A multi-objective power dispatch optimization method and system, and a storage medium

Abstract

The embodiment of the present application provides a kind of multi-target power scheduling optimization method, system and storage medium, belong to power dispatch field.The method comprises: obtaining the operation data of each load and load type, load type includes flexible adjustable load and non-flexible adjustable load;According to the power requirement in power limiting instruction, according to the preset priority and penalty coefficient, the switching state of each load and the flexible load set needing to be put into operation are determined;For the set, with the minimization of load adjustment cost as target, with the total power allowed to adjust as constraint, an optimization model is established;The optimization model is solved based on distributed iterative algorithm, and the power setting value of each flexible load is obtained;According to the switching state and power setting value, power dispatch instruction is generated and executed.The present application can realize the rapid autonomous scheduling of load under the condition of power limiting, effectively reduce production loss under the premise of meeting power constraint.

Description

Technical Field

[0001] This invention relates to the field of power dispatching technology, and more specifically to a multi-objective power dispatching optimization method, device, and storage medium. Background Technology

[0002] Steel companies frequently face power rationing orders or stringent demand targets from the power grid during production. To ensure the safe and stable operation of the power grid, companies need to quickly formulate and implement power rationing response strategies after the issuance of power rationing orders, in order to control the total power load of the entire plant within the prescribed range.

[0003] Traditional power curtailment methods rely primarily on manual experience and static load sequence tables. Operators manually execute power curtailment operations according to a predetermined load shedding sequence. This method is slow to respond and struggles to meet the demands of grid changes that occur on a minute or even second-by-second basis. Furthermore, static sequence tables lack a global optimization perspective and cannot incorporate real-time operating conditions of each production stage (such as current equipment status and process stage) for dynamic decision-making. This can easily lead to the mis-cutting of high-value production processes, resulting in significant production losses.

[0004] Therefore, how to achieve autonomous scheduling with second-level response and multi-objective coordination (safety, production continuity, and economy) while meeting power rationing orders is a technical problem currently facing steel companies. Summary of the Invention

[0005] The purpose of this invention is to provide a multi-objective power dispatch optimization method, system, and storage medium to solve the technical problems mentioned in the background.

[0006] To achieve the above objectives, embodiments of the present invention provide a multi-objective power dispatch optimization method. The optimization method includes: acquiring real-time operating data of each load in a target scenario and the load type of each load, wherein the load type includes flexibly adjustable loads and non-flexibly adjustable loads; determining the switching status of each load and the set of loads to be put into operation among the flexibly adjustable loads according to the power requirements in the power restriction order, based on preset priorities and penalty coefficients; establishing an optimization model for the set of loads to be put into operation with minimizing its load regulation cost as the optimization objective and the total allowable regulation power of the load set as the constraint condition; iteratively solving the optimization model based on a distributed iterative algorithm to obtain the power setting value of each load in the set of loads to be put into operation; and obtaining the power dispatch order for each load according to the switching status of the non-flexibly adjustable loads and the power setting value of each load in the set of loads to be put into operation, and executing the power dispatch order.

[0007] Optionally, the non-flexible load includes: uninterruptible critical loads related to safety or core processes, important interruptible loads involving important auxiliary facilities, and general interruptible loads involving non-critical auxiliary production facilities; wherein, a corresponding penalty coefficient is set for each load according to the load type, the magnitude of which reflects the production loss caused when the load is power-limited, the first penalty coefficient of the uninterruptible critical load is greater than the second penalty coefficient of the important interruptible load, the third penalty coefficient of the flexible load is greater than the fourth penalty coefficient of the general interruptible load, and the second penalty coefficient is greater than the third penalty coefficient.

[0008] Optionally, before determining the switching status of each load and the set of loads that need to be put into operation among the flexibly adjustable loads according to preset priorities and penalty coefficients, the optimization method further includes: determining the minimum guaranteed power based on the total power of all the uninterruptible critical loads and the minimum operating power of at least one of the flexibly adjustable loads; determining whether the power requirement in the power restriction command is greater than the minimum guaranteed power; if so, determining the switching status of each load and the set of loads that need to be put into operation among the flexibly adjustable loads according to preset priorities and penalty coefficients; if not, triggering an alarm and terminating the subsequent optimization process.

[0009] Optionally, the preset priorities, from high to low, are: general interruptible loads, flexible adjustable loads, important interruptible loads, and uninterruptible critical loads, wherein the higher the priority, the more likely it is to respond to the power limiting command and perform power limiting operations.

[0010] Optionally, determining the switching status of each load and the set of loads to be put into operation among the flexibly adjustable loads according to a preset priority and penalty coefficient includes: processing the general interruptible loads, the flexibly adjustable loads, and the important interruptible loads in sequence according to the preset priority. For the general interruptible loads, each load in the category currently in operation is disconnected in ascending order according to the unit power penalty cost. After disconnecting each load, the first total power of all currently operating loads is recalculated, and it is determined whether the first total power meets the power requirements in the power restriction order. If yes, the operation is terminated; otherwise, the power of the currently operating flexibly adjustable loads is reduced to their minimum operating power. The second total power of all currently operating loads is calculated, and it is determined whether the second total power meets the power requirements in the power restriction order. If the power requirement is met, the operation is terminated; otherwise, while ensuring that at least one of the flexibly adjustable loads remains operational, the remaining flexibly adjustable loads are disconnected one by one according to the unit power penalty cost. After each load is disconnected, the third total power of all currently operating loads is recalculated, and it is determined whether the third total power meets the power requirement in the power restriction command. If yes, the operation is terminated; otherwise, each of the important interruptible loads currently in operation is disconnected in ascending order according to the unit power penalty cost. After each load is disconnected, the fourth total power of all currently operating loads is recalculated, and it is determined whether the fourth total power meets the power requirement in the power restriction command. If yes, the operation is terminated; otherwise, an alarm is triggered. Wherein, the unit power penalty cost is the ratio of the penalty coefficient corresponding to each load to its power during normal operation.

[0011] Optionally, the formula for the optimization model is:

[0012] in, Indicates load i The aforementioned unit power penalty cost, This indicates the maximum allowable power value of the loads in the set of loads that need to be put into operation within the flexibly adjustable loads. This represents the minimum allowable power value for the loads in the set of loads that need to be put into operation within the flexibly adjustable loads. Indicates load i The power increment to be solved Indicates load i The running status, The time indicates that the operation has been cut off. This indicates that it has been put into operation. This represents the total allowable adjustable power of the load set. This refers to the set of loads that need to be put into operation within the flexibly adjustable load.

[0013] Optionally, the distributed iterative algorithm is a distributed fast consensus algorithm, which constructs an augmented Lagrangian function to solve the optimization model by pre-setting global dual variables and slack variables; the expression of the augmented Lagrangian function is:

[0014] in, Represents the global dual variable. Represents the slack variable, This represents the penalty parameter.

[0015] Optionally, the iterative solution of the optimization model based on the distributed iterative algorithm includes: according to the first... k The global dual variable of each load in the load set during round iteration slack variables and the total power increment of all loads in the load set. Based on the augmented Lagrangian function, the parallel solution is obtained for the first... k+ The power increment of each load in the load set during the first iteration According to the first k+ The total power increment of all loads in the load set during the first iteration. Determine the first k+ The slack variables in the first iteration According to the first k+ The total power increment in one iteration The slack variable and the allowable adjustment of total power Determine the first k+ The global dual variable in the first iteration When the original residual And dual residuals When the iteration stops, output the power increment of each load in the current load set. As the optimal solution; wherein, the power increment of each load in the current load set is... The sum of the minimum operating power of the load and the minimum operating power of the load shall be used as the power setting value of the load.

[0016] Optionally, the statement based on the first k The global dual variable of each load in the load set during round iteration slack variables and the total power increment of all loads in the load set. Based on the augmented Lagrangian function, the parallel solution is obtained for the first...k+ The power increment of each load in the load set during the first iteration The calculation formula is as follows: .

[0017] Optionally, the statement based on the first k+ The total power increment of all loads in the load set during the first iteration. Determine the first k+ The slack variables in the first iteration The calculation formula is as follows:

[0018] According to the first k+ The total power increment in one iteration The slack variable and the allowable adjustment of total power Determine the first k+ The global dual variable in the first iteration The calculation formula is as follows:

[0019] in, Indicates the first k+ The slack variables in the first iteration, Indicates the first k+ The total power increment in one iteration, Indicates the first k The global dual variable during round iteration, Indicates the first k+ The global dual variable in the first iteration, This represents the total allowable adjustable power of the load set. This represents the penalty parameter.

[0020] Secondly, the present invention provides a multi-objective power dispatch optimization system, the optimization system comprising: an industrial data acquisition module, used to acquire real-time operating data of each load in a target scenario and the load type of each load, wherein the load type includes flexibly adjustable loads and non-flexibly adjustable loads; an optimization module, used to determine the switching status of each load and the set of loads to be put into operation among the flexibly adjustable loads according to the power requirements in the power restriction order, and according to preset priority and penalty coefficient; for the set of loads to be put into operation, an optimization model is established with minimizing its load adjustment cost as the optimization objective and the total allowable adjustment power of the load set as the constraint condition; the optimization model is iteratively solved based on a distributed iterative algorithm to obtain the power setting value of each load in the set of loads to be put into operation; and a control module, used to obtain the power dispatch order of each load according to the switching status of the non-flexibly adjustable loads and the power setting value of each load in the set of loads to be put into operation, and control each load to execute the power dispatch order.

[0021] Thirdly, the present invention provides a machine-readable storage medium storing instructions that cause a machine to execute the multi-objective power dispatch optimization method described above.

[0022] Through the above technical solution, this invention first acquires real-time operating data and load types of each load in the target scenario, providing an accurate data foundation for subsequent decision-making. Upon receiving a power curtailment command, the switching status of each load and the set of loads to be put into operation from the flexibly adjustable loads are determined according to preset priorities and penalty coefficients. This process requires no manual intervention, avoiding time-consuming steps that rely on experience-based judgment, thereby compressing the overall decision-making time for power curtailment response to the second level, meeting the grid's requirements for rapid response. For the set of flexible loads to be put into operation, an optimization model is established with minimizing its load adjustment cost as the optimization objective and allowing total adjustable power as the constraint condition, and a distributed iterative algorithm is used to solve the model. This ensures that each flexible load can obtain an accurate power setpoint within its adjustable range, guaranteeing that the total power does not exceed the limit and minimizing production losses caused by power curtailment, avoiding excessive losses caused by simple disconnection methods in existing technologies. Finally, a complete power dispatch command is generated and executed based on the switching status of non-flexible adjustable loads and the power setpoint of flexible loads, forming an automated closed-loop process from data acquisition, status decision-making, optimization solution to command issuance, significantly improving the intelligence level and execution efficiency of power curtailment dispatch.

[0023] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0024] The accompanying drawings are provided to further illustrate embodiments of the present invention and form part of the specification. They are used together with the following detailed description to explain the embodiments of the present invention, but do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart illustrating a multi-objective power dispatch optimization method provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of a multi-objective power dispatch optimization system provided in an embodiment of the present invention. Detailed Implementation

[0025] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the scope of the present invention.

[0026] It should be noted that the acquisition, transmission, storage, use, and processing of data in the technical solution of this application all comply with relevant laws and regulations. In the embodiments of this application, certain existing industry solutions such as software, components, and models may be mentioned. These should be considered exemplary, intended only to illustrate the feasibility of implementing the technical solution of this application, and do not imply that the applicant has already used or necessarily used such solutions.

[0027] Figure 1 This is a flowchart illustrating a multi-objective power dispatch optimization method provided in an embodiment of the present invention. Figure 1 As shown, the optimization method includes steps S1-S5.

[0028] Step S1: Obtain real-time operating data of each load in the target scenario and the load type of each load, wherein the load type includes flexibly adjustable loads and non-flexibly adjustable loads; Step S2: Based on the power requirements in the power restriction order, and according to the preset priority and penalty coefficient, determine the switching status of each load and the set of loads that need to be put into operation among the flexibly adjustable loads; Step S3: For the set of loads that need to be put into operation, establish an optimization model with minimizing its load adjustment cost as the optimization objective and the total allowable adjustment power of the load set as the constraint condition; Step S4: Iteratively solve the optimization model based on the distributed iterative algorithm to obtain the power set value of each load in the set of loads to be put into operation; Step S5: Based on the switching status of the non-flexible load and the power setting value of each load in the set of loads to be put into operation, obtain the power dispatching instruction for each load and execute the power dispatching instruction.

[0029] For step S1, in this embodiment of the invention, the target scenario is the entire power system of a steel plant. The sources of real-time operational data include the steel plant's Energy Management System (EMS) and Manufacturing Execution System (MES). The EMS collects in real-time the total active power of all 10kV and above substations, power data of each outgoing circuit, and switch status data of each load. The MES collects in real-time the current operating status and technological stage (e.g., heating period, alloying period) of each production unit (e.g., converter, LF furnace, continuous casting machine, rolling mill, etc.). This data is used to subsequently determine the load's adjustability, identify penalty coefficients, and make power curtailment decisions.

[0030] The load types include flexibly adjustable loads and inflexibly adjustable loads. Flexibly adjustable loads (P3) refer to electrical equipment whose power can be continuously adjusted within a certain range, such as LF furnaces (ladle refining furnaces), whose operating power can be smoothly adjusted between minimum and maximum operating power as needed. Inflexibly adjustable loads, on the other hand, refer to loads that lack continuous adjustment capabilities and can only operate at rated power or be completely disconnected.

[0031] While acquiring data, a penalty coefficient is set for each load. The magnitude of this coefficient reflects the production loss incurred when the load is power-limited. A third penalty coefficient corresponds to the flexibly adjustable load P3. Set to a smaller value, generally between 100 and 500. Different values ​​can be selected according to the process stage (insulation, slag formation, refining, etc.). The smaller the value, the more inclined it is to reduce power or cut off.

[0032] In this embodiment of the invention, non-flexible adjustable loads are further subdivided into the following four categories: Critical loads P1 are related to safety or core processes and are uninterrupted critical loads, such as blast furnace blowers and oxygen generators. Power rationing is generally not allowed for these loads; power supply must be guaranteed. In the optimization method of this embodiment, these loads must operate normally, i.e., maintain normal operating power. The corresponding first penalty coefficient... Set it to a very large constant, for example, set it to 10. 6 The order of millions indicates that the cost of cutting off this type of load in any situation far exceeds the sum of other loads, meaning that power is not reduced in principle.

[0033] Critical interruptible loads P2 are those involving important auxiliary systems such as dust removal, water treatment, and raw material supply, or those related to environmental protection. Interruption of these loads could lead to significant production problems or environmental risks. These loads have two operating states: maintaining normal power output when power restrictions are not urgent, and being completely disconnected (power reduced to 0) when necessary. The corresponding second penalty coefficient... Set to a larger value, typically 5 to 10 times the load penalty factor for P3 or P4 levels, such as 1000 to 5000.

[0034] Generally, interruptible loads P4 involve non-critical auxiliary production facilities, such as parts of rolling mills and slag processing, and their interruption has minimal impact on mainline production. As the primary buffer resource for power rationing response, it has the highest priority and can be completely or partially cut off first. The corresponding fourth penalty coefficient... Set to a smaller value, usually less than the value for P3 type load, and the value range can be set between 1 and 100.

[0035] In step S2, based on the power requirements in the received power restriction order, and according to the preset priority and penalty coefficient, the switching status (i.e., put into operation or cut off) of all loads (including non-flexible loads and flexible loads) in the entire plant is determined, and the set of loads that need to be put into operation among the flexible loads is further determined for subsequent continuous power optimization.

[0036] Before determining the switching status of each load and the set of loads that need to be put into operation from the flexibly adjustable loads according to the power requirements in the power restriction order, preset priorities and penalty coefficients, the optimization method also includes: Summarize the power of all uninterruptible critical loads P1 The total power of the uninterruptible critical load P1 is obtained. The calculation formula is as follows:

[0037] in, i The encoded index representing the load. The set representing the uninterruptible critical load P1; Based on the total power of the uninterruptible critical load P1 and the minimum operating power of at least one flexibly adjustable load P3. Determine the minimum guaranteed power; The minimum guaranteed power refers to the minimum total power that the entire plant must maintain under the premise of ensuring production safety (i.e., all critical loads are not interrupted and at least one flexibly adjustable load is maintained to maintain minimum operation).

[0038] To determine whether the power requirement in the power rationing order is greater than the minimum guaranteed power, we need to check whether the following inequality holds:

[0039] in, This represents the total active power limit for the entire plant, i.e., the power requirement in the power curtailment order, indicating the value that the power grid requires the enterprise's total power consumption to not exceed. If it is greater than this value, it indicates that the power curtailment requirement has not exceeded the production safety bottom line, and subsequent switching decisions can continue to be executed. That is, according to the preset priority and penalty coefficient, the switching status of each load and the set of loads that need to be put into operation among the flexibly adjustable loads are determined. If it is not greater than this value, an alarm is triggered and the subsequent optimization process is terminated.

[0040] Specifically, the decision to switch loads is executed according to a preset priority order. In an optional embodiment of the present invention, the preset priorities from high to low are: general interruptible loads P4, flexibly adjustable loads P3, important interruptible loads P2, and uninterruptible critical loads P1, i.e., P4>P3>P2>P1. A higher priority indicates that the load responds to the power-limiting command and performs power-limiting operations (i.e., is switched off or has its power reduced first). The uninterruptible critical load P1 has the lowest priority and is generally not switched off in practice.

[0041] Furthermore, according to the preset priority and penalty coefficient, the switching status of each load and the set of loads that need to be put into operation in the flexibly adjustable load are determined, including steps S21-S25.

[0042] Step S21: Initial state setting, setting the initial switching state of all operable loads (i.e., general interruptible load P4, flexible adjustable load P3, and important interruptible load P2). All are set to be put into operation (let) This assumes that all these loads are engaged, while all uninterruptible critical loads P2 remain operational. Based on this, the calculation of all loads in the entire plant currently in operation is performed. The total power of the load is used as the initial total power. This initial total power reflects the overall plant power level before any power curtailment operations are implemented. Subsequent power cuts or reductions will be based on this level to gradually reduce power until the power requirements in the power curtailment order are met.

[0043] Step S22: For all currently running states The typical interruptible load P4 is calculated based on the unit power penalty cost. Arrange the loads in ascending order (i.e., prioritize those with the lowest cost), and disconnect them one by one in this order (even if the load is in operation). For each load disconnected, the first total power of all loads currently in operation throughout the plant is recalculated, and it is determined whether this first total power meets the power requirements of the power rationing order. If the conditions are met, the switching decision process is terminated, and no further loads of the following categories are processed. At this time, all unprocessed loads (including loads with subsequent priorities) remain in operation. If the total power of the entire plant still does not meet the power curtailment requirements after all general interruptible loads P4 are disconnected, then proceed to step S23.

[0044] Step S23: Reduce the power of all currently operational adjustable loads P3 to their respective preset minimum operating power. Then, the second total power of all loads in operation throughout the plant is recalculated. If the second total power already meets the power curtailment requirements, the decision is terminated. At this point, all flexibly adjustable loads remain in operation (but their power has been reduced to the minimum value), and subsequent priority loads (important interruptible loads and uninterruptible critical loads) also remain in operation.

[0045] Step S24: If the conditions are still not met, then, provided that at least one flexibly adjustable load P3 remains operational (e.g., at least one LF furnace is retained to maintain production continuity), the remaining flexibly adjustable loads, excluding the retained load, will be penalized according to the unit power cost penalty. The loads are arranged in ascending order and disconnected one by one. After each load is disconnected, the total plant power (third total power) is recalculated. Once the power curtailment requirements are met, subsequent operations cease. At this point, the disconnected flexible loads are in the "disconnected" state. The loads that remain operational constitute the set of loads that need to be put into operation, and their power is currently at its minimum.

[0046] If all detachable flexible loads are cut off (i.e., only one flexible load remains in operation with its power at the minimum value), and the total power of the entire plant still does not meet the power rationing requirements, then proceed to step S25.

[0047] Step S25: Apply a penalty cost per unit power to all critical interruptible loads P2 that are currently in operation. Arrange them in ascending order and cut them off one by one. After each one is cut off, recalculate the total power of the entire plant (fourth total power). Once the power rationing requirements are met, the process will terminate immediately.

[0048] If the total power output of the plant still does not meet the power rationing requirements after all critical interruptible loads have been disconnected, an alarm will be triggered, indicating that the power rationing demand exceeds the system's response capacity.

[0049] In this embodiment of the invention, the unit power penalty cost is defined as the ratio of the penalty coefficient of each load to its normal operating power. This indicator reflects the production loss corresponding to the removal of a unit of power and serves as the basis for deciding the order of load removal within the same priority range; that is, loads with lower unit power penalty costs are removed first.

[0050] After the switching decision process in step S2, the switching status (partial or complete disconnection) of all uninterruptible critical loads P1 is determined, as well as the switching status of general interruptible loads P4 and important interruptible loads P2. The set of loads that need to be put into operation from the flexible loads P3 is also determined (denoted as...). S Furthermore, the power of each flexible load in this set has been preset to its minimum operating power. .

[0051] For the set of loads that need to be put into operation in the flexibly adjustable load P3, the power curtailment that needs to be finely allocated is denoted as... Its calculation expression is:

[0052]

[0053] in, This indicates the total active power limit (power requirement) for the entire plant in the power rationing order. This represents the total power of all uninterruptible critical loads of type P1, i.e. ; Second item This indicates critical interruptible loads P2 and general interruptible loads P4 that have been removed or retained. Indicates the switching status (1 indicates that it is in operation, 0 indicates that it is disconnected). This indicates its rated power; the third item refers to the set of flexible loads that need to be put into operation. S ,in , This is the minimum operating power for this load.

[0054] This formula reflects the difference between the total plant power and the power curtailment target after the discrete switching decision is executed. This difference is the allowable power output from the set... S The total power that can be continuously adjusted by the flexibly adjustable load.

[0055] For step S3, by ensuring that the total regulating power does not exceed Under the premise of a set S Each flexible load in i Assign a precise power value (between the minimum and maximum operating power for each load) to minimize production losses due to power rationing. Since the power of these flexible loads (such as LF furnaces) can vary continuously within their regulation range, this step models the problem as a continuous optimization model.

[0056] For the load set that needs to be put into operation S Each flexible load in i ( Define the following known parameters: Indicates load i Minimum permissible operating power (determined by equipment physical characteristics or process requirements); Indicates load i Maximum permitted operating power; Indicates load i The unit power penalty cost is equal to the penalty coefficient of that load. With rated power The ratio, i.e. This value reflects the load. i The economic loss caused by reducing unit power; the higher the value, the higher the production value of that load.

[0057] definition Let represent the decision variable to be solved, and let represent the th... i The load is at minimum operating power The additional power required based on the existing load. i The final set power is Based on the equipment's adjustability, Must meet:

[0058] The objective function for minimizing the total production loss due to the flexibly adjustable load P3 caused by power rationing can be expressed as:

[0059] because ,but The above formula can be simplified to: The objective function represents the penalty cost per unit power. The larger the load (i.e., the higher the production value), the larger the amount of resources should be allocated. This is to reduce the amount of power cut, thereby reducing overall production losses.

[0060] The sum of the power increments of all operational flexible loads must not exceed the total power of flexible regulation. This serves as the total power constraint for the optimization model, namely:

[0061] because The above formula is This constraint ensures that the final total power output of the entire plant does not exceed the requirements of the power rationing order.

[0062] For each load in the set, its power increment is limited by its regulation capability, i.e.:

[0063] Therefore, the expression for the established optimization model is:

[0064] Solving this model ensures that the finite power margin is utilized. Prioritize allocation to flexible loads with the highest unit power penalty cost to minimize production losses.

[0065] In step S4, the optimization model is iteratively solved using a distributed iterative algorithm to obtain the power setpoint for each load in the set of loads to be put into operation. In this embodiment of the invention, the distributed iterative algorithm is a distributed fast consensus algorithm. The core of this algorithm is to decompose the global optimization problem into local subproblems that are solved independently by the controllers of each flexibly adjustable load to be put into operation, and to use a dynamic global dual variable as a coordinator to guide the local decisions of each flexible load controller to converge quickly to the global optimum. This achieves the goal of refined and robust global optimization of core flexible resources under strict time constraints.

[0066] In this embodiment of the invention, the controller of each operational flexible load (such as an LF furnace) is considered as an agent. Each agent only needs to obtain its own operating parameters (including minimum operating power, maximum operating power, and effective penalty coefficient) and the global information broadcast by the coordinator to independently solve the local subproblem and update its own power decision variables.

[0067] Specifically, introduce global dual variables. and slack variables s An augmented Lagrangian function is constructed and decomposed into an alternating iterative solution. Specifically, for slack variables... s ,

[0068] The constructed formula for the augmented Lagrange function is as follows:

[0069] in, This represents the global dual variable, which is used when the total power exceeds the limit. It will rise; Represents slack variables. This represents the penalty parameter, and This penalty parameter controls the severity of the penalty for violating global constraints and is also a key parameter affecting the convergence speed of the algorithm.

[0070] This augmented Lagrangian function adds the equality constraints from the original optimization model to the objective function in the form of linear and quadratic terms, thus transforming the constrained optimization problem into an unconstrained (or more tractable) problem. Based on this, the optimal solution is approximated by alternately updating the variables.

[0071] The optimization model is iteratively solved using a distributed fast consensus iterative algorithm. Each iteration includes the following four steps to sequentially update the power increment, slack variables, and global dual variables for each flexible load. The iterative process continues until the convergence condition is met.

[0072] Step S41: According to the first k The global dual variable of each load in the load set during round iteration slack variables and the total power increment of all loads in the load set. Based on the augmented Lagrangian function, the parallel solution is obtained for the first... k+ The power increment of each load in the load set during the first iteration ; In the In each round of iteration, each agent receives the current global dual variable. slack variables and the sum of the power increments of each load in the set of loads that need to be put into operation. Then, the following local optimization problems are solved independently and in parallel to update its own power increment.

[0073] Because this is about The strictly convex quadratic programming problem has a solution within a closed interval, therefore each agent can find its analytical or numerical solution extremely quickly. Physically, when A higher value indicates that the current total power is significantly exceeded, and the algorithm will guide each load to decrease. (i.e., reduce power); when the load itself A larger value indicates higher production value, and the algorithm will tend to assign it a larger value. .

[0074] Each agent in the first... k+ In one iteration, the power increment is updated by solving the following constrained local optimization problem. :

[0075] The constraints are: ;in, Indicates the first i The load in the current number k+ The power increment to be solved in one iteration is the decision variable of the local optimization problem. Indicates the first i The load in the first k Power increment in round iteration, For the firstk The sum of the power increments of all flexible loads in the round of iteration.

[0076] Step S42: According to the first k+ The total power increment of all loads in the load set during the first iteration. Determine the first k+ The slack variables in the first iteration ; The coordinator collects the updated power increments reported by all agents. Calculate the current total power increment Therefore, the slack variables are updated based on the following formula:

[0077] in, In the first k+ In one iteration, the updated slack variables, This is the global dual variable in the k-th iteration; when the sum of the power increments of all flexibly adjustable loads in the set is close to or exceeds... When the sum of power increments is much smaller, the slack variables take smaller values ​​or even zero; when the sum of power increments is much smaller than At that time, slack variables absorb excess power margin, ensuring the feasibility of equality constraints; max The function ensures that the slack variable is always non-negative; This represents the penalty parameter.

[0078] Step S43: According to the first k+ The total power increment in one iteration The slack variable and the allowable adjustment of total power Determine the first k+ The global dual variable in the first iteration ; According to the k+ Total power increment in 1 iteration Calculate relative to the total allowable adjustable power The deviation is updated by the following formula for the global dual variable. :

[0079] According to the above formula, if the current total power increment exceeds the allowable total power adjustment, that is... Then improve ,make In the next iteration, the linear penalty term in the local objective function of each agent increases, thereby prompting each load to reduce power (more significantly reducing power demand); conversely, if the total power increment is insufficient, i.e. Then reduce ,make The linear penalty is reduced, increasing the power of each load. Through this adjustment, the system gradually converges to the optimal allocation that satisfies the global power constraint.

[0080] Step S44: Repeat steps S41-S43 above, when the original residual And dual residuals When the iteration stops, output the power increment of each load in the current load set. As the optimal solution; When the algorithm converges, the current This refers to the optimal power increment for each flexibly adjustable load. Based on this, the... i The final power setting for each load is: .

[0081] It should be noted that in actual engineering implementation, the penalty parameter... The convergence threshold can be preset based on experience or offline simulation (e.g., a value between 0.1 and 10). , Generally set to a smaller number to represent the required precision, it can be flexibly configured according to the control precision requirements. In addition, global dual variables... The initial value is usually set to 0, slack variable s The initial value can be set to 0. During the iteration process, it is gradually adjusted through the update formulas in steps 42 and 43. and s This allows it to automatically converge to the optimal value. The choice of initial values ​​does not affect the final convergence result.

[0082] Using the aforementioned distributed iterative algorithm, all controllers capable of flexibly adjusting loads only need to perform simple local calculations and require minimal communication (receiving). , , Report It can converge to the optimal solution within several to dozens of iterations, meeting the industrial requirement of second-level response.

[0083] For step S5, after completing the load switching decision in step S2 and the distributed iterative solution in step S4, the final power dispatch scheme for all loads in the entire plant has been obtained. Specifically, it includes: For the uninterruptible critical load P1, keep it in operation at its normal operating power. ; For general interruptible load P4 and important interruptible load P2, the switching status is determined according to step S2. implement, This indicates that the load has been cut off during operation. This indicates that the load has been put into operation.

[0084] For the flexibly adjustable load P3, the switching state is determined according to step S2. Execution, in which flexible loads are kept running ( The final power setting value is , This is the optimal power increment obtained when step S4 converges.

[0085] The above dispatching scheme is converted into power dispatching instructions. Instruction types are distinguished based on load regulation characteristics: For continuously adjustable loads (i.e., flexibly adjustable loads, such as LF furnaces), generate analog setting instructions that include the target power value; For interruptible loads (including general interruptible loads, important interruptible loads, and disconnected flexible loads), generate switching control commands (closing / opening).

[0086] Instructions are issued to field execution devices via standard industrial protocols. The protocols used include OPC UA, Modbus TCP, or the IEC 60870-5-104 protocol. Instructions are sent from the computing server and transmitted through the core switches of the industrial data acquisition network and industrial data acquisition gateways deployed in each substation or production workshop to load control PLCs, substation protection devices, or intelligent controllers. These execution devices then perform switching operations or power regulation.

[0087] On the monitoring screen at the energy management center, the system displays key information on power curtailment scheduling in real time, including: current total power, curtailment target, algorithm convergence curve, adjustment status of loads of each priority, total penalty cost J, power deficit or margin, and average power utilization rate of LF boilers. This information helps operators monitor the power curtailment response process in real time and intervene promptly when alarms are triggered.

[0088] Through steps S1 to S5, this embodiment of the invention forms a complete automated closed loop from data acquisition, load classification, discrete switching decision-making, distributed optimization solution to instruction generation and issuance, achieving the goals of second-level response and minimizing production losses.

[0089] Example 1 The following example, a typical power grid demand exceeding the limit event of a large steel enterprise, will be used to specifically illustrate the technical solution of the present invention. The power grid dispatch center issued an instruction to the enterprise's energy management center, requiring it to reduce the total power load of the entire plant to below 100,000 kW within 2 minutes. The current total load composition of the entire plant is shown in Table 1.

[0090] Table 1. Composition of Total Plant Load

[0091] (1) Feasibility verification of power curtailment Summarize the power of all P1 loads: Ensure that at least one flexibly adjustable load P3 (LF furnace) is operating with its minimum operating power: The minimum guaranteed power is:

[0092] The minimum guaranteed power is lower than the power rationing target. The power rationing order is feasible; proceed with the next steps.

[0093] (2) Discrete switching decision All general interruptible loads P4 are disconnected from operation, and the power of all flexible loads P3 is preset to the minimum power of 8000kW. At this time, the total power of the entire plant becomes: 46900 + 40000 + 24000 = 110900 kW At this time, the total power value is greater than the power rationing target. It exceeds the limit by 10900kW.

[0094] (3) Remove the flexibly adjustable load P3 (retain at least one unit) Arrange the three LF furnaces in ascending order according to their penalty coefficients. At this point, LF furnace #1 ( =100) is in the heat preservation period, with the minimum penalty coefficient, 2#LF furnace ( =200) is in the slagging stage, with a moderate penalty coefficient, 3#LF furnace ( =300) is in the critical refining period, with the maximum penalty coefficient (must be retained).

[0095] The No. 1 LF boiler and No. 2 LF boiler, which have lower penalty coefficients, were shut down first. At this point, the total power was 94,900 kW, which is already below the power rationing target. At this point, the total power allowed for flexible adjustment is:

[0096] Using the fast consistency algorithm described in step S4, the total power can be flexibly adjusted and distributed across the P3 level load. The final power setpoint for furnace #3 (LF) is: 8000 + 5100 = 13100. kW .

[0097] (4) Final verification The total power of the entire plant is: 46900 + 40000 + 13100 = 100000 kW The power rationing target has been met. .

[0098] (5) Issuance of instructions The following power dispatch instructions are generated and issued: cut off all P4 type loads, cut off the load of boiler #1 LF, cut off the load of boiler #2 LF, adjust the power of boiler #3 LF to 13100kW, and keep the remaining loads (P1 and P2 equipment that remain in operation) unchanged.

[0099] The above instructions are issued to the integrated protection devices, PLCs or intelligent controllers of each substation via standard industrial protocols (such as 104 protocol, Modbus TCP or OPC UA) for execution.

[0100] As can be seen from this embodiment, the present invention can complete the entire process from data acquisition, discrete decision-making, continuous optimization to instruction issuance within seconds. Under the premise of meeting the power rationing instructions, it maximizes the preservation of high-value production links (the No. 3 LF furnace during the critical refining period) and minimizes production losses.

[0101] Figure 2 This is a schematic diagram of the structure of a multi-objective power dispatch optimization system provided in an embodiment of the present invention. Figure 2 As shown, the optimization system includes an industrial data acquisition module, an optimization module, and a control module.

[0102] The industrial data acquisition module is used to acquire real-time operating data of each load in the target scenario and the load type of each load, wherein the load type includes flexibly adjustable loads and non-flexibly adjustable loads.

[0103] In one embodiment of the present invention, the industrial data acquisition module includes industrial data acquisition gateways and core switches for the industrial data acquisition network deployed in various substations or production workshops. The industrial data acquisition gateways serve as a bridge between the system and external heterogeneous data sources, acquiring data from the Energy Management System (EMS), Manufacturing Execution System (MES), and field devices via standard industrial protocols such as OPC UA, Modbus TCP, or the 104 protocol. The acquired data includes power data such as the total active power of all 10kV and above substations in the plant, the power of each outgoing circuit, and the status of each circuit switch, as well as production data such as the current operating status and technological stage (e.g., heating period, alloying period) of each production unit (e.g., converter, LF furnace, continuous casting machine, rolling line).

[0104] The collected data, after protocol conversion, data standardization, and edge caching, is uniformly aggregated through the core switch of the industrial data acquisition network and pushed to the computing server through the dedicated network switch of the power dispatching and optimization system. The core switch of the industrial data acquisition network is the core device for data acquisition. It connects to multiple industrial data acquisition gateways at the lower layer and to the dedicated network switch of the power dispatching and optimization system at the upper layer, ensuring the efficiency and reliability of data acquisition and transmission.

[0105] The optimization module is used to determine the switching status of each load and the set of loads that need to be put into operation from the flexibly adjustable loads according to the power requirements in the power restriction order, based on preset priorities and penalty coefficients; for the set of loads that need to be put into operation, an optimization model is established with minimizing its load adjustment cost as the optimization objective and the total allowable adjustable power of the load set as the constraint condition; the optimization model is iteratively solved based on a distributed iterative algorithm to obtain the power setpoint value of each load in the set of loads that need to be put into operation.

[0106] In one embodiment of the present invention, the optimization module is a computing server deployed in an enterprise control center or data center. This computing server is the core decision-making and control device of the system, responsible for running the load classification model, discrete switching logic, and distributed optimization solution program based on the fast consensus algorithm.

[0107] The computing server receives real-time load data from the industrial data acquisition module and performs the following operations based on the power requirements in the power rationing order: First, it executes discrete switching decisions according to a preset priority (generally interruptible loads first, flexibly adjustable loads second, important interruptible loads third, and uninterruptible critical loads last) and penalty coefficient to determine the switching status of each load and the set of flexibly adjustable loads that need to be put into operation; then, for the set of flexibly adjustable loads that need to be put into operation, it establishes an optimization model with minimizing load adjustment costs as the optimization objective and allowing total adjustable power as the constraint; finally, it uses the Distributed Faster Consensus Algorithm (ADMM) to iteratively solve the optimization model to obtain the power setpoint for each flexibly adjustable load put into operation. Simultaneously, the computing server triggers an alarm when it determines that the power rationing order exceeds the system's response capacity.

[0108] The control module is used to obtain the power dispatching command for each load based on the switching status of the non-flexible load and the power setting value of each load in the set of loads to be put into operation, and to control each load to execute the power dispatching command.

[0109] In one embodiment of the present invention, the function of the control module is implemented collaboratively by a computing server and field execution equipment. The computing server generates power dispatch instructions based on the above decision results: for continuously adjustable loads (such as flexibly adjustable loads like LF boilers), it generates analog quantity setting instructions (including target power values); for interruptible loads (including general interruptible loads, important interruptible loads, and flexibly disconnected loads), it generates switch quantity control instructions (closing / opening).

[0110] The generated instructions are transmitted via the dedicated network switch of the power dispatch optimization system and the core switch of the industrial data acquisition network, and then through the industrial data acquisition gateway to the load control PLCs, substation integrated protection devices, or intelligent controllers in the field. These execution devices transmit the collected equipment data to the industrial data acquisition gateway on the one hand, and receive instructions from the computing server through the industrial data gateway on the other hand, and perform switching operations (closing / opening) or power regulation accordingly, thereby realizing the physical execution of power rationing dispatch instructions.

[0111] In one embodiment of the present invention, the optimization system further includes the following auxiliary devices: A storage server, installed in the center of the data center, is configured with a time-series database to store the time-series data (such as historical load data, optimization process data, event records, etc.) generated by the optimization system, providing a data foundation for subsequent analysis and auditing.

[0112] The system operation client, installed in the control room, allows operators to remotely access the system. Operators can view a real-time overview of power curtailment scheduling through the client, including current total power, curtailment targets, algorithm convergence curves, load adjustment status for each priority level, total penalty cost, power deficit or margin, and average power utilization rate of the LF boiler. Operators can also manually intervene through the client when the system triggers an alarm.

[0113] The dedicated network switch for the power dispatch optimization system is deployed in the information room to build a high-speed local area network for the core equipment of the system. It connects the computing server, storage server and system operation client to ensure highly reliable and low-latency interaction of global dual variables, power commands and screen data between the core equipment.

[0114] Specifically, after the optimization system is started, the industrial data acquisition gateway continuously collects load data from EMS, MES, and field devices, and pushes it to the computing server after standardization processing. When a power curtailment command is received, the computing server calls the load classification model to classify the load of the entire plant, performs discrete switching decisions according to preset priorities and penalty coefficients, determines the switching status of each load and the set of flexible loads that need to be put into operation, then establishes an optimization model and uses a distributed fast consensus algorithm for iterative solution to obtain the fine power setpoint for each flexible load. The computing server converts the decision results into power dispatch commands (analog setting or switch control), and sends them to the field execution equipment (PLC, integrated protection device, or intelligent controller) via the industrial data acquisition network to complete the power curtailment dispatch operation.

[0115] Throughout the process, the storage server continuously records system operation data, and the operation client displays scheduling status information in real time, forming a complete automated closed loop from data acquisition, status decision-making, optimization solution to instruction issuance and execution.

[0116] In other embodiments of the present invention, the functions of the industrial data acquisition module and some control modules can also be integrated into the computing server. That is, the computing server communicates directly with the field execution equipment through standard industrial protocols, without needing to go through a separate industrial data acquisition gateway. Furthermore, the above modules can be deployed on a single computing device or distributed across multiple computing devices, working collaboratively through high-speed network communication; all of these are within the scope of protection of the present invention.

[0117] On the other hand, embodiments of the present invention also provide a machine-readable storage medium storing instructions that cause a machine to execute the above-described multi-objective power dispatch optimization method.

[0118] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0119] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations 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, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0120] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0121] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0122] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0123] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0124] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0125] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0126] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A multi-objective power dispatch optimization method, characterized in that, The optimization method includes: Acquire real-time operating data of each load in the target scenario and the load type of each load, wherein the load type includes flexibly adjustable loads and non-flexibly adjustable loads; Based on the power requirements in the power restriction order, and according to the preset priority and penalty coefficient, the switching status of each load and the set of loads that need to be put into operation among the flexibly adjustable loads are determined. For the load set that needs to be put into operation, an optimization model is established with minimizing its load adjustment cost as the optimization objective and the total allowable adjustment power of the load set as the constraint condition. The optimization model is iteratively solved using a distributed iterative algorithm to obtain the power setpoint for each load in the set of loads to be put into operation; and Based on the switching status of the non-flexible load and the power setting value of each load in the set of loads to be put into operation, the power dispatching command for each load is obtained and executed.

2. The optimization method according to claim 1, characterized in that, The non-flexible adjustable load includes: Uninterruptible critical loads related to safety or core processes, significant interruptible loads involving important auxiliary facilities, and general interruptible loads involving non-critical auxiliary production facilities; Specifically, a penalty coefficient is set for each load according to the load type. The magnitude of this penalty coefficient reflects the production loss caused when the load is subject to power rationing. The first penalty coefficient for the uninterruptible critical load is greater than the second penalty coefficient for the important interruptible load, the third penalty coefficient for the flexibly adjustable load is greater than the fourth penalty coefficient for the general interruptible load, and the second penalty coefficient is greater than the third penalty coefficient.

3. The optimization method according to claim 2, characterized in that, Before determining the switching status of each load and the set of loads to be put into operation from the flexibly adjustable loads according to preset priorities and penalty coefficients, the optimization method further includes: The minimum guaranteed power is determined based on the total power of all the aforementioned uninterruptible critical loads and the minimum operating power of at least one of the aforementioned flexibly adjustable loads; Determine whether the power requirement in the power restriction order is greater than the minimum guaranteed power. If so, determine the switching status of each load and the set of loads that need to be put into operation among the flexibly adjustable loads according to the preset priority and penalty coefficient. If not, trigger an alarm and terminate the subsequent optimization process.

4. The optimization method according to claim 2, characterized in that, The preset priorities, from highest to lowest, are: general interruptible loads, flexible adjustable loads, important interruptible loads, and uninterruptible critical loads. The higher the priority, the more preferentially the power-limiting command is executed to perform power-limiting operations.

5. The optimization method according to claim 4, characterized in that, The step of determining the switching status of each load and the set of loads that need to be put into operation from the flexibly adjustable loads according to preset priorities and penalty coefficients includes: According to the preset priority, the general interruptible load, the flexibly adjustable load, and the important interruptible load are processed sequentially, wherein... For the aforementioned general interruptible loads, based on the unit power penalty cost, each load in this category that is currently in operation is disconnected one by one in ascending order. After disconnecting each load, the first total power of all currently operating loads is recalculated, and it is determined whether the first total power meets the power requirements in the power restriction order. If so, the operation is terminated. If not, the power of the currently operating adjustable load is reduced to its minimum operating power. The second total power of all currently operating loads is calculated, and it is determined whether the second total power meets the power requirements in the power restriction order. If so, the operation is terminated. If not, then, while ensuring that at least one of the aforementioned flexibly adjustable loads remains operational, the remaining flexibly adjustable loads are disconnected one by one according to the unit power penalty cost. After each load is disconnected, the third total power of all currently operating loads is recalculated, and it is determined whether the third total power meets the power requirements in the power restriction order. If so, the operation is terminated. If not, according to the unit power penalty cost, each load in the important interruptible load that is in operation will be cut off in ascending order. After each one is cut off, the fourth total power of all loads currently in operation will be recalculated, and it will be determined whether the fourth total power meets the power requirements in the power limiting command. If yes, the operation will be terminated; if no, an alarm will be triggered. The unit power penalty cost is the ratio of the penalty coefficient corresponding to each load to its power during normal operation.

6. The optimization method according to claim 1, characterized in that, The formula for the optimization model is: in, Indicates load i The aforementioned unit power penalty cost, This indicates the maximum allowable power value of the loads in the set of loads that need to be put into operation within the flexibly adjustable loads. This represents the minimum allowable power value for the loads in the set of loads that need to be put into operation within the flexibly adjustable loads. Indicates load i The power increment to be solved Indicates load i The running status, The time indicates that the operation has been cut off. This indicates that it has been put into operation. This represents the total allowable adjustable power of the load set. This refers to the set of loads that need to be put into operation within the flexibly adjustable load.

7. The optimization method according to claim 6, characterized in that, The distributed iterative algorithm is a distributed fast consensus algorithm, which constructs an augmented Lagrangian function to solve the optimization model by pre-setting global dual variables and slack variables; The expression for the augmented Lagrange function is: in, Represents the global dual variable. Represents the slack variable, This represents the penalty parameter.

8. The optimization method according to claim 7, characterized in that, The iterative solution of the optimization model based on the distributed iterative algorithm includes: According to the k The global dual variable of each load in the load set during round iteration slack variables and the total power increment of all loads in the load set. Based on the augmented Lagrangian function, the parallel solution is obtained for the first... k+ The power increment of each load in the load set during the first iteration ; According to the k+ The total power increment of all loads in the load set during the first iteration. Determine the first k+ The slack variables in the first iteration ; According to the k+ The total power increment in one iteration The slack variable and the allowable adjustment of total power Determine the first k+ The global dual variable in the first iteration ; When the original residual And dual residuals When the iteration stops, output the power increment of each load in the current load set. As the optimal solution; Specifically, the power increment of each load in the current load set will be... The sum of the minimum operating power of the load and the minimum operating power of the load shall be used as the power setting value of the load.

9. The optimization method according to claim 8, characterized in that, According to the first k The global dual variable of each load in the load set during round iteration slack variables and the total power increment of all loads in the load set. Based on the augmented Lagrangian function, the parallel solution is obtained for the first... k+ The power increment of each load in the load set during the first iteration The calculation formula is as follows: 。 10. The optimization method according to claim 8, characterized in that, According to the first k+ The total power increment of all loads in the load set during the first iteration. Determine the first k+ The slack variables in the first iteration The calculation formula is as follows: According to the first k+ The total power increment in one iteration The slack variable and the allowable adjustment of total power Determine the first k+ The global dual variable in the first iteration The calculation formula is as follows: in, Indicates the first k+ The slack variables in the first iteration, Indicates the first k+ The total power increment in one iteration, Indicates the first k The global dual variable during round iteration, Indicates the first k+ The global dual variable in the first iteration, This represents the total allowable adjustable power of the load set. This represents the penalty parameter.

11. A multi-objective power dispatch optimization system, characterized in that, The optimization system includes: An industrial data acquisition module is used to acquire real-time operating data of each load in the target scenario and the load type of each load, wherein the load type includes flexibly adjustable loads and non-flexibly adjustable loads; The optimization module is used to determine the switching status of each load and the set of loads that need to be put into operation from the flexibly adjustable loads, according to the power requirements in the power restriction order and a preset priority and penalty coefficient. For the set of loads that need to be put into operation, an optimization model is established with minimizing its load adjustment cost as the optimization objective and the total allowable adjustable power of the load set as the constraint. The optimization model is iteratively solved based on a distributed iterative algorithm to obtain the power setpoint value of each load in the set of loads that need to be put into operation. The control module is used to obtain the power dispatching command for each load based on the switching status of the non-flexible load and the power setting value of each load in the set of loads to be put into operation, and to control each load to execute the power dispatching command.

12. A machine-readable storage medium having instructions stored thereon for causing a machine to perform the optimization method as described in any one of claims 1-10.

Images