Multi-power-plant power regulation and control system and method, electronic equipment and medium

Abstract

The embodiment of the invention provides a multi-power-plant power regulation and control system and method, electronic equipment and a medium, and the system predicts the predicted output power of a target power plant in a future time period through a preset variable component sub-circuit and a space-time convolution network, and can effectively improve the power prediction accuracy of a new energy power plant. Furthermore, according to the operation characteristic parameters of the target power plant, the external environment data and a preset twin optimization engine, a power generation unit characteristic model and a resource control model of the target power plant are constructed, so that the equipment constraint of the target power plant under the specific equipment configuration is represented, and a safety boundary is established for subsequent power regulation and control of the target power plant. And finally, through the calculated predicted output power, the power generation unit characteristic model and the resource control model, determining a power regulation and control strategy for the target power plant, and sending a power regulation and control instruction to the target power plant, thereby realizing accurate power regulation and control for the target power plant.

Description

Technical Field

[0001] This application relates to the field of power plant power regulation technology, and in particular to a multi-power plant power regulation system, method, electronic device and medium. Background Technology

[0002] As the global energy structure gradually shifts towards low-carbon and intelligent transformation, the proportion of renewable energy power plants in the current power grid architecture is also increasing. Although renewable energy power plants can better mitigate excessive carbon emissions, they suffer from unstable power generation, such as wind and solar power plants. The output of these types of renewable energy power plants is often easily affected by external environmental factors and weather. Therefore, when the power grid predicts the output power of multiple power plants and allocates target output to each plant for future periods, the error in power allocation for renewable energy power plants is significant, and the grid cannot accurately allocate power to numerous different types of renewable energy power plants. Summary of the Invention

[0003] In view of the above problems, in order to improve the accuracy of power grid allocation to new energy power plants, this application provides a multi-power plant power regulation system, method, electronic device and medium.

[0004] The embodiments of this application disclose the following technical solutions:

[0005] In a first aspect, embodiments of this application provide a multi-power plant power regulation system applied to the grid side, wherein the grid is connected to multiple power plants, and the system includes: a parameter acquisition module, a power prediction module, a model building module, a strategy determination module, and a regulation module;

[0006] The parameter acquisition module is used to acquire at least one target power plant's operating characteristic parameters and external environmental data;

[0007] The power prediction module is used to predict the output power of the target power plant in a preset future time period based on operating characteristic parameters, external environmental data, preset variable quantum circuits and spatiotemporal convolutional networks.

[0008] The model building module is used to construct the power generation unit characteristic model and resource control model of the target power plant based on the operating characteristic parameters and the preset twin optimization engine.

[0009] The strategy determination module is used to determine the power regulation strategy for the target power plant based on the power generation unit characteristic model, resource control model and predicted output power of the target power plant.

[0010] The control module is used to send power control commands to the target power plant through power control strategies in order to control the power of the target power plant.

[0011] In one possible implementation, the power prediction module is specifically used for:

[0012] The operating characteristic parameters and external environment data are encoded to obtain encoded characteristic data, and the data chaos index of the encoded characteristic data is determined.

[0013] Based on the data chaos index, the quantum circuit depth of the preset variable quantum circuit is determined in order to construct the variable quantum circuit after depth adjustment.

[0014] The unitary transform is performed based on the operating characteristic parameters, external environment data, and a deeply adjusted variable quantum circuit to determine the unitary transform output state.

[0015] The quantum eigenvectors are obtained by performing positive operator measurement on the output state of the unitary transform;

[0016] Power prediction is performed using quantum eigenvectors and a spatiotemporal convolutional neural network to determine the predicted output power.

[0017] In one possible implementation, the power generation unit characteristic model includes: a wind turbine model and a photovoltaic module model; the resource control model includes: an active power control model and a reactive power control model; the wind turbine model includes: a power curve model and a time constant model; and the photovoltaic module model includes: a PV curve model and an inverter characteristic model.

[0018] In one possible implementation, the strategy determination module is specifically used for:

[0019] Cross-scale coupled calculations are performed based on wind turbine models, photovoltaic module models, active power models, and reactive power control models. The output value of the Ising model is solved by quantum annealing machine and predicted output power to determine the power dispatch strategy for the target power plant.

[0020] In one possible implementation, each power plant includes a separately associated FPGA; the FPGA is used to respond to power regulation commands from the grid side according to a preset task dynamic offloading strategy and a preset PBFT algorithm; the preset task dynamic offloading strategy is determined based on the energy consumption of the computation task, the energy consumption of the communication link, the task execution delay, and the maximum delay limit corresponding to the task execution.

[0021] Secondly, embodiments of this application provide a multi-power plant power regulation method, applied to the grid side, wherein the grid is connected to multiple power plants, and the method includes:

[0022] Acquire operational characteristic parameters and external environmental data of at least one target power plant;

[0023] Output power is predicted based on operating characteristic parameters, external environmental data, preset variable quantum circuits, and spatiotemporal convolutional networks to determine the predicted output power of the target power plant in a preset future time period.

[0024] Based on the operating characteristic parameters and the preset twin optimization engine, construct the power generation unit characteristic model and resource control model of the target power plant;

[0025] Based on the power generation unit characteristic model, resource control model and predicted output power of the target power plant, a power regulation strategy for the target power plant is determined.

[0026] The power regulation strategy sends power regulation commands to the target power plant to regulate the power of the target power plant.

[0027] In one possible implementation, output power prediction is performed based on operating characteristic parameters, external environmental data, a preset variable quantum circuit, and a spatiotemporal convolutional network to determine the predicted output power of the target power plant in a preset future time period, including:

[0028] The operating characteristic parameters and external environment data are encoded to obtain encoded characteristic data, and the data chaos index of the encoded characteristic data is determined.

[0029] Based on the data chaos index, the quantum circuit depth of the preset variable quantum circuit is determined in order to construct the variable quantum circuit after depth adjustment.

[0030] The unitary transform is performed based on the operating characteristic parameters, external environment data, and a deeply adjusted variable quantum circuit to determine the unitary transform output state.

[0031] The quantum eigenvectors are obtained by performing positive operator measurement on the output state of the unitary transform;

[0032] Power prediction is performed using quantum eigenvectors and a spatiotemporal convolutional neural network to determine the predicted output power.

[0033] In one possible implementation, the power generation unit characteristic model includes: a wind turbine model and a photovoltaic module model; the resource control model includes: an active power control model and a reactive power control model; the wind turbine model includes: a power curve model and a time constant model; and the photovoltaic module model includes: a PV curve model and an inverter characteristic model.

[0034] Thirdly, embodiments of this application provide an electronic device, which includes: a processor, a memory, and a system bus;

[0035] The processor and memory are connected via the system bus;

[0036] The memory is used to store one or more programs, which include instructions that, when executed by the processor, cause the processor to perform any of the possible multi-power plant power regulation methods in the first aspect.

[0037] Fourthly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the possible multi-power plant power regulation methods in the first aspect.

[0038] Compared with the prior art, this application has the following beneficial effects: This application provides a multi-power plant power regulation system, method, electronic device, and medium. The system is applied to the power grid side, with the power grid connected to multiple power plants. The system includes: a parameter acquisition module, a power prediction module, a model building module, a strategy determination module, and a regulation module. The parameter acquisition module is used to acquire the operating characteristic parameters and external environment data of at least one target power plant. The power prediction module is used to predict the output power of the target power plant in a preset future time period based on the operating characteristic parameters, external environment data, a preset variable quantum circuit, and a spatiotemporal convolutional network. The model building module is used to construct a power generation unit characteristic model and a resource control model of the target power plant based on the operating characteristic parameters and a preset twin optimization engine. The strategy determination module is used to determine a power regulation strategy for the target power plant based on the power generation unit characteristic model, resource control model, and predicted output power. The regulation module is used to send power regulation commands to the target power plant through the power regulation strategy to regulate the power of the target power plant. This system predicts the output power of a target power plant in future time periods using a pre-set variable quantum circuit and a spatiotemporal convolutional network, effectively improving the accuracy of power prediction for new energy power plants. Furthermore, based on the target power plant's operating characteristic parameters and a pre-set twin optimization engine, it constructs a generator unit characteristic model and a resource control model to characterize the equipment constraints of the target power plant under its specific equipment configuration, establishing a safety boundary for subsequent power regulation. Finally, by combining the calculated predicted output power, generator unit characteristic model, and resource control model, a power regulation strategy for the target power plant is determined, and power regulation commands are sent to the target power plant accordingly, thereby achieving precise power regulation. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of this application 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 some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 A schematic diagram of a multi-power plant power control system provided in this application embodiment;

[0041] Figure 2 A flowchart illustrating an output power prediction method provided in an embodiment of this application;

[0042] Figure 3 A schematic flowchart of a multi-power plant power regulation method provided in an embodiment of this application;

[0043] Figure 4 This is a schematic diagram of the structure of a multi-power plant power regulation electronic device provided in an embodiment of this application. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with specific embodiments and accompanying drawings. It should be particularly noted that the embodiments described in this application are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0045] It should be noted that, unless otherwise defined, the technical or scientific terms used in the embodiments of this application should have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms "first," "second," and similar terms used in the embodiments of this application do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed after the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are only used to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0046] As described earlier, while renewable energy power plants can better mitigate excessive carbon emissions, they suffer from unstable power generation, similar to wind and solar power plants. The output of these plants is often significantly affected by external environmental factors and weather. Therefore, when the power grid predicts the output power of multiple plants and allocates target output for future periods, the allocation error is substantial, making it difficult for the grid to accurately allocate output to numerous different types of renewable energy power plants.

[0047] To address the aforementioned issues, this application provides a multi-power plant power regulation system, method, electronic device, and medium. The system is applied to the power grid side, connecting multiple power plants. The system includes: a parameter acquisition module, a power prediction module, a model building module, a strategy determination module, and a regulation module. The parameter acquisition module acquires operating characteristic parameters and external environmental data of at least one target power plant. The power prediction module predicts output power based on the operating characteristic parameters, external environmental data, a preset variable quantum circuit, and a spatiotemporal convolutional network to determine the predicted output power of the target power plant in a preset future time period. The model building module constructs a generation unit characteristic model and a resource control model for the target power plant based on the operating characteristic parameters and a preset twin optimization engine. The strategy determination module determines a power regulation strategy for the target power plant based on the generation unit characteristic model, resource control model, and predicted output power. The regulation module sends power regulation commands to the target power plant according to the power regulation strategy to regulate its power output. This system predicts the predicted output power of the target power plant in the future time period using a preset variable quantum circuit and a spatiotemporal convolutional network, effectively improving the accuracy of power prediction for new energy power plants. Furthermore, based on the target power plant's operating characteristic parameters and a pre-set twin optimization engine, a generation unit characteristic model and a resource control model are constructed to characterize the equipment constraints of the target power plant under its specific equipment configuration, establishing a safety boundary for subsequent power regulation of the target power plant. Finally, by combining the calculated predicted output power, the generation unit characteristic model, and the resource control model, a power regulation strategy for the target power plant is determined, and power regulation commands are sent to the target power plant accordingly, thereby achieving precise power regulation of the target power plant.

[0048] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0049] See Figure 1 The figure is a schematic diagram of the structure of a multi-power plant power control system provided in an embodiment of this application, which specifically includes the following modules:

[0050] The parameter acquisition module 100 is used to acquire the operating characteristic parameters and external environment data of at least one target power plant.

[0051] The multi-power plant power regulation system provided in this application embodiment is deployed on the control terminal on the power grid side. The control terminal establishes communication connections with multiple power plants to regulate the power of various types of power plants.

[0052] First, it is necessary to acquire the operational characteristic parameters of at least one target power plant and its corresponding external environmental data. The target power plant is equipped with an IoT sensor array, with different types of sensing modules configured for power generation equipment, energy storage devices, and grid nodes. In this implementation, the target power plant can be various types of new energy power plants, such as wind power plants and photovoltaic power plants. Specifically, the power plant's operational characteristic parameters can be obtained by installing vibration sensors, temperature sensors, and power transmitters at the wind turbines and photovoltaic modules. Parameters collected in real time by the sensors, such as wind turbine blade stress, converter temperature, battery state of charge, and energy storage charging and discharging power, can all serve as the power plant's operational characteristic parameters. Additionally, frequency sensors and harmonic monitors can be deployed at key grid nodes to acquire frequency deviation and harmonic distortion rates in real time as representations of operational characteristic parameters.

[0053] For external environmental data acquisition, it can rely on detection equipment within the target power plant park, such as ultrasonic anemometers, high-precision irradiance sensors, and cloud optical thickness detectors, to collect wind speed in real time at a sampling rate of 1Hz, obtain irradiance at 0.5-second intervals, and monitor cloud movement vectors through lidar; at the same time, it can connect to satellite remote sensing data and regional meteorological forecasting systems to supplement long-term meteorological trends and extreme weather warning information.

[0054] The power prediction module 200 is used to predict the output power of the target power plant in a preset future period based on operating characteristic parameters, external environmental data, preset variable quantum circuits and spatiotemporal convolutional networks.

[0055] When predicting output power based on operating characteristic parameters and external environmental data, the aim is to extract features and perform nonlinear mapping on multi-source data through an architecture constructed using a pre-defined variable quantum circuit and a spatiotemporal convolutional network. This combines quantum computing with classical neural networks to build a prediction architecture that possesses both nonlinear feature extraction and temporal dynamic modeling capabilities, thereby improving the accuracy of power prediction. The following section will describe the power prediction process performed by the power prediction module, with reference to specific implementation diagrams.

[0056] See Figure 2 The figure is a flowchart illustrating an output power prediction method provided in an embodiment of this application, specifically including the following steps:

[0057] S101: Perform data encoding processing on the operating characteristic parameters and external environment data to obtain encoded characteristic data, and determine the data chaos index of the encoded characteristic data.

[0058] First, data encoding is implemented, using amplitude encoding technology to map continuous real-valued operational characteristic parameters to the probability amplitude space of quantum states, thereby constructing an initial quantum state that can be processed by quantum circuits. Specifically, the input vector containing parameters such as wind speed, irradiance, and battery state of charge is preprocessed using normalization to generate encoded characteristic data conforming to quantum state specifications, achieving efficient embedding of classical information into the quantum computation space.

[0059] The calculation formula for data encoding processing is as follows:

[0060] ;

[0061] In the formula, x represents operational characteristic parameters, such as wind speed and irradiance; x i Let i be the i-th component of the data.

[0062] After encoding is completed, the chaotic characteristics of the encoded feature data need to be quantitatively evaluated, i.e., the data chaos index needs to be calculated. This index characterizes the nonlinear complexity of the data by analyzing the dynamic behavior of the time series, specifically using the Maximum Lyapunov Exponent (MLE) estimation method:

[0063] In calculating the chaos index of data, the time series needs to be reconstructed in phase space first. The embedding dimension and time delay are determined using the delayed coordinate method to construct trajectory points in a high-dimensional phase space. Then, neighboring points with minimal initial distance are selected near each trajectory point, and the exponential growth rate of trajectory separation during evolution is tracked. Finally, the MLE value is obtained through time averaging. When MLE is greater than 0, it indicates that the data sequence has chaotic characteristics; the larger the value, the higher the degree of chaos (e.g., sudden changes in irradiance under extreme weather conditions, violent fluctuations in wind turbine output, etc.). This index not only provides a basis for the subsequent deep adaptive adjustment of variable quantum circuits—that is, high-chaotic data requires increasing the number of circuit layers to capture complex nonlinear correlations—but also ensures that quantum computing resources match the complexity of data features, avoiding efficiency imbalances or insufficient feature extraction when a fixed architecture processes data from different scenarios. This lays the foundation for data feature analysis for the efficient operation of the entire quantum-classical hybrid model.

[0064] S102: Determine the quantum circuit depth of the preset variable quantum circuit based on the data chaos index, so as to construct the variable quantum circuit after depth adjustment.

[0065] After determining the data chaos index, the depth of the preset variable quantum circuit needs to be dynamically adjusted based on this to construct a quantum computing architecture that adapts to the complexity of the data characteristics. As the core component connecting classical input and quantum computing, the depth of the variable quantum circuit (i.e., the number of quantum gate operation layers) directly determines its ability to extract high-dimensional nonlinear features. Among them, shallow circuits are suitable for linear or weakly nonlinear mappings of low-chaos data, while deep circuits capture higher-order correlations and nonlocal effects in strongly chaotic data through layer-by-layer superposition of unitary transformations.

[0066] The specific adjustment mechanism uses the data chaos index as the criterion: when the data chaos index is greater than a preset threshold (e.g., 0.3, indicating that the data exhibits significant chaotic characteristics, such as drastic fluctuations in wind and solar power output under sudden weather disturbances), a strategy to increase circuit depth is triggered. By superimposing a unitary transform layer containing a custom Hamiltonian on the original circuit, the circuit depth is gradually expanded from the initial layer 4 (assumed) to 8 layers. The parameters of each layer are dynamically optimized according to the data characteristics, enhancing the ability to modulate complex quantum states. If the data chaos index is lower than the threshold, it indicates that the data temporal correlation is relatively regular (e.g., a smooth power output curve under stable weather conditions). In this case, the circuit depth is maintained or appropriately reduced (e.g., maintaining 3-4 layers) to avoid resource waste caused by redundant calculations.

[0067] This adaptive adjustment strategy enables real-time matching of quantum circuit depth with the degree of data chaos: deep circuits break the linear superposition limitation of classical data through multi-layer nonlinear transformations, utilizing quantum superposition and entanglement effects to uncover hidden high-dimensional features (such as the high-order coupling between different wind turbine vibration frequencies and wind speeds, and the nonlocal correlation between photovoltaic module temperature and irradiance); shallow circuits complete basic feature mapping with efficient parameterized operations, balancing computational efficiency and feature extraction accuracy. In actual operation, this mechanism dynamically configures quantum computing resources to ensure that the variable quantum circuit has sufficient feature extraction capacity when processing highly chaotic data, while maintaining a lightweight architecture in simple scenarios, providing an adapted data processing channel for subsequent unitary transformation operations and quantum measurement stages, thereby improving the adaptability of the quantum-classical hybrid model to complex power generation scenarios from the underlying architecture level.

[0068] Specifically, the formula for adjusting the circuit depth of a variable quantum circuit is as follows:

[0069] ;

[0070] In the formula, L represents the quantum circuit depth, LE k Indicates the data chaos index;

[0071] S103: Perform unitary transformation based on operating characteristic parameters, external environment data, and deeply adjusted variable quantum circuit to determine the unitary transformation output state.

[0072] After completing the data encoding of the operational feature parameters and the deep adaptive adjustment of the variable quantum circuit, the unitary transform stage, as the core operation for quantum feature extraction, modulates the encoded quantum state layer by layer through parameterized quantum gate operations, ultimately generating an output state carrying deep data correlations. First, the encoded initial quantum state is injected into the deeply adjusted variable quantum circuit, which consists of multiple cascaded unitary transform modules, with each transform defined by a custom Hamiltonian and adjustable parameters.

[0073] Specifically, the formula for calculating the output state of the unitary transform is as follows:

[0074] ;

[0075] In the formula, L is the quantum circuit depth, k is the sequence index variable; θ k H is a trainable parameter vector; k For custom Hamiltonian.

[0076] As the quantum state evolves layer by layer in the circuit, each unitary transformation essentially performs nonlinear modulation on the probability amplitude distribution of the current quantum state. For low-chaos data, the unitary transformation of shallow circuits can map linear to weakly nonlinear features through a small number of rotations and entanglement operations. However, for strongly chaotic data (corresponding to deep circuits), the cascading effect of multiple unitary transformations gradually amplifies the higher-order correlations in the initial state. For example, when processing motion features involving sudden changes in wind speed and abrupt changes in wind turbine blade stress, deep circuits, through multiple U(θ) operations, can encode the wind speed-stress nonlinear coupling relationship implicit in the original data into the phase and entanglement degree of the quantum state, forming nonlocal features that are difficult to capture directly by classical computation. Finally, the output state after L layers of unitary transformations not only retains the original information of the operating characteristic parameters but also, through quantum superposition and entanglement effects, encodes complex features such as nonlinear correlations and multivariate couplings in the data into the high-dimensional Hilbert space of the quantum state. The unique aspect of this output state lies in the fact that its information is no longer limited to point-by-point mappings of classical data. Instead, through the unique parallel processing capabilities of quantum computing, it achieves a deep integration of global correlations and higher-order dependencies in the operational characteristic parameters, laying the foundation for the extraction of quantum feature vectors in the subsequent positive operator measurement stage. The entire unitary transformation process essentially leverages the nonlinear transformation advantages of quantum computing to "translate" classical operational characteristic parameters into quantum state representations containing deep physical correlations. This breaks through the bottleneck of traditional machine learning in extracting features from high-dimensional complex data, providing core technical support for accurately capturing the implicit coupling between equipment states and environmental variables in power generation systems.

[0077] S104: Perform positive operator measurement on the unitary transform output state to obtain the quantum eigenvector.

[0078] Positive operator measurement (POVM) of the unitary transform output state is a crucial step in converting quantum state information into classical eigenvectors. Its core lies in extracting the high-dimensional, complex features encoded in the quantum state through carefully designed measurement operators. After the deeply tuned variable quantum circuit performs a unitary transform on the operating characteristic parameters, the output state exists in quantum Hilbert space as a density matrix. This matrix contains nonlinear correlation information modulated by quantum superposition, entanglement, and other operations on the original data—information that cannot be directly processed by classical models and must be "decoded" into real-valued eigenvectors usable for subsequent predictions through the measurement process.

[0079] Specifically, the formula for calculating the quantum eigenvector is as follows:

[0080] ;

[0081] Among them, f i The measurement result or expected value; Tr is the trace operation, which operates on a matrix; ρ out It is the density matrix of the output state; E i Let i be the i-th positive operator value measure (POVM) element.

[0082] S105: Power prediction is performed based on quantum eigenvectors and a spatiotemporal convolutional neural network to determine the predicted output power.

[0083] After extracting quantum feature vectors, the process of combining them with a spatiotemporal convolutional neural network (STCNN) for power prediction is essentially a collaborative solution process that integrates high-order nonlinear features extracted by quantum computing with the spatiotemporal modeling capabilities of classical deep learning. First, the quantum feature vectors, as one of the core inputs, are fed into the input layer of the STCNN. This network architecture is specifically designed to handle power generation system data that exhibits both temporal dynamics and spatial correlations. In the temporal dimension, one-dimensional convolutional layers capture the long-term dependencies of power sequences (such as intraday load fluctuations and periodic changes in wind and solar power output), with their receptive field expanding as the number of network layers increases, ensuring feature extraction across different time scales. In the spatial dimension, considering the geographical distribution and electrical connections of multiple power generation units, graph convolutional layers or two-dimensional convolutional layers are used to process the spatial adjacency matrix, capturing the power coupling effects between devices.

[0084] The unique value of quantum feature vectors becomes apparent at this stage: the nonlinear correlation information extracted through quantum unitary transformation and positive operator measurement (such as chaotic features of equipment vibration and nonlocal coupling of meteorological variables) contained within them is embedded as a high-discrimination feature component into the network input, effectively compensating for the shortcomings of traditional spatiotemporal models in handling strongly nonlinear and nonstationary data. For example, when the input includes irradiance change data under extreme weather conditions, the chaotic dynamics information encoded in the quantum features can significantly enhance the network's ability to identify abnormal fluctuation patterns, avoiding prediction lag caused by relying solely on historical time-series data. The spatiotemporal feature extraction module in the middle layer of the network dynamically weights the contributions of quantum features and classical spatiotemporal features through cross-dimensional interactive operations—during periods of high data chaos, the weight of the quantum feature dimension automatically increases, dominating feature matching of abnormal patterns; while during stable operation, classical spatiotemporal features undertake the main modeling task, balancing computational efficiency.

[0085] Specifically, the formula for calculating the objective function is as follows:

[0086] ;

[0087] Where θ represents the quantum circuit parameters; W represents the classical neural network weights; η represents the mixing coefficient of the weight allocation between the quantum prediction result and the classical prediction result; y q (θ) represents the output of the quantum model; y c (W) represents the output of the classic model; y ture λ represents the target value in supervised learning; λ represents the regularization strength; and R(θ,W) represents the joint regularization term.

[0088] The above describes the output power prediction process of the power prediction module in the embodiments of this application. The following will continue to combine... Figure 1 The multi-power plant power control system provided in this embodiment will be introduced.

[0089] The model building module 300 is used to build the power generation unit characteristic model and resource control model of the target power plant based on the operating characteristic parameters and the preset twin optimization engine.

[0090] When constructing the power generation unit characteristic model and resource control model of the target power plant, it is necessary to use multi-dimensional operational characteristic parameters collected in real time by sensor networks, combined with the modeling framework integrated by the preset digital twin optimization engine, to achieve accurate mapping of the power generation unit characteristics and system operating status within the power plant. The preset twin optimization engine has built-in standardized multi-model templates covering power generation unit characteristic models and resource control models, and can dynamically call the corresponding calculation models according to the equipment type (such as wind turbines, photovoltaic modules, energy storage batteries, converters, etc.) and operating scenarios of different power plants.

[0091] The power generation unit characteristic model includes a wind turbine model and a photovoltaic module model, and the resource control model includes an active power control model and a reactive power control model. The wind turbine model is further subdivided into a power curve model and a time constant model, and the photovoltaic module model is further subdivided into a PV curve model and an inverter characteristic model.

[0092] Specifically, the power curve model is based on the actual power curve fitted from SCADA data, reflecting the mapping relationship between wind speed and power, and includes "inflection point" and "hysteresis" characteristics (such as start-stop wind speed). The commonly used piecewise double Sigmoid model formula is as follows:

[0093] ;

[0094] Hysteresis is achieved by adding a "hysteresis band" to the cut-in / cut-out wind speed:

[0095] ;

[0096] In the formula, P w (v) represents the actual active power output of the wind turbine at the current wind speed v, where v represents the instantaneous wind speed at the hub height, P rated Used to characterize the rated active power of a wind turbine, v ci Used to characterize the cut-in wind speed, v co Used to characterize the cut-out wind speed, Δ vhyst The hysteresis band width is characterized by k1 and a1, which are used to characterize the Sigmoid fitting parameters and control the inflection point shape. rated Used to characterize rated wind speed.

[0097] Furthermore, the formula for the time constant model (first-order inertia) is as follows:

[0098] ;

[0099] Virtual inertial support capability

[0100] The fan can release rotor kinetic energy during a frequency drop:

[0101] ;

[0102] In the formula, T start T represents the startup time constant (s), which is the inertial time from standby to target power. stop P represents the shutdown time constant, i.e., the inertial time from operation to complete power failure. wref H represents the power setting value. wt The constant of inertia of the fan is represented by f0, which is the ratio of the rotor kinetic energy to the rated capacity. The rated frequency of the system is f0, and the rate of change of the system frequency is df / dt. ΔP represents the system frequency. inertiaThis represents the instantaneous virtual inertial power increment that the wind turbine can provide.

[0103] On the other hand, the PV curve model in the photovoltaic module model is used to characterize the mapping relationship between irradiance, temperature, and output power. The specific formula is as follows:

[0104] ;

[0105] In the formula, I represents the output current of the photovoltaic cell. Indicates photocurrent, V represents the reverse saturation current of the diode, and V represents the output voltage of the photovoltaic cell. This indicates the parallel resistance of the photovoltaic cell. This represents the series resistance of a photovoltaic cell. This represents thermal voltage.

[0106] Furthermore, the inverter characteristic model includes its maximum power point tracking (MPPT) operating range, reactive power regulation capability, and overload capability, as shown in the following formulas:

[0107] MPPT operating range:

[0108] ;

[0109] Reactive power regulation capability:

[0110] ;

[0111] Short-term overload capacity:

[0112] ;

[0113] In the formula, P inv min ,P inv max This indicates that the inverter operates at the current DC bus voltage V. dc The minimum and maximum active power that can be output, Q inv S represents the reactive power of the inverter. inv rated Indicates the inverter's rated apparent power, k ol τ represents the short-time overload factor. th This represents the inverter's thermal time constant.

[0114] The active power control model is divided into the active wind / solar curtailment capability model, which describes the ability, rate and cost (i.e. power generation loss) to adjust power downward by adjusting the pitch angle (wind turbine) and causing the MPPT to deviate from the optimal operating point (solar).

[0115] The control formula for the fan is as follows:

[0116] ;

[0117] Adjustment rate constraint:

[0118] ;

[0119] Cost of wind curtailment (power generation loss):

[0120] ;

[0121] Photovoltaics: achieved by deviating from the voltage at the maximum power point, also defined as α:

[0122] ;

[0123] In the formula, α represents the power derating factor (0–1), 1 indicates no wind / solar curtailment, and 0 indicates complete power cut-off. w avail P represents the theoretical power output of the wind turbine at the current wind speed. w curt C represents the actual output power after wind curtailment. p (β,λ) represents the wind energy utilization coefficient, which is related to the blade pitch angle β (°) and the tip speed ratio λ. p max R represents the maximum wind energy utilization coefficient. α max c represents the maximum rate of change of the load reduction factor. e V represents the feed-in tariff or marginal cost of electricity generation. mpp V oc V op These represent the maximum power point voltage, open circuit voltage, and actual operating voltage, respectively.

[0124] On the other hand, the specific formula for the reactive voltage control model is as follows:

[0125] The reactive power curve describes the range of reactive power that a wind turbine / inverter can provide under different active power outputs. It forms the basis for power plant participation in grid voltage regulation.

[0126] The reactive power range that the wind turbine / inverter can provide under different active power outputs P:

[0127] ;

[0128] Taking an inverter as an example, the approximate elliptic equation is:

[0129] ;

[0130] Considering power factor constraints:

[0131] ;

[0132] In the formula, Q min (P),Q max (P) represents the minimum and maximum reactive power that the inverter can absorb and generate under the current active power P, respectively, and cosφ represents the power factor cosφ. min This indicates the minimum allowable power factor.

[0133] The strategy determination module 400 is used to determine the power regulation strategy for the target power plant based on the power generation unit characteristic model, resource control model and predicted output power of the target power plant.

[0134] The control module 500 is used to send power control commands to the target power plant through a power control strategy in order to control the power of the target power plant.

[0135] The core of determining the power regulation strategy based on the power generation unit characteristic model and resource control model of the target power plant, along with the predicted output power, lies in the deep integration of the cross-scale coupling constraints of various models with the wind and solar power prediction results. This constructs an optimization problem with multi-dimensional constraints and solves it efficiently using quantum computing technology. In determining the power regulation strategy for the target power plant, cross-scale coupling calculations are performed based on the power generation unit characteristic model and resource control model. The Ising model output value is then solved using a quantum annealing machine and the predicted output power to determine the power regulation strategy for the target power plant.

[0136] First, cross-scale coupled computation needs to break through the limitations of independent modeling and embed the integrated constraints and predicted output power into the quantum annealing solution framework. The specific execution process is to transform the power regulation problem into an energy minimization problem of the Ising model, where decision variables (such as energy storage charging and discharging power, wind curtailment rate, and solar curtailment rate) correspond to the qubit state, the physical limits of the device are transformed into the coupling strength Jij between qubits, and the economic objectives and reliability indicators constitute the external field strength.

[0137] Predicting output power not only provides time-series curves for wind and solar power generation, but also dynamically adjusts the Hamiltonian H of the Ising model through probability distribution characteristics (such as the power output fluctuation range within the confidence interval), incorporating the influence of prediction uncertainty into the solution process. The quantum annealing machine explores a massive solution space in parallel through the quantum tunneling effect, effectively improving the efficiency of determining the global optimal solution compared to traditional solvers.

[0138] Taking a typical scheduling scenario as an example: when it is predicted that the photovoltaic output will fluctuate by ±20% due to cloud changes at noon the next day, the quantum annealing machine will simultaneously verify the current carrying capacity of the converter (constrained by the electromagnetic field model), the temperature rise rate of the battery pack during high-power charging and discharging (constrained by the thermodynamic field model), and the upper limit of the adjustment frequency of the wind turbine pitch system (constrained by the mechanical stress model). Under the premise of satisfying all constraints, it will generate the charging and discharging timing strategy of the energy storage system---such as releasing part of the energy storage in advance before the peak output to avoid converter overload, supplementing charging during the low-frequency period of fluctuation using the off-peak electricity price, while controlling the battery temperature increase to not exceed 5°C and the number of stress cycles of the wind turbine blades to be within the fatigue limit.

[0139] The final output power scheduling strategy, after being verified by edge computing nodes, is distributed to each device via a 5G URLLC low-latency channel, achieving millisecond-level response and microsecond-level precision in control execution. This avoids the device damage risk caused by simplified constraints in traditional methods, and significantly improves the system's operational economy through the global optimization capabilities of quantum computing.

[0140] Specifically, the formula for this process is as follows:

[0141] ;

[0142] In the formula, H is the Hermitian operator, corresponding to the possible energy of the system; J ij h represents the coupling strength. i For external field strength; σ zi For qubit operators.

[0143] In one possible implementation, each power plant connected to the grid is equipped with a separately associated FPGA (Field Programmable Gate Array). The FPGA configured in each power plant is used to respond to power regulation commands from the grid side according to a preset task dynamic offloading strategy and a preset PBFT algorithm. The preset task dynamic offloading strategy is determined based on the energy consumption of the computing task, the energy consumption of the communication link, the task execution delay, and the maximum delay limit corresponding to the task execution.

[0144] Each power plant is equipped with a dedicated FPGA that undertakes the core task of localized intelligent response. Its operating mechanism deeply integrates dynamic task offloading strategy and preset PBFT consensus algorithm to ensure low latency and high reliability execution of power regulation commands on the grid side.

[0145] First, the reconfigurable computing architecture of FPGAs enables them to flexibly adapt to different computing tasks. A pre-defined dynamic task offloading strategy aims to minimize total system energy consumption while meeting task latency constraints, constructing a multi-objective optimization model that includes computing task energy consumption, communication link energy consumption, total task latency, and maximum allowable latency. When an edge node receives a power regulation command, it first evaluates the task attributes: for low-data-volume, high-real-time tasks such as real-time power adjustment of energy storage converters, execution is performed directly on the local FPGA to avoid network transmission latency. For complex optimization tasks involving coupled simulations (such as power allocation in wind turbine clusters), the dynamic offloading strategy determines whether to forward the command to adjacent edge nodes or the cloud.

[0146] If local computing resources are saturated and task latency margin is sufficient, nodes with idle computing power are selected for collaborative computing through the preset PBFT consensus mechanism. At the same time, intelligent algorithms are used to balance communication energy consumption and computing energy consumption during the offloading process. For example, the node with the closest distance and the best channel quality is selected first to reduce the overall energy consumption of a single task offloading.

[0147] Specifically, the objective function calculation formula for the preset task dynamic unloading strategy is as follows:

[0148] ;

[0149] ;

[0150] In the formula, E i comp Let E be the energy consumption of the i-th computational task, in kWh; i comm T represents the energy consumption of the i-th communication link, in kWh; Ti represents the total latency of the i-th task, in milliseconds; T i max The maximum allowed delay for the i-th task;

[0151] The default PBFT algorithm is as follows:

[0152] when ;

[0153] In the formula, v is the current view number, λ is the time decay coefficient, which controls the decay rate of the impact of historical faults; t is the runtime after the last view switch; δ i This is the identifier for abnormal behavior of node i (1 = abnormal, 0 = normal); threshold is the threshold value that triggers view switching.

[0154] In particular, in one possible implementation, the control system provided in this application embodiment also includes an adaptive communication middleware, which serves as the core link connecting the grid side, power plant edge nodes, and underlying equipment. This middleware supports multimodal communication protocol adaptation and can dynamically parse more than 12 communication protocols, including 5G URLLC, Industrial Ethernet, and LoRa (wide-area sensor coverage). Through a semantic conversion engine, it seamlessly maps Modbus and OPC UA data formats at the device layer with REST APIs and message queues in the cloud, solving the problem of interoperability between devices from multiple vendors.

[0155] In terms of security mechanisms, the middleware integrates quantum-secure key distribution and lightweight homomorphic encryption modules to implement end-to-end encryption of power control commands, ensuring that the commands cannot be tampered with even if intercepted during transmission. At the same time, it uses device digital twin fingerprint authentication technology to verify the identity and legitimacy of access devices in real time, preventing malicious node intrusion.

[0156] This application provides a multi-power plant power regulation system, method, electronic device, and medium. The system is applied to the power grid side, connecting multiple power plants. The system includes: a parameter acquisition module, a power prediction module, a model building module, a strategy determination module, and a regulation module. The parameter acquisition module acquires operating characteristic parameters and external environmental data of at least one target power plant. The power prediction module predicts the output power of the target power plant in a preset future time period based on the operating characteristic parameters, external environmental data, a preset variable quantum circuit, and a spatiotemporal convolutional network. The model building module constructs a generation unit characteristic model and a resource control model for the target power plant based on the operating characteristic parameters and a preset twin optimization engine. The strategy determination module determines a power regulation strategy for the target power plant based on the generation unit characteristic model, resource control model, and predicted output power. The regulation module sends power regulation commands to the target power plant according to the power regulation strategy to regulate its power output. This system predicts the output power of the target power plant in the future time period using a preset variable quantum circuit and a spatiotemporal convolutional network, effectively improving the accuracy of power prediction for new energy power plants. Furthermore, based on the target power plant's operating characteristic parameters and a pre-set twin optimization engine, a generation unit characteristic model and a resource control model are constructed to characterize the equipment constraints of the target power plant under its specific equipment configuration, establishing a safety boundary for subsequent power regulation of the target power plant. Finally, by combining the calculated predicted output power, the generation unit characteristic model, and the resource control model, a power regulation strategy for the target power plant is determined, and power regulation commands are sent to the target power plant accordingly, thereby achieving precise power regulation of the target power plant.

[0157] The following describes a multi-power plant power regulation method provided by an embodiment of this application. The multi-power plant power regulation method described below can be referred to in correspondence with the multi-power plant power regulation system described above.

[0158] See Figure 3 The figure is a flowchart illustrating a multi-power plant power regulation method provided in an embodiment of this application, specifically including the following steps:

[0159] S201: Obtain at least one target power plant's operating characteristic parameters and external environment data;

[0160] S202: Based on operating characteristic parameters, external environmental data, preset variable quantum circuits, and spatiotemporal convolutional networks, output power is predicted to determine the predicted output power of the target power plant in a preset future time period.

[0161] S203: Based on the operating characteristic parameters and the preset twin optimization engine, construct the power generation unit characteristic model and resource control model of the target power plant;

[0162] S204: Based on the power generation unit characteristic model, resource control model and predicted output power of the target power plant, determine the power regulation strategy for the target power plant;

[0163] S205: Send power control commands to the target power plant through a power control strategy to control the power of the target power plant.

[0164] In one possible implementation, output power prediction is performed based on operating characteristic parameters, external environmental data, a preset variable quantum circuit, and a spatiotemporal convolutional network to determine the predicted output power of the target power plant in a preset future time period, including:

[0165] The operating characteristic parameters and external environment data are encoded to obtain encoded characteristic data, and the data chaos index of the encoded characteristic data is determined.

[0166] Based on the data chaos index, the quantum circuit depth of the preset variable quantum circuit is determined in order to construct the variable quantum circuit after depth adjustment.

[0167] The unitary transform is performed based on the operating characteristic parameters, external environment data, and a deeply adjusted variable quantum circuit to determine the unitary transform output state.

[0168] The quantum eigenvectors are obtained by performing positive operator measurement on the output state of the unitary transform;

[0169] Power prediction is performed using quantum eigenvectors and a spatiotemporal convolutional neural network to determine the predicted output power.

[0170] In one possible implementation, the power generation unit characteristic model includes: a wind turbine model and a photovoltaic module model; the resource control model includes: an active power control model and a reactive power control model; the wind turbine model includes: a power curve model and a time constant model; and the photovoltaic module model includes: a PV curve model and an inverter characteristic model.

[0171] See Figure 4 The figure is a schematic diagram of the structure of a multi-power plant power regulation electronic device provided in an embodiment of this application, including:

[0172] Memory 11 is used to store computer programs;

[0173] The processor 12 is used to implement the steps of the multi-power plant power regulation method described in any of the above method embodiments when executing the computer program.

[0174] In this embodiment, the device can be an in-vehicle computer, a PC (Personal Computer), or a terminal device such as a smartphone, tablet computer, handheld computer, or portable computer.

[0175] The device may include a memory 11, a processor 12, and a bus 13.

[0176] The memory 11 includes at least one type of readable storage medium, such as flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 11 may be an internal storage unit of the device, such as the hard disk of the device. In other embodiments, the memory 11 may be an external storage device of the device, such as a plug-in hard disk, SmartMedia Card (SMC), Secure Digital (SD) card, Flash Card, etc. Furthermore, the memory 11 may include both internal and external storage units of the device. The memory 11 can be used not only to store application software and various types of data installed on the device, such as program code executing energy storage frequency modulation control methods, but also to temporarily store data that has been output or will be output. In some embodiments, the processor 12 may be a central processing unit (CPU).

[0177] In some embodiments, processor 12 may be a central processing unit (CPU), controller, microcontroller, microprocessor or other data processing chip, used to run program code stored in memory 11 or process data, such as program code for executing a fault prediction method.

[0178] This bus 13 can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 4 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0179] Furthermore, the device may also include a network interface 14, which may optionally include a wired interface and / or a wireless interface (such as a Wi-Fi interface, a Bluetooth interface, etc.), typically used to establish communication connections between the device and other electronic devices.

[0180] Optionally, the device may further include a user interface 15, which may include a display, an input unit such as a keyboard, and optionally, a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen, etc. The display may also be appropriately referred to as a screen or display unit, used to display information processed in the device and to display a visual user interface.

[0181] Figure 4 Only devices with components 11-15 are shown; those skilled in the art will understand that... Figure 4 The structure shown does not constitute a limitation on the device and may include fewer or more components than shown, or combine certain components, or have different component arrangements.

[0182] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, this application also provides a computer-readable storage medium storing computer instructions for causing the computer to execute the multi-power plant power regulation method as described in any of the above embodiments.

[0183] The computer-readable media in this application embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, program modules, 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-transfer medium that can be used to store information accessible by a computing device.

[0184] The computer instructions stored in the storage medium of the above embodiments are used to cause the computer to execute the multi-power plant power regulation method as described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0185] It should be noted that the various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for methods, systems, electronic devices, and media, since they are basically similar to the method embodiments, the descriptions are relatively simple, and relevant parts can be referred to the descriptions of the method embodiments. The methods, systems, electronic devices, and media described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components indicated as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of the solution in this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0186] The above description is merely one specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A multi-power plant power control system, characterized in that, Applied to the power grid side, where the power grid is connected to multiple power plants, the system includes: a parameter acquisition module, a power prediction module, a model building module, a strategy determination module, and a control module; The parameter acquisition module is used to acquire at least one target power plant's operating characteristic parameters and external environmental data; The power prediction module is used to predict the output power based on the operating characteristic parameters, the external environment data, the preset variable quantum circuit and the spatiotemporal convolutional network, so as to determine the predicted output power of the target power plant in a preset future time period. The model building module is used to build a power generation unit characteristic model and a resource control model of the target power plant based on the operating characteristic parameters and the preset twin optimization engine. The strategy determination module is used to determine a power regulation strategy for the target power plant based on the power generation unit characteristic model, the resource control model, and the predicted output power of the target power plant. The control module is used to send power control commands to the target power plant through the power control strategy in order to control the power of the target power plant.

2. The system according to claim 1, characterized in that, The power prediction module is specifically used for: The operational characteristic parameters and the external environment data are encoded to obtain encoded characteristic data, and the data chaos index of the encoded characteristic data is determined. Based on the data chaos index, the quantum circuit depth of the preset variable quantum circuit is determined in order to construct the variable quantum circuit after depth adjustment. Based on the operating characteristic parameters, the external environment data, and the deeply adjusted variable quantum circuit, a unitary transform is performed to determine the unitary transform output state. The unitary transform output state is subjected to positive operator measurement processing to obtain the quantum eigenvector; Power prediction is performed based on the quantum feature vector and the spatiotemporal convolutional neural network to determine the predicted output power.

3. The system according to claim 1, characterized in that, The power generation unit characteristic model includes: a wind turbine model and a photovoltaic module model; the resource control model includes: an active power control model and a reactive power control model; the wind turbine model includes: a power curve model and a time constant model; the photovoltaic module model includes: a PV curve model and an inverter characteristic model.

4. The system according to claim 3, characterized in that, The strategy determination module is specifically used for: Cross-scale coupled calculations are performed based on the wind turbine model, the photovoltaic module model, the active power model, and the reactive power control model. The output value of the Ising model is solved by the quantum annealing machine and the predicted output power to determine the power dispatch strategy for the target power plant.

5. The system according to claim 1, characterized in that, Each of the power plants includes a separately associated FPGA; the FPGA is used to respond to the power regulation command from the grid side according to a preset task dynamic offloading strategy and a preset PBFT algorithm; the preset task dynamic offloading strategy is determined based on the energy consumption of the computation task, the energy consumption of the communication link, the task execution delay, and the maximum delay limit corresponding to the task execution.

6. A method for power regulation of multiple power plants, characterized in that, Applied to the power grid side, where the power grid is connected to multiple power plants, the method includes: Acquire operational characteristic parameters and external environmental data of at least one target power plant; The output power is predicted based on the operating characteristic parameters, the external environment data, the preset variable quantum circuit, and the spatiotemporal convolutional network to determine the predicted output power of the target power plant in a preset future time period. Based on the aforementioned operational characteristic parameters and the preset twin optimization engine, a power generation unit characteristic model and a resource control model for the target power plant are constructed. Based on the power generation unit characteristic model, the resource control model, and the predicted output power of the target power plant, a power regulation strategy for the target power plant is determined. The power regulation strategy sends a power regulation command to the target power plant to regulate the power of the target power plant.

7. The method according to claim 6, characterized in that, The step of predicting output power based on the operating characteristic parameters, the external environment data, a preset variable quantum circuit, and a spatiotemporal convolutional network to determine the predicted output power of the target power plant in a preset future time period includes: The operational characteristic parameters and the external environment data are encoded to obtain encoded characteristic data, and the data chaos index of the encoded characteristic data is determined. Based on the data chaos index, the quantum circuit depth of the preset variable quantum circuit is determined in order to construct the variable quantum circuit after depth adjustment. Based on the operating characteristic parameters, the external environment data, and the deeply adjusted variable quantum circuit, a unitary transform is performed to determine the unitary transform output state. The unitary transform output state is subjected to positive operator measurement processing to obtain the quantum eigenvector; Power prediction is performed based on the quantum feature vector and the spatiotemporal convolutional neural network to determine the predicted output power.

8. The method according to claim 6, characterized in that, The power generation unit characteristic model includes: a wind turbine model and a photovoltaic module model; the resource control model includes: an active power control model and a reactive power control model; the wind turbine model includes: a power curve model and a time constant model; the photovoltaic module model includes: a PV curve model and an inverter characteristic model.

9. An electronic device, characterized in that, The device includes: a processor, a memory, and a system bus; The processor and the memory are connected via the system bus; The memory is used to store one or more programs, the one or more programs including instructions that, when executed by the processor, cause the processor to perform the multi-power plant power regulation method according to any one of claims 6-8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the multi-power plant power regulation method as described in any one of claims 6-8.

Images