Tensor flow-based comprehensive energy system multi-objective optimization scheduling method and system

Abstract

The invention provides an integrated energy system multi-objective optimization scheduling method and system based on tensor flow, and the method comprises the steps: determining a plurality of optimization objectives related to a scheduling demand according to the obtained scheduling demand of an integrated energy system; according to each optimization target, constructing a tensor flow topological structure comprising a target data node, a constraint data node, a directed edge and a computing node; based on the tensor flow topological structure, performing simplification conversion on the plurality of optimization objectives to obtain a single objective function of the integrated energy system and a constraint equation thereof; based on the single objective function and the constraint equation thereof, solving the single objective function to obtain a multi-objective scheduling scheme of the integrated energy system; according to the invention, the tensor topological structure is introduced into the multi-objective optimization scheduling of the integrated energy system, so that the relation between the high-dimensional data in the integrated energy system can be expressed and described more accurately, and a more comprehensive view is provided for the scheduling of the integrated energy system.

Description

Technical Field

[0001] The present invention relates to the field of integrated energy technology for power systems, and in particular to a multi-objective optimization scheduling method and system for integrated energy systems based on tensor flow. Background Art

[0002] As global warming intensifies, the development of a low-carbon economy is gaining increasing attention. As a major carbon emitter, the power industry plays a key role in achieving overall emission reduction targets. Therefore, reducing carbon emissions from the power industry is crucial to the development of a low-carbon economy.

[0003] The evolving integrated energy system (an integrated energy production, supply, and marketing system formed through the organic coordination and optimization of energy generation, transmission and distribution, conversion, storage, and consumption during planning, construction, and operation) continues to drive changes in traditional energy utilization models and plays an increasingly important role in the power system. By constructing and solving nonlinear multi-objective optimization problems based on system requirements through multi-objective optimal scheduling of integrated energy systems (emphasizing how to maximize overall system performance when multiple energy forms coexist and interact), the operation of the integrated energy system is optimized to adapt to dynamically changing energy demands and market conditions. This enables the integrated energy system to achieve complementary utilization of various energy sources, coordinate energy conversion and use, improve energy efficiency, promote regional economic and social development, and improve residents' living standards. This has become a key strategy designed for nonlinear multi-objective optimization problems in modern complex energy systems.

[0004] Unlike traditional single-objective optimization methods, multi-objective optimization scheduling methods can effectively cope with the complexity of diverse energy supply and demand. The core lies in constructing and solving nonlinear multi-objective optimization problems based on the interaction between different behaviors of system scheduling requirements, so as to optimize the operation of the energy system and adapt to dynamically changing energy demand and market conditions. However, the existing multi-objective optimization scheduling methods for integrated energy systems are unable to quickly process the complex coupling relationship between multidimensional data and multimodal data, resulting in slow solution calculation speed and difficulty in meeting real-time scheduling requirements. Summary of the Invention

[0005] In order to solve the problem in the prior art of multi-objective optimization scheduling of integrated energy systems that the complex coupling relationship between multi-dimensional data and multi-modal data cannot be quickly processed, resulting in slow calculation speed and difficulty in meeting real-time scheduling requirements, the present invention proposes a multi-objective optimization scheduling method for integrated energy systems based on tensor flow, comprising:

[0006] Determining, based on the obtained scheduling requirements of the integrated energy system, a plurality of optimization objectives associated with the scheduling requirements;

[0007] According to each optimization goal, a tensorflow topology structure including target data nodes, constraint data nodes, directed edges and computing nodes is constructed;

[0008] Based on the tensor flow topological structure of each optimization objective, the multiple optimization objectives are converted into a single objective function of the integrated energy system and a constraint equation of the single objective function;

[0009] Based on the single objective function and the constraint equation of the single objective function, the single objective function is solved to obtain a multi-objective scheduling plan for the integrated energy system.

[0010] Optionally, determining, based on the obtained scheduling requirements of the integrated energy system, a plurality of optimization objectives associated with the scheduling requirements may include:

[0011] According to the obtained scheduling requirements of the integrated energy system, selecting evaluation indicators corresponding to the scheduling requirements from a pre-set evaluation system;

[0012] According to the evaluation index corresponding to the scheduling requirement, a plurality of optimization objectives associated with the scheduling requirement are determined.

[0013] Optionally, the evaluation system includes: economic evaluation indicators, environmental evaluation indicators and energy efficiency evaluation indicators;

[0014] The secondary indicators corresponding to the economic evaluation indicators include one or more of the following: total investment cost, investment payback period, internal rate of return, financial net present value and unit energy consumption cost;

[0015] The secondary indicators corresponding to the environmental performance evaluation index include one or more of the following: annual carbon dioxide emission reduction, annual sulfur dioxide emission reduction, annual fossil energy consumption reduction, clean energy proportion and comprehensive waste utilization rate;

[0016] The secondary indicators corresponding to the energy efficiency evaluation index include one or more of the following: comprehensive energy utilization efficiency, primary energy utilization efficiency, distributed power self-use rate and equipment utilization efficiency;

[0017] The optimization objectives include: economic objectives, environmental protection objectives and energy efficiency objectives;

[0018] The economic objectives include one or more of the following: minimizing daily operation and maintenance costs, minimizing energy procurement costs, and minimizing equipment start-up and shutdown loss costs;

[0019] The environmental protection goals include one or more of the following: minimizing carbon emissions and maximizing the proportion of clean energy;

[0020] The energy efficiency objectives include one or more of the following: maximizing comprehensive energy utilization efficiency, maximizing equipment utilization rate, and maximizing primary energy utilization efficiency.

[0021] Optionally, constructing a tensor flow topology structure including target data nodes, constraint data nodes, directed edges, and computing nodes according to each optimization objective includes:

[0022] According to each optimization goal, determining the objective function expression, constraint equation and calculation node of the optimization goal;

[0023] Using the objective function expression as the target data node of the optimization target, and using the constraint equation as the constraint data node of the optimization target;

[0024] Transmitting the target data node to the computing node via a preset directed edge of the target data node;

[0025] Transmitting the constraint data node to the computing node via a preset directed edge of the constraint data node;

[0026] A tensor flow topology structure of the optimization target is constructed according to the target data nodes, constraint data nodes, directed edges of the target data nodes, directed edges of the constraint data nodes, and computing nodes of the optimization target.

[0027] Optionally, the tensor flow topology structure based on each optimization objective performs a simplification conversion on the multiple optimization objectives to obtain a single objective function of the integrated energy system and a constraint equation of the single objective function, including:

[0028] Outputting the calculation results of the calculation nodes of each optimization target according to the tensor flow topology structure of each optimization target;

[0029] Transmitting the calculation results of the calculation nodes of the optimization objectives and the preset system constraint equations to the preset integration calculation node through the preset integration directed edges, and outputting the calculation results of the integration calculation node;

[0030] According to the calculation results of the integrated calculation node, the weight coefficient of the objective function expression of each optimization goal is calculated using the anti-entropy weight method;

[0031] Obtaining a single objective function of the integrated energy system and a constraint equation of the single objective function according to weight coefficients of the objective function expressions of the optimization objectives;

[0032] The calculation result of the calculation node of each optimization target is the objective weight relationship coefficient of each optimization target; and the calculation result of the integrated calculation node is the sum of the objective weight relationship coefficients of each optimization target.

[0033] Optionally, the calculation formula of the anti-entropy weight method is as follows:

[0034]

[0035] Among them, ω i Represents the weight coefficient of the objective function expression of the i-th optimization goal; i = 1…m; m represents the total number of optimization goals; ε i Represents the subjective weight relationship coefficient of the i-th optimization objective; δ i Represents the objective weight relationship coefficient of the i-th optimization goal; ω si represents the predetermined subjective weight value of the i-th optimization objective; ω oi represents the predetermined objective weight value of the i-th optimization goal.

[0036] Optionally, the single objective function is solved based on the single objective function and the constraint equation of the single objective function to obtain the multi-objective scheduling scheme of the integrated energy system, including:

[0037] Obtaining a deviation tensor of the integrated energy system according to the single objective function, a constraint equation of the single objective function, and pre-calculated state update tensor data of the integrated energy system;

[0038] Constructing a Jacobian matrix according to the deviation tensor of the integrated energy system, and solving the Jacobian matrix to obtain a corrected state tensor;

[0039] When the corrected state tensor meets the preset accuracy requirement, the system parameters of the integrated energy system are updated, and a multi-objective scheduling plan of the integrated energy system is output based on the updated system parameters of the integrated energy system.

[0040] Optionally, the state update tensor data includes the following calculation process:

[0041] Initializing system parameters of the integrated energy system and setting initial values ​​of the state tensors of the integrated energy system;

[0042] Calculating the thermal power, active power, and required natural gas load flow of a combined heat and power (CHP) unit in the integrated energy system based on an initial value of the state tensor of the integrated energy system;

[0043] According to the thermal power, active power and required natural gas load flow of the cogeneration heat and power (CHP) unit, the initial value of the state tensor of the integrated energy system is updated to obtain state update tensor data.

[0044] Based on the same inventive concept, the present invention also provides a multi-objective optimization scheduling system for an integrated energy system based on tensor flow, comprising:

[0045] A multi-objective determination module, configured to determine, based on the obtained scheduling requirements of the integrated energy system, a plurality of optimization objectives associated with the scheduling requirements;

[0046] The tensor flow construction module is used to construct a tensor flow topology structure containing target data nodes, constraint data nodes, directed edges and computing nodes according to each optimization goal;

[0047] A single objective conversion module, configured to perform a simplification conversion on the multiple optimization objectives based on the tensor flow topology of each optimization objective, to obtain a single objective function of the integrated energy system and a constraint equation of the single objective function;

[0048] The multi-objective scheduling module is used to solve the single objective function based on the single objective function and the constraint equation of the single objective function to obtain a multi-objective scheduling plan for the integrated energy system.

[0049] Optionally, the multi-target determination module includes:

[0050] An indicator selection submodule is used to select an evaluation indicator corresponding to the dispatching demand from a pre-set evaluation system according to the dispatching demand of the obtained integrated energy system;

[0051] The target generation submodule is used to determine multiple optimization targets associated with the scheduling requirements based on evaluation indicators corresponding to the scheduling requirements.

[0052] Optionally, the evaluation system includes: economic evaluation indicators, environmental evaluation indicators and energy efficiency evaluation indicators;

[0053] The secondary indicators corresponding to the economic evaluation indicators include one or more of the following: total investment cost, investment payback period, internal rate of return, financial net present value and unit energy consumption cost;

[0054] The secondary indicators corresponding to the environmental performance evaluation index include one or more of the following: annual carbon dioxide emission reduction, annual sulfur dioxide emission reduction, annual fossil energy consumption reduction, clean energy proportion and comprehensive waste utilization rate;

[0055] The secondary indicators corresponding to the energy efficiency evaluation index include one or more of the following: comprehensive energy utilization efficiency, primary energy utilization efficiency, distributed power self-use rate and equipment utilization efficiency;

[0056] The optimization objectives include: economic objectives, environmental protection objectives and energy efficiency objectives;

[0057] The economic objectives include one or more of the following: minimizing daily operation and maintenance costs, minimizing energy procurement costs, and minimizing equipment start-up and shutdown loss costs;

[0058] The environmental protection goals include one or more of the following: minimizing carbon emissions and maximizing the proportion of clean energy;

[0059] The energy efficiency objectives include one or more of the following: maximizing comprehensive energy utilization efficiency, maximizing equipment utilization rate, and maximizing primary energy utilization efficiency.

[0060] Optionally, the tensor flow building module includes:

[0061] A function determination submodule is used to determine the objective function expression, constraint equation and calculation node of the optimization target according to each optimization target;

[0062] A node setting submodule, configured to use the objective function expression as the target data node of the optimization target and the constraint equation as the constraint data node of the optimization target;

[0063] A target node transmission submodule, configured to transmit the target data node to the computing node via a predetermined directed edge of the target data node;

[0064] A constraint node transmission submodule, configured to transmit the constraint data node to the computing node via a preset directed edge of the constraint data node;

[0065] The topology construction submodule is used to construct the tensor flow topology structure of the optimization target according to the target data nodes, constraint data nodes, directed edges of the target data nodes, directed edges of the constraint data nodes and computing nodes of the optimization target.

[0066] Optionally, the single target conversion module includes:

[0067] A node output submodule, configured to output the calculation results of the calculation nodes of each optimization target according to the tensor flow topology structure of each optimization target;

[0068] A node integration submodule, configured to transmit the calculation results of the calculation nodes of the optimization objectives and the preset system constraint equations to a preset integration calculation node through a preset integration directed edge, and output the calculation results of the integration calculation node;

[0069] A weight distribution submodule, configured to calculate the weight coefficients of the objective function expressions of the optimization objectives using an anti-entropy weight method according to the calculation results of the integrated calculation node;

[0070] A multi-objective integration submodule, configured to obtain a single objective function of the integrated energy system and a constraint equation of the single objective function according to weight coefficients of the objective function expressions of the optimization objectives;

[0071] The calculation result of the calculation node of each optimization target is the objective weight relationship coefficient of each optimization target; and the calculation result of the integrated calculation node is the sum of the objective weight relationship coefficients of each optimization target.

[0072] Optionally, the calculation formula of the anti-entropy weight method is as follows:

[0073]

[0074] Among them, ω i Represents the weight coefficient of the objective function expression of the i-th optimization goal; i = 1…m; m represents the total number of optimization goals; ε i Represents the subjective weight relationship coefficient of the i-th optimization objective; δ i Represents the objective weight relationship coefficient of the i-th optimization goal; ω si represents the predetermined subjective weight value of the i-th optimization objective; ω oi represents the predetermined objective weight value of the i-th optimization goal.

[0075] Optionally, the multi-objective scheduling module includes:

[0076] a deviation calculation submodule, configured to update tensor data according to the single objective function, the constraint equation of the single objective function, and the pre-calculated state of the integrated energy system to obtain a deviation tensor of the integrated energy system;

[0077] a matrix solving submodule, configured to construct a Jacobian matrix according to the deviation tensor of the integrated energy system, and solve the Jacobian matrix to obtain a corrected state tensor;

[0078] The scheme generation submodule is used to update the system parameters of the integrated energy system when the corrected state tensor meets the preset accuracy requirements, and output the multi-objective scheduling scheme of the integrated energy system based on the updated system parameters of the integrated energy system.

[0079] Optionally, the multi-objective scheduling module further includes: a tensor update submodule including:

[0080] A parameter initialization unit, used to initialize the system parameters of the integrated energy system and set the initial value of the state tensor of the integrated energy system;

[0081] a correlation data calculation unit, configured to calculate the thermal power, active power, and required natural gas load flow of the cogeneration heat and power (CHP) unit in the integrated energy system according to the initial value of the state tensor of the integrated energy system;

[0082] The data updating unit is used to update the initial value of the state tensor of the integrated energy system according to the thermal power, active power and required natural gas load flow of the cogeneration heat and power CHP unit to obtain state update tensor data.

[0083] In another aspect, the present invention further provides an electronic device, comprising: at least one processor and a memory; the memory and the processor are connected via a bus;

[0084] The memory is used to store one or more programs;

[0085] When the one or more programs are executed by the at least one processor, the aforementioned multi-objective optimization scheduling method for an integrated energy system based on tensor flow is implemented.

[0086] On the other hand, the present invention also provides a computer-readable storage medium having an execution program stored thereon. When the execution program is executed, the multi-objective optimization scheduling method for an integrated energy system based on tensor flow as described above is implemented.

[0087] Compared with the prior art, the present invention has the following beneficial effects:

[0088] The present invention provides a multi-objective optimization scheduling method and system for an integrated energy system based on tensor flow, comprising: determining multiple optimization objectives associated with the obtained scheduling requirements of the integrated energy system; constructing a tensor flow topology structure including target data nodes, constraint data nodes, directed edges and computing nodes according to each optimization objective; performing a simplification conversion on the multiple optimization objectives based on the tensor flow topology structure of each optimization objective to obtain a single objective function of the integrated energy system and a constraint equation of the single objective function; solving the single objective function based on the single objective function and the constraint equation of the single objective function to obtain a multi-objective scheduling scheme for the integrated energy system; the present application constructs a tensor flow topology structure in the multi-objective optimization scheduling process, which can efficiently process the complex coupling relationship between multidimensional data and multimodal data, and is conducive to improving data organization and transmission efficiency, thereby providing a more comprehensive perspective for the scheduling of the integrated energy system. BRIEF DESCRIPTION OF THE DRAWINGS

[0089] Figure 1 A schematic diagram of a process flow of a multi-objective optimization scheduling method for an integrated energy system based on tensor flow provided by the present invention;

[0090] Figure 2A sub-connection graph of each optimization objective in a multi-objective optimization scheduling method for an integrated energy system based on tensor flow provided by the present invention;

[0091] Figure 3 A sub-connection graph for transmitting multiple optimization objectives to an integrated computing node in a multi-objective optimization scheduling method for an integrated energy system based on tensor flow provided by the present invention;

[0092] Figure 4 A sub-connection graph combining the anti-entropy weight method and the Newton-Raphson method in a multi-objective optimization scheduling method for an integrated energy system based on tensor flow provided by the present invention;

[0093] Figure 5 A comparison diagram of pipeline flow rates of a natural gas subsystem before and after a heat load change in a multi-objective optimization scheduling method for an integrated energy system based on tensor flow provided in a specific embodiment;

[0094] Figure 6 A comparison diagram of gas pressure at nodes of a natural gas subsystem before and after a heat load change in a multi-objective optimization scheduling method for an integrated energy system based on tensor flow provided in a specific embodiment;

[0095] Figure 7 A schematic diagram of the structure of a multi-objective optimization scheduling system for an integrated energy system based on tensor flow provided by the present invention;

[0096] Figure 8 This is a structural diagram of an electronic device provided by the present invention. DETAILED DESCRIPTION

[0097] The present invention proposes a multi-objective optimization scheduling method, system, device and medium for an integrated energy system based on tensor flow. The specific implementation methods of the present invention are further described in detail below with reference to the accompanying drawings.

[0098] Example 1:

[0099] The present invention provides a multi-objective optimization scheduling method for an integrated energy system based on tensor flow. The flow chart is as follows: Figure 1 As shown, including:

[0100] Step 1: Based on the obtained dispatch requirements of the integrated energy system, determine multiple optimization objectives associated with the dispatch requirements;

[0101] Step 2: Based on each optimization goal, construct a TensorFlow topology structure containing target data nodes, constraint data nodes, directed edges, and computation nodes;

[0102] Step 3: Based on the tensor flow topology of each optimization objective, multiple optimization objectives are converted into a single objective function to obtain the single objective function and constraint equation of the integrated energy system;

[0103] Step 4: Based on the single objective function and the constraint equation of the single objective function, solve the single objective function to obtain the multi-objective scheduling plan of the integrated energy system.

[0104] Generally, when performing multi-objective optimization scheduling on an integrated energy system, obtaining scheduling requirements is a crucial step. Scheduling requirements include key information such as load demand, energy supply status, and equipment operating capacity. This information ensures the matching of the optimization plan with the actual operation, thereby enhancing the effectiveness of the optimization scheduling. In the prior art, the optimization target is determined by identifying and analyzing demand changes and considering multiple factors. On this basis, a corresponding multi-objective optimization model is constructed. The model usually includes multiple objective functions, each of which reflects a specific optimization target. Constraints are added within this framework to ensure the feasibility and rationality of each indicator. However, the prior art usually relies on predefined optimization targets and fails to fully consider the complexity and diversity of the system. This makes it difficult to fully reflect the needs of different stakeholders in specific applications, such as consumers' economic considerations, environmental protection goals, and the profitability of energy suppliers. This single-dimensional goal setting may result in the system being unable to achieve the optimal operating state under multiple constraints. In order to solve the above problems, the present invention considers constructing a flexible and dynamic evaluation system and selecting highly correlated evaluation indicators from the evaluation system to determine multiple optimization targets to adapt to the ever-changing scheduling needs. Specifically:

[0105] In one implementation, the process of determining multiple optimization objectives associated with the dispatching requirements based on the obtained dispatching requirements of the integrated energy system in step 1 may include:

[0106] According to the obtained dispatching requirements of the integrated energy system, select the evaluation indicators corresponding to the dispatching requirements from the pre-set evaluation system;

[0107] Determine multiple optimization objectives associated with the scheduling requirements based on evaluation indicators corresponding to the scheduling requirements;

[0108] For example, the above-mentioned scheduling requirements may include: real-time load demand, renewable energy supply, traditional energy generation capacity, system equipment status and operating efficiency, market electricity price fluctuations, user electricity usage habits and priorities, etc. The data in this example can be obtained through various methods, such as real-time data collection from smart meters and sensors, analysis of historical data, market data monitoring, and demand forecasting through a scheduling management system;

[0109] For example, the above evaluation system may include: economic evaluation index, environmental evaluation index and energy efficiency evaluation index;

[0110] The secondary indicators corresponding to the above-mentioned economic evaluation indicators may include one or more of the following: total investment cost, investment payback period, internal rate of return, financial net present value and unit energy consumption cost;

[0111] The secondary indicators corresponding to the above-mentioned environmental protection evaluation indicators may include one or more of the following: annual carbon dioxide emission reduction, annual sulfur dioxide emission reduction, annual fossil energy consumption reduction, clean energy proportion and comprehensive waste utilization rate;

[0112] The secondary indicators corresponding to the above energy efficiency evaluation indicators may include one or more of the following: comprehensive energy utilization efficiency, primary energy utilization efficiency, distributed power self-use rate and equipment utilization efficiency;

[0113] For example, the above optimization objectives may include: economic objectives, environmental protection objectives and energy efficiency objectives; each optimization objective includes an objective optimization function and a corresponding constraint equation;

[0114] The above economic objectives may include one or more of the following: minimizing daily operation and maintenance costs, minimizing energy procurement costs, and minimizing equipment start-up and shutdown loss costs;

[0115] The above environmental protection goals may include one or more of the following: minimizing carbon emissions and maximizing the proportion of clean energy;

[0116] The above energy efficiency objectives may include one or more of the following: maximizing comprehensive energy utilization efficiency, maximizing equipment utilization, and maximizing primary energy utilization efficiency;

[0117] In this implementation, in order to achieve a balance among multiple optimization objectives, it is necessary to select appropriate variables to construct a multi-objective evaluation system. This system can include multiple primary indicators, such as economy, environmental protection, and energy efficiency, and further subdivide these primary indicators into a series of specific secondary indicators. Secondary indicators can include total investment cost, annual carbon dioxide emission reduction, comprehensive energy utilization efficiency, etc., to ensure a comprehensive evaluation and optimization of each goal. The corresponding indicator division is shown in Table 1:

[0118] Table 1 Evaluation system indicator classification table

[0119]

[0120] According to the actual dispatching needs of comprehensive energy, appropriate evaluation indicators are selected to form a multi-objective optimization problem, in which each optimization goal can be established as shown in the attached Figure 2As shown in the sub-connection graph, for example, if the optimization problem takes the daily operating cost, carbon emission cost as the minimum and the comprehensive energy efficiency as the optimization objectives, three sub-connection graphs need to be established.

[0121] The above implementation method can significantly improve the optimization ability of the integrated energy system by systematically setting up an evaluation system, including multi-dimensional indicators of economy, environmental protection and energy efficiency. For example, the selected economic evaluation indicators can effectively evaluate the investment return and operating costs of the project, ensuring the economic feasibility of the system; the selected environmental evaluation indicators focus on carbon emissions and clean energy utilization, which is conducive to achieving the goal of environmental sustainable development; the setting of the selected energy efficiency evaluation indicators emphasizes energy utilization efficiency, which can further promote technological progress and rational allocation of resources; each optimization goal is made more scientific and reasonable in the formulation process through a clear target optimization function and corresponding constraints, effectively combining different stakeholders. This multi-level and multi-dimensional optimization approach enables the integrated energy system to respond flexibly to complex scheduling demands and market fluctuations, thereby improving the overall economic operation efficiency and environmental friendliness. This implementation method integrates the indicator systems of economy, environmental protection and energy efficiency to form a comprehensive evaluation structure, which is conducive to improving the coordination between various goals. This complex combination of indicators can not only cope with the single-dimensional challenges faced by traditional technical solutions, but also deal with the mutual constraints and comprehensive optimization of multiple goals in real applications. Moreover, this implementation method can enhance the system's resilience and decision-making accuracy in response to dynamic market demands, energy supply fluctuations and environmental policy pressures.

[0122] After determining the optimization objectives in the above steps, we can further consider concretizing these objectives into operational computational models. For example, we can build a comprehensive tensor flow topology for each optimization objective so that the integrated energy system can efficiently handle various data interactions and computational requirements related to the optimization process. Specifically:

[0123] In one implementation, the process of constructing a tensorflow topology structure including target data nodes, constraint data nodes, directed edges, and computation nodes according to each optimization objective in step 2 may include:

[0124] According to each optimization goal, determine the objective function expression, constraint equation and calculation node of the optimization goal;

[0125] The objective function expression is used as the target data node of the optimization target, and the constraint equation is used as the constraint data node of the optimization target;

[0126] The target data node is transferred to the computing node through the directed edge of the preset target data node;

[0127] Transmitting the constraint data nodes to the computing nodes through the preset directed edges of the constraint data nodes;

[0128] Construct a tensor flow topology structure of the optimization target based on the target data nodes, constraint data nodes, directed edges of the target data nodes, directed edges of the constraint data nodes, and computing nodes of the optimization target;

[0129] The concept of a tensor was first proposed by William Ron Hamilton in 1846. It is essentially a high-dimensional array, where the rank of a tensor is the dimension of its space. A 0th-order tensor is also called a scalar, a vector is a 1st-order tensor, a 2nd-order tensor is a matrix, a 3rd-order tensor is a space, a cumulative stack of multiple 2nd-order tensors, and so on. Considering that integrated energy systems involve multiple energy forms, multiple variables, and complex temporal relationships, traditional computational methods may not be able to effectively process multidimensional and multimodal data. Therefore, introducing tensors into multi-objective optimization scheduling methods can more accurately express and describe the data relationships in these complex systems. Tensors can not only process high-dimensional data but also capture high-order relationships between data, thus providing a more comprehensive perspective for integrated energy system scheduling.

[0130] In this implementation, the computational topology of the algorithm model is described by tensor flow, where each node is an abstract function mapping or mathematical expression, such as Figure 2 As shown in the figure, the data input of each optimization objective consists of two data nodes (for example, the target data node and the constraint data node mentioned above), two directed edges (one corresponding to the target data node and one corresponding to the constraint data node), and one computation node. The specific functions of each module are as follows:

[0131] (1) Data node: This node is used to describe the attributes of input data. In the present invention, data node a11 is used to record the objective function expression of the optimization target, and data node a21 is used to record the constraint equation of the optimization target. For example, if the objective function expression and constraint equation are designed with the goal of minimizing daily operation and maintenance cost, the expression of data node a11 can be:

[0132]

[0133] Where: C represents the daily operation and maintenance cost of the integrated energy system; j represents the type of equipment in the integrated energy system, including various energy conversion equipment and energy storage equipment; c j The operation and maintenance costs required for the energy storage device j to output unit power; is the actual output power of energy storage device j in unit time t; t = 1…T; T is the total time.

[0134] Correspondingly, the input of data node a21 is the constraint equation for the lowest daily operation and maintenance cost, which includes the operation constraints of the corresponding equipment. For example, the constraints of the energy storage equipment in the energy storage device are:

[0135] 1) Power balance constraints:

[0136]

[0137] Where: η j is the natural loss rate of energy storage device j; η c,j is the charging efficiency of energy storage device j; η f,j is the discharge efficiency of energy storage device j; is the actual capacity of energy storage device j in unit time t; is the actual charging power of energy storage device j in unit time t; is the actual discharge power of energy storage device j in unit time t; Δt is the time interval; S SOC It is the total SOC equipment collection of the integrated energy system.

[0138] 2) Energy storage equipment capacity constraints:

[0139]

[0140] Where: E j is the minimum value of the capacity of energy storage device j; is the maximum capacity of energy storage device j; E′ is the actual capacity of energy storage device j.

[0141] 3) Energy storage equipment charging and discharging power constraints:

[0142]

[0143] Where: represents the minimum charging power of energy storage device j in unit time t; represents the maximum charging power of energy storage device j in unit time t; represents the actual charging power of energy storage device j in unit time t; It represents the minimum discharge power of energy storage device j in unit time t; It represents the maximum discharge power of energy storage device j in unit time t; represents the actual discharge power of energy storage device j per unit time t;

[0144] (2) Directed edges: Directed edges represent data dependencies and usually transmit data in the form of tensors. By connecting two nodes, the output of the previous node becomes the input of the next node. In this implementation, directed edge a11 (i.e., the directed edge of the target data node) transmits the objective function expression of data node a11 (i.e., the target data node) to computation node a1, and directed edge a21 (i.e., the directed edge of the constraint data node) transmits the constraint equation of data node a21 (i.e., the constraint data node) to computation node a1.

[0145] (3) Computational nodes: These nodes often contain stateless computation or control operations and are primarily responsible for algorithmic logic expression or process control, such as logical operations, neural network operations, and equation solving operations. Considering that data nodes can obtain input data through Excel file import, URL reading, and other methods, and that data formats are diverse, computational node a1 reads various types of data formats and uniformly converts them into tensor expressions as data output to facilitate subsequent computational operations and ultimately determine the multi-objective type of the optimization problem.

[0146] This implementation efficiently handles the complex data relationships involved in the multi-objective optimization and scheduling of integrated energy systems by constructing a tensor flow topology consisting of target data nodes, constraint data nodes, directed edges, and compute nodes. The target data nodes store the objective function expression of the optimization objective, while the constraint data nodes store the constraint equations. These are transmitted to the compute nodes via directed edges, enabling modular organization and dynamic transmission of the objective function and constraint conditions. The compute nodes, acting as core processing units, convert the objective function and constraint equations into a tensor format, resolving the difficulty of integrating heterogeneous data in traditional approaches. This tensor flow topology not only efficiently processes multidimensional data (such as coupled energy flows of electricity, heat, and natural gas) but also captures high-order relationships between the data, significantly improving data organization and transmission efficiency. Furthermore, by connecting data nodes and compute nodes through directed edges, data can be automatically flowed and dynamically updated, avoiding the time-consuming and error-prone manual data integration required in traditional approaches. In integrated energy systems, optimization objectives often involve coupled energy flows such as electricity, heat, and natural gas, with complex and dynamically changing constraints. Traditional approaches struggle to efficiently handle these high-dimensional, multimodal data relationships. This solution addresses the inefficient data integration and redundant constraint processing issues inherent in traditional approaches by building a tensor flow topology that modularly stores objective functions and constraint equations and automatically transfers data via directed edges. Furthermore, compute nodes uniformly convert heterogeneous data into a tensor format, addressing the complex coupling of multidimensional and multimodal data.

[0147] The above steps, by constructing a tensor flow topology structure, can not only effectively organize the objective functions and constraint equations of each optimization goal, but also provide a clear path for subsequent calculations. The design of this structure enables efficient extraction and processing of the calculation information of each goal when performing multi-objective optimization. On this basis, it is possible to consider combining the anti-entropy weight method to perform a unified conversion of multiple optimization goals to form a unified objective function and corresponding constraint equations, thereby achieving the optimization of the efficiency of the integrated energy system. Specifically:

[0148] In one implementation, the process of performing a simplification conversion on multiple optimization objectives based on the tensor flow topology of each optimization objective in step 3 to obtain a single objective function of the integrated energy system and a constraint equation of the single objective function may include:

[0149] According to the tensor flow topology of each optimization target, the calculation results of the computing nodes of each optimization target are output;

[0150] The calculation results of the calculation nodes of each optimization target and the preset system constraint equation are transmitted to the preset integration calculation node through the preset integration directed edge, and the calculation results of the integration calculation node are output;

[0151] According to the calculation results of the integrated calculation nodes, the anti-entropy weight method is used to calculate the weight coefficient of the objective function expression of each optimization goal;

[0152] According to the weight coefficients of the objective function expressions of the optimization objectives, the single objective function of the integrated energy system and the constraint equation of the single objective function are obtained;

[0153] The calculation result of the calculation node of each optimization target is the objective weight relationship coefficient of each optimization target; the calculation result of the integrated calculation node is the sum of the objective weight relationship coefficients of each optimization target;

[0154] In this implementation, as shown in the attached Figure 3 As shown in the figure, the main part includes a data node, two directed edges, and two computation nodes. The data node b1 is used to input system constraint equations such as power system constraints and thermal network constraints. The power system constraints can include:

[0155] 1) Real-time power balance:

[0156]

[0157] Where: is the electrical load power of node k in unit time t; is the power supply power of node k in unit time t.

[0158] 2) Transmission constraints:

[0159]

[0160] Where: l li,t It represents the transmission power of transmission line li in unit time t; represents the transmission factor from the load node j0 to the corresponding transmission line li; j0∈S K ;S K Represents the numbered set of all load nodes in the power subsystem; represents the net injected power of load node j0 in unit time t; represents the local fixed load power of load node j0 in unit time t; S L It is a set with all node numbers of the power subsystem as elements.

[0161] 3) Capacity constraints of transmission lines:

[0162]

[0163] Where: l li,t It represents the upper limit of the forward power of the transmission line li in unit time t; It represents the upper limit of the reverse power of the transmission line li in unit time t; represents the actual power flow from the load node j0 to the corresponding transmission line li; represents the net injected power of load node j0 in unit time t; represents the local fixed load power of load node j0 at unit time t; j0∈S K ;S K Represents the numbered set of all load nodes in the power subsystem; S E represents a collection of transmission lines.

[0164] Considering that the equipment included in each optimization target in the integrated energy system is repeated, for example, the two optimization targets of minimum daily operation and maintenance cost and maximum comprehensive energy efficiency both include energy storage equipment, so the constraint equations of both include the power balance constraint, capacity constraint and charge and discharge power constraint of energy storage equipment. In order to avoid the impact of repeated judgment constraints on the overall calculation time and improve the efficiency of subsequent energy flow calculation, the computing node b1 obtains the computing nodes in step 2 (which can be expressed as a i Represents the calculation result of the i-th optimization objective), deletes the repeated constraint equations of the device, adds the system constraint equation obtained by data node b1, and finally transmits the sorted objective function expression and constraint equation to computing node b2 through directed edge b2.

[0165] The anti-entropy weight method is an objective weight determination method obtained by improving the entropy weight method. The greater the difference in indicators, the smaller the entropy value obtained and the larger the weight coefficient. This method can improve the degree of uncertainty in measurement and reduce the volatility of weight values. The specific calculation steps are as follows:

[0166]

[0167] Where: d i Represents the anti-entropy value of the i-th optimization objective; ω oi represents the predetermined objective weight value of the i-th optimization goal; Indicates the network index value corresponding to the i-th optimization target in the j1-th sample Normalized value of; j1 = 1…n; n represents the total number of samples; i = 1…m; m represents the total number of optimization targets;

[0168] Since the importance of subjective weights is inconsistent with that of objective weights, this paper uses the basic theory of matrix theory to find the relative importance of subjective and objective weights, and then finds the final weight value. Assuming that the relative importance of subjective weights and objective weights are ε and δ respectively, the subjective weight relationship coefficient ε of the i-th optimization objective is i and objective weight relationship coefficient δ i They can be expressed as follows:

[0169]

[0170] Where: ω zi represents the intermediate weight of the i-th optimization objective; ω oi represents the objective weight value of the i-th optimization goal; k0 is the dimension of the normalized feature vector. For example, if the optimization goals are “cost”, “reliability”, and “environmental protection”, then k0 = 3; ω si represents the predetermined subjective weight value of the i-th optimization objective.

[0171] Combined with the calculated weight relationship coefficient, the final combined weight can be expressed as:

[0172]

[0173] Among them, ω i Represents the weight coefficient of the objective function expression of the i-th optimization goal; i = 1…m; m represents the total number of optimization goals; ε i Represents the subjective weight relationship coefficient of the i-th optimization objective; δ i Represents the objective weight relationship coefficient of the i-th optimization goal; ω si represents the predetermined subjective weight value of the i-th optimization objective; ω oirepresents the predetermined objective weight value of the i-th optimization goal;

[0174] When the anti-entropy weight method is used to convert the multi-objective optimization problem into a single-objective optimization problem, the anti-entropy weight method is used by computing node b2 to determine the weight of each optimization objective. The final expression (assuming it contains three optimization objectives) can be expressed as:

[0175] F=ω1*F1+ω2*F2+ω3*F3;

[0176] Where: ω1 represents the weight coefficient of the first optimization target obtained by anti-entropy method; F1 represents the objective function expression of the first optimization target; ω2 represents the weight coefficient of the second optimization target obtained by anti-entropy method; F2 represents the objective function expression of the second optimization target; ω3 represents the weight coefficient of the third optimization target obtained by anti-entropy method; F3 represents the objective function expression of the third optimization target;

[0177] In the above implementation, the calculation node of each optimization target outputs its calculation result (i.e., the objective weight relationship coefficient), and transmits the calculation result and the system constraint equation to the integration calculation node by integrating directed edges. The integration calculation node summarizes the objective weight relationship coefficient of each optimization target and uses the anti-entropy weight method to calculate the weight coefficient of each optimization target. This data transmission and integration mechanism based on the tensor flow topology structure of each optimization target can solve the problem that the multi-target weight distribution in the traditional method relies on manual experience and is difficult to adjust dynamically. The anti-entropy weight method realizes the scientific distribution of multi-target weights by quantifying the objective weight relationship coefficient of each target, avoiding subjective bias. Finally, a single objective function and constraint equation are generated according to the weight coefficient, which can significantly reduce the complexity of the optimization problem. Although the anti-entropy weight method and data integration The concept may be used alone in the existing technology, but combining it with the tensor flow topology structure and applying it to the multi-objective optimization scheduling scenario of the integrated energy system is not an easy solution. Especially in the integrated energy system, the optimization objectives usually involve multiple dimensions such as economy, environmental protection, and energy efficiency. There are complex coupling relationships between the objectives. Traditional methods are difficult to efficiently handle this multi-objective weight allocation problem. This implementation method dynamically integrates the calculation results of each optimization objective through the tensor flow topology structure, and uses the anti-entropy weight method to scientifically calculate the weight coefficient, which solves the problem that the weight allocation in the traditional method is highly subjective and difficult to dynamically adjust; and the integrated computing node can realize the automatic calculation of multi-objective weights by summarizing the objective weight relationship coefficients of each optimization objective, avoiding errors caused by manual intervention. In addition, this implementation method can also add technical features for redundant constraint identification and elimination to further improve the efficiency of multi-objective optimization scheduling. In the integrated computing node, through the preset constraint identification algorithm (such as constraint redundancy detection based on graph theory, constraint conflict detection based on matrix rank analysis), the repeated constraints between the optimization objectives (such as the power balance constraints of energy storage equipment) are automatically detected and eliminated or merged. This technical feature can significantly reduce the number of constraints in the optimization problem and reduce the computational complexity, thereby speeding up the solution. The identification and elimination of redundant constraints can also avoid the waste of resources caused by repeated calculations, further improving computing efficiency. The combination of this technical feature with the anti-entropy weight method and the tensor flow topology structure is conducive to the scientific distribution of multi-objective weights and can also optimize the constraints.

[0178] In the above steps, the anti-entropy weight method is used to transform multiple optimization objectives in the integrated energy system into a single problem, and finally a single objective function and constraint equation are obtained. This is conducive to accurately determining the importance of each optimization objective in the overall objective, and can also effectively integrate different indicators to form a unified decision-making basis. On this basis, the Newton-Raphson method can be considered to solve the single objective function, thereby obtaining a multi-objective scheduling plan for the integrated energy system, providing a practical solution for practical applications. Specifically:

[0179] In one implementation, the process of solving the single objective function based on the single objective function and the constraint equation of the single objective function in step 4 to obtain the multi-objective scheduling solution for the integrated energy system may include:

[0180] updating tensor data according to the single objective function, the constraint equation of the single objective function, and the pre-calculated state of the integrated energy system to obtain a deviation tensor of the integrated energy system;

[0181] According to the deviation tensor of the integrated energy system, the Jacobian matrix is ​​constructed and solved to obtain the corrected state tensor;

[0182] When the corrected state tensor meets the preset accuracy requirement, the system parameters of the integrated energy system are updated, and a multi-objective scheduling plan of the integrated energy system is output based on the updated system parameters of the integrated energy system.

[0183] In this implementation, the above-mentioned state update tensor data may include the following calculation process:

[0184] Initialize the system parameters of the integrated energy system and set the initial value of the state tensor of the integrated energy system;

[0185] Based on the initial value of the state tensor of the integrated energy system, the thermal power, active power and required natural gas load flow of the cogeneration CHP unit in the integrated energy system are calculated;

[0186] According to the thermal power, active power and required natural gas load flow of the cogeneration heat and power (CHP) unit, the initial value of the state tensor of the integrated energy system is updated to obtain the state update tensor data.

[0187] Specifically, as attached Figure 4 As shown, the main part of the implementation method includes two directed edges (directed edges c1 and c2) and a computing node c1, wherein the directed edge c1 connects the computing node b2 of step 2, and transmits the preprocessed single-objective optimization problem and constraint equation to the computing node c1; the computing node c1 performs energy flow calculation based on the Newton-Raphson method, and transmits the calculation results to the output node through another directed edge.

[0188] The basic principle of using the Newton-Raphson method to solve the objective function in computing node c1 is as follows:

[0189] Remember the initial value of the state tensor is x (0) , then the Taylor expansion approximation is:

[0190] F(x (0) +Δx)≈F(x (0) )+J (0) Δx=0;

[0191] Where: F represents the system deviation tensor; Δx represents the correction state tensor; F(x (0) ) represents the system deviation tensor of the initial value of the state tensor; J (0) For F(x) in the state tensor x=x (0) The Jacobian Matrix when .

[0192] F(x (0) )≈-J (0) Δx;

[0193] Δx≈-[J (0) ] -1 F(x (0) );

[0194] Note Δx (0) =-[J (0) ] -1 F(x (0) ), then the new solution of the state tensor is:

[0195] x (1) =x (0) +Δx (0) ;

[0196] x (1) Use it as the new initial value to continue iterating and complete the calculation and solution.

[0197] The corresponding node calculation steps are:

[0198] (1) Initialize system parameters and set the initial value of the integrated energy system state tensor;

[0199] (2) Calculate the thermal power generated by the CHP unit, the active power generated in the power system, and the required natural gas load flow, and update the relevant tensor data;

[0200] (3) Based on the single-objective optimization function and constraint equations obtained in step 3, calculate the system deviation tensor F;

[0201] (4) Calculate the partial derivatives of the power, heat, and natural gas subsystems to obtain the tensor form of the Jacobian matrix J and solve it to obtain the modified state tensor Δx;

[0202] (5) Update the state tensor x, the expression is x (1) =x (0) +Δx;

[0203] (6) Determine whether the system state tensor meets the iteration accuracy. If Δx < λ, where λ represents the preset accuracy requirement, stop the iteration; otherwise, repeat steps (2) to (5) above.

[0204] (7) Solve other relevant parameters of the integrated energy system and output a multi-objective scheduling plan, where the multi-objective scheduling plan may include: energy flow distribution results, operating cost data (including fuel costs, electricity purchase and sales costs, operation and maintenance costs, etc.), environmental impact data (carbon emissions, sulfur dioxide emissions and other environmental indicators), efficiency indicator data (primary energy utilization efficiency, equipment utilization efficiency), constraint satisfaction (such as power balance constraints, capacity constraints, charging and discharging power constraints, etc.) and resource allocation results.

[0205] In this implementation, based on a single objective function, constraint equations, and state update tensor data, the deviation tensor of the integrated energy system is calculated to quantify the gap between the current state and the target state. A Jacobian matrix containing the coupled partial derivatives of the power, thermal, and natural gas subsystems is then constructed. Solving this matrix yields a revised state tensor to guide the update of state variables. This tensor-based deviation calculation and matrix solution mechanism addresses the low efficiency and slow convergence of traditional methods for solving high-dimensional nonlinear problems. The state update tensor data calculation process initializes system parameters, calculates the thermal power, active power, and natural gas load flow of the CHP units, and updates the initial state tensor value, enabling dynamic modeling of multiple energy flow coupling. This technical approach not only accurately captures the operating state of the integrated energy system but also rapidly approaches the optimal solution through iterative updates, significantly improving solution accuracy and computational efficiency. Ultimately, when the revised state tensor meets the preset accuracy, a multi-objective scheduling solution is output, providing reliable technical support for the economic operation and low-carbon scheduling of the integrated energy system. Although the Newton-Raphson method and the concept of state update used in this implementation may be used separately in existing technologies, combining them with tensor data and applying them to multi-objective optimization scheduling scenarios of integrated energy systems is not an easy-to-think-of solution. In integrated energy systems, optimization problems usually involve the coupling of multiple energy flows such as electricity, heat, and natural gas. The constraints are complex and dynamically changing. Traditional methods are difficult to efficiently handle such high-dimensional nonlinear problems. This solution dynamically models the multi-energy flow output of the CHP unit through state update tensor data, and uses the Jacobian matrix to quantify the relationship between the deviation and the state variable, which can solve the problems of low solution efficiency and slow convergence speed in traditional methods. In addition, the deviation calculation and matrix solution mechanism based on tensor data can significantly improve the calculation accuracy and efficiency, providing reliable technical support for the real-time scheduling of large-scale integrated energy systems.

[0206] In summary, the present invention addresses the problem that in the prior art, when performing multi-objective optimization scheduling on an integrated energy system, the complex coupling relationship between multidimensional data and multimodal data cannot be quickly processed, resulting in slow calculation speed and inability to meet real-time scheduling requirements. The present invention proposes a multi-objective optimization scheduling method for an integrated energy system based on tensor flow, which realizes multi-objective optimization scheduling of the integrated energy system through tensor flow. First, based on the scheduling requirements, the integrated energy system evaluation system such as environmental protection, energy efficiency, and economy is used as the optimization target, and the objective function expression and constraint equation of each optimization target are determined as the multi-node input of the tensor flow. Secondly, the objective function and constraint equation are transferred to the preprocessing node using directed edges. The node converts the multi-objective optimization into a single-objective optimization based on the anti-entropy weight method. Finally, the preprocessed optimization problem is also transferred to the computing node using directed edges. The node performs energy flow calculation using the Newton-Raphson method and uses the result as the output of the tensor flow. The technical solution used in the present invention has strong adaptability and flexibility, and has significant technical effects and application value in improving system operation efficiency, reducing operating costs, and reducing environmental impact.

[0207] Example 2:

[0208] A specific example illustrates the proposed multi-objective optimization scheduling method for an integrated energy system based on tensor flow. For example, a nine-node heating network in northern my country is selected. The heating demand is indoor heating for buildings. Based on user comfort, the heat load is adjustable within a 10% range, while the return water temperature is limited to 75-95°C. Node N1 is represented as the heat source node, and nodes N2-N6 are represented as heat exchange stations. Mass regulation is used, and the heat network flow is fixed. The cooling demand is for cold storage, and the indoor temperature limit for the cold storage is -14 to -16°C.

[0209] The embodiment takes the minimum daily operating cost as the optimization scheduling requirement and constructs a multi-objective optimization problem of minimizing the sum of energy cost, operation and maintenance cost, and start-up and shutdown cost (daily operating cost = the sum of energy cost + operation and maintenance cost + start-up and shutdown cost).

[0210] (2) Analysis of the results of the embodiment

[0211] To verify the effectiveness of the multi-objective optimization scheduling method of the present invention, two scheduling scenarios are proposed for comparison: 1) Scenario I, normal scheduling, without considering virtual energy storage in the hot and cold networks; and 2) Scenario II, taking virtual energy storage into account and utilizing the method of the present invention.

[0212] Table 2 shows the system operating costs for each of the two dispatch methods over a single day. Scenario II saves 22,336 yuan in total operating costs compared to Scenario I during the daily dispatch cycle. Of this, the cost of electricity purchase and sale from the upper-level power grid in Scenario II is 35,967 yuan, 69.8% less than in Scenario I, contributing significantly to the cost reduction.

[0213] Table 2 Operation costs under two scheduling modes

[0214]

[0215] By attaching Figure 5 -Attached Figure 6 The results show the heating and cooling output of the heating and cooling grid systems under two scheduling methods. In Scenario I, the output of the heating and cooling grid systems is always equal to the system's heating and cooling loads. In Scenario II, the output of both the heating and cooling grids is decoupled from the load and is no longer always equal. Specifically, in the heating grid, from 10:00 PM to 8:00 AM, the system's thermal unit output is low, and the thermal output is less than the thermal load. During this period, the virtual energy storage system releases energy, and the heat energy stored in the heating grid during other periods fills this imbalance. After 9:00 AM, unit output increases, and the system's thermal output exceeds the thermal load. This imbalance is then stored in the heating grid, completing the energy storage function of the virtual energy storage system. Similarly, in the cooling grid, due to fluctuations in electricity prices and ambient temperature, the output of the cooling grid units fluctuates more frequently. When the cooling output is not equal to the cooling load, energy is stored in the cooling grid, or the shortfall is filled by energy stored in the cooling grid from a previous period.

[0216] This specific embodiment illustrates that the multi-objective optimization scheduling method for an integrated energy system based on tensor flow provided by the present invention can achieve decoupling between output and load, and play the role of virtual energy storage. This decoupling mechanism enables the integrated energy system to flexibly adjust the supply and demand balance of energy in different time periods, avoiding waste and improving energy efficiency. In addition, because it can adapt to the fluctuations in electricity prices and ambient temperature, the entire integrated energy system can operate more economically and efficiently. Therefore, it can be illustrated that the scheduling method of the present invention can achieve a dual improvement in economy and energy efficiency through comprehensive management and technical integration, providing an excellent solution and practical value for regional energy management.

[0217] Example 3:

[0218] The present invention based on the same inventive concept also provides a multi-objective optimization scheduling system for an integrated energy system based on tensor flow, the structural composition diagram is as follows: Figure 7 As shown, including:

[0219] A multi-objective determination module is used to determine multiple optimization objectives associated with the dispatching requirements based on the obtained dispatching requirements of the integrated energy system;

[0220] The tensor flow construction module is used to construct a tensor flow topology structure containing target data nodes, constraint data nodes, directed edges and computing nodes according to each optimization goal;

[0221] The single objective conversion module is used to convert multiple optimization objectives into a single function based on the tensor flow topology of each optimization objective, and obtain the single objective function and constraint equation of the single objective function of the integrated energy system;

[0222] The multi-objective scheduling module is used to solve the single objective function based on the single objective function and the constraint equation of the single objective function to obtain the multi-objective scheduling plan of the integrated energy system.

[0223] In one implementation, the multi-objective determination module may include:

[0224] The indicator selection submodule is used to select evaluation indicators corresponding to the scheduling requirements from a pre-set evaluation system based on the obtained scheduling requirements of the integrated energy system;

[0225] The target generation submodule is used to determine multiple optimization targets associated with the scheduling requirements based on the evaluation indicators corresponding to the scheduling requirements.

[0226] For example, the above evaluation system may include: economic evaluation index, environmental evaluation index and energy efficiency evaluation index;

[0227] The secondary indicators corresponding to the above-mentioned economic evaluation indicators may include one or more of the following: total investment cost, investment payback period, internal rate of return, financial net present value and unit energy consumption cost;

[0228] The secondary indicators corresponding to the above-mentioned environmental protection evaluation indicators may include one or more of the following: annual carbon dioxide emission reduction, annual sulfur dioxide emission reduction, annual fossil energy consumption reduction, clean energy proportion and comprehensive waste utilization rate;

[0229] The secondary indicators corresponding to the above energy efficiency evaluation indicators may include one or more of the following: comprehensive energy utilization efficiency, primary energy utilization efficiency, distributed power self-use rate and equipment utilization efficiency;

[0230] The above optimization objectives may include: economic objectives, environmental protection objectives and energy efficiency objectives;

[0231] The above economic objectives may include one or more of the following: minimizing daily operation and maintenance costs, minimizing energy procurement costs, and minimizing equipment start-up and shutdown loss costs;

[0232] The above environmental protection goals may include one or more of the following: minimizing carbon emissions and maximizing the proportion of clean energy;

[0233] The above energy efficiency objectives may include one or more of the following: maximizing comprehensive energy utilization efficiency, maximizing equipment utilization, and maximizing primary energy utilization efficiency.

[0234] In one implementation, the tensorflow building block may include:

[0235] The function determination submodule is used to determine the objective function expression, constraint equation and calculation node of the optimization target according to each optimization target;

[0236] The node setting submodule is used to use the objective function expression as the target data node of the optimization target and the constraint equation as the constraint data node of the optimization target;

[0237] The target node transmission submodule is used to transmit the target data node to the computing node through the directed edge of the preset target data node;

[0238] The constraint node transmission submodule is used to transmit the constraint data node to the computing node through the directed edge of the preset constraint data node;

[0239] The topology construction submodule is used to construct the tensor flow topology structure of the optimization target based on the target data nodes, constraint data nodes, directed edges of the target data nodes, directed edges of the constraint data nodes, and computing nodes of the optimization target.

[0240] In one implementation, the above-mentioned single target conversion module may include:

[0241] The node output submodule is used to output the calculation results of the computing nodes of each optimization target according to the tensor flow topology structure of each optimization target;

[0242] The node integration submodule is used to transmit the calculation results of the calculation nodes of each optimization target and the preset system constraint equation to the preset integration calculation node through the preset integration directed edge, and output the calculation results of the integration calculation node;

[0243] The weight distribution submodule is used to calculate the weight coefficient of the objective function expression of each optimization target based on the calculation results of the integrated calculation nodes and the anti-entropy weight method;

[0244] The multi-objective integration submodule is used to obtain the single objective function and the constraint equation of the single objective function of the integrated energy system according to the weight coefficient of the objective function expression of each optimization objective;

[0245] Among them, the calculation result of the calculation node of each optimization target is the objective weight relationship coefficient of each optimization target; the calculation result of the integrated calculation node is the sum of the objective weight relationship coefficients of each optimization target.

[0246] For example, the calculation formula of the above anti-entropy weight method can be as follows:

[0247]

[0248] Among them, ω i Represents the weight coefficient of the objective function expression of the i-th optimization goal; i = 1…m; m represents the total number of optimization goals; ε i Represents the subjective weight relationship coefficient of the i-th optimization objective; δ i Represents the objective weight relationship coefficient of the i-th optimization goal; ω si represents the predetermined subjective weight value of the i-th optimization objective; ω oi represents the predetermined objective weight value of the i-th optimization goal.

[0249] In one implementation, the multi-objective scheduling module may include:

[0250] a deviation calculation submodule, configured to update tensor data according to a single objective function, a constraint equation of the single objective function, and a pre-calculated state of the integrated energy system to obtain a deviation tensor of the integrated energy system;

[0251] The matrix solving submodule is used to construct the Jacobian matrix based on the deviation tensor of the integrated energy system and solve the Jacobian matrix to obtain the corrected state tensor;

[0252] The scheme generation submodule is used to update the system parameters of the integrated energy system when the corrected state tensor meets the preset accuracy requirements, and output the multi-objective scheduling scheme of the integrated energy system based on the updated system parameters of the integrated energy system.

[0253] In one implementation, the multi-objective scheduling module may further include a tensor update submodule, which may specifically include:

[0254] A parameter initialization unit, used to initialize the system parameters of the integrated energy system and set the initial value of the state tensor of the integrated energy system;

[0255] a correlation data calculation unit, configured to calculate the thermal power, active power, and required natural gas load flow of the cogeneration heat and power (CHP) unit in the integrated energy system according to the initial value of the state tensor of the integrated energy system;

[0256] The data updating unit is used to update the initial value of the state tensor of the integrated energy system according to the thermal power, active power and required natural gas load flow of the cogeneration heat and power CHP unit to obtain state update tensor data.

[0257] Example 4:

[0258] like Figure 8As shown, the present invention also provides an electronic device, which may be a computer, a single-chip microcomputer, a smart mobile device, or the like. The electronic device in this embodiment may include a processor, a memory, a transceiver component, and the like. The memory, processor, and transceiver component are connected via a bus; the memory may be used to store an execution program, which may include instructions; and the processor may be used to execute the instructions stored in the memory. The memory may also be used to store data, which may be accessed and / or modified during the execution of the instructions.

[0259] The processor may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in the storage medium to implement the corresponding method flow or corresponding function, so as to realize the steps of a multi-objective optimization scheduling method of an integrated energy system based on tensor flow in the above embodiment.

[0260] Example 5:

[0261] Based on the same inventive concept, the present invention also provides a readable storage medium, specifically an electronic device readable storage medium (Memory), which is a memory device in an electronic device for storing programs and data. It can be understood that the storage medium here can include both built-in storage media in the electronic device and, of course, extended storage media supported by the electronic device. The storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space. These instructions can be one or more execution programs (including program codes). It should be noted that the storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor loads and executes one or more instructions stored in the storage medium, which can implement the steps of a multi-objective optimization scheduling method for an integrated energy system based on tensor flow in the above embodiment.

[0262] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0263] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0264] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0265] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0266] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit its scope of protection. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that after reading the present invention, those skilled in the art may still make various changes, modifications or equivalent substitutions to the specific implementation methods of the application, but these changes, modifications or equivalent substitutions are all within the scope of protection of the pending claims.

Claims

1. A multi-objective optimization scheduling method for an integrated energy system based on tensor flow, characterized in that: include: Determining, based on the obtained scheduling requirements of the integrated energy system, a plurality of optimization objectives associated with the scheduling requirements; According to each optimization goal, a tensorflow topology structure including target data nodes, constraint data nodes, directed edges and computing nodes is constructed; Based on the tensor flow topological structure of each optimization objective, the multiple optimization objectives are converted into a single objective function of the integrated energy system and a constraint equation of the single objective function; Based on the single objective function and the constraint equation of the single objective function, the single objective function is solved to obtain a multi-objective scheduling plan for the integrated energy system.

2. The method according to claim 1, wherein The step of determining, based on the obtained scheduling requirements of the integrated energy system, a plurality of optimization objectives associated with the scheduling requirements includes: According to the obtained scheduling requirements of the integrated energy system, selecting evaluation indicators corresponding to the scheduling requirements from a pre-set evaluation system; According to the evaluation index corresponding to the scheduling requirement, a plurality of optimization objectives associated with the scheduling requirement are determined.

3. The method according to claim 2, wherein The evaluation system includes: economic evaluation indicators, environmental evaluation indicators and energy efficiency evaluation indicators; The secondary indicators corresponding to the economic evaluation indicators include one or more of the following: total investment cost, investment payback period, internal rate of return, financial net present value and unit energy consumption cost; The secondary indicators corresponding to the environmental performance evaluation index include one or more of the following: annual carbon dioxide emission reduction, annual sulfur dioxide emission reduction, annual fossil energy consumption reduction, clean energy proportion and comprehensive waste utilization rate; The secondary indicators corresponding to the energy efficiency evaluation index include one or more of the following: comprehensive energy utilization efficiency, primary energy utilization efficiency, distributed power self-use rate and equipment utilization efficiency; The optimization objectives include: economic objectives, environmental protection objectives and energy efficiency objectives; The economic objectives include one or more of the following: minimizing daily operation and maintenance costs, minimizing energy procurement costs, and minimizing equipment start-up and shutdown loss costs; The environmental protection goals include one or more of the following: minimizing carbon emissions and maximizing the proportion of clean energy; The energy efficiency objectives include one or more of the following: maximizing comprehensive energy utilization efficiency, maximizing equipment utilization rate, and maximizing primary energy utilization efficiency.

4. The method according to claim 1, wherein The tensor flow topology structure including target data nodes, constraint data nodes, directed edges and computing nodes is constructed according to each optimization goal, including: According to each optimization goal, determining the objective function expression, constraint equation and calculation node of the optimization goal; Using the objective function expression as the target data node of the optimization target, and using the constraint equation as the constraint data node of the optimization target; Transmitting the target data node to the computing node via a preset directed edge of the target data node; Transmitting the constraint data node to the computing node via a preset directed edge of the constraint data node; A tensor flow topology structure of the optimization target is constructed according to the target data nodes, constraint data nodes, directed edges of the target data nodes, directed edges of the constraint data nodes, and computing nodes of the optimization target.

5. The method according to claim 4, wherein The tensor flow topology structure based on each optimization objective performs a simplification conversion on the multiple optimization objectives to obtain a single objective function of the integrated energy system and a constraint equation of the single objective function, including: Outputting the calculation results of the calculation nodes of each optimization target according to the tensor flow topology structure of each optimization target; Transmitting the calculation results of the calculation nodes of the optimization objectives and the preset system constraint equations to the preset integration calculation node through the preset integration directed edges, and outputting the calculation results of the integration calculation node; According to the calculation results of the integrated calculation node, the weight coefficient of the objective function expression of each optimization goal is calculated using the anti-entropy weight method; Obtaining a single objective function of the integrated energy system and a constraint equation of the single objective function according to weight coefficients of the objective function expressions of the optimization objectives; The calculation result of the calculation node of each optimization target is the objective weight relationship coefficient of each optimization target; and the calculation result of the integrated calculation node is the sum of the objective weight relationship coefficients of each optimization target.

6. The method according to claim 5, wherein The calculation formula of the weight coefficient of the objective function expression of each optimization goal is as follows: Among them, ω i Represents the weight coefficient of the objective function expression of the i-th optimization goal; i = 1…m; m represents the total number of optimization goals; ε i Represents the subjective weight relationship coefficient of the i-th optimization objective; δ i Represents the objective weight relationship coefficient of the i-th optimization goal; ω si represents the predetermined subjective weight value of the i-th optimization objective; ω oi represents the predetermined objective weight value of the i-th optimization goal.

7. The method according to claim 1, wherein The single objective function is solved based on the single objective function and the constraint equation of the single objective function to obtain the multi-objective scheduling scheme of the integrated energy system, including: Obtaining a deviation tensor of the integrated energy system according to the single objective function, a constraint equation of the single objective function, and pre-calculated state update tensor data of the integrated energy system; Constructing a Jacobian matrix according to the deviation tensor of the integrated energy system, and solving the Jacobian matrix to obtain a corrected state tensor; When the corrected state tensor meets the preset accuracy requirement, the system parameters of the integrated energy system are updated, and a multi-objective scheduling plan of the integrated energy system is output based on the updated system parameters of the integrated energy system.

8. The method according to claim 7, wherein The state update tensor data includes the following calculation process: Initializing system parameters of the integrated energy system and setting initial values ​​of the state tensors of the integrated energy system; Calculating the thermal power, active power, and required natural gas load flow of a combined heat and power (CHP) unit in the integrated energy system based on an initial value of the state tensor of the integrated energy system; According to the thermal power, active power and required natural gas load flow of the cogeneration heat and power (CHP) unit, the initial value of the state tensor of the integrated energy system is updated to obtain state update tensor data.

9. A multi-objective optimization scheduling system for integrated energy systems based on tensor flow, characterized in that: include: A multi-objective determination module, configured to determine, based on the obtained scheduling requirements of the integrated energy system, a plurality of optimization objectives associated with the scheduling requirements; The tensor flow construction module is used to construct a tensor flow topology structure containing target data nodes, constraint data nodes, directed edges and computing nodes according to each optimization goal; A single objective conversion module, configured to perform a simplification conversion on the multiple optimization objectives based on the tensor flow topology of each optimization objective, to obtain a single objective function of the integrated energy system and a constraint equation of the single objective function; The multi-objective scheduling module is used to solve the single objective function based on the single objective function and the constraint equation of the single objective function to obtain a multi-objective scheduling plan for the integrated energy system.

10. The system according to claim 9, wherein: The multi-target determination module includes: An indicator selection submodule is used to select an evaluation indicator corresponding to the dispatching demand from a pre-set evaluation system according to the dispatching demand of the obtained integrated energy system; The target generation submodule is used to determine multiple optimization targets associated with the scheduling requirements based on evaluation indicators corresponding to the scheduling requirements.

11. The system according to claim 10, wherein: The evaluation system includes: economic evaluation indicators, environmental evaluation indicators and energy efficiency evaluation indicators; The secondary indicators corresponding to the economic evaluation indicators include one or more of the following: total investment cost, investment payback period, internal rate of return, financial net present value and unit energy consumption cost; The secondary indicators corresponding to the environmental performance evaluation index include one or more of the following: annual carbon dioxide emission reduction, annual sulfur dioxide emission reduction, annual fossil energy consumption reduction, clean energy proportion and comprehensive waste utilization rate; The secondary indicators corresponding to the energy efficiency evaluation index include one or more of the following: comprehensive energy utilization efficiency, primary energy utilization efficiency, distributed power self-use rate and equipment utilization efficiency; The optimization objectives include: economic objectives, environmental objectives and energy efficiency objectives; the economic objectives include one or more of the following: minimization of daily operation and maintenance costs, minimization of energy procurement costs and minimization of equipment start-up and shutdown loss costs; The environmental protection goals include one or more of the following: minimizing carbon emissions and maximizing the proportion of clean energy; The energy efficiency objectives include one or more of the following: maximizing comprehensive energy utilization efficiency, maximizing equipment utilization rate, and maximizing primary energy utilization efficiency.

12. The system according to claim 9, wherein The tensor flow building block includes: A function determination submodule is used to determine the objective function expression, constraint equation and calculation node of the optimization target according to each optimization target; A node setting submodule, configured to use the objective function expression as the target data node of the optimization target and the constraint equation as the constraint data node of the optimization target; A target node transmission submodule, configured to transmit the target data node to the computing node via a predetermined directed edge of the target data node; A constraint node transmission submodule, configured to transmit the constraint data node to the computing node via a preset directed edge of the constraint data node; The topology construction submodule is used to construct the tensor flow topology structure of the optimization target according to the target data nodes, constraint data nodes, directed edges of the target data nodes, directed edges of the constraint data nodes and computing nodes of the optimization target.

13. The system according to claim 12, wherein: The single target conversion module includes: A node output submodule, configured to output the calculation results of the calculation nodes of each optimization target according to the tensor flow topology structure of each optimization target; A node integration submodule, configured to transmit the calculation results of the calculation nodes of the optimization objectives and the preset system constraint equations to a preset integration calculation node through a preset integration directed edge, and output the calculation results of the integration calculation node; A weight distribution submodule, configured to calculate the weight coefficients of the objective function expressions of the optimization objectives using an anti-entropy weight method according to the calculation results of the integrated calculation node; A multi-objective integration submodule, configured to obtain a single objective function of the integrated energy system and a constraint equation of the single objective function according to weight coefficients of the objective function expressions of the optimization objectives; The calculation result of the calculation node of each optimization target is the objective weight relationship coefficient of each optimization target; and the calculation result of the integrated calculation node is the sum of the objective weight relationship coefficients of each optimization target.

14. The system according to claim 13, wherein: The calculation formula of the anti-entropy weight method is as follows: Among them, ω i Represents the weight coefficient of the objective function expression of the i-th optimization goal; i = 1…m; m represents the total number of optimization goals; ε i Represents the subjective weight relationship coefficient of the i-th optimization objective; δ i Represents the objective weight relationship coefficient of the i-th optimization goal; ω si represents the predetermined subjective weight value of the i-th optimization objective; ω oi represents the predetermined objective weight value of the i-th optimization goal.

15. The system according to claim 9, wherein: The multi-objective scheduling module includes: a deviation calculation submodule, configured to update tensor data according to the single objective function, the constraint equation of the single objective function, and the pre-calculated state of the integrated energy system to obtain a deviation tensor of the integrated energy system; a matrix solving submodule, configured to construct a Jacobian matrix according to the deviation tensor of the integrated energy system, and solve the Jacobian matrix to obtain a corrected state tensor; The scheme generation submodule is used to update the system parameters of the integrated energy system when the corrected state tensor meets the preset accuracy requirements, and output the multi-objective scheduling scheme of the integrated energy system based on the updated system parameters of the integrated energy system.

16. The system according to claim 15, wherein: The multi-objective scheduling module further includes a tensor update submodule including: A parameter initialization unit, used to initialize the system parameters of the integrated energy system and set the initial value of the state tensor of the integrated energy system; a correlation data calculation unit, configured to calculate the thermal power, active power, and required natural gas load flow of the cogeneration heat and power (CHP) unit in the integrated energy system according to the initial value of the state tensor of the integrated energy system; The data updating unit is used to update the initial value of the state tensor of the integrated energy system according to the thermal power, active power and required natural gas load flow of the cogeneration heat and power CHP unit to obtain state update tensor data.

17. An electronic device, characterized in that: include: at least one processor and memory; The memory and the processor are connected via a bus; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, a multi-objective optimization scheduling method for an integrated energy system based on tensor flow as described in any one of claims 1 to 8 is implemented.

18. A computing device readable storage medium, characterized in that: An execution program is stored thereon, and when the execution program is executed, a multi-objective optimization scheduling method for an integrated energy system based on tensor flow as described in any one of claims 1 to 8 is implemented.

Images