Methods for protecting the privacy of integrated energy systems in multiple parks

By employing localized differential privacy processing and the Laplace mechanism in a multi-park integrated energy system, perturbation data is generated for optimized scheduling, solving the problem of privacy data leakage in existing technologies and achieving stronger privacy protection.

CN117454423BActive Publication Date: 2026-11-17EAST CHINA ELECTRIC POWER TEST & RES INST +2
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202311504573.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-11-13
Publication Date
2026-11-17
Estimated Expiration
2043-11-13

AI Technical Summary

Technical Problem

Existing multi-park integrated energy systems are unable to effectively defend against external eavesdropping and semi-honest attacks during boundary information exchange, leading to the leakage of privacy data.

Method used

Localized differential privacy processing is used to generate disturbance data. By constructing a low-carbon collaborative scheduling model for a multi-park integrated energy system, load demand curves are generated and privacy protection is implemented under the Laplace mechanism. The scheduling strategy is optimized to resist eavesdropping and speculation.

Benefits of technology

This system protects the privacy of park parameters and user behavior during boundary information exchange, prevents external eavesdropping and third-party speculation, and enhances the system's privacy protection capabilities.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117454423B_ABST
    Figure CN117454423B_ABST
Patent Text Reader

Abstract

A method for protecting the privacy of entities in a multi-park integrated energy system is proposed. This method constructs a low-carbon collaborative scheduling model for the multi-park integrated energy system. Based on the load demand curve generated by the model, localized differential privacy processing under the Laplace mechanism is applied. Then, based on the energy purchase strategies returned by each park, optimized scheduling is performed to obtain the energy sales strategy, and the residuals are calculated. When the residuals are less than a threshold, the privacy-preserving optimized scheduling strategy for each participating entity is obtained. This invention generates perturbation data through localized differential privacy processing, preventing external eavesdroppers from obtaining accurate data while also preventing third parties from inferring park parameters and user behavior based on the data.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to a technology in the field of information security, specifically a method for protecting the privacy of entities in a multi-park integrated energy system. Background Technology

[0002] The coupling between various existing energy networks is becoming increasingly tight, forming integrated energy systems containing multiple energy networks. In multi-park integrated energy systems, each park belongs to different stakeholders, and the privacy data of each park entity is not disclosed to other park entities. Therefore, only distributed collaborative scheduling can achieve economic scheduling of multi-park integrated energy systems. Most existing multi-park integrated energy systems adopt distributed algorithms such as the Alternating Direction Multiplier Method (ADMM), which achieves collaborative scheduling of multi-park integrated energy systems through the interaction of boundary variables and gradient information between agents. This method decomposes the problem into a two-level form, and through information iteration between the upper and lower levels, enables parks to exchange boundary information with third-party means, achieving the goal of optimal scheduling solution through multiple iterations.

[0003] However, existing technologies cannot defend against external eavesdroppers who obtain agent privacy data through eavesdropping and use methods such as fiber optic eavesdropping to obtain agent privacy. They also cannot defend against semi-honest attackers, namely the third-party agent itself. Although the third-party agent strictly follows the process and performs its own tasks during system operation, it will try to infer user behavior based on the information it receives, such as power generation cost parameters and user electricity consumption habits. Summary of the Invention

[0004] This invention addresses the issue that existing technologies can only achieve partial privacy protection, and that boundary information interaction may still lead to privacy leaks. It proposes a method for protecting the main privacy of a multi-park integrated energy system. By generating perturbation data through localized differential privacy processing, it ensures that external eavesdroppers cannot eavesdrop on accurate data, while third parties cannot infer park parameters and user behavior based on the data.

[0005] This invention is achieved through the following technical solution:

[0006] This invention relates to a method for protecting the privacy of entities in a multi-park integrated energy system. By constructing a low-carbon collaborative scheduling model for the multi-park integrated energy system, and performing localized differential privacy processing under the Laplace mechanism on the load demand curve generated by the model, the energy sales strategy is obtained by optimizing the scheduling based on the energy purchase strategy returned by each park, and the residual is calculated. When the residual is less than a threshold, the privacy-protected optimized scheduling strategy for each participating entity is obtained.

[0007] This invention relates to a system for implementing the above method, comprising: an integrated energy park service provider unit, park units, and a localized differential privacy unit, wherein: the integrated energy park service provider unit performs scheduling processing based on the load demand information after localized differential privacy to obtain scheduling instruction results for each park; the park units perform intra-park scheduling processing based on the scheduling instruction information to obtain scheduling and load demand results; and the localized differential privacy unit performs localized differential privacy processing based on the load demand information of each park to obtain the load demand results of each park after localized differential privacy, and uploads them to the integrated energy park service provider unit. Attached Figure Description

[0008] Figure 1 A schematic diagram of a collaborative scheduling model for a multi-park integrated energy system applying differential privacy;

[0009] Figure 2 This is a flowchart of the present invention. Detailed Implementation

[0010] like Figure 1 As shown, the multi-park integrated energy system collaborative scheduling model constructed in this embodiment includes: a park module, a load demand module, a localized differential privacy module, and an integrated energy park service provider module. The integrated energy park service provider module issues scheduling instructions to each park module, and the park modules return their respective load demands. The localized differential privacy module protects the privacy of the load demands of each park module by adding noise to the data before returning their respective load demands to the integrated energy park service provider module.

[0011] like Figure 2 As shown, this embodiment illustrates the subject privacy protection method based on the aforementioned multi-park integrated energy system collaborative scheduling model, which includes:

[0012] Step 1) Construct and initialize model parameters, specifically including:

[0013] 1.1) Taking the minimum total economic operating cost of the integrated energy system in each park as the objective function, the total economic operating cost function of the integrated energy system in multiple parks is defined. Wherein: S T For a time period set; S P Let f(j,t) be the set of parks, including park 1, park 2, and park 3; f(j,t) is the operating cost function of park j in the t-th time period.

[0014] 1.2) The operating cost function of park j in the t-th time period is f(j,t)=f1(j,t)+f2(j,t), where: equipment operating cost function f1(j,t) and carbon trading cost function f2(j,t).

[0015] 1.3) Set the initial energy sales power and energy price for the park service providers, and the initial equipment parameters and load parameters for each park, and set the initial iteration number k=1;

[0016] Step 2) Set model constraints, including:

[0017] a) Constraints on energy trading between industrial parks in: These represent the imported electrical power, thermal power, and gas power of park j during the t-th time period (a negative value indicates the exported power).

[0018] b) Internal constraints of the park, including:

[0019] i) System-level constraints, namely, the supply and demand balance of electricity, heat, and gas:

[0020] Wherein: S N For a collection of devices; P g (i,t) represents the gas supply of device i during the t-th time period; P eL (i,t), P hL (i,t), P gl (i,t) represent the electricity, heat, and gas consumption (load) of device i in the t-th time period, respectively; P EL (t), P HL (t), P GL (t) represents the electrical load, heat load, and gas load of the park during the t-th time period, respectively.

[0021] ii) Equipment-level constraints, including: energy production and consumption equipment and energy conversion equipment, wherein:

[0022] The energy generation and consumption equipment includes:

[0023] 1) Carbon Capture Power Plant (CCPP): Where: P(i,t) is the total power of device i in the t-th time period; P CO2 (i,t) represents the carbon capture power of carbon capture power plant i during the t-th time period; P base (i) represents the fixed power loss of device i during the t-th time period; ω i(CCPP) Let be the carbon capture operation energy consumption coefficient of carbon capture power plant i; Let be the amount of carbon captured by carbon capture power plant i in the t-th time period; β be the carbon capture efficiency. A collection of equipment that generates CO2 emissions, including coal-fired power plants and combined heat and power (CHP) units; E CO2 (i,t) represents the CO2 emissions of device i during the t-th time period; P min (i), P max(i) represents the minimum and maximum output power of device i; c coal Cost of coal per unit of electricity generated.

[0024] 2) Wind farm WT: Where: P WT (i,t) represents the actual wind power generated by wind farm i during the t-th time period, P qf (i,t) represents the wind curtailment of wind farm i during the t-th time period.

[0025] The energy conversion device includes:

[0026] 1) Operating constraints of the electro-gas conversion (P2G): The CO2 feedstock used in this P2G equipment is all derived from carbon capture, specifically: Where: η h,i(P2G) The efficiency of the electro-gas conversion equipment in converting electrical energy into heat energy; η g,i(P2G) The efficiency of the electro-gas conversion equipment (i) in converting electrical energy into gas energy; P eL,max (i) represents the maximum power consumption of device i; ω CO2,i(P2G) is the CO2 consumption coefficient of the electro-gas conversion device i.

[0027] 2) Combined Heat and Power (CHP) system: The industrial park adopts a CHP unit model with adjustable heat-to-power ratio, specifically: Where: η i(CHP) The overall efficiency of the combined heat and power (CHP) equipment i; κ min,i(CHP) κ max,i(CHP) These represent the upper and lower limits of the heat-to-power ratio for cogeneration equipment i; P gL,max (i) represents the maximum air load of device i.

[0028] 3) Energy purchase: f(i,t)=c ebuy P e (i,t)+c hbuy P h (i,t)(2.15), where: c ebuy For electricity price; c gbuy This refers to the price of gas.

[0029] 4) Energy Storage: A unified model is used for electric, thermal, and gas energy storage devices. Here, we take electric energy storage devices as an example; thermal and hydrogen energy storage devices are similar. Specifically: Where: P cha (i,t), P dis (i,t) represent the charging and discharging power of energy storage device i during the t-th time period; B cha (i,t), B dis (i,t) are both binary variables, representing the charging and discharging state parameters of energy storage device i in the t-th time period, respectively. cha (i,t)=1、B dis(i,t) = 0 indicates that the device is in a charging state, B cha (i,t)=0、B dis (i,t) = 1 indicates that the state is in a state of energy release; η cha,i(ES) η dis,i(ES) Let Si and St be the charging and discharging efficiencies of energy storage device i, respectively; S(i,t) is the percentage of capacity of energy storage device i in the t-th time period; Pi S (i) represents the rated capacity of energy storage device i; S min (i), S max (i) represents the upper and lower limits of the capacity percentage of energy storage device i.

[0030] Step 3) Use an optimized scheduling algorithm to solve the optimized scheduling problem for each park according to the scheduling instructions, and obtain the energy purchase strategy for each park, specifically including:

[0031] 3.1) Decouple the multi-park integrated energy system from the internal parks by adding a replication variable for inlet and outlet power. The original variable with superscript I is optimized by the integrated energy supplier, and a replication variable with superscript P is introduced and optimized by the service providers of each park. The original inlet and outlet power in the model... The power output of electricity, heat, and gas sold by the integrated energy supplier to each park is transformed into a newly introduced replication variable with the superscript P. Let's define the power output of the park's purchases of electricity, heat, and gas from integrated energy suppliers; then, we add constraints on the replication variables to ensure that the power output sold equals the power output purchased.

[0032] 3.2) Apply the augmented Lagrangian function to relax the constraint back to the original objective function:

[0033] Where: P I The decision variables representing the optimization of integrated energy suppliers include and j∈S P , t∈S T ;P P Decision variables optimized for park service providers include and j∈S P , t∈S T ;λ=[λ e ,λ h ,λ g ] represents the corresponding Lagrange multiplier; ρ is the penalty factor.

[0034] 3.3) Based on the participating entities, the problem is decomposed into lower-level sub-problems and upper-level problems. The j-th sub-problem is defined as follows: PThe objective functions for each sub-problem solved by the park service provider are as follows:

[0035] 3.4) Based on the optimization results of each park service provider, the decision variables for optimizing the park service providers are derived: Then update the Lagrange multipliers λ based on the decision variable results obtained from solving the upper and lower level problems. k+1 =λ k +ρ(P I,k+1 -P P,k+1 ).

[0036] Step 4) Perform localized differential privacy processing on the load demand of each park to obtain the park energy purchase strategy after localized differential privacy processing, so that for each park, the load demand vector X P Load demand vector after localized differential privacy The indistinguishability is α; the optimal scheduling cost obtained after using localized differential privacy data satisfies... Specifically: for adjacent datasets D~ α D' and the query function f, The differential privacy output under the Laplace mechanism is M(D) = f(D) + r, where f(D) and f(D') satisfy the indistinguishability requirement. r is noise, which follows a Laplace distribution. ε represents the privacy budget. Generally speaking, the smaller the privacy budget, the greater the noise, the lower the usability of the results, and the better the privacy protection; f i * Let P represent the optimal scheduling cost function for park i. i P This is the original load demand vector of park i. It is the load demand vector after localized differential privacy in the park.

[0037] Step 5) The park service provider optimizes the scheduling based on the energy purchase strategies returned by each park to obtain the energy sales strategy. Specifically, based on the solution results of the optimized decision variables by the park service provider, the upper-level problem of the integrated energy service provider is solved: Based on the optimization solution results of the integrated energy service provider, the decision variables for the integrated energy service provider's optimization are derived:

[0038] Step 6) The park service provider calculates the residual r, and outputs the scheduling results of each participating entity when the residual is less than the threshold; otherwise, it updates the energy price λ. k+1 =λ k +ρ(P I,k+1 -P P,k+1 Then return to step 3) and iterate again, specifically: |r k+1||2=||P I,k+1 -P P,k+1 ||2.

[0039] Compared to existing ADMM distributed scheduling solutions, this method can protect the privacy of entities in each park of a multi-park integrated energy system to a certain extent. However, this method involves direct interaction of marginal information, which may still lead to privacy leaks.

[0040] Unlike existing studies that generally do not include dedicated privacy protection, this invention adds localized differential privacy protection based on the load requirements of each park, in order to meet the privacy protection needs of the main entities in each park.

[0041] The above-described specific implementations can be partially adjusted by those skilled in the art in different ways without departing from the principles and purpose of the present invention. The scope of protection of the present invention is defined by the claims and is not limited to the above-described specific implementations. All implementation schemes within the scope of the claims are bound by the present invention.

Claims

1. A method for protecting the privacy of entities in a multi-park integrated energy system, characterized in that, By constructing a low-carbon collaborative scheduling model for a multi-park integrated energy system, and after localized differential privacy processing under the Laplace mechanism based on the load demand curve generated by the model, the energy sales strategy is obtained by optimizing the scheduling based on the energy purchase strategy returned by each park, and the residual is calculated. When the residual is less than a threshold, the privacy-preserving optimized scheduling strategy for each participating entity is obtained, which specifically includes: Step 1) Build the model and initialize the model parameters, specifically including: 1.1) Taking the minimum total economic operating cost of the integrated energy system in each park as the objective function, the total economic operating cost function of the integrated energy system in multiple parks is defined. ,in: For time periods; This is a collection of parks, including Park 1, Park 2, and Park 3; For the park In the Operating cost function over a given time period; 1.2) Park In the Operating cost function within a time period Among them: equipment operating cost function Carbon trading cost function ; 1.3) Set the initial energy sales capacity and energy price for the park service providers, and the initial equipment parameters and load parameters for each park, and set the initial number of iterations. ; Step 2) Set model constraints; Step 3) Use an optimized scheduling algorithm to optimize the scheduling of each park according to the scheduling instructions, and obtain the energy purchase strategy of each park; Step 4) Perform localized differential privacy processing on the load demand of each park to obtain the park energy purchase strategy after localized differential privacy processing, so that for each park, the load demand vector... Load demand vector after localized differential privacy Satisfying the indistinguishability is The optimal scheduling cost obtained after using localized differential privacy data satisfies ; Step 5) The park service provider optimizes the scheduling based on the energy purchase strategies returned by each park to obtain the energy sales strategy; Step 6) Park service provider calculates residuals If the value is less than the threshold, the scheduling results of each participating entity will be output; otherwise, the energy price will be updated. Then return to step 3) and iterate again.

2. The method for protecting the privacy of the main body of a multi-park integrated energy system according to claim 1, characterized in that, Step 2 specifically includes: a) Constraints on inter-park energy trading ,in: , , The park In the The imported electrical power, thermal power, and gas power within a time period; when the gas power is negative, it indicates the exported power. b) Internal constraints of the park, including: i) System-level constraints, namely, the supply and demand balance of electricity, heat, and gas: ,in: A collection of devices; For equipment In the Gas supply volume within a specific time period; , , respectively equipment In the Electricity, heat, and gas consumption within a specific time period; , , The park is in the Electricity load, heat load, and gas load within a given time period; ii) Equipment-level constraints, including: energy production and consumption equipment and energy conversion equipment, wherein: The energy generation and consumption equipment includes: 1) Carbon Capture Power Plant (CCPP): ,in: For equipment In the Total power over a given time period; For carbon capture power plants In the Carbon capture power over a given period; For equipment In the Fixed power loss within a time period; For carbon capture power plants The energy consumption coefficient of carbon capture operation; For carbon capture power plants In the Carbon capture volume over a given period; For carbon capture efficiency; A collection of equipment that generates CO2 emissions, including coal-fired power plants and combined heat and power (CHP) units; For equipment In the Within a time period Emissions; , For equipment Minimum and maximum output power; Cost of coal per unit of electricity generated; 2) Wind farm WT: ,in: For wind farm In the Actual wind power within a given time period For wind farm In the Wind curtailment during specific time periods; The energy conversion device includes: 1) Operating constraints of the electro-gas conversion (P2G): The CO2 feedstock used in this P2G equipment is all derived from carbon capture, specifically: ,in: For electro-gas conversion equipment The efficiency of converting electrical energy into heat energy; For electro-gas conversion equipment The efficiency of converting electrical energy into gas energy; For equipment The maximum power consumption; For electro-gas conversion equipment CO2 consumption coefficient; 2) Combined Heat and Power (CHP) system: The industrial park adopts a CHP unit model with adjustable heat-to-power ratio, specifically: ,in: For combined heat and power equipment Overall efficiency; , These are combined heat and power (CHP) equipment The upper and lower limits of the thermoelectric ratio; For equipment Maximum gas load; 3) Energy purchase: (2.15), where: For electricity price; For gas prices; 4) Energy Storage: A unified model is used for electric, thermal, and gas energy storage devices. When the energy storage device is an electric energy storage device, the specific model is as follows: ,in: , Energy storage devices In the Charging and discharging power within a time period; , All are binary variables, representing energy storage devices. In the Charge and discharge status parameters within a time period , This indicates that it is in a charging state. , This indicates that the device is in a state of energy release. , Energy storage devices The charging and discharging efficiency; For energy storage devices In the Percentage of capacity within a time period; For energy storage devices Rated capacity; , Energy storage devices The upper and lower limits of the capacity percentage.

3. The method for protecting the privacy of the main body of a multi-park integrated energy system according to claim 1, characterized in that, Step 3 specifically includes: 3.1) Decouple the multi-park integrated energy system from the internal parks by adding a replication variable for inlet and outlet power. The original variable with superscript I is optimized by the integrated energy supplier, and a replication variable with superscript P is introduced and optimized by the service providers of each park. The original inlet and outlet power in the model... , , The power output of electricity, heat, and gas sold by the integrated energy supplier to each park is transformed into a newly introduced replication variable with the superscript P. , , Let's define the power output of the park's purchases of electricity, heat, and gas from integrated energy suppliers; then, we add constraints on the replication variables to ensure that the power output sold equals the power output purchased. ; 3.2) Apply the augmented Lagrangian function to relax the constraint back to the original objective function: ,in: The decision variables representing the optimization of integrated energy suppliers include , and , , ; Decision variables optimized for park service providers include , and , , ; Indicates the corresponding Lagrange multiplier; As a penalty factor; 3.3) Based on the participating entities, the problem is decomposed into lower-level sub-problems and upper-level problems. The objective functions for each sub-problem solved by the park service provider are as follows: ; 3.4) Based on the optimization results of each park service provider, the decision variables for optimizing the park service providers are derived: Then update the Lagrange multipliers based on the decision variable results obtained from solving the problems at the upper and lower levels: .

4. The method for protecting the privacy of the main body of a multi-park integrated energy system according to claim 1, characterized in that, Step 4) specifically involves: for adjacent datasets and query functions , , Differential privacy output results under the Laplace mechanism ,in: and Satisfying indistinguishability , It is noise and follows a Laplace distribution. , For privacy budget, Indicates the park The optimal scheduling cost function, It is the park The original load demand vector, It is the park Load demand vector after localized differential privacy.

5. The method for protecting the privacy of the main body of a multi-park integrated energy system according to claim 1, characterized in that, Step 5) specifically includes: Based on the optimization solution results of the integrated energy service provider, the decision variables for the optimization of the integrated energy service provider are obtained by inverse solution: .

6. The method for protecting the privacy of the main body of a multi-park integrated energy system according to claim 1, characterized in that, The residuals are calculated as follows: .

7. A privacy protection system for a multi-park integrated energy system implementing the method of any one of claims 1-6, characterized in that, include: The system comprises an integrated energy park service provider unit, park units, and a localized differential privacy unit. The integrated energy park service provider unit performs scheduling processing based on the load demand information after localization and differential privacy, obtaining scheduling instructions for each park. The park units perform intra-park scheduling processing based on the scheduling instructions, obtaining scheduling and load demand results. The localized differential privacy unit performs localized differential privacy processing based on the load demand information of each park, obtaining the load demand results of each park after localization and differential privacy, and uploads them to the integrated energy park service provider unit.

Citation Information

Patent Citations

  • Market-driven multi-park integrated energy system low-carbon cooperative scheduling method

    CN117455262A