the phase distribution of the high-deviation historical nodes exhibits a single-peak or weakly multi-peak characteristic. The average phase of the high-deviation network is then calculated according to the original procedure, and the subsequent confidence assessment step is initiated.

[0044] If the weighted synchronization degree is less than the effective resonance threshold, it indicates that there may be vector cancellation due to multi-peak divergence. In this case, the phase angle of the current node is read, and all nodes in the high-deviation historical node set are traversed. The absolute difference between the phase angle θj of each historical node j and the phase angle θc of the current node is calculated. All historical nodes are sorted in ascending order of absolute difference, and the top N nodes with the smallest difference are selected to form a preliminary local coherent subcluster, where N is a preset subcluster size parameter. In this embodiment, it is set to 30% of the total number of high-deviation historical nodes and no less than 5 nodes.

[0045] The system checks whether the number of nodes in the initial local coherent subcluster meets the preset minimum statistical sample size. If it does, the system proceeds to the recalculation step. If it does not, it indicates that the number of historical malicious nodes with phases close to the current node is insufficient, which may lead to insufficient statistical reliability of the recalculation. In this case, the system backtracks to the screening step of constructing the high-deviation historical node set, reads the anomaly threshold used in the original screening, multiplies the anomaly threshold by a relaxation factor of 0.8, and obtains a temporary relaxation threshold. From the entire historical node set, nodes with phase angles greater than the screening threshold and amplitudes greater than the temporary relaxation threshold are re-screened. The newly screened nodes are added to the initial local coherent subcluster until the number of nodes in the subcluster reaches the minimum statistical sample size.

[0046] After completing the construction of the local coherent subclusters, the complex sequence parameter calculation process is re-executed based on the local coherent subclusters to obtain the weighted synchronization degree and the average phase of the high deviation network.

[0047] This mechanism ensures that even when there are multiple malicious user groups with different time preferences within the site, the system can still accurately identify the matching relationship between the current event and its actual malicious cluster, significantly reducing the false negative rate caused by multimodal distribution.

[0048] If the weighted synchronization degree is not less than the preset effective resonance threshold, then the calculated weighted synchronization degree and the high deviation network average phase are retained.

[0049] If the weighted synchronization degree is not less than the preset effective resonance threshold, it means that there are no multi-peaks in the set of high-deviation historical nodes. In this case, the previously calculated weighted synchronization degree and the average phase of the high-deviation network are directly used to participate in the subsequent calculation.

[0050] The charging intention confidence score is calculated based on weighted synchronization degree, high deviation network average phase, and standardized deviation. A corresponding countdown control command is generated based on this score and mapped to a digital twin model. The digital twin model mapped in this invention is pre-constructed based on real charging stations. By acquiring the charging pile location distribution of the target charging station, a 3D visualization model of the vehicle and charging piles with the same geometric dimensions and spatial layout as the physical station is established in virtual space. Simultaneously, a data binding relationship is established between the physical charging pile state machine (including charging in progress, completed, offline, faulty, and anti-occupancy countdown states) and the virtual model state machine. Through an IoT gateway, the real-time output charging power sequence, battery status parameters, and remaining countdown timer from the physical charging pile are used as driving data to update the state of the corresponding virtual components in the digital twin model in real time, achieving a state mirroring between the physical world and the virtual space. Based on this, the system synchronizes the countdown control command and the currently determined confidence score to the digital twin model, and the digital twin model updates the countdown display status of the corresponding parking space in the virtual environment. If a low-confidence fake charging is identified, the virtual model can trigger a high-brightness warning, enabling site managers to identify malicious nodes engaging in evasion behavior on the digital twin system.

[0051] In this embodiment, calculating the confidence score of charging intention includes: Based on the difference between the phase angle of the current node and the average phase of the target high deviation network, the phase resonance score is calculated in combination with the target weighted synchronization degree. Based on the standardized deviation degree and the phase resonance score, the comprehensive anomaly score is calculated. The comprehensive anomaly score is modulated using the amplitude of the current node to obtain the modulated anomaly score. The modulated anomaly score is then mapped and transformed to obtain the charging intention confidence score.

[0052] First, the argument of the weighted complex sequence parameter is calculated to obtain the average phase of the high-deviation network. The average phase of the high-deviation network represents the center position of the phase distribution of high-deviation historical nodes, reflecting the typical trigger time of malicious charging events within the station during the countdown period. Next, the difference between the current node's phase angle and the average phase of the high-deviation network is calculated to obtain the phase deviation angle. The phase deviation angle represents the degree of deviation of the trigger time of the current charging event from the center of the historical malicious pattern; the smaller the value, the more consistent the current event is with the typical trigger time of the historical malicious pattern. Finally, the weighted synchronization degree is multiplied by the cosine value of the phase deviation angle to obtain the phase resonance score. The phase resonance score represents the degree of resonance between the current event and the high-deviation historical event group. When the weighted synchronization degree is high and the phase deviation angle is small, the cosine function value is close to 1, and the resonance score is close to the value of the weighted synchronization degree.

[0053] The standardized deviation is normalized, and then multiplied by a preset first fusion weight. The phase resonance score is multiplied by a preset second fusion weight, and the two products are added together to obtain the comprehensive anomaly score. The amplitude of the current node is normalized, and then the modulation factor is calculated using the following formula: Modulation factor = Normalized amplitude × 0.5 + 0.5. The comprehensive anomaly score is multiplied by the modulation factor to obtain the modulated anomaly score. When the amplitude is large, the normalized amplitude is close to 1, the modulation factor is close to 1, and the anomaly score is not compressed. When the amplitude of the current event is close to 0, the normalized amplitude is close to 0, the modulation factor is 0.5, and the anomaly score is halved. The purpose of this modulation is that even if the current event is at the end of a countdown and synchronized with historical high-deviation events, if the physical characteristics show that the energy transfer of the event is reasonable, the power is continuous, and the temperature change is normal, the system reduces the anomaly score, avoiding misjudgment of genuine recharging.

[0054] The mapping slope parameter and the judgment threshold are preset. The mapping slope parameter is set to 5, and the judgment threshold is set to 0.5. The difference between the modulated anomaly score and the judgment threshold is calculated. This difference is multiplied by the mapping slope parameter to obtain the input value. This input value is then input into the Sigmoid function to obtain the output value. The charging intention confidence score is then calculated as 1 - the output value. When the modulated anomaly score is greater than the judgment threshold, the output of the Sigmoid function is close to 1, and after inversion, the confidence score is close to 0, indicating a low level of confidence in the charging behavior. When the modulated anomaly score is less than the judgment threshold, the output of the Sigmoid function is close to 0, and after inversion, the confidence score is close to 1, indicating a high level of confidence in the charging behavior.

[0055] In this embodiment, if the confidence score of charging intention is greater than the preset high confidence threshold, the remaining time of the anti-occupancy countdown is restored to the initial countdown time. If the confidence score of charging intention is greater than or equal to the preset low confidence threshold and less than or equal to the high confidence threshold, the remaining time of the anti-occupancy countdown is extended by a preset penalty time based on the current value. If the confidence score of charging intention is less than the low confidence threshold, the remaining time of the anti-occupancy countdown continues to decrease.

[0056] The charging intention confidence score is compared with preset high-confidence and low-confidence thresholds. The high-confidence threshold is set to 0.75, and the low-confidence threshold is set to 0.4. When the charging intention confidence score is greater than 0.75, the system determines that the current charging event is a high-confidence genuine charging event and generates a countdown complete reset command, which restores the remaining time of the anti-occupancy countdown to the initial countdown time.

[0057] When the confidence score of the charging intention is greater than or equal to 0.4 and less than or equal to 0.75, the system determines the current charging event as a medium-confidence charging event. A countdown partial reset command is generated, which extends the remaining duration of the anti-occupancy countdown by 15 minutes based on the current value.

[0058] When the confidence score of the charging intention is less than 0.4, the system determines that the current charging event is a low-confidence spurious charging event. A countdown maintenance command is generated, which keeps the remaining time of the anti-occupancy countdown decreasing without performing any reset or extension operations.

[0059] like Figure 2 As shown, in scenarios where malicious users continuously probe, the fixed threshold rules used in existing technologies become easily figured out by users after several attempts, leading to an increased success rate of avoidance. However, this invention, by constructing feature vectors and calculating the final phase coupling strength between the current node and historical nodes, maintains an extremely low avoidance rate even when users continuously change parameters. This is because the user's behavior consistently resonates highly with the historical high-deviation network in the phase space, demonstrating extremely high dynamic robustness.

[0060] like Figure 3 As shown, the present invention also provides a digital twin-based integrated charging and shutdown control system, which is used to implement the above-described method. The system includes: The feature acquisition module obtains the charging power sequence output by the charging pile, constructs a feature vector based on the charging power sequence and battery state parameters, and calculates the Euclidean norm of the feature vector to obtain the amplitude.

[0061] The associated network construction module, when receiving a request to recharge during the anti-occupancy countdown, defines the current charging process of the charging pile as a charging event, maps the current charging event to the current node, maps the historical charging events to historical nodes, maps the ratio of the remaining countdown time of the current node to the initial countdown time to the phase angle of the current node, and calculates the final phase coupling strength between the current node and the historical node based on the amplitude and phase angle of the current node and the historical node.

[0062] The deviation calculation module obtains the phase angle distribution of normal charging events in historical nodes, calculates the standardized deviation degree based on the phase angle of the current node and the phase angle distribution, filters high deviation historical nodes from the historical nodes, constructs a complex vector based on the amplitude and phase angle of the high deviation historical nodes, and calculates the weighted synchronization degree and the average phase of the high deviation network based on the final phase coupling strength and the complex vector.

[0063] The decision mapping module calculates the charging intention confidence score based on weighted synchronization degree, high deviation network average phase and standardized deviation degree, generates corresponding countdown control commands based on the charging intention confidence score, and maps them to the digital twin model.

[0064] It should be noted that the preset thresholds, algorithm parameters, and specific values ​​of normalization processing involved in the embodiments of the present invention are merely illustrative examples and are not intended to limit the present invention. Those skilled in the art can determine appropriate parameter values ​​through limited experimental testing and statistical analysis based on the historical data distribution of actual sites, equipment characteristics, and management needs; this process requires no creative effort. Furthermore, regarding the denominator in the formulas, when the denominator is zero in the extreme case, those skilled in the art will employ conventional numerical processing methods, such as adding a small positive number to the denominator or performing conditional branch judgments to avoid computational anomalies. These processing methods are conventional techniques. To make the technical solutions and formulas of the present invention clearer and more concise, each possible boundary case will not be described in detail here.

Claims

1. A method for integrated control of charging and stopping based on digital twins, characterized in that, include: Obtain the charging power sequence output by the charging pile, construct a feature vector based on the charging power sequence and battery state parameters, and calculate the Euclidean norm of the feature vector to obtain the amplitude. When a recharging request is received during the anti-occupancy countdown, the current charging process of the charging pile is defined as a charging event, the current charging event is mapped to the current node, and the historical charging events are mapped to historical nodes. The ratio of the remaining countdown time of the current node to the initial countdown time is mapped to the phase angle of the current node. Based on the amplitude and phase angle of the current node and the historical nodes, the final phase coupling strength between the current node and the historical nodes is calculated. The phase angle distribution of normal charging events in historical nodes is obtained. The standardized deviation is calculated based on the phase angle of the current node and the phase angle distribution. High-deviation historical nodes are selected from the historical nodes. A complex vector is constructed based on the amplitude and phase angle of the high-deviation historical nodes. The weighted synchronization degree and the average phase of the high-deviation network are calculated based on the final phase coupling strength and the complex vector. Among them, events in the historical charging events with a charging duration greater than a preset duration and a battery state of charge increment greater than a preset proportion are selected as normal event samples. The phase angle standard deviation of all normal event samples is calculated. The standardized deviation is calculated based on the phase angle of the current node, the mean phase angle of normal events, and the standard deviation of the phase angle of normal events. Nodes whose phase angle exceeds the screening threshold and whose amplitude is greater than the preset abnormal threshold are selected from all historical nodes to form a high deviation historical node set. Complex numbers are constructed based on the amplitude and phase angle of each node in the high deviation historical node set. The complex numbers are then weighted and summed with the final phase coupling strength of the corresponding nodes to obtain a weighted complex number sum. The weighted complex number sum is then combined with the final phase coupling strength of the high deviation historical node set to obtain the weighted complex order parameter. The weighted synchronization degree and the average phase of the high deviation network are calculated using the weighted complex order parameter. If the weighted synchronization degree is less than the preset effective resonance threshold, the absolute difference between the phase angle of the current node and the phase angle of each historical node in the high deviation historical node set is calculated. Based on the absolute difference, local coherent subclusters are selected from the high deviation historical node set. A complex vector is constructed based on the amplitude and phase angle of each historical node in the local coherent subcluster. Based on the final phase coupling strength and the complex vector of the local coherent subcluster, the local weighted synchronization degree and the local high deviation network average phase are recalculated. The local weighted synchronization degree is used as the weighted synchronization degree, and the local high deviation network average phase is used as the high deviation network average phase. If the weighted synchronization degree is not less than the preset effective resonance threshold, the calculated weighted synchronization degree and the average phase of the high deviation network are retained. Based on the weighted complex number and the sum of the final phase coupling strength with the set of high deviation historical nodes, the weighted complex order parameter is obtained, and the weighted synchronization degree and the average phase of the high deviation network are calculated through the weighted complex order parameter. The charging intention confidence score is calculated based on the weighted synchronization degree, the average phase of the high deviation network, and the standardized deviation degree. The corresponding countdown control command is generated according to the charging intention confidence score and mapped to the digital twin model.

2. The method according to claim 1, characterized in that, A feature vector is constructed based on the charging power sequence and battery state parameters, including: The change in battery state of charge during the charging process is obtained from the battery state parameters. The change in battery state of charge is multiplied by the preset nominal battery capacity to obtain the theoretical energy transfer amount. The actual energy transfer amount is calculated based on the charging power sequence. The energy transfer index is calculated based on the actual energy transfer amount and the theoretical energy transfer amount. Calculate the sum of squares of power changes in the charging power sequence, and perform a negative exponential mapping on the sum of squares of power changes to obtain the power index; Get the remaining time of the anti-occupancy countdown, get the preset initial countdown time, calculate the ratio of the remaining time to the initial countdown time, and if the ratio is less than the preset time boundary ratio, set the time indicator to the first value; otherwise, set it to the second value. Obtain the battery temperature change rate from the battery status parameters. If the temperature change rate is less than the preset temperature change threshold and the charging duration is less than the preset time threshold, set the thermal balance index to the first value; otherwise, set it to the second value. The normalized energy transfer index, power index, time index, and heat balance index are used as feature vectors.

3. The method according to claim 1, characterized in that, Calculate the final phase coupling strength between the current node and historical nodes, including: Based on the amplitude of the current node, the amplitude of the historical nodes, the phase angle of the current node, and the phase angle of the historical nodes, the basic phase coupling strength is calculated. The user identifiers stored in the current node and the historical nodes are obtained. If the user identifiers of the current node and the historical nodes are the same, the indicator function is set to the first weight value; otherwise, it is set to the second weight value. The user home gain coefficient is calculated based on the indicator function and the preset gain constant. The basic phase coupling strength is multiplied by the user home gain coefficient to obtain the final phase coupling strength.

4. The method according to claim 1, characterized in that, Locally coherent subclusters are selected from the set of historical nodes with high deviation based on absolute difference, including Calculate the absolute difference between the phase angle of each historical node in the high deviation historical node set and the phase angle of the current node. Sort all historical nodes in the high deviation historical node set in ascending order of absolute difference, and select the first preset number of historical nodes to form a local coherent sub-cluster.

5. The method according to claim 1, characterized in that, Calculate the confidence score for charging intention, including: Based on the difference between the phase angle of the current node and the average phase of the target high deviation network, the phase resonance score is calculated in combination with the target weighted synchronization degree. Based on the standardized deviation degree and the phase resonance score, the comprehensive anomaly score is calculated. The comprehensive anomaly score is modulated using the amplitude of the current node to obtain the modulated anomaly score. The modulated anomaly score is then mapped and transformed to obtain the charging intention confidence score.

6. The method according to claim 1, characterized in that, Based on the confidence score of the charging intention, corresponding countdown control commands are generated, including: If the confidence score of charging intention is greater than the preset high confidence threshold, the remaining time of the anti-occupancy countdown will be restored to the initial countdown time. If the confidence score of charging intention is greater than or equal to the preset low confidence threshold and less than or equal to the high confidence threshold, the remaining time of the anti-occupancy countdown will be extended by a preset penalty time based on the current value. If the confidence score of charging intention is less than the low confidence threshold, the remaining time of the anti-occupancy countdown will continue to decrease.

7. A digital twin-based integrated charging and stopping control system, used to implement the method described in any one of claims 1-6, characterized in that, The system includes: The feature acquisition module obtains the charging power sequence output by the charging pile, constructs a feature vector based on the charging power sequence and battery state parameters, and calculates the Euclidean norm of the feature vector to obtain the amplitude. The associated network construction module, when receiving a recharging request during the anti-occupancy countdown, defines the current charging process of the charging pile as a charging event, maps the current charging event to the current node, maps the historical charging events to historical nodes, maps the ratio of the remaining countdown time of the current node to the initial countdown time to the phase angle of the current node, and calculates the final phase coupling strength between the current node and the historical node based on the amplitude and phase angle of the current node and the historical node. The deviation calculation module obtains the phase angle distribution of normal charging events in historical nodes, calculates the standardized deviation degree based on the phase angle of the current node and the phase angle distribution, filters high deviation historical nodes from historical nodes, constructs a complex vector based on the amplitude and phase angle of the high deviation historical nodes, and calculates the weighted synchronization degree and the average phase of the high deviation network based on the final phase coupling strength and the complex vector. The decision mapping module calculates the charging intention confidence score based on weighted synchronization degree, high deviation network average phase and standardized deviation degree, generates corresponding countdown control commands based on the charging intention confidence score, and maps them to the digital twin model.

Images