A control method and device for low-carbon scheduling of a multi-energy system, an electronic device, and a storage medium

Abstract

This invention discloses a control method, device, electronic equipment, and storage medium for low-carbon scheduling of multi-energy systems, belonging to the technical field of low-carbon scheduling control for multi-energy systems. The method includes: acquiring real-time operating data of each node in the multi-energy system; constructing node carbon intensity constraints based on network topology parameters, carbon emission intensity parameters, and real-time operating data, and dividing the multi-energy system into regions to generate multiple independent sub-regions; constructing a region boundary coupling model for each independent sub-region, and establishing a region augmented Lagrangian function in conjunction with the node carbon intensity constraints; generating scheduling results for each independent sub-region through distributed collaborative optimization; integrating the scheduling results of each independent sub-region to generate an optimal system scheduling control command, and adjusting the operating output of each physical device in the multi-energy system accordingly. By implementing this invention, the problem of generating scheduling control commands with global optimization performance in existing technologies can be solved.

Description

Technical Field

[0001] This invention relates to the field of low-carbon scheduling and control technology for multi-energy systems, and specifically to a control method, device, electronic equipment, and storage medium for low-carbon scheduling of multi-energy systems. Background Technology

[0002] With the deepening of low-carbon goals, low-carbon scheduling and control of multi-energy systems has become a key technology for improving comprehensive energy utilization efficiency and achieving energy conservation and emission reduction. Multi-energy systems typically couple multiple energy forms such as electricity, heat, and gas, as well as a large number of underlying physical devices. By constructing a scientific low-carbon scheduling and control architecture, it is possible to rationally allocate energy flow and conversion within the system by combining the real-time operating conditions of each node and the network topology.

[0003] However, existing low-carbon scheduling and control methods for multi-energy systems suffer from technical shortcomings when facing ever-expanding system scales, including low scheduling computation efficiency and large deviations in control command optimization. This is because, to achieve high-precision low-carbon control, scheduling models typically require complex network topology parameters and detailed node carbon intensity constraints. Traditional methods often employ centralized information processing and optimization frameworks, requiring the control center to simultaneously process large-scale real-time operational data from all network nodes, easily leading to a surge in computational dimensionality and communication congestion. Conventional solutions that attempt to use partitioned scheduling often lack detailed modeling of the interaction boundaries between adjacent regions, making it difficult to accurately characterize the complex physical connections and carbon flow coupling states between different sub-regions during decoupling computation. This lack of boundary coordination mechanisms causes existing distributed algorithms to easily get stuck in local optima or fail to converge during mathematical derivation and iterative optimization, ultimately failing to integrate into scheduling and control commands with global optimization performance, and hindering effective carbon reduction regulation of underlying equipment. Summary of the Invention

[0004] This invention provides a control method, apparatus, electronic device, and storage medium for low-carbon scheduling of multi-energy systems. It can solve the problems in the prior art where multi-energy system scheduling faces large-scale data, resulting in a surge in dimensionality and communication congestion due to centralized computing, and where conventional distributed schemes are prone to getting stuck in local optima or non-convergence due to a lack of refined modeling of the physical boundary coupling state between regions, thus making it difficult to generate scheduling control instructions with global optimization performance.

[0005] An embodiment of the present invention provides a control method for low-carbon scheduling of multi-energy systems, comprising: Acquire real-time node operation data of each node in a multi-energy system; Based on preset network topology parameters and preset carbon emission intensity parameters, combined with real-time node operation data, node carbon intensity constraints are generated; based on preset system partitioning rules, the multi-energy system is divided into regions, generating multiple independent sub-regions. For each independent sub-region, the neighboring sub-regions of the current independent sub-region are determined based on the network topology parameters, the physical connection relationship between the current independent sub-region and the neighboring sub-regions is extracted, and the region boundary coupling model of the current independent sub-region is generated. Based on the regional boundary coupling model, nodal carbon intensity constraints, and a preset optimization objective function, the regional augmented Lagrangian function of the current independent sub-region is generated; the regional augmented Lagrangian function is then subjected to distributed optimization calculation to generate the optimization result of the current independent sub-region. After generating the optimization results for all independent sub-regions, the optimization results for all independent sub-regions are integrated to generate the optimal system scheduling and control command; the operating output values ​​of the underlying physical devices in the multi-energy system are adjusted according to the optimal system scheduling and control command.

[0006] Furthermore, the preset carbon emission intensity parameters include the carbon emission intensity parameters of generators and natural gas at each node in the multi-energy system. Based on preset network topology parameters and preset carbon emission intensity parameters, combined with real-time node operation data, node carbon intensity constraints are generated, including: Based on real-time node operation data, extract the node active power input, node active power output, and node natural gas output of each node; Based on the preset network topology parameters, extract the absolute active power flow distribution matrix; For each node, carbon emission calculation is performed on the active power input of the current node based on the current node's generator carbon emission intensity parameters to obtain the node generator carbon emission of the current node. By combining the preset cogeneration conversion coefficient with the current node's natural gas carbon emission intensity parameters, the equivalent carbon emission of the current node's natural gas output is calculated to obtain the node's natural gas conversion carbon emission. Based on the network topology parameters, determine the neighboring nodes that have a transmission relationship with the current node; Based on the power flow components associated with the current node in the absolute active power flow distribution matrix, and the real-time node carbon intensity of the adjacent nodes, carbon flow interaction is performed to obtain the node power flow injection carbon emissions of the current node. The carbon emissions from the node's generators, natural gas conversion, and power flow injection are aggregated to generate the total carbon input of the current node. Based on the cogeneration conversion coefficient, the node natural gas output of the current node is equivalently converted into the node natural gas conversion energy of the current node. The node natural gas conversion energy, node active power output, and power flow components associated with the current node are aggregated to generate the total node output energy of the current node. Based on the total output energy of the current node, the total input carbon of the current node is normalized to generate the real-time node carbon intensity of the current node. After completing the calculation steps for all nodes in the multi-energy system, a node carbon intensity constraint condition is constructed, with the real-time node carbon intensity of each node not exceeding the preset maximum node carbon intensity upper limit threshold.

[0007] Furthermore, based on preset system partitioning rules, the multi-energy system is divided into regions, generating multiple independent sub-regions, including: Based on the preset system partitioning rules, the nodes in the multi-energy system are divided into multiple sets of non-overlapping sub-nodes; For each set of child nodes, based on the network topology parameters, extract the physical connection relationships within the current set of child nodes and generate the region-specific connection relationships for the current set of child nodes. Combine the current set of child nodes with the internal connection relationships within the current set of child nodes to generate the current independent sub-region.

[0008] Furthermore, based on network topology parameters, neighboring sub-regions of the current independent sub-region are determined, the physical connection relationships between the current independent sub-region and its neighboring sub-regions are extracted, and a region boundary coupling model of the current independent sub-region is generated, including: Based on network topology parameters, identify the physical connection relationships between nodes in the current independent sub-region and nodes in other independent sub-regions of the multi-energy system, and use these relationships as the cross-regional boundary connections of the current independent sub-region. Other independent sub-regions associated with the cross-regional boundary connection relationship are identified as adjacent sub-regions of the current independent sub-region; Extract the nodes connected by the cross-regional boundary connection within the current independent sub-region, and use them as the boundary nodes of the current independent sub-region; Based on the boundary nodes of the current independent sub-region, generate the boundary information extraction matrix of the current independent sub-region; Using the boundary information extraction matrix of the current independent sub-region, the boundary node state data corresponding to the boundary node is extracted from the real-time node operation data and real-time node carbon intensity of the current independent sub-region. Based on the cross-regional boundary connection relationship, define temporary boundary variables to characterize the interaction state between regions; Based on a pre-defined temporary variable association matrix, the boundary temporary variables are associated with the boundary node state data to generate the current independent sub-region's regional boundary coupling model.

[0009] Furthermore, based on the regional boundary coupling model, nodal carbon intensity constraints, and a pre-defined optimization objective function, the regional augmented Lagrangian function for the current independent sub-region is generated, including: Based on the preset optimization objective function, determine the local objective function of the current independent sub-region; Based on the region boundary coupling model, dual multiplier penalty terms and quadratic penalty terms corresponding to the current independent sub-regions are generated. Linearize the non-convex constraints of the node carbon intensity constraints to generate the linearized carbon intensity constraints of the current independent sub-regions. The local objective function of the region, the dual multiplier penalty term, and the quadratic penalty term are superimposed to generate the main body of the augmented objective function of the current independent sub-region; Using the linearized carbon intensity constraint as the local solution boundary of the augmented objective function, the region augmented Lagrangian function of the current independent sub-region is generated.

[0010] Furthermore, a distribution optimization calculation is performed on the augmented Lagrangian function of the region to generate the optimization results for the current independent sub-regions, including: Obtain the initial set of scheduling variables, initial boundary temporary variables, initial dual multiplier parameters, and initial penalty coefficients for the current independent subregion; Repeatedly perform the distributed optimization iteration operation until the preset distributed optimization convergence condition is met; wherein, the distributed optimization convergence condition is that the current original convergence residual is not greater than the preset original residual threshold, and the current dual convergence residual is not greater than the preset dual residual threshold. The set of target scheduling variables generated when the iteration stops is determined as the optimization result of the current independent sub-region; The distributed optimization iterative operation includes: The target scheduling variable set is generated by performing a minimum value solution based on the current baseline scheduling variable set, the current baseline boundary temporary variable, the current baseline dual multiplier parameter, the current baseline penalty coefficient, and the regional augmented Lagrangian function; wherein, the initial baseline scheduling variable set is the initial scheduling variable set; the initial baseline boundary temporary variable is the initial boundary temporary variable; the initial baseline dual multiplier parameter is the initial dual multiplier parameter; and the initial baseline penalty coefficient is the initial penalty coefficient. Generate privacy obfuscation information based on the target set of scheduling variables and the current baseline dual multiplier parameters; Cross-regional interaction processing is performed based on privacy-obfuscated information to generate temporary variables for the target boundary. Parameter update calculations are performed based on the target scheduling variable set and the target boundary temporary variables to generate the target dual multiplier parameters; The residual quantization process is performed based on the target scheduling variable set and the target boundary temporary variable to generate the current original convergence residual and the current dual convergence residual, and the target penalty coefficient is generated based on the original convergence residual and the dual convergence residual. When it is determined that the convergence condition of distributed optimization is not met, the target scheduling variable set, target boundary temporary variable, target dual multiplier parameter and target penalty coefficient are updated to the current baseline scheduling variable set, current baseline boundary temporary variable, current baseline dual multiplier parameter and current baseline penalty coefficient, respectively.

[0011] Furthermore, after generating the optimization results for all independent sub-regions, the optimization results for all independent sub-regions are integrated to generate the optimal system scheduling and control command; the operating output values ​​of the underlying physical devices in the multi-energy system are adjusted according to the optimal system scheduling and control command, including: After generating the optimization results for all independent sub-regions, the internal scheduling variables contained in all the optimization results are globally aggregated and concatenated to generate the optimal system scheduling and control instructions. From the optimal system scheduling and control instructions, the target operation setting parameters corresponding to each underlying physical device in the multi-energy system are extracted; The target operation settings parameters are sent to the local control terminal of the underlying physical device. Through the local control terminal, the underlying physical devices are driven to operate according to the corresponding target parameters and adjust the corresponding operating output values.

[0012] Based on the above method embodiments, the present invention provides corresponding apparatus embodiments.

[0013] One embodiment of the present invention provides a control device for low-carbon scheduling of a multi-energy system, comprising: a data acquisition module, a system preprocessing module, a boundary coupling modeling module, a distributed optimization calculation module, and a scheduling control execution module; The data acquisition module is used to acquire real-time node operation data of each node in the multi-energy system; The system preprocessing module is used to generate node carbon intensity constraints based on preset network topology parameters and preset carbon emission intensity parameters, combined with real-time node operation data; and to divide the multi-energy system into regions based on preset system partitioning rules, generating multiple independent sub-regions. The boundary coupling modeling module is used to determine the neighboring sub-regions of the current independent sub-region based on network topology parameters for each independent sub-region, extract the physical connection relationship between the current independent sub-region and the neighboring sub-regions, and generate the region boundary coupling model of the current independent sub-region. The distributed optimization calculation module is used to generate the regional augmented Lagrangian function of the current independent sub-region based on the regional boundary coupling model, the node carbon intensity constraint condition and the preset optimization objective function; and to perform distributed optimization calculation on the regional augmented Lagrangian function to generate the optimization result of the current independent sub-region. The scheduling and control execution module is used to integrate the optimization results of all independent sub-regions after generating the optimization results of all independent sub-regions, and generate the optimal system scheduling and control command; and adjust the operating output value of the underlying physical equipment in the multi-energy system according to the optimal system scheduling and control command.

[0014] Based on the above method embodiments, the present invention provides corresponding electronic device embodiments.

[0015] An embodiment of the present invention provides an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the control method for low-carbon scheduling of a multi-energy system as described in any of the above-described method embodiments.

[0016] Based on the above method embodiments, the present invention provides corresponding storage medium embodiments.

[0017] One embodiment of the present invention provides a storage medium storing a computer program thereon, wherein, when the computer program is running, it controls the device where the storage medium is located to execute the control method for low-carbon scheduling of a multi-energy system as described in any of the above-described method embodiments.

[0018] Compared with the prior art, the present invention has the following beneficial effects: This invention provides a control method, apparatus, electronic device, and storage medium for low-carbon scheduling of a multi-energy system. The method acquires real-time node operation data of each node in the multi-energy system; constructs node carbon intensity constraints based on preset network topology parameters, carbon emission intensity parameters, and real-time node operation data; divides the multi-energy system into regions according to preset system partitioning rules, generating multiple independent sub-regions; for each independent sub-region, determines its adjacent sub-regions and the physical connections between regions, constructing a region boundary coupling model; based on the region boundary coupling model, node carbon intensity constraints, and preset optimization objective function, establishes a region augmented Lagrangian function and performs distributed collaborative optimization to generate the optimal scheduling result for each independent sub-region; after obtaining the optimal scheduling results for all independent sub-regions, performs global coordination and result integration to generate the optimal system scheduling control command; and adjusts and controls the operating output of various underlying physical devices in the multi-energy system according to the optimal system scheduling control command.

[0019] The technical solution of this invention effectively resolves the problems of dimensionality surge and communication congestion caused by traditional centralized scheduling when processing large-scale node data across the entire network by dividing a multi-energy system into multiple independent sub-regions and performing distributed optimization calculations. Simultaneously, addressing the deficiency of boundary coordination mechanisms in conventional distributed solutions, this invention constructs a regional boundary coupling model by extracting the physical connection relationships between adjacent sub-regions and fusing it with node carbon intensity constraints to generate a regional augmented Lagrangian function. This feature accurately characterizes the physical boundaries and carbon flow coupling states between regions, overcoming the problem of traditional optimization calculations easily getting trapped in local optima or failing to converge, thereby ensuring that the final issued control commands for underlying devices possess reliable global optimization performance. Attached Figure Description

[0020] Figure 1 This is a schematic flowchart of a control method for low-carbon scheduling of a multi-energy system provided in an embodiment of the present invention.

[0021] Figure 2 This is a schematic diagram of the structure of a control device for low-carbon scheduling of a multi-energy system provided in an embodiment of the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] like Figure 1As shown, to address the problems in existing technologies where multi-energy system scheduling faces massive data, leading to a surge in dimensionality and communication congestion due to centralized computation, and the tendency of conventional distributed schemes to fall into local optima or fail to converge due to a lack of refined modeling of the physical boundary coupling states between regions, thus making it difficult to generate scheduling control instructions with global optimization performance, an embodiment of the present invention provides a control method for low-carbon scheduling of multi-energy systems, comprising at least the following steps: Step S1: Obtain real-time node operation data of each node in the multi-energy system; Specifically, real-time node operation data of each node in the multi-energy system is acquired as the physical foundation of the entire control method. This acquired real-time node operation data includes power network operation data, natural gas network operation data, and energy conversion node operation data. To accurately characterize the physical flow patterns and coupling evolution states of energy within the multi-energy system, each underlying network within the system follows a strict mathematical coupling model, and the extracted data indicators are constrained by fixed physical conservation laws.

[0024] For the power network portion, a DC optimal power flow model is constructed to quantify node power balance relationships, line power transmission relationships, and node injected power boundary limits. The power dimension data in the acquired real-time node operation data follows the following power network active power balance relationship model: In the formula, Input the value for the active power of the node; This represents the active power output value of the node. For the power lines under its jurisdiction, starting from the node of origin Flowing to the final node The active power flow; It is the set of all physical nodes in the multi-energy system.

[0025] Meanwhile, the active power flow transmitted through the line is constrained by the node phase angle and line impedance parameters, satisfying the following line power transmission model: In the formula, The phase angle value is the value of the starting node; The phase angle value at the endpoint; This refers to the line impedance value of the power line to which it belongs; This refers to the set of all physical circuits in the multi-energy system. To avoid ambiguity caused by repetition of symbols in the original disclosure document, a unified symbol will be used. Replace the original node set symbol with Replace the original line set symbol to ensure the uniqueness of the global variable interpretation.

[0026] The power input to each node must be limited within a safe physical threshold, satisfying the following node injection power constraint model: In the formula, To set the lower limit of active power input safety for a node; To set the safe upper limit of active power input for a node.

[0027] For the natural gas network portion, parameters related to the natural gas supply and demand balance, pipeline transmission status, and natural gas injection volume boundaries are obtained. The natural gas dimension data in the acquired real-time node operation data follows the following natural gas supply and demand balance model: In the formula, Input energy values ​​for natural gas at the node; The value of natural gas output energy at the node; For the natural gas pipeline, along the starting node To the destination node Numerical value of flowing natural gas energy flow.

[0028] The natural gas energy input to the node is limited within the capacity limits of the underlying physical equipment, and has the following natural gas injection constraint model: In the formula, The lower limit value for natural gas input energy at the node; This represents the upper limit value for the natural gas input energy of the node.

[0029] The natural gas transmission status is directly affected by the pipeline operating pressure and the natural gas energy conversion coefficient. The natural gas energy flow value must satisfy the following segmented pipeline transmission characteristic model determined by the flow direction: Based on the operating condition where the starting node pressure is greater than or equal to the ending node pressure, the calculation formula for the forward transmission of the first pipeline is satisfied: Based on the operating condition where the pressure at the starting node is less than the pressure at the ending node, the calculation formula for reverse transmission in the second pipeline is satisfied: In the formula, To determine the pipeline friction coefficient corresponding to the natural gas pipeline; The pre-set upper heating value energy coefficient of natural gas; This represents the pressure value at the starting node; This represents the pressure value at the endpoint.

[0030] For the energy conversion node portion, the real-time node operation data also needs to cover the multiple coupling relationships between electrical energy and natural gas energy conversion to heat demand within the energy hub. For an energy hub including combined heat and power units and gas-fired boilers, the following energy matrix conversion constraint model must be satisfied: In the formula, To set the power load demand value for the node; To set the thermal load demand value for the node; This refers to the operating efficiency value of the transformer equipment; The physical power generation efficiency value of a combined heat and power unit; The physical heating efficiency value of the combined heat and power unit; The heating efficiency value of the gas furnace equipment; The newly introduced natural gas energy allocation coefficient parameter is used to accurately characterize the physical proportion of energy output from the natural gas network allocated to the combined heat and power unit and the gas furnace equipment, and its value range is between zero and one.

[0031] By comprehensively collecting node operation data covering multiple dimensions such as power flow, natural gas flow, and heat flow, and modeling physical conversion relationships, the basic state benchmark of the underlying physical energy network operation is fully established. This ensures that the subsequent construction of dynamic carbon intensity constraint models and the execution of large-scale distributed optimization calculations have an absolutely accurate data source and a data foundation that strictly conforms to objective physical conservation laws.

[0032] Step S2: Based on preset network topology parameters and preset carbon emission intensity parameters, and combined with real-time node operation data, generate node carbon intensity constraints; divide the multi-energy system into regions based on preset system partitioning rules to generate multiple independent sub-regions; In a preferred embodiment, the preset carbon emission intensity parameters include the generator carbon emission intensity parameters and natural gas carbon emission intensity parameters of each node in the multi-energy system. Based on preset network topology parameters and preset carbon emission intensity parameters, combined with real-time node operation data, node carbon intensity constraints are generated, including: Based on real-time node operation data, extract the node active power input, node active power output, and node natural gas output of each node; Based on the preset network topology parameters, extract the absolute active power flow distribution matrix; For each node, carbon emission calculation is performed on the active power input of the current node based on the current node's generator carbon emission intensity parameters to obtain the node generator carbon emission of the current node. By combining the preset cogeneration conversion coefficient with the current node's natural gas carbon emission intensity parameters, the equivalent carbon emission of the current node's natural gas output is calculated to obtain the node's natural gas conversion carbon emission. Based on the network topology parameters, determine the neighboring nodes that have a transmission relationship with the current node; Based on the power flow components associated with the current node in the absolute active power flow distribution matrix, and the real-time node carbon intensity of the adjacent nodes, carbon flow interaction is performed to obtain the node power flow injection carbon emissions of the current node. The carbon emissions from the node's generators, natural gas conversion, and power flow injection are aggregated to generate the total carbon input of the current node. Based on the cogeneration conversion coefficient, the node natural gas output of the current node is equivalently converted into the node natural gas conversion energy of the current node. The node natural gas conversion energy, node active power output, and power flow components associated with the current node are aggregated to generate the total node output energy of the current node. Based on the total output energy of the current node, the total input carbon of the current node is normalized to generate the real-time node carbon intensity of the current node. After completing the calculation steps for all nodes in the multi-energy system, a node carbon intensity constraint condition is constructed, with the real-time node carbon intensity of each node not exceeding the preset maximum node carbon intensity upper limit threshold.

[0033] In a preferred embodiment, the multi-energy system is divided into regions based on a preset system partitioning rule, generating multiple independent sub-regions, including: Based on the preset system partitioning rules, the nodes in the multi-energy system are divided into multiple sets of non-overlapping sub-nodes; For each set of child nodes, based on the network topology parameters, extract the physical connection relationships within the current set of child nodes and generate the region-specific connection relationships for the current set of child nodes. Combine the current set of child nodes with the internal connection relationships within the current set of child nodes to generate the current independent sub-region.

[0034] Specifically, in step S2, based on preset network topology parameters and preset carbon emission intensity parameters, and combined with real-time node operation data, node carbon intensity constraints are generated; the multi-energy system is divided into regions based on preset system partitioning rules, generating multiple independent sub-regions. For the generation phase of node carbon intensity constraints, pre-defined carbon emission intensity parameters need to be introduced as basic quantitative indicators. These pre-defined carbon emission intensity parameters cover the generator carbon emission intensity parameters and natural gas carbon emission intensity parameters of each node in the multi-energy system. To accurately calculate the carbon emission flow path within the multi-energy system, the active power input, active power output, and natural gas output of each node are extracted synchronously based on real-time node operating data. Simultaneously, an absolute active power flow distribution matrix is ​​extracted based on pre-defined network topology parameters. To avoid algebraic sign interference from the physical power flow direction, the elements of the absolute active power flow distribution matrix undergo absolute value rectification according to the direction state of the basic physical power flow, satisfying the following defined absolute power flow distribution transformation calculation formula: In the formula, The absolute active power flow distribution component that is determined from the starting node to the ending node; This refers to the absolute active power flow distribution component that flows back from the endpoint node to the origin node. Through the above segmented conversion model, regardless of the positive or negative direction of the actual physical power flow, it can be converted into an absolutely positive carbon flow driving the basic carrier.

[0035] After acquiring the basic variables, a rigorous calculation of the carbon flow intensity value is performed for each node. Based on the current node's generator carbon emission intensity parameters, a direct physical carbon emission calculation is performed on the current node's active power input to obtain the node's generator carbon emission. To comprehensively cover the coupled carbon emission effects of multiple energy networks, an equivalent carbon emission conversion is performed on the current node's natural gas output, combining a preset cogeneration conversion coefficient with the current node's natural gas carbon emission intensity parameters to obtain the current node's natural gas conversion carbon emission. Next, based on the network topology parameters, neighboring nodes with transmission relationships with the current node are identified. Based on the power flow components associated with the current node in the absolute active power flow distribution matrix and the real-time node carbon intensity of neighboring nodes, carbon flow interaction deduction is performed to obtain the current node's power flow injection carbon emission. The current node's generator carbon emission, natural gas conversion carbon emission, and power flow injection carbon emission are linearly aggregated to generate the current node's total node input carbon.

[0036] To obtain a standardized intensity index, the energy flow benchmark of the current node must be defined simultaneously. Based on the cogeneration conversion coefficient, the node's natural gas output is equivalently converted into the node's natural gas conversion energy. The node's natural gas conversion energy, node active power output, and associated power flow components are linearly aggregated to generate the node's total output energy. Subsequently, using the node's total output energy as a benchmark, the node's total input carbon is normalized to generate the real-time node carbon intensity. The real-time node carbon intensity calculation process follows the following carbon emission flow conservation ratio model: In the formula, To calculate the real-time node carbon intensity value of the current node; This represents the current node's generator carbon emission intensity parameter value; This represents the current node's natural gas carbon emission intensity parameter value; This represents the real-time node carbon intensity values ​​of the neighboring nodes adjacent to the current node. In the above formula, the numerator polynomial represents the total carbon input of the generated nodes, and the denominator polynomial represents the total energy output of the generated nodes.

[0037] After completing the computational steps for all nodes in the multi-energy system, mandatory emission reduction limits are imposed on all physical nodes in the network. A node carbon intensity constraint is constructed using the real-time node carbon intensity of each node not exceeding a preset maximum node carbon intensity threshold as a hard criterion boundary. The mathematical expression for the node carbon intensity constraint is as follows: In the formula, The upper limit threshold parameter for the maximum carbon intensity that a specific node is allowed to accommodate.

[0038] Based on the obtained global carbon emission constraints, a physical dimensionality reduction and decoupling process is performed on the multi-energy system. This process begins by dividing the nodes in the multi-energy system into multiple non-overlapping sets of sub-nodes based on preset system partitioning rules. These partitioning rules include geospatial distribution attribute rules and control function attribute rules, ensuring a high degree of physical compactness among nodes within each region. For each sub-node set, based on the overall network topology parameters of the multi-energy system, the physical connections between nodes within the current sub-node set are extracted, generating the intra-regional connectivity relationships for that sub-node set. Finally, the current sub-node set and its intra-regional connectivity relationships are topologically combined to generate the current independent sub-region.

[0039] By constructing carbon intensity constraints through in-depth multi-energy coupling and flow processes, and dividing the large-scale network into non-overlapping sub-regions according to the spatial physical morphology, not only is a clear and rigorous quantitative monitoring benchmark provided for the coordinated emission reduction of multi-energy networks, but also the high-dimensional computational barrier faced by centralized control is completely dismantled from the system topology dimension, which greatly improves the solution convergence efficiency of the distributed optimization architecture under the massive node scale.

[0040] Step S3: For each independent sub-region, determine the neighboring sub-regions of the current independent sub-region based on the network topology parameters, extract the physical connection relationship between the current independent sub-region and the neighboring sub-regions, and generate the region boundary coupling model of the current independent sub-region. In a preferred embodiment, neighboring sub-regions of the current independent sub-region are determined based on network topology parameters, the physical connection relationships between the current independent sub-region and its neighboring sub-regions are extracted, and a region boundary coupling model of the current independent sub-region is generated, including: Based on network topology parameters, identify the physical connection relationships between nodes in the current independent sub-region and nodes in other independent sub-regions of the multi-energy system, and use these relationships as the cross-regional boundary connections of the current independent sub-region. Other independent sub-regions associated with the cross-regional boundary connection relationship are identified as adjacent sub-regions of the current independent sub-region; Extract the nodes connected by the cross-regional boundary connection within the current independent sub-region, and use them as the boundary nodes of the current independent sub-region; Based on the boundary nodes of the current independent sub-region, generate the boundary information extraction matrix of the current independent sub-region; Using the boundary information extraction matrix of the current independent sub-region, the boundary node state data corresponding to the boundary node is extracted from the real-time node operation data and real-time node carbon intensity of the current independent sub-region. Based on the cross-regional boundary connection relationship, define temporary boundary variables to characterize the interaction state between regions; Based on a pre-defined temporary variable association matrix, the boundary temporary variables are associated with the boundary node state data to generate the current independent sub-region's regional boundary coupling model.

[0041] Specifically, for each independent sub-region that has been initially divided, the physical interaction boundary between the current independent sub-region and the remaining part of the multi-energy system is further established. Based on the pre-acquired network topology parameters, the physical transmission lines connecting nodes within the current independent sub-region and nodes within other independent sub-regions of the multi-energy system are identified in the global network topology map. The extracted physical connections that cross the regional division boundaries are taken as the cross-regional boundary connections of the current independent sub-region. Topological path tracing is performed along the cross-regional boundary connections, and other independent sub-regions directly associated with the other end of the cross-regional boundary connections are identified as the adjacent sub-regions of the current independent sub-region.

[0042] After establishing adjacent sub-regions, it is necessary to pinpoint the specific anchor points where physical coupling occurs. Within the current independent sub-region, extract the physical nodes directly connected to the cross-regional boundary as the boundary nodes of the current independent sub-region. Based on the topological distribution of the extracted boundary nodes, generate a boundary information extraction matrix specifically for the current independent sub-region. The purpose of constructing the boundary information extraction matrix is ​​to extract the core parameters involved in cross-regional interactions from the complex internal state data of the region. Using the generated boundary information extraction matrix, perform matrix mapping operations on the real-time node operation data and real-time node carbon intensity of the current independent sub-region to accurately extract the boundary node state data that strictly corresponds to the boundary nodes.

[0043] To achieve decoupling and privacy-preserving information exchange between adjacent sub-regions in a distributed computing architecture, a core mathematical bridge must be introduced. Based on the identified cross-regional boundary connections, temporary boundary variables are defined to characterize the interaction between energy flow and carbon emission transfer between different independent sub-regions in a multi-energy system. A pre-defined temporary variable correlation matrix is ​​introduced, and based on this matrix, a consistent association is established between the newly defined temporary boundary variables and the extracted boundary node state data, thereby generating the regional boundary coupling model for the current independent sub-region. The constructed regional boundary coupling model satisfies the following mathematical equation constraints: In the formula, The boundary information extraction matrix for the current independent sub-regions generated in the preceding steps; The region's internal state column vector is formed by sequentially arranging all real-time node operation data and real-time node carbon intensity within the current independent sub-region. The region's internal state column vector includes the node phase angle value, node pressure value, and real-time node carbon intensity value corresponding to the nodes within the region. The result of the matrix multiplication operation is equivalent to the extracted boundary node state data; The temporary variable correlation matrix is ​​set; Temporary variables defined for boundaries.

[0044] By identifying cross-regional topological connections and introducing boundary temporary variables to construct a strict equality coupling constraint model, the energy convergence state and carbon flow fusion state between adjacent independent sub-regions are accurately depicted at the physical level. At the algorithmic level, the globally high-dimensional strongly coupled network is successfully decoupled into a low-dimensional regional mathematical model that can be solved independently through boundary temporary variables. This lays a seamless architectural foundation for subsequent high-concurrency distributed privacy optimization computation.

[0045] Step S4: Based on the regional boundary coupling model, node carbon intensity constraints, and the preset optimization objective function, generate the regional augmented Lagrangian function for the current independent sub-region; perform distributed optimization calculation on the regional augmented Lagrangian function to generate the optimization result for the current independent sub-region. In a preferred embodiment, based on the region boundary coupling model, nodal carbon intensity constraints, and a preset optimization objective function, the region augmented Lagrangian function of the current independent sub-region is generated, including: Based on the preset optimization objective function, determine the local objective function of the current independent sub-region; Based on the region boundary coupling model, dual multiplier penalty terms and quadratic penalty terms corresponding to the current independent sub-regions are generated. Linearize the non-convex constraints of the node carbon intensity constraints to generate the linearized carbon intensity constraints of the current independent sub-regions. The local objective function of the region, the dual multiplier penalty term, and the quadratic penalty term are superimposed to generate the main body of the augmented objective function of the current independent sub-region; Using the linearized carbon intensity constraint as the local solution boundary of the augmented objective function, the region augmented Lagrangian function of the current independent sub-region is generated.

[0046] In a preferred embodiment, a distribution optimization calculation is performed on the region augmentation Lagrangian function to generate the optimization results for the current independent sub-regions, including: Obtain the initial set of scheduling variables, initial boundary temporary variables, initial dual multiplier parameters, and initial penalty coefficients for the current independent subregion; Repeatedly perform the distributed optimization iteration operation until the preset distributed optimization convergence condition is met; wherein, the distributed optimization convergence condition is that the current original convergence residual is not greater than the preset original residual threshold, and the current dual convergence residual is not greater than the preset dual residual threshold. The set of target scheduling variables generated when the iteration stops is determined as the optimization result of the current independent sub-region; The distributed optimization iterative operation includes: The target scheduling variable set is generated by performing a minimum value solution based on the current baseline scheduling variable set, the current baseline boundary temporary variable, the current baseline dual multiplier parameter, the current baseline penalty coefficient, and the regional augmented Lagrangian function; wherein, the initial baseline scheduling variable set is the initial scheduling variable set; the initial baseline boundary temporary variable is the initial boundary temporary variable; the initial baseline dual multiplier parameter is the initial dual multiplier parameter; and the initial baseline penalty coefficient is the initial penalty coefficient. Generate privacy obfuscation information based on the target set of scheduling variables and the current baseline dual multiplier parameters; Cross-regional interaction processing is performed based on privacy-obfuscated information to generate temporary variables for the target boundary. Parameter update calculations are performed based on the target scheduling variable set and the target boundary temporary variables to generate the target dual multiplier parameters; The residual quantization process is performed based on the target scheduling variable set and the target boundary temporary variable to generate the current original convergence residual and the current dual convergence residual, and the target penalty coefficient is generated based on the original convergence residual and the dual convergence residual. When it is determined that the convergence condition of distributed optimization is not met, the target scheduling variable set, target boundary temporary variable, target dual multiplier parameter and target penalty coefficient are updated to the current baseline scheduling variable set, current baseline boundary temporary variable, current baseline dual multiplier parameter and current baseline penalty coefficient, respectively.

[0047] Specifically, in step S4, based on the regional boundary coupling model, the nodal carbon intensity constraint, and the preset optimization objective function, the regional augmented Lagrangian function of the current independent sub-region is generated; the regional augmented Lagrangian function is then subjected to distributed optimization calculation to generate the optimization result of the current independent sub-region. To solve complex low-carbon economic scheduling problems in a distributed architecture, a dedicated mathematical optimization model needs to be constructed for each independent sub-region. Based on a pre-defined optimization objective function, a local objective function for the current independent sub-region is determined. This local objective function is mathematically aggregated from three core operating costs: electricity operating costs, natural gas operating costs, and carbon emission penalty costs. The electricity operating cost is represented as a quadratic function of generator active power output; the natural gas operating cost is represented as a linear function of natural gas supply; and the carbon emission penalty cost is calculated based on the system's actual carbon emissions and the set unit carbon tax price. The local objective function follows the following cost aggregation formula: In the formula, For the generated current independent subregion, the region-local objective function; This is the set of all generator nodes contained within the current independent subregion; These are the secondary cost coefficient, primary cost coefficient, and constant cost coefficient set for the corresponding generator nodes; This is the set of all natural gas source nodes contained within the current independent sub-region; The unit gas purchase cost value set for the corresponding natural gas source node; The unit carbon tax price parameter is set in advance; This is the set of all physical nodes contained within the current independent sub-region.

[0048] Based on the region boundary coupling model generated by the aforementioned steps, the Lagrange duality principle is introduced to generate dual multiplier penalty terms and quadratic penalty terms corresponding to the current independent sub-regions. The dual multiplier penalty term is used to relax the equality constraint boundaries between regions, and the quadratic penalty term is used to enhance the internal convexity of the algorithm's iterative convergence process.

[0049] For node carbon intensity constraints with strong non-convex characteristics, linearization of the non-convex constraints is implemented to generate linearized carbon intensity constraints for the current independent sub-regions. The linearization process relies on a continuous dual Taylor expansion algorithm. First, the iterative variables within the non-convex constraints are separated using the Jacobi iteration method to construct basic iterative separation constraints. Then, a first-order Taylor expansion strategy is used to perform two linearization decompositions. The first expansion operation fixes the node carbon intensity values ​​output from the previous Taylor iteration step, and the second expansion operation corrects the change measures of the associated running variables. The final generated linearized carbon intensity constraints satisfy the following dual expansion calculation formula: In the formula, The sequence number of the expansion iterations for the defined continuous dual Taylor expansion algorithm; with Variables with superscripts represent known state parameters that were fixed in the previous expansion iteration; those with superscripts represent known state parameters. The variables with superscripts represent the unknown parameters that need to be solved in the current iteration. The above expansion process eliminates strong non-convexity while fully preserving the objective physical meaning of the carbon emission flow process.

[0050] The local objective function, dual multiplier penalty term, and quadratic penalty term are algebraically superimposed to generate the augmented objective function body of the current independent sub-region. Then, using the linearized carbon intensity constraint as the local solution boundary of the augmented objective function body, the region-augmented Lagrangian function of the current independent sub-region is generated. The generated region-augmented Lagrangian function satisfies the following augmented construction formula: In the formula, For the augmented Lagrangian function of the combined region; The transpose of the current benchmark dual multiplier parameters is the vector matrix. This is the current baseline penalty coefficient; Let be the set of scheduling variables to be solved; This is a temporary variable for the current baseline boundary.

[0051] After establishing the underlying mathematical optimization model, distributed optimization calculations are performed on the augmented Lagrangian function of the region to generate the optimization results for the current independent sub-region. The distributed optimization calculation begins in the algorithm state initialization phase, obtaining the initial set of scheduling variables, initial boundary temporary variables, initial dual multiplier parameters, and initial penalty coefficients for the current independent sub-region. Subsequently, the distributed optimization iterative operation is repeated until the preset distributed optimization convergence conditions are met. The distributed optimization convergence conditions are strictly defined as the current original convergence residual not exceeding a preset original residual threshold, and the current dual convergence residual not exceeding a preset dual residual threshold. The final set of target scheduling variables generated at the end of the iteration is determined as the optimization result for the current independent sub-region.

[0052] The distributed optimization iterative operation encompasses a series of alternating solution computations and privacy data interaction actions. Based on the current baseline scheduling variable set, the current baseline boundary temporary variables, the current baseline dual multiplier parameters, the current baseline penalty coefficients, and the regional augmented Lagrangian function, a minimum value solution is performed to find the variable analytic point that minimizes the value of the regional augmented Lagrangian function, generating the target scheduling variable set. In the initial state of the first iteration, each baseline parameter is equal to its corresponding initially acquired parameter.

[0053] To achieve cross-regional collaborative interaction without disclosing privacy-preserving operational data within a region, privacy-obfuscating information is generated based on the target scheduling variable set and the current baseline dual multiplier parameters. The generated privacy-obfuscating information satisfies the following state mask superposition calculation formula: In the formula, The generated privacy obfuscation information column vector; This is the set of target scheduling variables generated in the preliminary minimum solution step.

[0054] Cross-regional interaction processing is performed based on privacy obfuscation information to generate temporary variables for the target boundary. This cross-regional interaction processing integrates homomorphic encryption and dynamic random weight masking strategies. Utilizing the additive homomorphic property of homomorphic encryption, communication links are established between adjacent sub-regions, transmitting encrypted privacy obfuscation information carrying random weight factors. At the ciphertext level, a weighted sum of adjacent boundary state variables is directly calculated to aggregate the ciphertext. Upon receiving the aggregated ciphertext, a local decryption operation is performed using a dedicated region private key. This is then divided by the sum of the injected random weight values, thereby calculating the temporary variables for the target boundary under secure conditions that completely block the reverse calculation path of plaintext parameters.

[0055] Parameter update calculations are performed based on the target scheduling variable set and the target boundary temporary variables to generate target dual multiplier parameters. The parameter update calculations satisfy the following multiplier gradient step calculation formula controlled by the penalty coefficient: In the formula, To update the accumulated target dual multiplier parameters; Temporary target boundary variables generated for cross-regional interactive decryption.

[0056] Residual quantization is performed on the target scheduling variable set and the target boundary temporary variables to generate the current original convergence residual and the current dual convergence residual. Based on these two residuals, a target penalty coefficient is generated. The original convergence residual measures the degree of physical violation of inter-regional equality constraints, while the dual convergence residual measures the fluctuation drift difference of the boundary dual variables between adjacent iterations. The two iterative residuals satisfy the following second-norm quantitative evaluation formula: In the formula, The current raw convergence residuals are obtained from quantitative statistics; This is the current dual convergence residual obtained from quantitative statistics. Subsequently, an adaptive residual balancing control strategy is executed within the algorithm. By comparing the order of magnitude ratio between the original convergence residual and the dual convergence residual, the value of the baseline penalty coefficient is dynamically increased or decreased to generate a target penalty coefficient that strictly balances the convergence speed of the original state with the convergence speed of the dual space.

[0057] After generating all parameters within a single loop, the underlying iterative convergence endpoint is evaluated. If the distributed optimization convergence condition is not met, the target scheduling variable set, target boundary temporary variable, target dual multiplier parameter, and target penalty coefficient are replaced and updated with the current baseline scheduling variable set, current baseline boundary temporary variable, current baseline dual multiplier parameter, and current baseline penalty coefficient, respectively, to support the next round of distributed optimization alternating calculations.

[0058] By introducing Lagrange duality theory and homomorphic encryption cross-regional interaction mechanism to perform alternating direction multiplier solving, the system computing power bottleneck of high-dimensional non-convex carbon strength constraints that are difficult to implement centralized large-scale solution is not only completely overcome, but also achieves optimal distributed collaborative scheduling optimization of carbon emission reduction penalty costs and basic economic operating costs within the huge energy network under the premise of absolutely guaranteeing the privacy and security of the underlying equipment operation data in each physical sub-region.

[0059] Step S5: After generating the optimization results of all independent sub-regions, integrate the optimization results of all independent sub-regions to generate the optimal system scheduling and control command; adjust the operating output values ​​of the underlying physical devices in the multi-energy system according to the optimal system scheduling and control command.

[0060] In a preferred embodiment, after generating the optimization results for all independent sub-regions, the optimization results for all independent sub-regions are integrated to generate the optimal system scheduling and control command; the operating output values ​​of the underlying physical devices in the multi-energy system are adjusted according to the optimal system scheduling and control command, including: After generating the optimization results for all independent sub-regions, the internal scheduling variables contained in all the optimization results are globally aggregated and concatenated to generate the optimal system scheduling and control instructions. From the optimal system scheduling and control instructions, the target operation setting parameters corresponding to each underlying physical device in the multi-energy system are extracted; The target operation settings parameters are sent to the local control terminal of the underlying physical device. Through the local control terminal, the underlying physical devices are driven to operate according to the corresponding target parameters and adjust the corresponding operating output values.

[0061] Specifically, once the distributed optimization convergence condition is met, the data scattered across various computing nodes needs to be aggregated and restored to the global physical control dimension. First, the optimization results output after iterative calculations of all independent sub-regions in the multi-energy system are obtained. Then, the internal scheduling variables contained in all the optimization results are globally aggregated and concatenated. These internal scheduling variables encompass multiple mathematical state quantities spanning both power flow and natural gas flow dimensions. Through multi-dimensional reverse restoration and reassembly operations, an optimal system scheduling control command covering the entire multi-energy system is generated. This optimal system scheduling control command essentially constitutes the physical action reference surface for the next operating cycle, internally containing the globally optimal power scheduling scheme and natural gas allocation scheme.

[0062] To accurately translate macro-level optimization strategies into underlying hardware actions, target operating parameters corresponding precisely to each underlying physical device in the multi-energy system are further extracted from the optimal system scheduling and control commands. These target operating parameters precisely map to specific generator unit active power output targets, natural gas compressor power settings in the natural gas pipeline network, and physical opening settings of natural gas pipeline valves. After command parsing and decomposition, the target operating parameters are distributed to the local control terminals of their respective underlying physical devices via industrial communication links. The local control terminal, acting as the central interface between the algorithm architecture and hardware, is responsible for receiving and parsing data parameters from the upper-level decision-making channel.

[0063] The final physical control actions are executed through local control terminals distributed throughout the multi-energy system. These terminals directly drive the corresponding underlying physical devices to operate according to the set parameters for the target, strictly adjusting the corresponding output values. Specifically, the hardware actions involve adjusting the active power output of the generator set and regulating the operating power of the corresponding compressor and the fluid flow area of ​​the corresponding valve in the natural gas pipeline network. This process of thoroughly converting the results of the distributed algorithm optimization into control commands for physical hardware not only breaks the application limitations of pure mathematical algorithms confined to virtual computation but also achieves a globally optimal physical technology effect—a deep synergy between the economic cost of operating the underlying physical entities of the multi-energy system and low-carbon emission reduction—while meeting extremely stringent carbon emission intensity constraints.

[0064] Based on the above method embodiments, the present invention provides corresponding apparatus embodiments.

[0065] like Figure 2 As shown, an embodiment of the present invention provides a control device for low-carbon scheduling of a multi-energy system, including: a data acquisition module, a system preprocessing module, a boundary coupling modeling module, a distributed optimization calculation module, and a scheduling control execution module; The data acquisition module is used to acquire real-time node operation data of each node in the multi-energy system; The system preprocessing module is used to generate node carbon intensity constraints based on preset network topology parameters and preset carbon emission intensity parameters, combined with real-time node operation data; and to divide the multi-energy system into regions based on preset system partitioning rules, generating multiple independent sub-regions. The boundary coupling modeling module is used to determine the neighboring sub-regions of the current independent sub-region based on network topology parameters for each independent sub-region, extract the physical connection relationship between the current independent sub-region and the neighboring sub-regions, and generate the region boundary coupling model of the current independent sub-region. The distributed optimization calculation module is used to generate the regional augmented Lagrangian function of the current independent sub-region based on the regional boundary coupling model, the node carbon intensity constraint condition and the preset optimization objective function; and to perform distributed optimization calculation on the regional augmented Lagrangian function to generate the optimization result of the current independent sub-region. The scheduling and control execution module is used to integrate the optimization results of all independent sub-regions after generating the optimization results of all independent sub-regions, and generate the optimal system scheduling and control command; and adjust the operating output value of the underlying physical equipment in the multi-energy system according to the optimal system scheduling and control command.

[0066] It should be noted that the embodiments of the device described above correspond to the embodiments of the present invention described above, and can realize the control method for low-carbon scheduling of multi-energy systems as described in any one of the present invention. Furthermore, the embodiments of the device described above are merely illustrative. The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Additionally, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between modules indicates that they have a communication connection, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without creative effort.

[0067] Based on the above-described method embodiments of the present invention, a corresponding embodiment of an electronic device is provided.

[0068] An embodiment of the present invention provides an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the control method for low-carbon scheduling of multi-energy systems according to any one of the present invention, or, when the processor executes the computer program, it implements the functions of each module in the above-described device embodiments.

[0069] For example, the computer program may be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program in the terminal device.

[0070] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0071] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0072] The memory can be used to store the computer programs and / or modules. The processor implements various functions of the terminal device by running or executing the computer programs and / or modules stored in the memory and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, applications required for at least one function, etc.; the data storage area may store data created based on the use of the mobile phone, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital card (SD card), flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0073] Based on the above method embodiments, the present invention provides corresponding storage medium embodiments; Another embodiment of the present invention provides a storage medium including a stored computer program, wherein, when the computer program is running, the device where the storage medium is located controls the execution of any one of the above-described multi-energy system low-carbon scheduling control methods of the present invention.

[0074] The aforementioned storage medium is a computer-readable storage medium, and the computer program includes computer program code, which may be in the form of source code, object code, executable file, or certain intermediate forms. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0075] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples.

[0076] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A control method for low-carbon scheduling of a multi-energy system, characterized in that, include: Acquire real-time node operation data of each node in a multi-energy system; Based on preset network topology parameters and preset carbon emission intensity parameters, combined with real-time node operation data, node carbon intensity constraints are generated. The multi-energy system is divided into regions based on preset system partitioning rules, generating multiple independent sub-regions; For each independent sub-region, the neighboring sub-regions of the current independent sub-region are determined based on the network topology parameters, the physical connection relationship between the current independent sub-region and the neighboring sub-regions is extracted, and the region boundary coupling model of the current independent sub-region is generated. Based on the regional boundary coupling model, nodal carbon intensity constraints, and a preset optimization objective function, the regional augmented Lagrangian function of the current independent sub-region is generated; the regional augmented Lagrangian function is then subjected to distributed optimization calculation to generate the optimization result of the current independent sub-region. After generating the optimization results for all independent sub-regions, the optimization results for all independent sub-regions are integrated to generate the optimal system scheduling and control command; the operating output values ​​of the underlying physical devices in the multi-energy system are adjusted according to the optimal system scheduling and control command.

2. The control method of claim 1, wherein, The preset carbon emission intensity parameters include the carbon emission intensity parameters of generators and natural gas at each node in the multi-energy system; Based on preset network topology parameters and preset carbon emission intensity parameters, combined with real-time node operation data, node carbon intensity constraints are generated, including: Based on real-time node operation data, extract the node active power input, node active power output, and node natural gas output of each node; Based on the preset network topology parameters, extract the absolute active power flow distribution matrix; For each node, carbon emission calculation is performed on the active power input of the current node based on the current node's generator carbon emission intensity parameters to obtain the node generator carbon emission of the current node. By combining the preset cogeneration conversion coefficient with the current node's natural gas carbon emission intensity parameters, the equivalent carbon emission of the current node's natural gas output is calculated to obtain the node's natural gas conversion carbon emission. Based on the network topology parameters, determine the neighboring nodes that have a transmission relationship with the current node; Based on the power flow components associated with the current node in the absolute active power flow distribution matrix, and the real-time node carbon intensity of the adjacent nodes, carbon flow interaction is performed to obtain the node power flow injection carbon emissions of the current node. The carbon emissions from the node's generators, natural gas conversion, and power flow injection are aggregated to generate the total carbon input of the current node. Based on the cogeneration conversion coefficient, the node natural gas output of the current node is equivalently converted into the node natural gas conversion energy of the current node. The node natural gas conversion energy, node active power output, and power flow components associated with the current node are aggregated to generate the total node output energy of the current node. Based on the total output energy of the current node, the total input carbon of the current node is normalized to generate the real-time node carbon intensity of the current node. After completing the calculation steps for all nodes in the multi-energy system, a node carbon intensity constraint condition is constructed, with the real-time node carbon intensity of each node not exceeding the preset maximum node carbon intensity upper limit threshold.

3. The control method for low-carbon scheduling of multi-energy systems as described in claim 2, characterized in that, The multi-energy system is divided into regions based on preset system partitioning rules, generating multiple independent sub-regions, including: Based on the preset system partitioning rules, the nodes in the multi-energy system are divided into multiple sets of non-overlapping sub-nodes; For each set of child nodes, based on the network topology parameters, extract the physical connection relationships within the current set of child nodes and generate the region-specific connection relationships for the current set of child nodes. Combine the current set of child nodes with the internal connection relationships within the current set of child nodes to generate the current independent sub-region.

4. The control method for low-carbon dispatching of multi-energy systems as described in claim 3, characterized in that, Based on network topology parameters, the neighboring sub-regions of the current independent sub-region are determined, the physical connection relationships between the current independent sub-region and its neighboring sub-regions are extracted, and a region boundary coupling model of the current independent sub-region is generated, including: Based on network topology parameters, identify the physical connection relationships between nodes in the current independent sub-region and nodes in other independent sub-regions of the multi-energy system, and use these relationships as the cross-regional boundary connections of the current independent sub-region. Other independent sub-regions associated with the cross-regional boundary connection relationship are identified as adjacent sub-regions of the current independent sub-region; Extract the nodes connected by the cross-regional boundary connection within the current independent sub-region, and use them as the boundary nodes of the current independent sub-region; Based on the boundary nodes of the current independent sub-region, generate the boundary information extraction matrix of the current independent sub-region; Using the boundary information extraction matrix of the current independent sub-region, the boundary node state data corresponding to the boundary node is extracted from the real-time node operation data and real-time node carbon intensity of the current independent sub-region. Based on the cross-regional boundary connection relationship, define temporary boundary variables to characterize the interaction state between regions; Based on a pre-defined temporary variable association matrix, the boundary temporary variables are associated with the boundary node state data to generate the current independent sub-region's regional boundary coupling model.

5. The control method for low-carbon scheduling of multi-energy systems as described in claim 4, characterized in that, Based on the regional boundary coupling model, nodal carbon intensity constraints, and a pre-defined optimization objective function, the regional augmented Lagrangian function for the current independent sub-region is generated, including: Based on the preset optimization objective function, determine the local objective function of the current independent sub-region; Based on the region boundary coupling model, dual multiplier penalty terms and quadratic penalty terms corresponding to the current independent sub-regions are generated. The node carbon intensity constraints are linearized from non-convex constraints to generate the linearized carbon intensity constraints for the current independent sub-regions. The local objective function of the region, the dual multiplier penalty term, and the quadratic penalty term are superimposed to generate the main body of the augmented objective function of the current independent sub-region; Using the linearized carbon intensity constraint as the local solution boundary of the augmented objective function, the region augmented Lagrangian function of the current independent sub-region is generated.

6. The control method for low-carbon scheduling of multi-energy systems as described in claim 5, characterized in that, The distribution optimization calculation is performed on the augmented Lagrangian function of the region to generate the optimization results for the current independent sub-regions, including: Obtain the initial set of scheduling variables, initial boundary temporary variables, initial dual multiplier parameters, and initial penalty coefficients for the current independent subregion; Repeatedly perform the distributed optimization iteration operation until the preset distributed optimization convergence condition is met; wherein, the distributed optimization convergence condition is that the current original convergence residual is not greater than the preset original residual threshold, and the current dual convergence residual is not greater than the preset dual residual threshold. The set of target scheduling variables generated when the iteration stops is determined as the optimization result of the current independent sub-region; The distributed optimization iterative operation includes: The target scheduling variable set is generated by performing a minimum value solution based on the current baseline scheduling variable set, the current baseline boundary temporary variable, the current baseline dual multiplier parameter, the current baseline penalty coefficient, and the regional augmented Lagrangian function; wherein, the initial baseline scheduling variable set is the initial scheduling variable set; the initial baseline boundary temporary variable is the initial boundary temporary variable; the initial baseline dual multiplier parameter is the initial dual multiplier parameter; and the initial baseline penalty coefficient is the initial penalty coefficient. Generate privacy obfuscation information based on the target set of scheduling variables and the current baseline dual multiplier parameters; Cross-regional interaction processing is performed based on privacy-obfuscated information to generate temporary variables for the target boundary. Parameter update calculations are performed based on the target scheduling variable set and the target boundary temporary variables to generate the target dual multiplier parameters; The residual quantization process is performed based on the target scheduling variable set and the target boundary temporary variable to generate the current original convergence residual and the current dual convergence residual, and the target penalty coefficient is generated based on the original convergence residual and the dual convergence residual. When it is determined that the convergence condition of distributed optimization is not met, the target scheduling variable set, target boundary temporary variable, target dual multiplier parameter and target penalty coefficient are updated to the current baseline scheduling variable set, current baseline boundary temporary variable, current baseline dual multiplier parameter and current baseline penalty coefficient, respectively.

7. The control method for low-carbon scheduling of multi-energy systems as described in claim 6, characterized in that, After generating the optimization results for all independent sub-regions, the optimization results for all independent sub-regions are integrated to generate the optimal system scheduling and control instructions; Adjusting the operating output values ​​of the underlying physical devices in a multi-energy system according to the optimal system scheduling and control commands, including: After generating the optimization results for all independent sub-regions, the internal scheduling variables contained in all the optimization results are globally aggregated and concatenated to generate the optimal system scheduling and control instructions. From the optimal system scheduling and control instructions, the target operation setting parameters corresponding to each underlying physical device in the multi-energy system are extracted; The target operation settings parameters are sent to the local control terminal of the underlying physical device. Through the local control terminal, the underlying physical devices are driven to operate according to the corresponding target parameters and adjust the corresponding operating output values.

8. A control device for low-carbon dispatching of a multi-energy system, characterized in that, include: The system comprises a data acquisition module, a system preprocessing module, a boundary coupling modeling module, a distributed optimization calculation module, and a scheduling control execution module. The data acquisition module is used to acquire real-time node operation data of each node in the multi-energy system; The system preprocessing module is used to generate node carbon intensity constraints based on preset network topology parameters and preset carbon emission intensity parameters, combined with real-time node operation data. The multi-energy system is divided into regions based on preset system partitioning rules, generating multiple independent sub-regions; The boundary coupling modeling module is used to determine the neighboring sub-regions of the current independent sub-region based on network topology parameters for each independent sub-region, extract the physical connection relationship between the current independent sub-region and the neighboring sub-regions, and generate the region boundary coupling model of the current independent sub-region. The distributed optimization calculation module is used to generate the regional augmented Lagrangian function of the current independent sub-region based on the regional boundary coupling model, the node carbon intensity constraint condition and the preset optimization objective function; and to perform distributed optimization calculation on the regional augmented Lagrangian function to generate the optimization result of the current independent sub-region. The scheduling and control execution module is used to integrate the optimization results of all independent sub-regions after generating the optimization results of all independent sub-regions, and generate the optimal system scheduling and control command; and adjust the operating output value of the underlying physical equipment in the multi-energy system according to the optimal system scheduling and control command.

9. An electronic device, characterized in that, The system includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements the control method for low-carbon scheduling of a multi-energy system as described in any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium includes a stored computer program, wherein, when the computer program is executed, it controls the device where the storage medium is located to perform the control method for low-carbon scheduling of a multi-energy system as described in any one of claims 1 to 7.

Images