Power system dynamic network resource allocation method and system
Patent Information
- Application Number
- CN202610725300.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-05-25
- Publication Date
- 2026-09-29
AI Technical Summary
[0003]目前，现有的电力系统的动态网络资源分配方案，存着如下问题：第一，现有方案往往基于固定的设定前提，比如资源总量固定或者网络节点固定，但是这种假设前提与实际情况严重不符，因此现有方案的精确性较差；第二，现有方案往往采用集中式的更新和分配策略，这在小规模电力网络中应用较好，但是已经不再适用于现今的大规模电力系统；第三，现有方案在应对资源需求的不确定性过程中，其鲁棒性较差
[0046]本发明提供的这种电力系统动态网络资源分配方法及系统，通过对目标电力系统的动态网络资源分配过程进行建模，并引入拉个朗日函数、拉个朗日乘子和对偶方案对构建的模型进行转化和计算，不仅实现了电力系统的动态网络资源分配，而且可靠性更高，精确性更高。
Smart Images
Figure CN122840461A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of electrical automation, specifically relating to a method and system for dynamic network resource allocation in power systems. Background Technology
[0002] Traditional power systems are rigid systems where power sources follow loads, with controllable power sources, predictable loads, and unidirectional power flow. However, with the addition of new energy power generation systems, the power generation side of the power system has gradually become a system with strong uncertainty and volatility. At the same time, the load side is also shifting from "passive acceptance" to "active response" due to the development of electric vehicles, smart homes, and other technologies. This has led to deep coupling and mutual influence among the various links in the "power source-grid-load-storage" chain.
[0003] Currently, existing dynamic network resource allocation schemes for power systems have the following problems: First, existing schemes are often based on fixed assumptions, such as a fixed total amount of resources or a fixed number of network nodes. However, these assumptions are seriously inconsistent with the actual situation, resulting in poor accuracy. Second, existing schemes often adopt centralized update and allocation strategies, which work well in small-scale power networks but are no longer suitable for today's large-scale power systems. Third, existing schemes are not robust enough in dealing with the uncertainty of resource demand. Summary of the Invention
[0004] One of the objectives of this invention is to provide a highly reliable and accurate method for dynamic network resource allocation in power systems.
[0005] The second objective of this invention is to provide a system for implementing the aforementioned dynamic network resource allocation method for power systems.
[0006] The dynamic network resource allocation method for power systems provided by this invention includes the following steps:
[0007] S1. Obtain data information about the target power system;
[0008] S2. Based on the data obtained in step S1, construct a dynamic network resource allocation model for the target power system;
[0009] S3. Based on the Lagrangian function and dual scheme, transform the model constructed in step S2;
[0010] S4. Based on the distributed self-optimizing Lagrange iteration scheme, iteratively solve the transformation model obtained in step S3;
[0011] S5. Based on the results obtained in step S4, complete the dynamic network resource allocation of the target power system.
[0012] Step S1, which involves acquiring data information about the target power system, specifically includes the following steps:
[0013] The target power system is defined as having several resource units, each capable of independent calculations; and capable of data exchange between resource units.
[0014] Acquire data information from the target power system;
[0015] The data information includes cost data and resource data of the target power system.
[0016] Step S2, which involves constructing a dynamic network resource allocation model for the target power system based on the data information obtained in step S1, specifically includes the following steps:
[0017] The dynamic network resource allocation model of the target power system is represented by the following formula:
[0018] In the formula This represents a dynamic network resource allocation model for the target power system. This represents the resource allocation variable for the i-th resource unit; The total number of resource units; Let i be the cost function of the i-th resource unit; Let i be the set of feasible resource allocations for the i-th resource unit; This is the initial resource value allocated to the i-th resource unit.
[0019] Step S3, which involves transforming the model constructed in step S2 based on the Lagrangian function and dual scheme, specifically includes the following steps:
[0020] Lagrange multipliers are introduced to process the constraints in the model in order to construct the Lagrange function corresponding to model P;
[0021] By employing a dual approach, the obtained Lagrangian function is transformed into the corresponding dual problem, thus completing the transformation process of the model constructed in step S2.
[0022] Step S3 specifically includes the following steps:
[0023] For the model constructed in step S2, Lagrange multipliers are introduced to handle the constraints, and the Lagrange function corresponding to model P is constructed, expressed as:
[0024] In the formula Let P be the Lagrangian function corresponding to model P; These are the Lagrange multipliers corresponding to the constraints;
[0025] The dual function of the Lagrangian function corresponding to model P Defined as minimizing the Lagrangian function with respect to the original decision variables while satisfying the constraints of each resource unit, it is expressed as:
[0026] In the formula The extended cost function for the i-th resource unit is... The convex conjugate function value at point X; X is the Cartesian product of the constraint sets of all resource units;
[0027] The dual problem of P is represented by DP, which is:
[0028] In the formula, R is the set of real numbers, i.e., Lagrange multipliers. The set of possible values;
[0029] By analyzing Lagrange multipliers Optimize the solution to achieve the optimal overall system objective;
[0030] The DP problem is equivalently transformed into a minimization form, represented as: .
[0031] Step S4, based on the distributed self-optimizing Lagrange iteration scheme, iteratively solves the transformation model obtained in step S3, specifically including the following steps:
[0032] Each resource unit performs collaborative updates locally to jointly achieve a decentralized solution to the transformation model obtained in step S3;
[0033] During the solution process, each resource unit calculates a locally consistent estimate of the global Lagrange multipliers based on the Lagrange multiplier data of its neighboring resource units.
[0034] Each resource unit, based on the obtained local consistency estimate, performs local optimization calculations on the decision variables of the transformed model under the condition of satisfying its own constraints;
[0035] Based on the local optimization calculation results, each resource unit updates its own Lagrange multipliers;
[0036] Repeat the above steps until the set conditions are met to complete the iterative solution of the transformation model obtained in step S3.
[0037] Step S4 specifically includes the following steps:
[0038] During the initialization phase, each resource unit i initializes a local copy of a Lagrange multiplier. The initialization processes of each resource unit are independent of each other.
[0039] In the k-th iteration, each resource unit i interacts with its neighboring resource units to obtain a local copy of the Lagrange multipliers of the neighboring resource units; the local consistency estimate of the global Lagrange multipliers is calculated using the following formula:
[0040] In the formula The local consistency estimate of the global Lagrange multipliers calculated for the i-th resource unit; The weight coefficient assigned to the j-th resource unit by the i-th resource unit during the k-th iteration; For the local copy of the Lagrange multiplier of the j-th resource unit in the k-th iteration;
[0041] Based on the obtained local consistency estimate Under the condition of satisfying its own constraints, each resource unit i performs local optimization calculations on the decision variables of the transformed model to obtain the local decision variables of the i-th resource unit, denoted as: ;
[0042] Based on local optimization calculation results Each resource unit i updates its own Lagrange multipliers using the following formula:
[0043] In the formula For the local copy of the Lagrange multiplier updated for the (k+1)th iteration process of the i-th resource unit; The set step size parameter; The baseline amount of resources that the i-th resource unit can obtain during the k-th iteration. Noisy observations , The baseline amount of resources that the i-th resource unit can obtain during the k-th iteration. The random perturbation term;
[0044] Repeat the above steps until the set conditions are met to complete the iterative solution of the transformation model obtained in step S3.
[0045] This invention also provides a system for implementing the dynamic network resource allocation method for power systems, comprising a data acquisition module, a model building module, a model transformation module, an iterative solution module, and a resource allocation module; the data acquisition module, model building module, model transformation module, iterative solution module, and resource allocation module are connected in series; the data acquisition module is used to acquire data information of the target power system and upload the data information to the model building module; the model building module is used to construct a dynamic network resource allocation model of the target power system based on the received data information and the acquired data information, and upload the data information to the model transformation module; the model transformation module is used to transform the constructed model based on the received data information and a Lagrangian function and dual scheme, and upload the data information to the iterative solution module; the iterative solution module is used to iteratively solve the obtained transformed model based on a distributed self-optimizing Lagrangian iterative scheme based on the received data information, and upload the data information to the resource allocation module; the resource allocation module is used to complete the dynamic network resource allocation of the target power system based on the received data information and the obtained results.
[0046] The dynamic network resource allocation method and system for power systems provided by this invention, by modeling the dynamic network resource allocation process of the target power system and introducing a Lagrange function, a Lagrange multiplier and a dual scheme to transform and calculate the constructed model, not only realizes the dynamic network resource allocation of the power system, but also has higher reliability and higher accuracy. Attached Figure Description
[0047] Figure 1 This is a schematic diagram of the method flow of the present invention.
[0048] Figure 2 This diagram illustrates the comparison between total cost and optimal cost under different time lengths in an embodiment of the method of the present invention.
[0049] Figure 3 This is a schematic diagram comparing the total power and load demand under different time lengths in an embodiment of the method of the present invention.
[0050] Figure 4 This is a schematic diagram of the functional modules of the system of the present invention. Detailed Implementation
[0051] like Figure 1 The diagram shown is a flowchart of the method of the present invention: The dynamic network resource allocation method for power systems disclosed in this invention includes the following steps:
[0052] S1. Obtain data information of the target power system; specifically including the following steps:
[0053] The target power system is defined as having several resource units, each capable of independent calculations; and capable of data exchange between resource units.
[0054] Acquire data information from the target power system;
[0055] The data information includes cost data and resource data of the target power system;
[0056] S2. Based on the data obtained in step S1, construct a dynamic network resource allocation model for the target power system; specifically, this includes the following steps:
[0057] The dynamic network resource allocation model of the target power system is represented by the following formula:
[0058] In the formula This represents a dynamic network resource allocation model for the target power system. This represents the resource allocation variable for the i-th resource unit; The total number of resource units; Let be the cost function of the i-th resource unit, used to characterize the operating cost or comprehensive cost of the resource unit under different resource configuration conditions. This cost function is a convex function. Let i be the set of feasible resource allocations for the i-th resource unit; The initial resource value allocated to the i-th resource unit; the set is a compact convex set and is held only by the corresponding resource unit itself;
[0059] In practical implementation, the cost function The cost function can be set to the operating cost, regulation cost, resource utilization deviation cost, network loss cost, voltage deviation penalty cost, load shedding cost, or renewable energy curtailment penalty cost of the i-th resource unit; as a preferred option, the cost function... It can be in the following form:
[0060] In the formula Let be the quadratic cost coefficient of the i-th resource unit, and ; Let be the linear cost coefficient of the i-th resource unit; This is the fixed cost item for the i-th resource unit; because Therefore, the cost function is a convex function;
[0061] The model P (also known as problem P) is a network resource allocation problem. It is used to optimize the allocation of a fixed total amount of resources in a system composed of multiple resource units, so that each node can achieve the optimal overall objective function of the system while satisfying its own constraints.
[0062] S3. Based on the Lagrangian function and dual scheme, transform the model constructed in step S2; specifically, the following steps are included:
[0063] Lagrange multipliers are introduced to process the constraints in the model in order to construct the Lagrange function corresponding to model P;
[0064] By employing a dual approach, the obtained Lagrangian function is transformed into the corresponding dual problem, thus completing the transformation process of the model constructed in step S2.
[0065] In practice, the following steps can be taken:
[0066] For the model constructed in step S2, Lagrange multipliers are introduced to handle the constraints, and the Lagrange function corresponding to model P is constructed, expressed as:
[0067] In the formula Let P be the Lagrangian function corresponding to model P; These are the Lagrange multipliers corresponding to the constraints; this expression introduces Lagrange multipliers into the original objective function. The global resource balancing constraints of the system are incorporated into the objective function, thereby transforming the original constrained optimization problem into an unconstrained form, providing a mathematical basis for subsequent distributed solutions.
[0068] The dual function of the Lagrangian function corresponding to model P Defined as minimizing the Lagrangian function with respect to the original decision variables while satisfying the constraints of each resource unit, it is expressed as:
[0069] In the formula The extended cost function for the i-th resource unit is... The convex conjugate function value at point X; X is the Cartesian product of the constraint sets of all resource units;
[0070] The dual problem of P is represented by DP, which is:
[0071] In the formula, R is the set of real numbers, i.e., Lagrange multipliers. The set of possible values;
[0072] By analyzing Lagrange multipliers Optimize the solution to achieve the optimal overall system objective;
[0073] The DP problem is equivalently transformed into a minimization form, represented as: ;
[0074] Among them, due to Since the functions are convex, each function is also convex, thus ensuring the optimizability of the dual problem; at the same time, transforming the original problem into the dual problem facilitates the solution of distributed problems.
[0075] S4. Based on the distributed self-optimizing Lagrange iteration scheme, iteratively solve the transformation model obtained in step S3; specifically including the following steps:
[0076] Each resource unit performs collaborative updates locally to jointly achieve a decentralized solution to the transformation model obtained in step S3;
[0077] During the solution process, each resource unit calculates a locally consistent estimate of the global Lagrange multipliers based on the Lagrange multiplier data of its neighboring resource units.
[0078] Each resource unit, based on the obtained local consistency estimate, performs local optimization calculations on the decision variables of the transformed model under the condition of satisfying its own constraints;
[0079] Based on the local optimization calculation results, each resource unit updates its own Lagrange multipliers;
[0080] Repeat the above steps until the set conditions are met to complete the iterative solution of the transformation model obtained in step S3;
[0081] In practice, the following steps can be taken:
[0082] During the initialization phase, each resource unit i initializes a local copy of a Lagrange multiplier. The initialization processes of each resource unit are independent of each other.
[0083] In the k-th iteration, each resource unit i interacts with its neighboring resource units to obtain a local copy of the Lagrange multipliers of the neighboring resource units; the local consistency estimate of the global Lagrange multipliers is calculated using the following formula:
[0084] In the formula The local consistency estimate of the global Lagrange multipliers calculated for the i-th resource unit; The weight coefficient assigned to the j-th resource unit by the i-th resource unit in the k-th iteration is used to reflect the communication relationship and information reliability between nodes; For the local copy of the Lagrange multiplier of the j-th resource unit in the k-th iteration;
[0085] Based on the obtained local consistency estimate Under the condition of satisfying its own constraints, each resource unit i performs local optimization calculations on the decision variables of the transformed model to obtain the local decision variables of the i-th resource unit, denoted as: The process relies on the cost function and constraint information held by the nodes themselves, ensuring the fully distributed execution of the original variable update process.
[0086] Based on local optimization calculation results Each resource unit i updates its own Lagrange multipliers using the following formula:
[0087] In the formula For the local copy of the Lagrange multiplier updated for the (k+1)th iteration process of the i-th resource unit; The step size parameter is set to adjust the iteration update magnitude; The baseline amount of resources that the i-th resource unit can obtain during the k-th iteration. Noisy observations , The baseline amount of resources that the i-th resource unit can obtain during the k-th iteration. The random disturbance term is used to characterize resource measurement errors or random load fluctuations;
[0088] Repeat the above steps until the set conditions are met to complete the iterative solution of the transformation model obtained in step S3. In specific implementation, the set termination condition can be that the relative error between the local Lagrange multiplier and the dual optimal Lagrange multiplier of each resource unit is less than the preset error threshold, which is expressed as follows:
[0089] In the formula Let be the local Lagrange multiplier of the i-th resource unit in the k-th iteration. For the optimal Lagrange multiplier of the dual problem, To preset the error threshold, It is a collection of resource units; as an optional implementation method, The preferred value is 0.1;
[0090] The convergence analysis of the above iterative optimization problem is as follows:
[0091] Assuming step size For non-incremental, And satisfy and Then the sequence and For all i, the following (1) and (2) are satisfied:
[0092] (1): ,at this time This is the optimal solution for DP;
[0093] (2): This is the optimal solution for P.
[0094] S5. Based on the results obtained in step S4, complete the dynamic network resource allocation of the target power system.
[0095] To verify the effectiveness of the method of the present invention, the scheme of the present invention was specifically applied to the dispatching process of a power system. Taking the IEEE 14-node test system as an example, the cost functions of the five generators in the system are all quadratic convex functions. The total system load is set to 200MW, and each generator exchanges information with neighboring nodes based on a dynamic communication topology. A step size sequence is set in the simulation. And introduced bounded, zero-mean random noise to simulate load fluctuations.
[0096] Simulation results are as follows Figure 2 and Figure 3 As shown. (Through) Figure 2 and Figure 3 As can be seen, under both constant and random load scenarios, the proposed solution enables the total system generation cost to converge to the theoretical optimal value, which is obtained through independent centralized calculation and serves as a verification benchmark. The error between the algorithm result and this benchmark is less than 10%. Simultaneously, the dual variables of each node, i.e., the incremental costs, reach consensus, and the total power generation of the entire network strictly meets the 200MW load demand, verifying the correctness of the proposed method. Under the influence of random noise, the total system cost still converges rapidly to the optimal range within a finite number of iterations, the total power generation accurately tracks the load demand, and the incremental costs of each generator quickly reach a consensus. Compared to traditional methods, the proposed solution can guarantee the expected convergence and operational stability of the system's optimization performance, significantly improving the reliability, security, and robustness of the power system.
[0097] like Figure 4The diagram shows the functional modules of the system of this invention: The system for implementing the dynamic network resource allocation method of the power system disclosed in this invention includes a data acquisition module, a model building module, a model transformation module, an iterative solution module, and a resource allocation module; the data acquisition module, model building module, model transformation module, iterative solution module, and resource allocation module are connected in series; the data acquisition module is used to acquire data information of the target power system and upload the data information to the model building module; the model building module is used to construct a dynamic network resource allocation model of the target power system based on the received data information and the acquired data information, and upload the data information to the model transformation module; the model transformation module is used to transform the constructed model based on the received data information and the Lagrangian function and dual scheme, and upload the data information to the iterative solution module; the iterative solution module is used to iteratively solve the obtained transformed model based on the received data information and the distributed self-optimizing Lagrangian iterative scheme, and upload the data information to the resource allocation module; the resource allocation module is used to complete the dynamic network resource allocation of the target power system based on the received data information and the obtained results.
Claims
1. A method for dynamic network resource allocation in a power system, comprising the following steps: S1. Obtain data information about the target power system; S2. Based on the data obtained in step S1, construct a dynamic network resource allocation model for the target power system; S3. Based on the Lagrangian function and dual scheme, transform the model constructed in step S2; S4. Based on the distributed self-optimizing Lagrange iteration scheme, iteratively solve the transformation model obtained in step S3; S5. Based on the results obtained in step S4, complete the dynamic network resource allocation of the target power system.
2. The power system dynamic network resource allocation method according to claim 1, characterized in that... Step S1, which involves acquiring data information about the target power system, specifically includes the following steps: The target power system is defined as having several resource units, each capable of independent calculations; and capable of data exchange between resource units. Acquire data information from the target power system; The data information includes cost data and resource data of the target power system.
3. The method for dynamic network resource allocation in a power system according to claim 2, characterized in that... Step S2, which involves constructing a dynamic network resource allocation model for the target power system based on the data information obtained in step S1, specifically includes the following steps: The dynamic network resource allocation model of the target power system is represented by the following formula: In the formula This represents a dynamic network resource allocation model for the target power system. This represents the resource allocation variable for the i-th resource unit; The total number of resource units; Let i be the cost function of the i-th resource unit; Let i be the set of feasible resource allocations for the i-th resource unit; This is the initial resource value allocated to the i-th resource unit.
4. The power system dynamic network resource allocation method according to claim 3, characterized in that... Step S3, which involves transforming the model constructed in step S2 based on the Lagrangian function and dual scheme, specifically includes the following steps: Lagrange multipliers are introduced to process the constraints in the model in order to construct the Lagrange function corresponding to model P; By employing a dual approach, the obtained Lagrangian function is transformed into the corresponding dual problem, thus completing the transformation process of the model constructed in step S2.
5. The method for dynamic network resource allocation in a power system according to claim 4, characterized in that... Step S3 specifically includes the following steps: For the model constructed in step S2, Lagrange multipliers are introduced to handle the constraints, and the Lagrange function corresponding to model P is constructed, expressed as: In the formula Let P be the Lagrangian function corresponding to model P; These are the Lagrange multipliers corresponding to the constraints; The dual function of the Lagrangian function corresponding to model P Defined as minimizing the Lagrangian function with respect to the original decision variables while satisfying the constraints of each resource unit, it is expressed as: In the formula The extended cost function for the i-th resource unit is... The convex conjugate function value at point X; X is the Cartesian product of the constraint sets of all resource units; The dual problem of P is represented by DP, which is: In the formula, R is the set of real numbers; By analyzing Lagrange multipliers Optimize the solution to achieve the optimal overall system objective; The DP problem is equivalently transformed into a minimization form, represented as: .
6. The method for dynamic network resource allocation in a power system according to claim 5, characterized in that... Step S4, based on the distributed self-optimizing Lagrange iteration scheme, iteratively solves the transformation model obtained in step S3, specifically including the following steps: Each resource unit performs collaborative updates locally to jointly achieve a decentralized solution to the transformation model obtained in step S3; During the solution process, each resource unit calculates a locally consistent estimate of the global Lagrange multipliers based on the Lagrange multiplier data of its neighboring resource units. Each resource unit, based on the obtained local consistency estimate, performs local optimization calculations on the decision variables of the transformed model under the condition of satisfying its own constraints; Based on the local optimization calculation results, each resource unit updates its own Lagrange multipliers; Repeat the above steps until the set conditions are met to complete the iterative solution of the transformation model obtained in step S3.
7. The power system dynamic network resource allocation method according to claim 4, characterized in that... Step S4 specifically includes the following steps: During the initialization phase, each resource unit i initializes a local copy of a Lagrange multiplier. The initialization processes of each resource unit are independent of each other. In the k-th iteration, each resource unit i interacts with its neighboring resource units to obtain a local copy of the Lagrange multipliers of the neighboring resource units; the local consistency estimate of the global Lagrange multipliers is calculated using the following formula: In the formula The local consistency estimate of the global Lagrange multipliers calculated for the i-th resource unit; The weight coefficient assigned to the j-th resource unit by the i-th resource unit during the k-th iteration; For the local copy of the Lagrange multiplier of the j-th resource unit in the k-th iteration; Based on the obtained local consistency estimate Under the condition of satisfying its own constraints, each resource unit i performs local optimization calculations on the decision variables of the transformed model to obtain the local decision variables of the i-th resource unit, denoted as: ; Based on local optimization calculation results Each resource unit i updates its own Lagrange multipliers using the following formula: In the formula For the local copy of the Lagrange multiplier updated for the (k+1)th iteration process of the i-th resource unit; The set step size parameter; The baseline amount of resources that the i-th resource unit can obtain during the k-th iteration. Noisy observations , The baseline amount of resources that the i-th resource unit can obtain during the k-th iteration. The random perturbation term; Repeat the above steps until the set conditions are met to complete the iterative solution of the transformation model obtained in step S3.
8. A system for implementing the dynamic network resource allocation method for power systems according to any one of claims 1 to 7, characterized in that... It includes a data acquisition module, a model building module, a model transformation module, an iterative solution module, and a resource allocation module; these modules are connected in series. The data acquisition module acquires data information of the target power system and uploads it to the model building module. The model building module constructs a dynamic network resource allocation model of the target power system based on the received and acquired data information and uploads the model transformation model to the model transformation module. The model transformation module is used to transform the constructed model based on the received data information, using the Lagrange function and dual scheme, and then upload the data information to the iterative solution module. The iterative solution module is used to iteratively solve the obtained transformation model based on the received data information and the distributed self-optimizing Lagrange iterative scheme, and then upload the data information to the resource allocation module. The resource allocation module is used to complete the dynamic network resource allocation of the target power system based on the received data and the results obtained.