A homomorphic encryption-based vehicle-network interaction data collaborative calculation privacy protection system

Abstract

This application relates to the field of homomorphic encryption technology, specifically to a privacy protection system for vehicle-to-grid interactive data collaborative computing based on homomorphic encryption. The system includes: a vehicle-side data encryption module for collecting real-time active power of vehicles and calculating its increment, generating a payload through homomorphic encryption, and encapsulating it with a grid command number into a packet; a cloud-based validity verification module for calculating oscillation based on the magnitude of grid command changes during the data packet lag period, generating a validity mask, and cleaning the encrypted payload using the homomorphic scalar multiplication property; and a decryption and power compensation adjustment module for aggregating and decrypting the cleaned payload to obtain the effective total power adjustment, while simultaneously statistically analyzing the sample effectiveness to implement linear gain compensation, thereby obtaining the compensation feedback power. This application solves the problem of difficulty in removing lagging reverse adjustment data in encrypted environments, improving the security and stability of grid power control.

Description

Technical Field

[0001] This application relates to the field of homomorphic encryption technology, specifically to a privacy protection system for collaborative computing of vehicle-to-everything (V2X) interactive data based on homomorphic encryption. Background Technology

[0002] Vehicle-to-grid (V2G) interaction refers to electric vehicles acting as distributed energy storage units interacting with the power grid to assist in peak shaving; homomorphic encryption is a cryptographic technique that allows specific algebraic operations to be performed directly on ciphertext data and the decryption result is consistent with the same operation performed on plaintext; in the V2G automatic generation control system, electric vehicles, as distributed energy storage units, need to adjust their charging and discharging power according to real-time instructions from the power grid dispatch center to maintain the stability of the power grid frequency.

[0003] Because vehicle terminals access cloud platforms via public networks, data transmission faces unpredictable random delays. While the platform can handle delays using timestamps in plaintext communication mode, in scenarios employing homomorphic encryption to protect user privacy, the cloud can only perform algebraic operations on the ciphertext and cannot decrypt it to obtain specific values ​​or directly determine the data's timeliness. If the grid regulation command reverses direction (e.g., from charging to discharging), delayed data packets, if still responding to old commands and mixed into aggregation calculations, will generate reverse regulation components. This not only negates effective regulation capabilities but may even cause control loop oscillations. In summary, existing technologies lack a mechanism to effectively identify and eliminate reverse regulation data caused by communication delays in a ciphertext computing environment without decrypting user data, thus reducing the security and stability of grid power control. Summary of the Invention

[0004] To address the aforementioned technical issues, the purpose of this application is to provide a privacy protection system for collaborative computing of vehicle-to-everything (V2X) interactive data based on homomorphic encryption. The specific technical solution adopted is as follows: This application proposes a privacy protection system for collaborative computing of vehicle-to-everything (V2X) interactive data based on homomorphic encryption, the system comprising: The vehicle-side data encryption module is used to acquire power grid commands and collect the active power of the vehicle terminal in real time; calculate the real-time active power increment of the vehicle terminal based on the real-time active power of the vehicle terminal; perform homomorphic encryption on the active power increment to generate an encrypted payload, and encapsulate the encrypted payload and its corresponding power grid command number into a data packet. The cloud-based validity verification module is used to calculate the oscillation degree of the power grid command based on the change amplitude of the power grid command within a preset lag period of the data packet, and to determine whether the data packet is valid; the module uses the scalar multiplication property of the homomorphic encryption algorithm and the validity mask to perform homomorphic projection operation on the encrypted payload in the data packet to clean the encrypted payload. The decryption and power compensation module is used to aggregate and decrypt the cleaned encrypted payload to obtain the effective total power; to count the number of effective data packets in each preset control cycle to calculate the sample effectiveness rate; and to use the sample effectiveness rate to perform linear gain compensation on the effective total power to obtain the compensation feedback power.

[0005] Preferably, the process of calculating the real-time active power increment of the vehicle terminal includes: The difference between the active power of the vehicle terminal at each time step and the active power of the vehicle terminal at the adjacent previous time step is taken as the active power increment of the vehicle terminal at each time step, wherein the active power increment at the initial time step is 0.

[0006] Preferably, the process of generating the encrypted payload is as follows: Multiply the active power increment of the vehicle terminal at each time point by a preset integerization factor and perform a rounding operation to obtain the integer field value of the active power increment. The integer field value is encrypted using the addition homomorphic encryption public key to generate the encrypted payload.

[0007] Preferably, the preset lag period is the time interval from the time the power grid command in the data packet is issued to the time the data packet is received by the cloud.

[0008] Preferably, the calculation process for the power grid command oscillation is as follows: Obtain the target power value of the entire network for each power grid command, calculate the sum of the absolute differences between the target power values ​​of the entire network of all adjacent power grid commands in the time sequence within the lag period, and record it as the cumulative adjustment range of the command. The power grid command includes two fields: power grid command number and target power value of the entire network. The absolute difference between the target power value of the initial grid command and the target power value of the final grid command during the lag period is recorded as the net adjustment range of the command. The sum of the net adjustment range of the command and the preset factor is calculated, and the ratio of the cumulative adjustment range of the command to the sum is taken as the power grid command oscillation degree, wherein the preset factor is a constant greater than 0.

[0009] Preferably, the process of determining whether a data packet is valid is as follows: If the current power grid command oscillation is less than or equal to the preset tolerance threshold, the data packet is valid; otherwise, the data packet is invalid.

[0010] Preferably, the process of cleaning the encrypted payload includes: Define homomorphic projection operation In the formula, This represents the cleaned and encrypted payload within the data packet; This indicates the encrypted payload within the data packet; This is the validity mask for the data packet; Represents homomorphic scalar multiplication; Among them, the validity mask when the data packet is valid. Set to 1, which is the validity mask when the data packet is invalid. Set to 0; when When the value is 1, the original encrypted payload is maintained; when... When the value is 0, the corresponding plaintext value in the encrypted payload is set to 0.

[0011] Preferably, the method for obtaining the effective total power adjustment is as follows: Perform a homomorphic summation operation on all cleaned encrypted payloads in all current data packets, and record the result as the total amount of regional encryption adjustment; The total amount of the region encryption adjustment is decrypted using the private key of the additive homomorphic encryption to obtain the plaintext and value; The result of dividing the plaintext sum by the preset integer multiplier is used as the effective total power for adjustment.

[0012] Preferably, the sample validity rate is the percentage of data packets with a validity mask of 1 received by the cloud within a preset control period in the total number of data packets.

[0013] Preferably, the compensation feedback power In the formula, This indicates that the total power is effectively regulated; This indicates the sample effectiveness rate; This represents the preset sample efficiency threshold; max() represents the maximum value function.

[0014] This application has the following beneficial effects: This application first obtains power grid commands and collects real-time active power of vehicles, calculates power increments to extract dynamic response components, and then uses homomorphic encryption technology to encapsulate the data for privacy. The encrypted payload is associated with the logically anchored power grid command number to form a data packet. This design not only effectively hides the user's electricity load characteristics and achieves end-to-end data confidentiality protection, but also gives the data packet the self-verification capability of logical causality by binding the command number instead of relying on unreliable physical timestamps. This ensures that even in a network environment with random latency, the cloud can still accurately trace the data status and provide a reliable basis for subsequent risk blocking and precise aggregation. Thus, while protecting privacy, it maintains the effectiveness and security of vehicle-to-grid interaction control. Furthermore, this application constructs the lag period of data packets and analyzes the geometric characteristics of the power grid command trajectory during this period. It calculates the command oscillation degree to quantify the consistency between the lag data and the current power grid demand. Then, it generates a validity mask based on the dynamic threshold mechanism of the sliding window. This module utilizes the scalar multiplication characteristics of homomorphic encryption to perform homomorphic projection cleaning on the load in the ciphertext domain according to the mask. It physically removes the reverse adjustment data without decrypting the user data, thereby effectively eliminating the risk of power grid frequency control caused by command flipping. While strictly protecting user privacy, it realizes the security and logical validity of power grid aggregation calculation. Ultimately, this application utilizes homomorphic encryption technology to securely aggregate and decrypt the cleaned ciphertext payload, obtaining the effectively regulated total power. Linear gain compensation is then applied to the power using statistical samples to generate the final feedback power. This module not only achieves accurate calculation of privacy data in the ciphertext domain, ensuring user data security, but also effectively solves the problem of low regional regulation capacity statistics caused by the removal of reverse risk data through the introduction of a compensation mechanism. This prevents over-regulation of the power grid while restoring the regulation capacity gain, thereby providing the power grid dispatch center with a safe, reliable, and statistically representative actual adjustable capacity. This completes the closed-loop control of vehicle-grid interaction, improving the security and stability of power grid control. Attached Figure Description

[0015] Figure 1 A block diagram of a vehicle-to-everything (V2X) interactive data collaborative computing privacy protection system based on homomorphic encryption, provided as an embodiment of this application; Figure 2 A flowchart illustrating the validity mask generation process provided in one embodiment of this application. Detailed Implementation

[0016] The following description, in conjunction with the accompanying drawings, details a specific scheme for a vehicle-to-everything (V2X) interactive data collaborative computing privacy protection system based on homomorphic encryption, as provided in this application.

[0017] Please see Figure 1 The diagram illustrates a privacy protection system for vehicle-to-everything (V2X) interactive data collaborative computing based on homomorphic encryption, according to an embodiment of this application. The system includes: a vehicle-side data encryption module 101, a cloud-based validity verification module 102, and a decryption and power compensation adjustment module 103.

[0018] The vehicle-side data encryption module 101 is used to acquire power grid commands and collect the active power of the vehicle terminal in real time; calculate the real-time active power increment of the vehicle terminal based on the real-time active power of the vehicle terminal; perform homomorphic encryption on the active power increment to generate an encrypted payload, and encapsulate the encrypted payload and the corresponding power grid command number into a data packet.

[0019] Because data transmission delays after vehicle terminals connect to the public network are random, physical timestamps are insufficient for the cloud to determine whether data conforms to the current power grid control logic. Therefore, a logical correspondence must be established between the data generation time and the power grid command issuance time. The real-time active power is read from the vehicle terminal metering device, i.e., the BMS. In this embodiment, the active power sampling frequency is set to 10Hz. In practical applications, implementers can set it according to specific circumstances; this embodiment does not impose special restrictions.

[0020] Furthermore, as a prerequisite for performing risk assessments, the cloud aggregation platform must establish a time baseline for backtracking historical states. The cloud platform allocates a fixed-length buffer in memory to store grid commands within the last 300 seconds, designated as the command history buffer pool. This buffer pool is used to store all Automatic Generation Control (AGC) command records issued by the grid dispatch center in real time. Each record contains two fields: ,in, Assign power grid command number, This represents the target power value for the entire network corresponding to the power grid command. The cloud platform updates this buffer pool at a rate no less than the sampling frequency of the vehicle terminal to ensure coverage of all possible sampling time points at the edge, avoiding trajectory reconstruction distortion due to undersampling. Furthermore, the vehicle terminal associates and stores the active power grid command in memory, completing the anchoring of physical data to logical commands.

[0021] To avoid directly uploading the total power of vehicle terminals and thus revealing users' basic electricity load characteristics, such as charging habits and battery capacity, and to reduce the magnitude of encrypted data, the absolute power value is not transmitted directly. Instead, the real-time active power increment of the vehicle terminals is calculated based on the real-time active power of the vehicle terminals. This allows the extraction of dynamic response components reflecting the grid's regulation needs. Specifically: The difference between the active power of the vehicle terminal at each time step and the active power of the vehicle terminal at the adjacent previous time step is taken as the active power increment of the vehicle terminal at each time step, wherein the active power increment at the initial time step is 0.

[0022] Furthermore, to ensure the confidentiality of active power increment data transmitted over untrusted public networks, and to meet the cloud's requirement for subsequent aggregation calculations in an encrypted state, this embodiment generates an encrypted payload by homomorphically encrypting the active power increment, and encapsulates the encrypted payload and the corresponding grid command number into a data packet. Specifically: Since the general additive homomorphic encryption algorithm is defined on the integer ring, it cannot directly handle physical power values ​​containing decimals. In order to preserve calculation precision, floating-point data needs to be extended. In this embodiment, the active power increment of the vehicle terminal at each time step is multiplied by a preset integerization factor and rounded to obtain the integer field value of the active power increment. The integer field value is encrypted using the additive homomorphic encryption public key to generate the encrypted payload. In this embodiment, the additive homomorphic encryption algorithm adopts the Paillier algorithm. Assuming the active power is 1.5KW, 1.5KW is multiplied by 1000 to convert it to 1500W. At this time, the value of the preset integerization factor is 1000. Therefore, the value of the integerization factor is determined by the power unit.

[0023] The process of using homomorphic encryption public keys to encrypt data is a well-known technique and will not be elaborated further.

[0024] Furthermore, in order for the cloud platform to still assess the timeliness of data and aggregate it by region even when the payload cannot be decrypted, the encrypted payload and the associated plaintext index credentials must be structurally encapsulated. Specifically: in: This represents the encrypted payload, used for subsequent ciphertext calculations; This represents the grid command number corresponding to the loaded load, serving as a logical key value for tracing historical command trajectories in the cloud. This indicates the timestamp that corresponds to the active power increment and is only used for statistical analysis of network transmission delay, not as a basis for control logic judgment. This is the identifier for the communication subdomain to which the vehicle terminal belongs, used for domain aggregation in the cloud.

[0025] The constructed data packet is sent to the cloud aggregation platform. Thereafter, the data packet is transmitted in the network as an independent unit of information. Even if out-of-order delivery or delays occur, the information it carries will remain intact. It always maintains a causal link to the state of the power grid at the time the data was generated.

[0026] Thus, this embodiment obtains power grid commands and collects real-time active power of vehicles, calculates power increments to extract dynamic response components, and then uses homomorphic encryption technology to encapsulate the data for privacy. The encrypted payload is associated with the logically anchored power grid command number to form a data packet. This design not only effectively hides the user's electricity load characteristics and achieves end-to-end data confidentiality protection, but also gives the data packet the self-verification capability of logical causality by binding the command number instead of relying on unreliable physical timestamps. This ensures that even in a network environment with random latency, the cloud can still accurately trace the data status and provide a reliable basis for subsequent risk blocking and precise aggregation. Thus, while protecting privacy, it maintains the effectiveness and security of vehicle-to-grid interaction control.

[0027] The cloud-based validity verification module 102 is used to calculate the oscillation degree of the power grid command based on the change amplitude of the power grid command within a preset lag period of the data packet, and to determine whether the data packet is valid; it uses the scalar multiplication characteristics of the homomorphic encryption algorithm and the validity mask to perform homomorphic projection operation on the encrypted payload in the data packet to clean the encrypted payload.

[0028] Due to the random delays of public networks, the arrival time of data packets will inevitably lag behind their generation time. In order to assess the changes in power grid demand during this lag period, this embodiment uses the time interval from the time the power grid command is issued to the time the data packet is received by the cloud as the lag period of the data packet. The lag period includes the time the power grid command is issued and the time the data packet is received by the cloud. If the data packet lag time exceeds the depth of the buffer pool, the data packet is determined to be invalid and discarded.

[0029] Due to network transmission delays, the grid's regulation demand may have changed significantly, or even reversed, during the lag period. Blindly using lagging historical data for current grid regulation could lead to a reverse regulation effect—that is, increasing power generation when the grid needs to reduce its load, and vice versa, thus exacerbating grid fluctuations rather than mitigating them. Therefore, this embodiment analyzes the magnitude of changes in grid commands during the lag period, calculates the grid command oscillation, and assesses the consistency between the lagging data and current grid demand. Specifically: The sum of the absolute differences between the target power values ​​of all adjacent grid commands in time sequence within the lag period is calculated and denoted as the cumulative adjustment range of the command. The grid command includes two fields: grid command number and target power value of the whole grid. The absolute difference between the target power value of the initial grid command and the target power value of the final grid command during the lag period is recorded as the net adjustment range of the command. Furthermore, the sum of the net adjustment amplitude of the command and the preset factor is calculated, and the ratio of the cumulative adjustment amplitude of the command to the sum is used as the power grid command oscillation degree, wherein the preset factor is a constant greater than 0.

[0030] It should be noted that the preset factor is set manually; in this embodiment, the preset factor is set to... Provided that the denominator is not zero and does not excessively affect the calculation result, the implementer may set it according to the specific situation. This embodiment does not impose any special restrictions.

[0031] Based on the grid command oscillation, it can be understood that the grid command oscillation quantifies the geometric tortuosity of the grid command trajectory within the lag period, representing the proportion of invalid fluctuations relative to effective displacements in the grid command change path. In other words, it reflects the severity of directional reversals or repeated oscillations in grid regulation demand during transmission. The calculation of grid command oscillation is mainly affected by two factors: the cumulative adjustment amplitude and the net adjustment amplitude. A larger cumulative adjustment amplitude and a smaller net adjustment amplitude result in a larger grid command oscillation value, reflecting severe back-and-forth oscillations or directional reversals in the grid command within the lag period. This means that the lag data uploaded by vehicles is highly likely to be contrary to the current grid control logic. Directly using this data would trigger serious reverse regulation risks, exacerbating grid fluctuations and even causing control oscillations. Conversely, a smaller cumulative adjustment amplitude and a larger net adjustment amplitude result in a smaller grid command oscillation, reflecting a relatively stable and monotonic command change path. This means that the adjustment direction of the lag data is basically consistent with the current grid demand. Although the data is late, the logic remains valid, and it can safely participate in aggregation calculations without negatively impacting system stability.

[0032] After quantifying the hysteresis risk characteristics of a single data packet, i.e., the oscillation degree of power grid commands, through the above steps, the next step is to address how to set reasonable judgment criteria to distinguish between "valid data" and "risky data." If a fixed threshold is used, when the power grid is in a stable adjustment phase, an overly strict threshold may lead to the misjudgment and discarding of a large amount of usable data; while when the power grid is in a high-frequency and violent fluctuation phase, an overly wide threshold may introduce "reverse data" that seriously interferes with the adjustment. In order to solve the problem that a single static threshold cannot adapt to dynamic and time-varying operating conditions, this embodiment introduces a dynamic threshold mechanism based on a sliding time window. This allows the system to automatically tighten the filtering criteria when facing network congestion or rapid fluctuations in power grid commands, while appropriately relaxing the restrictions during normal operation. This maximizes the utilization of valid data resources while ensuring the safety of regulation. Specifically: The cloud platform maintains a length of In this embodiment, the value of W is 100, which is used to store the data packets processed beforehand. In this embodiment, the 90th percentile of the power grid command oscillation is used as the tolerance threshold for each data packet.

[0033] It should be noted that this embodiment uses the 90th percentile as the tolerance threshold to construct an adaptive filtering mechanism that "removes significant outliers." In a dynamically changing network and power grid environment, only the 10% of extreme abnormal data with the highest oscillations are identified as high-risk and blocked. This achieves a balance between preventing reverse regulation interference and maximizing the utilization of effective data resources, avoiding both resource waste caused by overly strict thresholds and risks introduced by overly broad thresholds. This value is not fixed and can be adjusted according to the different emphases on security and data utilization in specific application scenarios. For example, in scenarios with extremely high requirements for power grid stability, it can be increased to 95% to strengthen filtering, while in scenarios with poor network environment or where regulation potential needs to be explored, it can be reduced to 80% or 85% to reduce false judgments.

[0034] Furthermore, the power grid command oscillation of each data packet is compared with the corresponding tolerance threshold to generate a validity mask: Validity mask for data packet n. The expression is: In the formula, This represents the power grid command oscillation degree of data packet n; This represents the tolerance threshold for data packet n, where This indicates that the data is within the acceptable range and will be retained as a valid data packet; This indicates that the data has a high risk of being reversed, is an invalid data packet, and needs to be blocked.

[0035] Preferably, the flowchart of the validity mask generation process provided in this embodiment is as follows: Figure 2 As shown.

[0036] Furthermore, in order to avoid decrypting the encrypted payload Under the premise of physically removing high-risk data, logical blocking is performed using the scalar multiplication property (Ciphertext-Plaintext Multiplication) of the additive homomorphic encryption algorithm. That is, the scalar multiplication property and validity mask of the homomorphic encryption algorithm are used to perform homomorphic projection operation on the encrypted payload in the data packet to clean the encrypted payload. Specifically: Define homomorphic projection operation In the formula, This represents the cleaned and encrypted payload within the data packet; This indicates the encrypted payload within the data packet; This is the validity mask for the data packet; Represents homomorphic scalar multiplication; when When the value is 0, the corresponding plaintext value in the encrypted payload is set to 0; when When the value is 1, the original encrypted payload is maintained. This operation achieves physical zeroing of the reverse adjustment data in the ciphertext state without leaking any numerical information.

[0037] Thus, this embodiment constructs the lag period of data packets and analyzes the geometric characteristics of the power grid command trajectory during this period. It calculates the command oscillation degree to quantify the consistency between the lag data and the current power grid demand. Then, it generates a validity mask based on the dynamic threshold mechanism of the sliding window. This module utilizes the scalar multiplication characteristics of homomorphic encryption to perform homomorphic projection cleaning on the load in the ciphertext domain according to the mask. It physically removes the reverse adjustment data without decrypting the user data, thereby effectively eliminating the risk of power grid frequency control caused by command flipping. While strictly protecting user privacy, it achieves the security and logical validity of power grid aggregation calculation.

[0038] The decryption and adjustment power compensation module 103 is used to aggregate and decrypt the cleaned encrypted payload to obtain the effective total power; count the number of effective data packets in each preset control cycle to calculate the sample effectiveness rate; and use the sample effectiveness rate to perform linear gain compensation on the effective total power to obtain the compensation feedback power.

[0039] After verifying and cleaning the validity of individual data packets, the core task in the cloud is to aggregate vehicle response data and transform it into control commands usable by the power grid dispatch. However, direct aggregation faces two major challenges: First, how to perform efficient computation on the cleaned ciphertext data while ensuring data privacy. Homomorphic encryption can be used to convert invalid data into zero values, allowing for lossless accumulation of valid components directly in the ciphertext domain. Second, how to address the issue of low regional regulation capacity statistics caused by the removal of reverse risk data during cleaning. If the aggregated results after cleaning are directly reported, the power grid cannot perceive the actual hardware regulation capabilities of the removed vehicles, potentially leading to control loop gain attenuation and over-regulation. Therefore, this embodiment aggregates and decrypts the cleaned encrypted payload to obtain the total effective regulation power; counts the number of data packets with a validity mask of 1 within each preset control cycle to calculate the sample effectiveness rate; and uses the sample effectiveness rate to perform linear gain compensation on the total effective regulation power to obtain the compensation feedback power. The specific process is as follows: First, the cleaned encrypted payloads are aggregated and decrypted to obtain the effective total adjustment power. Specifically, a homomorphic accumulation operation is performed on all cleaned encrypted payloads in all current data packets, and the result is recorded as the total regional encryption adjustment amount. Furthermore, the total amount of regional encryption adjustment is decrypted using the additive homomorphic encryption private key to obtain the plaintext and value; The result of dividing the plaintext sum by the preset integerization factor is taken as the effective total power, and its unit is KW.

[0040] The process of homomorphic accumulation and the process of decryption using the homomorphic encryption private key are well-known techniques and will not be elaborated further.

[0041] The effective total regulation power can be understood as the sum of active power obtained by homomorphically accumulating and decrypting the regulation capabilities of all vehicle terminals aligned with the grid command direction after high-risk data has been cleaned and removed by the cloud-based validity verification module. It characterizes the scale of safe and phase-aligned effective response resources actually obtained by the regional grid within a specific control cycle. The calculation of the effective total regulation power is mainly affected by the sum of active power corresponding to the cleaned and encrypted loads and the preset integer multiplier. A larger sum of active power corresponding to the cleaned and encrypted loads indicates a larger effective total regulation power, reflecting a larger number of participating vehicle terminals with high response output, indicating sufficient regional regulation resources to strongly support the frequency stability regulation of the grid. Conversely, a smaller sum of cleaned and encrypted loads indicates a smaller effective total regulation power, reflecting a large amount of data being removed due to lag risk or low actual vehicle response output, resulting in a scarcity of effective regional regulation resources. Directly feeding back this value would lead the grid dispatch center to underestimate the regulation potential of the region, potentially causing over-regulation or resource shortages in other regions.

[0042] Furthermore, because module 102 physically removes invalid encrypted payloads, the number of valid encrypted payload samples is less than the total number of encrypted payload samples. Direct feedback would lead to an underestimation of the regional regulation capacity's statistical value, causing the power grid AGC controller to misjudge insufficient regulation resources and trigger over-regulation. Therefore, it is necessary to statistically reconstruct the removed capacity, that is, to calculate the sample effectiveness by counting the number of data packets with an effectiveness mask of 1 in each preset control cycle. Specifically: The percentage of data packets with a validity mask of 1 received by the cloud within a preset control period is taken as the sample validity rate. In this embodiment, the preset control period is 4 seconds because the standard regulation commands of the automatic generation control (AGC) system are usually issued in cycles of several seconds, such as the common 4 seconds or 16 seconds. Aligning the control period to 4 seconds is sufficient to collect a sufficient number of vehicle terminal response data packets for aggregation calculation, ensuring the smoothness and representativeness of the calculation results, and minimizing the data processing delay. This ensures that the regulation capacity information fed back to the power grid has high timeliness, thereby meeting the real-time requirements of the power grid's rapid frequency response. In actual applications, implementers can also set it according to specific circumstances. This embodiment does not impose any special restrictions.

[0043] Furthermore, this embodiment utilizes the sample effectiveness rate to perform linear gain compensation on the total effective regulation power to obtain the compensation feedback power. Specifically: Compensation feedback power In the formula, This indicates that the total power is effectively regulated; This indicates the sample effectiveness rate; This represents the preset sample efficiency threshold; max() represents the maximum value function.

[0044] It should be noted that the preset sample effectiveness threshold is set to 0.5. This is so that when the proportion of effective samples is less than 50%, the denominator in the compensation formula will be locked at 0.5, that is, the maximum compensation ratio is limited to 2 times. This is done to prevent excessive estimation errors or severe oscillations in the system output caused by the compensation ratio approaching infinity in extreme conditions such as extreme network congestion or severe fluctuations in power grid commands, which would result in very few effective samples. This ensures that while the system recovers some of the lost capacity through compensation, it will not feed back severely distorted false regulation capabilities to the power grid due to excessive inference amplification, thereby maintaining the overall stability of the control system.

[0045] Ultimately, the compensation feedback power will be... The actual adjustable capacity is fed back to the power grid dispatch center to complete closed-loop control. Through this compensation mechanism, the system recovers the adjustable capacity lost due to data cleaning while ensuring the correct adjustment direction, and maintains the stability of the control loop gain.

[0046] Thus, this embodiment utilizes homomorphic encryption technology to securely aggregate and decrypt the cleaned ciphertext payload, obtaining the effectively regulated total power. Linear gain compensation is then applied to the power using statistical samples to generate the final feedback power. This module not only achieves accurate calculation of privacy data in the ciphertext domain, ensuring user data security, but also effectively solves the problem of low regional regulation capacity statistics caused by the removal of reverse risk data through the introduction of a compensation mechanism. This prevents over-regulation of the power grid while restoring the regulation capacity gain, thereby providing the power grid dispatch center with a safe, reliable, and statistically representative actual adjustable capacity, completing the closed-loop control of vehicle-grid interaction.

[0047] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the principles of this application should be included within the protection scope of this application.

Claims

1. A privacy protection system for vehicle-to-everything (V2X) interactive data collaborative computing based on homomorphic encryption, characterized in that, The system includes: The vehicle-side data encryption module is used to acquire power grid commands and collect the active power of the vehicle terminal in real time; calculate the real-time active power increment of the vehicle terminal based on the real-time active power of the vehicle terminal; perform homomorphic encryption on the active power increment to generate an encrypted payload, and encapsulate the encrypted payload and its corresponding power grid command number into a data packet. The cloud-based validity verification module is used to calculate the oscillation degree of the power grid command based on the change amplitude of the power grid command within a preset lag period of the data packet, and to determine whether the data packet is valid; the module uses the scalar multiplication property of the homomorphic encryption algorithm and the validity mask to perform homomorphic projection operation on the encrypted payload in the data packet to clean the encrypted payload. The decryption and power compensation module is used to aggregate and decrypt the cleaned encrypted payload to obtain the effective total power; to count the number of effective data packets in each preset control cycle to calculate the sample effectiveness rate; and to use the sample effectiveness rate to perform linear gain compensation on the effective total power to obtain the compensation feedback power.

2. The privacy protection system for vehicle-to-everything (V2X) interactive data collaborative computing based on homomorphic encryption according to claim 1, characterized in that, The process of calculating the real-time active power increment of the vehicle terminal includes: The difference between the active power of the vehicle terminal at each time step and the active power of the vehicle terminal at the adjacent previous time step is taken as the active power increment of the vehicle terminal at each time step, wherein the active power increment at the initial time step is 0.

3. The privacy protection system for vehicle-to-everything (V2X) interactive data collaborative computing based on homomorphic encryption according to claim 1, characterized in that, The process of generating the encrypted payload is as follows: Multiply the active power increment of the vehicle terminal at each time point by a preset integerization factor and perform a rounding operation to obtain the integer field value of the active power increment. The integer field value is encrypted using the addition homomorphic encryption public key to generate the encrypted payload.

4. The privacy protection system for vehicle-to-everything (V2X) interactive data collaborative computing based on homomorphic encryption according to claim 1, characterized in that, The preset lag period is the time interval from the moment the power grid command in the data packet is issued to the moment the data packet is received by the cloud.

5. A privacy protection system for vehicle-to-everything (V2X) interactive data collaborative computing based on homomorphic encryption as described in claim 1, characterized in that, The calculation process for the power grid command oscillation is as follows: Obtain the target power value of the entire network for each power grid command, calculate the sum of the absolute differences between the target power values ​​of the entire network of all adjacent power grid commands in the time sequence within the lag period, and record it as the cumulative adjustment range of the command. The power grid command includes two fields: power grid command number and target power value of the entire network. The absolute difference between the target power value of the initial grid command and the target power value of the final grid command during the lag period is recorded as the net adjustment range of the command. The sum of the net adjustment range of the command and the preset factor is calculated, and the ratio of the cumulative adjustment range of the command to the sum is taken as the power grid command oscillation degree, wherein the preset factor is a constant greater than 0.

6. The privacy protection system for vehicle-to-everything (V2X) interactive data collaborative computing based on homomorphic encryption according to claim 1, characterized in that, The process for determining whether a data packet is valid is as follows: If the current power grid command oscillation is less than or equal to the preset tolerance threshold, the data packet is valid; otherwise, the data packet is invalid.

7. A privacy protection system for vehicle-to-everything (V2X) interactive data collaborative computing based on homomorphic encryption as described in claim 6, characterized in that, The process of cleaning the encrypted payload includes: Define homomorphic projection operation In the formula, This represents the cleaned and encrypted payload within the data packet; This indicates the encrypted payload within the data packet; This is the validity mask for the data packet; Represents homomorphic scalar multiplication; Among them, the validity mask when the data packet is valid. Set to 1, which is the validity mask when the data packet is invalid. Set to 0; when When the value is 1, the original encrypted payload is maintained; when When the value is 0, the corresponding plaintext value in the encrypted payload is set to 0.

8. A privacy protection system for vehicle-to-everything (V2X) interactive data collaborative computing based on homomorphic encryption as described in claim 1, characterized in that, The method for obtaining the effective total power adjustment is as follows: Perform a homomorphic summation operation on all cleaned encrypted payloads in all current data packets, and record the result as the total amount of regional encryption adjustment; The total amount of the region encryption adjustment is decrypted using the private key of the additive homomorphic encryption to obtain the plaintext and value; The result of dividing the plaintext sum by the preset integer multiplier is used as the effective total power for adjustment.

9. A privacy protection system for vehicle-to-everything (V2X) interactive data collaborative computing based on homomorphic encryption as described in claim 1, characterized in that, The sample validity rate is the percentage of data packets with a validity mask of 1 received by the cloud within a preset control period in the total number of data packets.

10. A privacy protection system for vehicle-to-everything (V2X) interactive data collaborative computing based on homomorphic encryption as described in claim 1, characterized in that, Compensation feedback power In the formula, This indicates that the total power is effectively regulated; This indicates the sample effectiveness rate; This represents the preset sample efficiency threshold; max() represents the maximum value function.

Images