A multi-type power resource output control method, device, equipment and storage medium

Abstract

The application discloses a multi-type power resource output control method, device and equipment and a storage medium, and is applied to the field of new-type power system operation control, a joint control state is composed of a resource output setting quantity and a network margin adjustment quantity, an auxiliary control state is generated based on a predicted marginal signal of a current period, and a public operation response quantity is acquired after the output setting of the current period is executed; single-resource response reconstruction is carried out based on the public operation response quantity, average reconstruction marginal adjustment signals are calculated in batches in combination with new energy prediction scenarios; the joint control state and the predicted marginal signal of the next control period are output by correcting the joint control state and the predicted marginal signal of the current period based on the average reconstruction marginal adjustment signals. The multi-type power resource output online control based on limited feedback reconstruction correction solves the problem that it is difficult to simultaneously realize online tracking, deviation suppression and network margin maintenance when feedback information is limited, resource coupling constraints are strong and prediction errors exist continuously.

Description

Technical Field

[0001] This invention relates to the field of novel power system operation control, and in particular to a method for controlling the output of multiple types of power resources, a device for controlling the output of multiple types of power resources, electronic equipment, and a computer-readable storage medium. Background Technology

[0002] As various types of power resources are continuously integrated into the energy system, system operation is gradually shifting from single centralized control to multi-resource collaborative control. New energy output exhibits volatility and prediction errors, while conventional adjustable power generation resources face constraints related to output limits, ramp-up, and operational status. Controllable loads also have comfort and recovery constraints. When multiple resources work together on the same power grid, they are further constrained by shared network margins such as line power flow, voltage, transformer capacity, or regional exchange. Therefore, the control process not only needs to set the output for each resource but also needs to suppress deviations caused by prediction errors within continuous control cycles and prevent output adjustments from further compressing the already strained network margins.

[0003] Among existing control methods, rolling optimization methods typically require a relatively complete objective function, constraint model, and multiple solutions, resulting in insufficient real-time performance when the control cycle is short or communication feedback is limited. Data-driven or disturbance feedback correction methods are highly dependent on the coverage of historical samples, the quality of disturbance response records, and the completeness of online feedback, leading to unstable correction effects when feedback results are limited or operating boundaries change. Single-resource optimization methods struggle to handle the coupling of new energy sources, conventional adjustable power generation, and controllable loads under shared network margins. Multi-objective optimization methods focus on overall solutions and rarely provide guidance on how to reconstruct an actual common feedback to an unexecuted auxiliary control state. While general online optimization methods can provide an iterative framework, they typically do not combine new energy prediction scenario batching, common response slope filtering, and network margin projection correction to form a closed loop oriented towards engineering control. Summary of the Invention

[0004] The purpose of this invention is to provide a method, device, equipment, and storage medium for controlling the output of multiple types of power resources, which can be applied to the field of new power system operation control. This method is based on online control of the output of multiple types of power resources with limited feedback reconstruction correction, which solves the problem that existing control methods are difficult to achieve online tracking, deviation suppression, and network margin maintenance simultaneously when feedback information is limited, resource coupling constraints are strong, and prediction errors persist.

[0005] To address the aforementioned technical problems, this invention provides a method for controlling the output of multiple types of power resources, comprising: A joint control state of multiple types of power resources, consisting of resource output setting and network margin adjustment, is constructed. Based on the predicted marginal signal of the current control cycle, the joint control state of the current control cycle is projected and predicted to obtain the auxiliary control state. Obtain the common operation response quantity under the joint control state in the current control cycle, and determine the single resource response estimate of each power resource under the auxiliary control state based on the common operation response quantity; Multiple output prediction scenarios are established, and the reconstruction marginal adjustment signal under each output prediction scenario is determined based on the single resource response estimate. The average reconstruction marginal adjustment signal is obtained by averaging the reconstruction marginal adjustment signals of all output prediction scenarios. The joint control state and the predicted marginal signal of the current control cycle are updated based on the average reconstructed marginal adjustment signal, and used as the joint control state and the predicted marginal signal of the next control cycle. In each control cycle, the output control of multiple types of power resources is performed based on the joint control state.

[0006] Optionally, the joint control state of the current control period is projected and predicted based on the predicted marginal signal of the current control period to obtain an auxiliary control state, including: The joint control state of the current control cycle is moved in the opposite direction of the predicted marginal signal of the current control cycle and projected onto the feasible set of the joint control states of the current control cycle to obtain the auxiliary control state. The expression for the auxiliary control state is: ; In the formula, This refers to the auxiliary control state. For the first projection operator, The joint control state is the state during the current control period t. The predicted marginal signal for the current control cycle, Update the step size when the state is positive.

[0007] Optionally, determining the individual resource response estimate of each power resource under the auxiliary control state based on the common operational response quantity includes: The response slope signal is determined, and the response slope signal is extrapolated along the direction of change of the power resources based on the common operation response quantity to obtain the single resource response estimate of each power resource under the auxiliary control state. The expression for the single-resource response estimate is: ; In the formula, The setpoint for the output of the nth candidate power resource under the auxiliary control state. This is the single-resource response estimate of the power resource in the nth time during the current control period t, under the output setpoint of the candidate resource. The setpoint for the power output of the power resource in the nth cycle of the current control period t. The response slope signal is the current control period t. This refers to the common operational response quantity for the current control period t.

[0008] Optionally, the response slope signal is determined, including: Obtain the response slope setting value, and determine the response slope setting value as the response slope signal; Alternatively, the response slope signal may be determined based on the common operating response quantity of the current control cycle and the previous control cycle and the corresponding total net injection quantity; The expression for the response slope signal is: ; In the formula, The common operational response quantity is the current control period t. The common operational response value of the previous control cycle is referred to as clip, where clip is the limiting operator. This is the total net injection amount from the previous control cycle. This refers to the total net injection amount for the current control cycle. This is the lower limit value of the response slope signal. This is the upper limit of the response slope signal. This is the initial response slope signal for the current control cycle. For the response slope filter coefficients, The response slope signal for the current control cycle. This refers to the response slope signal from the previous control cycle. It is a small positive quantity used to make the denominator non-zero.

[0009] Optionally, multiple output prediction scenarios are established, and the reconstructed marginal adjustment signal under each output prediction scenario is determined based on the single resource response estimate, including: Multiple output prediction scenarios are constructed based on the predicted distribution of available power output from new energy sources and / or prediction residual resampling. The reconstructed marginal adjustment signal for each of the output prediction scenarios is determined based on the single resource response estimate. The expression for the reconstructed marginal adjustment signal is: ; In the formula, This refers to the reconfigured marginal adjustment signal of the nth power resource in the current control period t under the s-th output prediction scenario. Set the output amount for the nth candidate power resource. This is the response slope signal for the current control cycle. For the s-th output prediction scenario in the current control cycle t, for gradient value, Let n be the local regulation function of the power resource in the current control period t under the s-th output prediction scenario. This is the single-resource response estimate of the nth power resource in the current control period t under the output setting of the candidate resource.

[0010] Optionally, updating the joint control state and the predicted marginal signal of the current control cycle based on the average reconstructed marginal adjustment signal, as the joint control state and the predicted marginal signal of the next control cycle, includes: A joint correction signal is obtained based on the combination of the average reconstructed marginal adjustment signal and the network margin signal; The joint control state of the current control cycle is updated based on the joint correction signal as the joint control state of the next control cycle; The predicted marginal signal of the current control cycle is updated based on the joint correction signal as the predicted marginal signal of the next control cycle.

[0011] Optionally, the expression for the joint control state in the next control cycle is: ; In the formula, The joint control state is for the next control cycle t+1. For the first projection operator, The joint control state in the current control cycle. This refers to the auxiliary control state. The joint correction signal, Update the step size when the state is positive; The expression for the predicted marginal signal in the next control cycle is: ; In the formula, The predicted marginal signal for the next control period t+1, For the second projection operator, The updated weights for the predicted marginal signal, The predicted marginal signal is for the current control period t. This refers to the joint correction signal.

[0012] To solve the above-mentioned technical problems, the present invention provides a multi-type power resource output control device, comprising: The first module is used to construct a joint control state of multiple types of power resources consisting of resource output setting and network margin adjustment. Based on the predicted marginal signal of the current control cycle, the joint control state of the current control cycle is projected and predicted to obtain the auxiliary control state. The second module is used to obtain the common operation response quantity under the joint control state in the current control cycle, and to determine the single resource response estimate of each power resource under the auxiliary control state based on the common operation response quantity. The third module is used to establish multiple output prediction scenarios, determine the reconstruction marginal adjustment signal under each output prediction scenario based on the single resource response estimate, and average the reconstruction marginal adjustment signals of all output prediction scenarios to obtain the average reconstruction marginal adjustment signal. The fourth module is used to update the joint control state and the predicted marginal signal of the current control cycle based on the average reconstructed marginal adjustment signal, as the joint control state and the predicted marginal signal of the next control cycle. The fifth module is used to control the output of multiple types of power resources based on the joint control state in each control cycle.

[0013] To solve the above-mentioned technical problems, the present invention provides an electronic device, comprising: Memory, used to store computer programs; A processor is used to implement the multi-type power resource output control method when executing the computer program.

[0014] To address the aforementioned technical problems, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the multi-type power resource output control method.

[0015] As can be seen, this invention constructs a joint control state for multiple types of power resources, consisting of resource output setting and network margin adjustment. Based on the predicted marginal signal of the current control cycle, it projects and predicts the joint control state of the current control cycle to obtain an auxiliary control state. It then acquires the common operating response quantity under the joint control state of the current control cycle and determines the single-resource response estimate of each power resource under the auxiliary control state based on the common operating response quantity. Multiple output prediction scenarios are established, and the reconfiguration marginal adjustment signal under each output prediction scenario is determined based on the single-resource response estimate. The reconfiguration marginal adjustment signals of all output prediction scenarios are averaged to obtain the average reconfiguration marginal adjustment signal. The joint control state and predicted marginal signal of the current control cycle are updated based on the average reconfiguration marginal adjustment signal, serving as the joint control state and predicted marginal signal for the next control cycle. In each control cycle, output control of multiple types of power resources is performed based on the joint control state.

[0016] This invention provides online control of multi-type power resource output based on finite feedback reconstruction correction, which solves the problem that existing control methods are difficult to achieve online tracking, deviation suppression and network margin maintenance simultaneously when feedback information is limited, resource coupling constraints are strong and prediction errors persist. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0018] Figure 1 A flowchart of a method for controlling the output of multiple types of power resources provided in an embodiment of the present invention; Figure 2 This is a structural block diagram of a multi-type power resource output control device provided in an embodiment of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] As new energy power generation resources, conventional adjustable power generation resources, controllable loads, and other adjustable resources are continuously integrated into distribution networks, microgrids, and integrated energy systems in industrial parks, system operation is gradually shifting from single centralized control to multi-resource collaborative control. New energy output exhibits volatility and prediction errors; conventional adjustable power generation resources face constraints related to output limits, ramp-up, and operational status; and controllable loads also have comfort and recovery constraints. When multiple resources act on the same power grid, they are also subject to shared network margin constraints such as line power flow, voltage, transformer capacity, or regional exchange. Therefore, the control process not only needs to set the output for each resource but also needs to suppress deviations caused by prediction errors within continuous control cycles and prevent output adjustments from further compressing the already strained network margin.

[0021] Existing technologies can be broadly categorized as follows: The first category is the source-grid-load-storage rolling optimization method, which obtains the output plan by constructing a multi-resource scheduling model and solving it repeatedly in each time period; the second category is the data-driven or disturbance feedback correction method, which relies on historical operating samples, disturbance response records, or trained predictors to correct the control output; the third category is the single-unit, single-type new energy, or local equipment output optimization method, which focuses on the power output or load status of a certain type of resource; the fourth category is the multi-energy multi-objective optimization method, which usually solves the problem comprehensively with objectives such as energy consumption, emissions, operating constraints, or safety margins; and the fifth category is the general online optimization or gradient tracking method, which focuses on convergence and feasibility in the continuous iteration process.

[0022] While the aforementioned methods can address some power output regulation issues, they still have shortcomings: Rolling optimization methods typically require a relatively complete objective function, constraint model, and multiple solutions, resulting in insufficient real-time performance when the control cycle is short or communication feedback is limited; Data-driven or disturbance feedback correction methods are highly dependent on the coverage of historical samples, the quality of disturbance response records, and the completeness of online feedback, leading to unstable correction effects when feedback results are limited or operating boundaries change; Single-resource optimization methods struggle to handle the coupling of new energy sources, conventional adjustable power generation, and controllable loads under shared network margins; Multi-objective optimization methods focus on overall solutions and rarely provide guidance on how to reconstruct an actual common feedback to an unexecuted auxiliary control state; Although general online optimization methods can provide an iterative framework, they typically do not combine new energy prediction scenario batching, common response slope filtering, and network margin projection correction to form a closed loop oriented towards engineering control.

[0023] The purpose of this invention is to provide an online control method for the output of multiple types of power resources based on finite feedback reconfiguration correction, so as to solve the problem that existing control methods are difficult to achieve online tracking, deviation suppression and network margin maintenance simultaneously when feedback information is limited, resource coupling constraints are strong and prediction errors persist.

[0024] The following combination Figure 1 , Figure 1 A flowchart of a method for controlling the output of multiple types of power resources provided in an embodiment of the present invention, the method may include: S101: Construct a joint control state of multiple types of power resources consisting of resource output setting and network margin adjustment. Based on the predicted marginal signal of the current control cycle, project and predict the joint control state of the current control cycle to obtain the auxiliary control state.

[0025] This embodiment does not limit the specific types of multiple types of power resources. Generally, they can include new energy power generation resources, conventional adjustable power generation resources, and controllable load resources, etc. There are no limitations here, and the settings can be based on actual applications.

[0026] This embodiment can collect operational data of various types of power resources and network operation constraint data, and establish a joint control state based on the collected data.

[0027] This embodiment can acquire the available output prediction, prediction variance, and historical residuals of each new energy power generation resource through a data acquisition interface; it can also acquire the current output, output upper and lower limits, ramping capability, and adjustable state parameters of each conventional adjustable power generation resource; and it can acquire the aggregated adjustable capacity, response upper and lower limits, recovery time window, or aggregated execution status of each controllable load resource. Simultaneously, it acquires network constraint data such as line power flow, voltage, transformer capacity, or regional switching, and forms a shared network margin matrix and residual margin vector based on linearized sensitivity. Furthermore, it can read the common operating response and summarized output change records saved in the previous control cycle as historical feedback records for subsequent response slope estimation.

[0028] The aforementioned data typically originates from resource controllers, power prediction systems, dispatch automation systems, energy management systems, network security verification functions, or monitoring systems. For the controllable load side, the collection of individual household identities, billing information, or individual household electricity consumption details is not required; aggregated adjustable capacity boundaries and execution status can be used. The data acquisition output of this embodiment may include resource-side output boundaries, network margin, historical common response records, and prediction residual information required for generating new energy prediction scenarios. These data then enter the subsequent feasible set construction, feedback reconstruction, and scenario batch correction calculation processes.

[0029] The common operational feedback for the current control cycle is not generated in this embodiment, but is obtained after the resource output setting of the current cycle is executed. If the resource output setting of a certain power resource is missing in the current cycle communication, the resource output setting of the previous control cycle can be used as a reference, and the adjustable range of that power resource can be temporarily narrowed to a safe range.

[0030] This embodiment can construct a joint control state consisting of resource output setting and network margin adjustment based on the upper and lower limits of each power resource output, its adjustable capacity, and the network margin, and determine the feasible set of joint control states.

[0031] Specifically, in this embodiment, the output settings of each power resource can be stacked into a resource-side control vector. The output increase direction of new energy power generation resources and conventional adjustable power generation resources is taken as positive, and controllable load resources are included in the same direction according to the equivalent adjustable amount, which facilitates the subsequent processing of multiple types of resources using a unified projection calculation.

[0032] The resource-side control vector can be represented as: ; Where t represents the index of the control cycle, Let represent the resource-side control vector for the t-th control cycle. In this embodiment, the current control cycle can be set to t. This represents the stacking operation of column vectors. This represents the output setting of new energy power generation resource i in the current control period t. Index for new energy power generation resources For the collection of new energy power generation resources, Indicates conventional adjustable power generation resources In the cycle The output setting, This is an index for conventional adjustable power generation resources. It is a collection of conventional adjustable power generation resources. Indicates controllable load resources In the current control cycle The adjustment amount, Index for controllable load resources For a controllable load resource set, This represents the feasible set of resources, which is formed by the upper and lower limits of local output, ramping limits, and adjustable capacity of each resource.

[0033] The joint control state can be represented as: ; in, This represents the joint control state of the current control cycle, consisting of the resource-side control vector and the network margin adjustment. This represents the resource-side control vector for the current control cycle. This represents the network margin adjustment vector for the current control cycle, used to describe the tightness of shared network constraints. Represents the feasible set of joint control states. This represents the stacking operation of column vectors. This represents the feasible set on the resource side. This represents the set of non-negative limits for network margin adjustment. This represents the Cartesian product.

[0034] Based on the above definition, the resource local output boundary is determined by the set The shared network margin is constrained by both inequalities and margin adjustment. All subsequent prediction and correction steps are performed within the feasible set of the joint control state. The projection is performed on the data source, so the updated resource output settings will not deviate from the adjustable range of resources and network security margin requirements. In specific calculations, projection can be understood as finding the nearest feasible joint control state from the candidate joint control states that satisfies the local output boundary and network margin constraints.

[0035] This embodiment can project and predict the joint control state of the current control cycle based on the predicted marginal signal of the current control cycle to obtain the auxiliary control state.

[0036] In this embodiment, the predicted marginal signal is a vector used to predict the joint correction direction of the current control cycle. Its dimension is consistent with the joint control state, including predicted correction components from the resource side and the network side. The network margin adjustment is a network-side adjustment state used to describe the tightness of shared network constraints. This signal is updated by weighting the joint correction signal from the previous control cycle with the predicted marginal signal used in this cycle, and then obtaining it through amplitude-limited projection. The specific update method is described in the subsequent predicted marginal signal update steps. In the first control cycle, this signal can be initialized as a zero vector.

[0037] Specifically, in this embodiment, the joint control state of the current control cycle can be moved in the opposite direction of the predicted marginal signal of the current control cycle and projected onto the feasible set of the joint control states of the current control cycle to obtain an auxiliary control state. This auxiliary control state is not directly issued for execution, but is used to evaluate the correction direction of the current control cycle.

[0038] The expression for the auxiliary control state is: ; In the formula, To assist in controlling the state, This only represents the intermediate calculation index. Let be the first projection operator, representing the feasible set for projection onto the joint control state. The projection operator, This represents the joint control state during the current control period t. This is the predicted marginal signal for the current control cycle. Update the step size when the state is positive.

[0039] This step provides a candidate state for calculation but not for actual execution. After obtaining common operational feedback for the current cycle, the actual feedback information is reconstructed into this candidate state, thus obtaining a direction that can be used to correct the output setting for the next cycle. If the predicted marginal signal of the previous cycle is limited or initialized to a zero vector, the auxiliary control state will degenerate to the current feasible state or its neighboring feasible state, without affecting the continued execution of subsequent feedback reconstruction steps.

[0040] S102: Obtain the common operation response quantity under the joint control state of the current control cycle, and determine the single resource response estimate of each power resource under the auxiliary control state based on the common operation response quantity.

[0041] This embodiment can obtain the common operational response quantity under the joint control state of the current control cycle. That is, after executing the resource output setting in the joint control state in the current control cycle, the monitoring system will issue or calculate a common operational feedback result, i.e., the common operational response quantity. The common operational feedback result includes at least the actual aggregated response quantity and a response slope signal used to reflect the intensity of the impact of changes in aggregated output.

[0042] This embodiment does not limit the specific type of common operational response quantity. It can generally be a system adjustment marginal index, power balance response index, network congestion intensity index, or an equivalent common response index defined by the control platform.

[0043] To describe the impact of changes in aggregate output near the current operating point on the common operating response, an approximate linear relationship is established between the common operating response and the changes in aggregate output, expressed as: ; in, Indicates the current control cycle Observed public operation response volume, This represents the baseline item in the public operation response quantity. This represents the response slope near the current operating point, and is taken as a positive value. Indicates the current control cycle The total net injection volume. The vicinity of the current operating point refers to the area covered by observable output changes within the current control cycle or adjacent short-term windows, using the current actual resource output setting and current common operation feedback results as references. This relationship is used to approximately describe the impact of changes in the total net injection volume within the aforementioned range on the common operation response.

[0044] The expression for the total net injection amount is: ; In the formula, For the current control cycle The total net injection amount This indicates that all elements are 1 and the dimension is equal to the sum of all elements and the sum of all elements. Consistent column vectors This indicates the transpose operation. This represents the resource-side control vector for the current control period t. This represents the baseline quantity after conversion of external load and uncontrollable injection.

[0045] This embodiment does not require obtaining the complete response function; it only needs to utilize the common operational feedback result and the estimated response slope to reconstruct the single-resource response estimate. The direction of the common operational response quantity is kept consistent during configuration; that is, an increase in the response quantity indicates an increase in the operational deviation that needs to be suppressed. If the control platform uses the opposite definition, then the opposite number is uniformly taken during access.

[0046] This embodiment can determine the response slope signal, and use the common operation response quantity as a benchmark to extrapolate along the change direction of power resources using the response slope signal to obtain the single resource response estimate of each power resource under the auxiliary control state.

[0047] This embodiment does not limit the method of determining the response slope signal. It can generally be given directly by the system or estimated based on the common operating response quantity and the summary output change record within a short time window.

[0048] For example, obtain the response slope setting value and determine the response slope setting value as the response slope signal.

[0049] Alternatively, the response slope signal can be determined based on the common operating response quantity of the current control cycle and the previous control cycle, and the corresponding total net injection quantity. The expression for the response slope signal is: ; In the formula, This represents the common operational response quantity for the current control period t. The clip is the common operating response value of the previous control cycle, where clip is the clipping operator. This represents the total net injection amount from the previous control period. This represents the total net injection amount for the current control cycle. To respond to the lower limit of the slope signal, To respond to the upper limit of the slope signal, This is the initial response slope signal for the current control cycle. The slope filter coefficients range from 0 to 1. This is the response slope signal for the current control cycle. This is the response slope signal from the previous control cycle. A small positive quantity used to make the denominator non-zero. `clip` restricts the input of the first term within the parentheses to a specific value. to between.

[0050] To suppress abrupt changes, the estimated response slope signal is clipped and smoothed. In the above formula, clip is a clipping operator, which means that when the input is less than the lower limit, the lower limit is taken; when the input is greater than the upper limit, the upper limit is taken; and when the input is between the upper and lower limits, the input value remains unchanged.

[0051] Furthermore, after obtaining the initial response slope signal, filtering can be performed to obtain the filtered response slope signal. .

[0052] This embodiment uses the common operation response as a benchmark and extrapolates along the direction of power resource change using the response slope signal to obtain the individual resource response estimate for each power resource under auxiliary control conditions. Therefore, even if the auxiliary control state is not actually executed, an individual resource response estimate that can be used for marginal adjustment calculations can still be generated. When the aggregate net injection change is too small or the slope estimate fluctuates too much, amplitude limiting and smoothing processes can prevent single measurement disturbances from being directly amplified into excessive output corrections.

[0053] The expression for the single-resource response estimate can be: ; In the formula, This is the setpoint for the output of the nth power resource as a candidate resource under auxiliary control conditions. This is the single-resource response estimate for the nth power resource in the current control period t, under the candidate resource output setpoint. This is the setpoint for the resource output of the nth power resource in the current control cycle t. This is the response slope signal for the current control period t. This represents the common operational response quantity for the current control cycle t.

[0054] S103: Establish multiple output prediction scenarios, determine the reconstruction marginal adjustment signal under each output prediction scenario based on the single resource response estimate, and average the reconstruction marginal adjustment signals of all output prediction scenarios to obtain the average reconstruction marginal adjustment signal.

[0055] This embodiment can establish multiple new energy power output prediction scenarios in batches. This embodiment does not limit the specific method of establishing the scenarios. Generally, multiple power output prediction scenarios can be constructed based on the predicted distribution of available new energy power output and / or prediction residual resampling. For each power output prediction scenario, the reconstructed marginal adjustment signal of each power resource under auxiliary control state is calculated.

[0056] Specifically, the reconstructed marginal adjustment signal for each output prediction scenario is determined based on the single resource response estimate. The expression for the reconstructed marginal adjustment signal is as follows: ; In the formula, This represents the reconstructed marginal adjustment signal for the nth power resource in the current control period t under the sth power output prediction scenario. Set the output amount for the candidate power resource of the nth power resource. This is the response slope signal for the current control cycle. For the s-th output prediction scenario in the current control period t, for gradient value, Let be the local regulation function of the nth power resource in the current control period t under the s-th output prediction scenario. This is the single-resource response estimate for the nth power resource in the current control period t, under the candidate resource output setpoint. This is the slope compensation term.

[0057] The local regulation function is a differentiable function that describes the technical regulation cost or constraint deviation of power resources under the corresponding output prediction scenario, and is differentiable with respect to the output setting of candidate resources. By calculating the gradient of the local regulation function with respect to the output setting of candidate resources, a local gradient term is obtained for constructing the reconstructed marginal regulation signal. In this calculation, the local regulation function can be selected according to the resource type: new energy power generation resources can include available output deviation and curtailment penalty; conventional adjustable power generation resources can include output change and ramp-up penalty; and controllable load resources can include regulation amount and recovery constraint penalty. The above function is used to describe the technical regulation relationship and constraint deviation on the resource side.

[0058] The local adjustment function should be differentiable with respect to the output setting of the candidate resource in order to calculate the local gradient term in the aforementioned reconstructed marginal adjustment signal. In one implementation, a quadratic penalty function for reference output deviation and output change can be used: ; in, Indicates the first The power resources in the current control cycle , No. Reference output for an output prediction scenario and These represent the reference output deviation penalty coefficient and the output change penalty coefficient, respectively, both of which are non-negative. The subscripts 1 and 2 are only used to distinguish between the two types of penalty coefficients. The meanings of the other symbols are the same as above. The reference output and the corresponding resource output setting adopt the same dimensions and positive / negative direction conventions. The reference output of new energy resources can be determined by the predicted available output of the scenario. Other resources can use the reference setting in the local operation plan. The reference setting of controllable load adopts the adjustment quantity caliber.

[0059] In the example above, the gradient of the local adjustment function is: ; When calculating this gradient, the resource output setpoint, scene reference output, and penalty coefficient already executed in the current cycle remain unchanged. This gradient is substituted into the aforementioned expression for reconstructing the marginal adjustment signal, and then combined with the single resource response estimate and slope compensation term to complete the calculation. The aforementioned quadratic function is one optional form; the local adjustment function can also adopt other differentiable forms depending on the resource characteristics.

[0060] Furthermore, the average reconstructed marginal adjustment signal is obtained by averaging the reconstructed marginal adjustment signals of all output prediction scenarios. The expression for the average reconstructed marginal adjustment signal is as follows: ; In the formula, To reconstruct the marginal adjustment signal on average, B represents the number of all output prediction scenarios. This represents the summation of the reconstructed marginal adjustment signals across all output prediction scenarios. The reconstructed marginal adjustment signal vector is composed of the reconstructed marginal adjustment signals of all power resources in the current control period t under the s-th power output prediction scenario.

[0061] The single-scenario reconstruction marginal adjustment signal reflects the direction of further adjustments to various resources under a certain new energy forecast scenario, while the averaged signal weakens the impact of single forecast bias and measurement noise. This averaged signal serves as the resource-side correction direction input for subsequent correction processes. The scenario batch size can be set according to the allowable computation time of the control cycle. When the computation time is short, historical residual sampling can be used to form a smaller batch to maintain online speed.

[0062] S104: Update the joint control state and predicted marginal signal of the current control cycle based on the average reconstructed marginal adjustment signal, and use it as the joint control state and predicted marginal signal of the next control cycle.

[0063] This embodiment can obtain a joint correction signal based on the combination of the average reconstructed marginal adjustment signal and the network margin signal: ; In the formula, For joint correction signals, To reconstruct the marginal adjustment signal on average, Represents the network margin sensitivity matrix transpose, This represents a vector representing the stacked network margin adjustment values. This represents the network residual margin vector for the current control period t. This represents the linear impact of changes in resource-side output on network constraints. This represents the remaining margin component after deducting the resource-side impact, and h is the candidate resource-side control vector obtained by stacking the candidate resource output settings. This represents the column vector stacking operator, which stacks the resource-side correction component and the network margin component into a joint correction vector in this formula.

[0064] In the above formula, `col` is a column vector stacking operator, which means that the first component within the parentheses is used as the upper half of the joint correction signal, the second component is used as the lower half of the joint correction signal, and they are concatenated into a column vector in top-to-bottom order. When the margin of a certain shared network approaches the lower limit or exceeds the limit, the adjustment amount corresponding to the margin is fed back to the output direction of each resource through the sensitivity matrix to suppress further output adjustments that compress the margin.

[0065] Among them, the shared network margin sensitivity matrix and the network residual margin vector need to satisfy: ; ; in, This represents the shared network margin sensitivity matrix, where each row corresponds to network constraints such as lines, transformers, voltage, or area switching. Let represent the resource-side control vector of the current control period t, and h represent the candidate resource-side control vector for the auxiliary control state. Indicates the current control cycle The network residual margin vector, This indicates that none of the network constraint components exceed the corresponding residual margin.

[0066] This embodiment can calculate the linear impact of candidate resource output on network constraints based on the candidate resource output setting and the shared network margin sensitivity matrix in the auxiliary control state. The linear impact is then subtracted from the network remaining margin in the current control cycle to obtain the network margin signal. The network margin signal is used as the network-side correction component of the joint correction signal to update the network margin adjustment.

[0067] The network margin adjustment in the joint control state is a different quantity from the aforementioned network margin signal. The former is the network-side adjustment state updated with the control cycle, achieved through... Participate in resource-side correction. The latter is... , represents the remaining margin component after deducting the influence of candidate resources, and is used as the network-side correction component to update the network margin adjustment.

[0068] In this embodiment, the joint control state of the current control cycle can be updated based on the joint correction signal as the joint control state of the next control cycle. The expression for the joint control state of the next control cycle is: ; In the formula, This represents the joint control state for the next control cycle t+1. For the first projection operator, This represents the joint control state for the current control cycle. To assist in controlling the state, For joint correction signals, Update the step size when the state is positive.

[0069] The aforementioned projection correction step processes two types of information simultaneously: on the resource side, the reconstructed marginal adjustment signal provides the deviation suppression direction, and on the network side, the remaining margin component provides the safety constraint direction. The projection operation ensures that the corrected joint control state falls back into the feasible set; therefore, this embodiment can directly extract the output setting for the next cycle from the corrected joint control state. The upper part of the joint correction signal acts on the resource output setting direction, and the lower part acts on the network margin adjustment direction; both parts jointly determine whether to increase, decrease, or maintain the output of each resource in the next cycle.

[0070] The output setpoints of each resource are extracted from the projected and corrected joint control state, and after safety checks, are sent to the new energy controller, conventional adjustable generator controller, and controllable load controller via the command issuance interface. The updated resource-side control vector for the next control cycle can be: ; in, This represents the resource-side control vector extracted from the joint control state after projection correction for the next cycle. Indicates new energy power generation resources The output setting for the next cycle, Indicates conventional adjustable power generation resources The output setting for the next cycle, Indicates controllable load resources The adjustment amount in the next cycle, This represents the stacking operation of column vectors. , and represents the set of new energy power generation resources, the set of conventional adjustable power generation resources, and the set of controllable load resources, respectively. i, g, and d represent the indexes of new energy power generation resources, conventional adjustable power generation resources, and controllable load resources, respectively.

[0071] Furthermore, this embodiment can also use the joint correction signal of the current period to update the prediction marginal signal of the next period. The prediction marginal signal of the current control period is updated based on the joint correction signal as the prediction marginal signal of the next control period. The expression for the prediction marginal signal of the next control period is: ; In the formula, This is the predicted marginal signal for the next control period t+1. The second projection operator represents the projection onto the limiting set. The operator, This represents the norm-limited set of the predicted marginal signal. To predict the updated weights of the marginal signal, This is the predicted marginal signal for the current control period t. This is a joint correction signal.

[0072] The updated predicted marginal signal is used in the next control cycle, allowing the feedback reconstruction and network margin correction results of this cycle to be carried over to the auxiliary state generation of the next cycle. At this point, the online closed loop of a control cycle ends, and execution repeats in the next cycle. If the safety check detects any instruction exceeding the controller's allowable range, the resource output setting is truncated to the allowable range, and the state is updated with the actual executed value in the data acquisition of the next cycle.

[0073] S105: In each control cycle, output control of multiple types of power resources is performed based on the joint control state.

[0074] In this embodiment, during each control cycle, the response can be reconstructed based on limited feedback data, and the joint control state is updated based on the reconstructed response. The joint control state is corrected through iterative iteration in each control cycle, and the output control of various types of power resources is performed based on the corrected joint control state.

[0075] This embodiment does not require the control platform to repeatedly test at the candidate output setting. Instead, it uses the actual common feedback and filter response slope of the current cycle to reconstruct the common response quantity to the auxiliary control state, which is suitable for online operating environments where feedback information is limited.

[0076] This embodiment integrates new energy sources, conventional adjustable power generation, and controllable loads into the joint control state, and maps shared network constraints to the resource output direction through network margin adjustment, so that the output update takes into account both resource-side deviation and network-side constraints.

[0077] In this embodiment, a batch of new energy prediction residual scenarios is introduced in the correction step to average the reconstructed marginal adjustment signal, thereby reducing the impact of single prediction error, measurement noise and common feedback disturbance on the control direction.

[0078] This embodiment uses a weighted update and limiting mechanism based on predicted marginal signals to enable the control direction to follow the time-varying changes in load, available renewable energy output, and network margin, while avoiding overcorrection caused by abnormal feedback.

[0079] Based on the above embodiments, the present invention provides online control of multi-type power resource output based on finite feedback reconstruction correction, which solves the problem that existing control methods are difficult to simultaneously achieve online tracking, deviation suppression and network margin maintenance when feedback information is limited, resource coupling constraints are strong and prediction errors persist.

[0080] The following combination Figure 2 , Figure 2 This is a structural block diagram of a multi-type power resource output control device provided in an embodiment of the present invention. The device may include: The first module is used to construct the joint control state of multiple types of power resources, which consists of resource output setting and network margin adjustment. Based on the predicted marginal signal of the current control cycle, the joint control state of the current control cycle is projected and predicted to obtain the auxiliary control state. The second module is used to obtain the common operation response quantity under the joint control state of the current control cycle, and to determine the single resource response estimate of each power resource under the auxiliary control state based on the common operation response quantity. The third module is used to establish multiple output prediction scenarios, determine the reconstruction marginal adjustment signal under each output prediction scenario based on the single resource response estimate, and average the reconstruction marginal adjustment signals of all output prediction scenarios to obtain the average reconstruction marginal adjustment signal. The fourth module is used to update the joint control state and predicted marginal signal of the current control cycle based on the average reconstructed marginal adjustment signal, and to serve as the joint control state and predicted marginal signal for the next control cycle. The fifth module is used to control the output of multiple types of power resources based on the joint control state in each control cycle.

[0081] Based on the above embodiments, the present invention provides online control of multi-type power resource output based on finite feedback reconstruction correction, which solves the problem that existing control methods are difficult to simultaneously achieve online tracking, deviation suppression and network margin maintenance when feedback information is limited, resource coupling constraints are strong and prediction errors persist.

[0082] Based on the above embodiments, the first module 100 may include: The first unit is used to move the joint control state of the current control cycle in the opposite direction of the predicted marginal signal of the current control cycle and project it onto the feasible set of the joint control states of the current control cycle to obtain the auxiliary control state. The expression for the auxiliary control state is: ; In the formula, To assist in controlling the state, For the first projection operator, This represents the joint control state during the current control period t. This is the predicted marginal signal for the current control cycle. Update the step size when the state is positive.

[0083] Based on the above embodiments, the second module 200 may include: The second unit is used to determine the response slope signal. Based on the common operation response quantity, the response slope signal is extrapolated along the direction of change of power resources to obtain the single resource response estimate of each power resource under the auxiliary control state. The expression for the single-resource response estimate is: ; In the formula, This is the setpoint for the output of the nth power resource as a candidate resource under auxiliary control conditions. This is the single-resource response estimate for the nth power resource in the current control period t, under the candidate resource output setpoint. This is the setpoint for the resource output of the nth power resource in the current control cycle t. This is the response slope signal for the current control period t. This represents the common operational response quantity for the current control cycle t.

[0084] Based on the above embodiments, the second unit may include: The first subunit is used to acquire the response slope setpoint and determine the response slope setpoint as the response slope signal; The second subunit is used to determine the response slope signal based on the common operating response quantity of the current control cycle and the previous control cycle and the corresponding total net injection quantity. The expression for the response slope signal is: ; In the formula, This represents the common operational response quantity for the current control period t. The clip is the common operating response value of the previous control cycle, where clip is the clipping operator. This represents the total net injection amount from the previous control period. This represents the total net injection amount for the current control cycle. To respond to the lower limit of the slope signal, To respond to the upper limit of the slope signal, This is the initial response slope signal for the current control cycle. For the response slope filter coefficients, This is the response slope signal for the current control cycle. This is the response slope signal from the previous control cycle. It is a small positive quantity used to make the denominator non-zero.

[0085] Based on the above embodiments, the third module 300 may include: The third unit is used to construct multiple power output prediction scenarios based on the predicted distribution of available power output from new energy sources and / or prediction residual resampling. The fourth unit is used to determine the reconstructed marginal adjustment signal under each output prediction scenario based on the single resource response estimate; The expression for reconstructing the marginal adjustment signal is: ; In the formula, This represents the reconstructed marginal adjustment signal for the nth power resource in the current control period t under the sth power output prediction scenario. Set the output amount for the candidate power resource of the nth power resource. This is the response slope signal for the current control cycle. For the s-th output prediction scenario in the current control period t, for gradient value, Let be the local regulation function of the nth power resource in the current control period t under the s-th output prediction scenario. This is the single-resource response estimate for the nth power resource in the current control period t under the candidate resource output setting.

[0086] Based on the above embodiments, the fourth module 400 may include: The fifth unit is used to obtain the joint correction signal based on the combination of the average reconstructed marginal adjustment signal and the network margin signal; The sixth unit is used to update the joint control state of the current control cycle based on the joint correction signal as the joint control state of the next control cycle. The seventh unit is used to update the prediction marginal signal of the current control cycle as the prediction marginal signal of the next control cycle based on the joint correction signal.

[0087] Based on the above embodiments, the expression for the joint control state in the next control cycle is: ; In the formula, This represents the joint control state for the next control cycle t+1. For the first projection operator, This represents the joint control state for the current control cycle. To assist in controlling the state, For joint correction signals, Update the step size when the state is positive; The expression for the predicted marginal signal of the next control cycle is: ; In the formula, This is the predicted marginal signal for the next control period t+1. For the second projection operator, To predict the updated weights of the marginal signal, This is the predicted marginal signal for the current control period t. This is a joint correction signal.

[0088] Based on the above embodiments, the present invention also provides an electronic device, which may include a memory and a processor. The memory stores a computer program, and when the processor calls the computer program in the memory, it can implement the steps provided in the above embodiments. Of course, the device may also include various necessary network interfaces, a power supply, and other components.

[0089] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by an execution terminal or processor, can implement the method provided in the embodiments of the present invention; the storage medium may include various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0090] In this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, without necessarily requiring or implying any such actual relationship or order between these entities or operations. Furthermore, 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 a process, method, article, or apparatus. Without further limitation, 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 said element.

Claims

1. A method for controlling the output of multiple types of power resources, characterized in that, include: A joint control state of multiple types of power resources, consisting of resource output setting and network margin adjustment, is constructed. Based on the predicted marginal signal of the current control cycle, the joint control state of the current control cycle is projected and predicted to obtain the auxiliary control state. Obtain the common operation response quantity under the joint control state in the current control cycle, and determine the single resource response estimate of each power resource under the auxiliary control state based on the common operation response quantity; Multiple output prediction scenarios are established, and the reconstruction marginal adjustment signal under each output prediction scenario is determined based on the single resource response estimate. The average reconstruction marginal adjustment signal is obtained by averaging the reconstruction marginal adjustment signals of all output prediction scenarios. The joint control state and the predicted marginal signal of the current control cycle are updated based on the average reconstructed marginal adjustment signal, and used as the joint control state and the predicted marginal signal of the next control cycle. In each control cycle, the output control of multiple types of power resources is performed based on the joint control state.

2. The multi-type power resource output control method according to claim 1, characterized in that, Based on the predicted marginal signal of the current control cycle, the joint control state of the current control cycle is projected and predicted to obtain the auxiliary control state, including: The joint control state of the current control cycle is moved in the opposite direction of the predicted marginal signal of the current control cycle and projected onto the feasible set of the joint control states of the current control cycle to obtain the auxiliary control state. The expression for the auxiliary control state is: ； In the formula, This refers to the auxiliary control state. For the first projection operator, The joint control state is the state during the current control period t. The predicted marginal signal for the current control cycle, Update the step size when the state is positive.

3. The multi-type power resource output control method according to claim 1, characterized in that, Determining the individual resource response estimate of each power resource under the auxiliary control state based on the common operation response quantity includes: The response slope signal is determined, and the response slope signal is extrapolated along the direction of change of the power resources based on the common operation response quantity to obtain the single resource response estimate of each power resource under the auxiliary control state. The expression for the single-resource response estimate is: ； In the formula, The setpoint for the output of the nth candidate power resource under the auxiliary control state. This is the single-resource response estimate of the power resource in the nth time during the current control period t, under the output setpoint of the candidate resource. The setpoint for the power output of the power resource in the nth cycle of the current control period t. The response slope signal is the current control period t. This refers to the common operational response quantity for the current control period t.

4. The multi-type power resource output control method according to claim 3, characterized in that, Determine the response slope signal, including: Obtain the response slope setting value, and determine the response slope setting value as the response slope signal; Alternatively, the response slope signal may be determined based on the common operating response quantity of the current control cycle and the previous control cycle and the corresponding total net injection quantity; The expression for the response slope signal is: ； In the formula, The common operational response quantity is the current control period t. The common operational response value of the previous control cycle is referred to as clip, where clip is the limiting operator. This is the total net injection amount from the previous control cycle. This refers to the total net injection amount for the current control cycle. This is the lower limit value of the response slope signal. This is the upper limit of the response slope signal. This is the initial response slope signal for the current control cycle. For the response slope filter coefficients, The response slope signal for the current control cycle. This refers to the response slope signal from the previous control cycle. It is a small positive quantity used to make the denominator non-zero.

5. The multi-type power resource output control method according to claim 1, characterized in that, Multiple output prediction scenarios are established, and the reconstructed marginal adjustment signal under each output prediction scenario is determined based on the single resource response estimate, including: Multiple output prediction scenarios are constructed based on the predicted distribution of available power output from new energy sources and / or prediction residual resampling. The reconstructed marginal adjustment signal for each of the output prediction scenarios is determined based on the single resource response estimate. The expression for the reconstructed marginal adjustment signal is: ； In the formula, This refers to the reconfigured marginal adjustment signal of the nth power resource in the current control period t under the s-th output prediction scenario. Set the output amount for the nth candidate power resource. This is the response slope signal for the current control cycle. For the s-th output prediction scenario in the current control cycle t, for gradient value, Let n be the local regulation function of the power resource in the current control period t under the s-th output prediction scenario. This is the single-resource response estimate of the nth power resource in the current control period t under the output setting of the candidate resource.

6. The method for controlling the output of multiple types of power resources according to claim 1, characterized in that, The joint control state and the predicted marginal signal of the current control cycle are updated based on the average reconstructed marginal adjustment signal, and used as the joint control state and the predicted marginal signal of the next control cycle, including: A joint correction signal is obtained based on the combination of the average reconstructed marginal adjustment signal and the network margin signal; The joint control state of the current control cycle is updated based on the joint correction signal as the joint control state of the next control cycle; The predicted marginal signal of the current control cycle is updated based on the joint correction signal as the predicted marginal signal of the next control cycle.

7. The multi-type power resource output control method according to claim 6, characterized in that, The expression for the joint control state in the next control cycle is: ； In the formula, The joint control state is for the next control cycle t+1. For the first projection operator, The joint control state in the current control cycle. This refers to the auxiliary control state. The joint correction signal, Update the step size when the state is positive; The expression for the predicted marginal signal in the next control cycle is: ； In the formula, The predicted marginal signal for the next control period t+1, For the second projection operator, The updated weights for the predicted marginal signal, The predicted marginal signal is for the current control period t. This refers to the joint correction signal.

8. A multi-type power resource output control device, characterized in that, include: The first module is used to construct a joint control state of multiple types of power resources consisting of resource output setting and network margin adjustment. Based on the predicted marginal signal of the current control cycle, the joint control state of the current control cycle is projected and predicted to obtain the auxiliary control state. The second module is used to obtain the common operation response quantity under the joint control state in the current control cycle, and to determine the single resource response estimate of each power resource under the auxiliary control state based on the common operation response quantity. The third module is used to establish multiple output prediction scenarios, determine the reconstruction marginal adjustment signal under each output prediction scenario based on the single resource response estimate, and average the reconstruction marginal adjustment signals of all output prediction scenarios to obtain the average reconstruction marginal adjustment signal. The fourth module is used to update the joint control state and the predicted marginal signal of the current control cycle based on the average reconstructed marginal adjustment signal, as the joint control state and the predicted marginal signal of the next control cycle. The fifth module is used to control the output of multiple types of power resources based on the joint control state in each control cycle.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor is configured to implement the multi-type power resource output control method as described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, implement the multi-type power resource output control method as described in any one of claims 1 to 7.

Images