A new energy consumption control method, device and storage medium

Abstract

This application discloses a method, apparatus, equipment, and storage medium for controlling the absorption of new energy sources. The method includes: determining the predicted output of new energy sources in a given region, the predicted state of charge (SOC) of energy storage in that region, and the predicted load in that region, thereby determining a global absorption rate target; for each region, determining the excess power that needs to be absorbed by that region based on the global absorption rate target, etc.; controlling the new energy output of the region based on the excess power that needs to be absorbed, the new energy regulation weight, and the predicted output of new energy sources in that region; controlling the charging and discharging power of the energy storage in that region based on the excess power that needs to be absorbed, the energy storage regulation weight, the predicted load in that region, and the predicted output of new energy sources in that region; and controlling the controllable load of the region based on the excess power that needs to be absorbed; wherein the charging and discharging power of the energy storage in that region is constrained based on the predicted SOC of the energy storage in that region. Therefore, it is possible to quickly adapt to fluctuations in new energy sources and improve the absorption rate of new energy sources.

Description

Technical Field

[0001] This application relates to the field of power system technology, and in particular to a new energy consumption control method, equipment and storage medium. Background Technology

[0002] Distribution transformer substations are key nodes for the integration of new energy sources (distributed photovoltaic and wind power) into the distribution network, and their capacity to absorb these new energy sources directly affects the safe and economical operation of the distribution network. Currently, with the vigorous promotion of the "dual carbon" target, the penetration rate of distributed new energy in distribution transformer substations continues to increase. However, existing methods for controlling the absorption of new energy still suffer from low absorption rates, failing to consider the differences between different zones within the distribution transformer substation, and needing to improve the speed of adapting to fluctuations in new energy demand. Summary of the Invention

[0003] This application provides a new energy consumption control method, device, and storage medium that takes into account the differences between various zones of the distribution transformer area, can quickly adapt to new energy fluctuations, and can improve the new energy consumption rate.

[0004] In a first aspect, embodiments of this application provide a method for controlling the consumption of new energy sources, including:

[0005] Determine the predicted values ​​of renewable energy output, state of charge (SOC) of energy storage, and load of each zone within the distribution transformer area, and determine the predicted values ​​of renewable energy output and load of the distribution transformer area.

[0006] The global absorption rate target is determined based on the predicted new energy output and load of the distribution transformer area.

[0007] For each zone, the excess power that the zone needs to absorb is determined based on the global absorption rate target, the zone weight coefficient, the predicted value of the renewable energy output of the distribution transformer area, and the renewable energy penetration rate of the zone.

[0008] The system controls the renewable energy output of a region based on the excess power to be absorbed by that region, the renewable energy regulation weight, and the predicted renewable energy output of that region. It also controls the energy storage charging and discharging power of that region based on the excess power to be absorbed by that region, the energy storage regulation weight, the predicted load of that region, and the predicted renewable energy output of that region. Furthermore, it controls the controllable load of that region based on the excess power to be absorbed by that region. The energy storage charging and discharging power of that region is constrained by the predicted energy storage SOC of that region.

[0009] Monitor the operating status data of the distribution transformer area. If the operating status data does not meet the preset safety constraints, adjust the distribution transformer area accordingly.

[0010] Secondly, embodiments of this application provide an electronic device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the method provided in embodiments of this application.

[0011] Thirdly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to perform the method of embodiments of this application.

[0012] The technical solution provided in this application determines the predicted values ​​of renewable energy output, energy storage SOC, and load within a distribution transformer area, and establishes a global absorption target for the distribution transformer area. For each area, the global absorption target is broken down to determine the excess power that the area needs to absorb. Based on the excess power to be consumed by the area, the renewable energy control weight, and the predicted renewable energy output, the renewable energy output of the area is controlled. Furthermore, based on the excess power to be consumed by the area, the energy storage control weight, the predicted load, and the predicted renewable energy output, the energy storage charging and discharging power of the area is controlled. By controlling the controllable load of a zone based on the excess power it needs to consume, the differences between zones are taken into account. The excess power that each zone needs to absorb is determined, and each zone is controlled based on the excess power it needs to absorb and other corresponding indicators. This allows for rapid adaptation to fluctuations in renewable energy and can significantly improve the renewable energy absorption rate. Furthermore, by constraining the energy storage charging and discharging power of a zone through the predicted value of its energy storage SOC, and by monitoring the operating status data of the distribution transformer area, the distribution transformer area can be regulated when preset safety constraints are not met, thereby improving the safety of the distribution transformer area. Attached Figure Description

[0013] Figure 1 A flowchart of a new energy consumption control method provided for the implementation of this application;

[0014] Figure 2 A flowchart of a new energy consumption control method provided for the implementation of this application;

[0015] Figure 3 A structural block diagram of a new energy consumption control device provided for the implementation of this application;

[0016] Figure 4 This is a schematic diagram of an electronic device structure provided in an embodiment of this application. Detailed Implementation

[0017] The present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0018] Figure 1This is a flowchart illustrating a new energy consumption control method provided in an embodiment of this application. The method can be executed by a new energy consumption control device, which can be implemented by software and / or hardware. The device can be configured in electronic devices such as computers. Figure 1 As shown, the method provided in this application embodiment includes the following steps:

[0019] S110: Determine the predicted values ​​of renewable energy output, state of charge (SOC) of energy storage, and load of each zone within the distribution transformer area, and determine the predicted values ​​of renewable energy output and load of the distribution transformer area.

[0020] In this embodiment, optionally, determining the predicted values ​​of renewable energy output, state of charge (SOC) of energy storage, and load of each zone within the distribution transformer area includes: collecting renewable energy data, load data, energy storage data, and meteorological data for each zone within the distribution network area; extracting features from the renewable energy data, load data, and energy storage data to obtain time-series feature data; inputting the time-series feature data into a dynamic prediction model to obtain the predicted values ​​of renewable energy output, load, and SOC of energy storage for each zone; wherein, the dynamic prediction model includes a multi-head attention layer and a bidirectional LSTM layer; during the training process, the dynamic prediction model uses the loss value obtained from the loss function to perform dynamic prediction... The model is adjusted; the loss function of the dynamic prediction model is determined by a weighted combination of new energy output loss, energy storage SOC loss, L2 regularization term, smoothing loss, and load loss; wherein, the new energy output loss is determined based on the actual value of the regional new energy output and the predicted value of the regional new energy output; the energy storage SOC loss is determined based on the actual value of the regional energy storage SOC and the predicted value of the regional energy storage SOC; the smoothing loss is determined based on the predicted value of the regional new energy output and the predicted value of the regional energy storage SOC; the L2 regularization term is determined based on the mapping weight matrix in the dynamic prediction model and the weight matrix of the bidirectional LSTM layer; and the load loss is determined based on the actual value of the regional load and the predicted value of the regional load.

[0021] In this embodiment, the predicted values ​​of new energy output, load, and energy storage SOC in the distribution transformer area can also be predicted based on the dynamic prediction model, and the processing method is the same as that for the predicted values ​​of each zone.

[0022] Specifically, various types of data can be collected, cleaned, and feature extracted through a multi-source sensing layer. Sensor terminals can be deployed to collect renewable energy output data, load data, energy storage data, and meteorological data from distribution transformer areas or different zones, simultaneously completing data cleaning and feature extraction. The data range and accuracy constraints can be found in Table 1.

[0023] Table 1

[0024]

[0025] Specifically, outlier removal (3σ criterion) can be performed on the collected data: the data sequence formed by the collected data... Calculate the mean with standard deviation ,like This value was identified as an outlier and replaced with the mean of the previous 5 samples.

[0026] ;

[0027] in, This refers to the replacement data for the outlier when the i-th data is an outlier.

[0028] Specifically, time-series feature data is extracted, which includes 3D meteorological features, 4D renewable energy output features, and 5D load and energy storage features; among them, the 3D meteorological features include irradiance. Wind speed ,temperature The 4-dimensional renewable energy output characteristics include the hourly average / peak value of photovoltaic (PV) output / wind turbine output / renewable energy output volatility; the 5-dimensional load and energy storage characteristics include real-time load value, load volatility, energy storage SOC, energy storage charging and discharging power, and load peak-valley difference. The 4-dimensional renewable energy output characteristics can be historical renewable energy output characteristics. The peak value of PV output is also included. Load peak-valley difference ;in, This represents the peak load. This represents the load trough value. Time-series characteristic data can be uploaded every 5 minutes, and abnormal data can be (e.g., ...). (Sudden drop > 20% / s) Instant upload.

[0029] In this embodiment, the dynamic prediction model can be an improved LSTM-fused meteorological coupled prediction model. The data processing procedure can be:

[0030] Multi-source perception layer → Multi-head attention layer: 12-dimensional temporal feature data generate matrix;

[0031] Multi-head attention layer → Bidirectional LSTM layer: 64-dimensional fused features Inputting into a bidirectional LSTM layer, historical and future information are fused through bidirectional computation to solve the problem of "incomplete temporal dependency";

[0032] Bidirectional LSTM layer → Co-optimization layer: 64-dimensional hidden state They are used for new energy output forecasting, energy storage SOC forecasting, and load forecasting, respectively. The new energy output forecast value feeds back into the energy storage SOC forecast value, solving the problem of "weak coupling between new energy output and energy storage SOC".

[0033] Specifically, multi-head attention layers can capture the spatiotemporal dependencies of input temporal feature data. The multi-head attention layer captures these dependencies in parallel using multiple sets of attention heads, as shown in the following formula:

[0034] ;

[0035] ;

[0036] ;

[0037] ;

[0038] in, The 12-dimensional feature vector input to the multi-source sensing layer includes 3-dimensional meteorological features, 4-dimensional new energy output features, and 5-dimensional load and energy storage features. These are the query matrix, key matrix, and value matrix;

[0039] in, Input data to Linear transformation weights of the matrix (dimension 12×64). These correspond to the bias vectors (dimension 64×1);

[0040] in, Let i be the value of the attention head. The first Weight matrix of attention heads (dimension 64×64). For the number of attention heads;

[0041] in, , where is the dimension of the key matrix. This is a scaling factor to prevent the attention coefficient from becoming numerically saturated due to excessive dimensionality; The projection matrix (512×64) output by the multi-head attention layer compresses the 512-dimensional fusion features of the 8 attention heads to 64 dimensions, which are then used as the input to the bidirectional LSTM layer.

[0042] The bidirectional LSTM layer, through time-series computation in both forward and reverse directions, integrates historical and future information, overcoming the limitation of traditional unidirectional LSTM layers in not being able to utilize "future context." The formula system is as follows:

[0043]

[0044]

[0045]

[0046] in, This is the 64-dimensional feature vector output by the multi-head attention layer;

[0047] in, These are the hidden states of the forward / backward LSTM layers at time t (dimension 64×1), corresponding to the temporal dependencies of "history → present" and "future → present", respectively. The hidden state of the forward LSTM layer at time t-1; The hidden state of the inverse LSTM layer at time t+1;

[0048] in, The cell states of the forward / reverse LSTM layers at time t (dimension 64×1) are used for long-term memory retention. The cell state of the forward LSTM layer at time t; The cell state of the inverse LSTM layer at time t+1;

[0049] in, The output vector of the input gate in the forward gated recurrent unit (GRU) at time t is used to control the proportion of input information flowing in at the current time.

[0050] The output vector of the input gate in the reverse gated loop unit at time t is used to control the proportion of input information flowing in at the current time during backpropagation;

[0051] is the output vector of the forget gate in the forward-gated recurrent unit at time t, used to control the forgetting ratio of cell state information from the previous time step;

[0052] is the output vector of the forget gate in the backward-gated recurrent unit at time t, used to control the proportion of cell state information forgotten during backpropagation;

[0053] is the output vector of the output gate in the forward-gated recurrent unit at time t, used to control the output ratio of cell state information at the current time;

[0054] is the output vector of the output gate in the reverse gated recurrent unit at time t, used to control the output ratio of the cell state information at the current time during backpropagation;

[0055] in, : These are the bias terms for each gate and candidate cell state in the forward and reverse gated recurrent units, respectively, used to adjust the results of the linear transformation.

[0056] in, and These are the weight matrices (64×64) for the forward LSTM layer and the reverse LSTM layer, respectively. and These are the hidden layer weight matrices for the forward LSTM layer and the reverse LSTM layer, respectively (dimension 64×64).

[0057] in, The fusion weight for the positive hidden state is slightly higher than the fusion weight for the negative hidden state (0.45), prioritizing the influence of historical data.

[0058] In this embodiment, for the prediction of new energy output, it is necessary to simultaneously capture short-term fluctuations (such as 15-minute-level changes caused by cloud cover) and long-term trends (such as the shift in daily peak hours caused by seasonal changes). The formula system is as follows:

[0059]

[0060]

[0061]

[0062]

[0063]

[0064] in, Let t+i be the predicted output value (kW) of new energy sources at time t+i, where i=1,2,…,96 (corresponding to the next 24 hours, with an interval of 15 minutes).

[0065] in, It is the weighted historical average power output (kW) of new energy sources at time t+i over the same period of the past 7 days. The weighting coefficient corresponds to the actual value of daily renewable energy output in history (weighting coefficient for the most recent 3 days > 0.8, weighting coefficient for the 7th day ≈ 0.3). The current date; The actual historical renewable energy output value at time t+i over the past d days;

[0066] in, The linear trend term (kW) is fitted by least squares using the actual output of new energy sources at time t+i over the past 30 days. k is the trend slope (e.g., k=0.2kW / 15min in summer, indicating that the output increases by 0.2kW every 15 minutes).

[0067] in, Let be the time variable, and b be the intercept term of the trend line, used to characterize the baseline level of the long-term trend. For a point in time in a historical time series, This represents the actual output value of new energy sources at time t+i on day d in the past. It is the average value at historical time points. This is the historical average output of new energy sources;

[0068] in, This is the real-time error correction term, based on the actual output of new energy sources at time t. Predicted output of new energy sources at time t The deviation calculation is corrected by a factor of 0.12.

[0069] in, This represents the 64-dimensional hidden state output by the bidirectional LSTM layer. This is the feature mapping function (linear transformation). The mapping weight matrix corresponding to the prediction of new energy power output. The activation function constrains the output to [-1, 1] to prevent the predicted output of new energy sources from exceeding the rated power of the equipment; This is a bias term with a dimension of 64×1, used to adjust the results of the linear transformation.

[0070] In this embodiment, for energy storage SOC prediction, overcharging (>80%) and over-discharging (<20%) need to be avoided, while also incorporating the predicted output value of new energy sources. The formula system is as follows:

[0071]

[0072]

[0073]

[0074]

[0075] in, It is a concatenated vector (65×1) of the hidden state (64-dimensional) of the bidirectional LSTM layer and the predicted value of new energy output (1-dimensional). , These are the maximum and minimum SOC values ​​for energy storage, respectively.

[0076] As a safety margin correction term, when the predicted value of energy storage SOC is close to the safety boundary (>75% or <25%), a correction of ±2% is introduced to prevent overcharging and over-discharging due to prediction deviation in actual operation. Let be the predicted SOC value of the energy storage at time t+i;

[0077] in, Feature mapping function (linear transformation) This is the mapping weight matrix corresponding to the predicted SOC value of energy storage, with a weight of 0.35 (higher than 0.35). (average weight 0.015) As the activation function, the mapping result is constrained to Ensure that the predicted SOC value of energy storage is within [ , Within the range; The bias term has the following dimensions: .

[0078] In this embodiment, the load forecast value is calculated based on the following formula:

[0079]

[0080] ;

[0081] in: : Load forecast at any given time; For the first A 64-dimensional fused feature vector output from each partition via a multi-head attention layer; for Forecast values ​​of renewable energy output at any given time;

[0082] in, The feature mapping function is of the form: ;in, , is the corresponding mapping weight matrix ( (Corresponding weight is 0.45) For bias terms; The activation function constrains the predicted values ​​to be non-negative; This is the load power scaling factor, which is taken as the historical load peak-to-valley difference; This is the baseline load power, and its value is taken as the historical minimum load value.

[0083] In this embodiment, the loss function of the dynamic prediction model is a multi-task loss function. The multi-task loss function needs to balance the prediction accuracy of new energy output and energy storage SOC, while preventing the dynamic prediction model from overfitting. The formula system is as follows:

[0084]

[0085]

[0086]

[0087]

[0088]

[0089]

[0090]

[0091]

[0092] in, For the output loss of new energy sources, MAE (mean absolute error) and SMAPE (symmetric mean absolute percentage error) are integrated. MAE focuses on absolute error (such as kW-level deviation), while SMAPE focuses on relative error (such as percentage deviation).

[0093] in, For energy storage SOC loss, MSE (mean square error) is used, which has a more significant penalty for large deviations (such as energy storage SOC prediction deviation > 5%).

[0094] in, For L2 regularization terms, This is the weight matrix of the bidirectional LSTM layer. This is a regularization coefficient to prevent overfitting in the dynamic prediction model; This is the mapping weight matrix corresponding to the new energy forecast. This is the mapping weight matrix corresponding to the predicted SOC value of energy storage.

[0095] in, To smooth out losses, the continuity of the predicted values ​​of renewable energy output and regional energy storage SOC is constrained at adjacent time points to avoid sudden changes in the predicted values ​​of regional renewable energy output or regional energy storage SOC (such as rising from 50% to 60% within 15 minutes).

[0096] in For load loss, for The actual value of the load at any given time. for Forecasted load values ​​at any time This represents the number of time periods to be predicted.

[0097] in, The weight of the loss item is slightly higher than that of the loss of energy storage SOC (0.35), so as to give priority to ensuring the accuracy of the prediction of energy output.

[0098] In this embodiment, the training of the dynamic prediction model can be performed offline (system initialization), specifically:

[0099] Dataset: Historical data from the past 12 months (sample size ≥ 100,000 records), divided into training set / validation set / test set in a 7:2:1 ratio;

[0100] Optimizer: Adam (learning rate) , , );

[0101] Iteration termination condition: The loss function value of the validation set does not decrease for 50 consecutive iterations, and the optimal parameters are saved.

[0102] Online updates (2:10-2:25 AM daily):

[0103] Incremental training: Freeze the parameters of the multi-head attention layer and the two-layer LSTM layer, and co-optimize the parameters of the layer. The training time is ≤30 minutes.

[0104] Error verification: If the SMAPE error increases by more than 5% after the update, it will automatically roll back to the historical optimal parameters.

[0105] S120: Determine the global absorption rate target based on the predicted output of new energy sources in the distribution transformer area and the predicted load of the distribution transformer area.

[0106] In this embodiment, the distribution transformer area can be divided into three zones based on user type and renewable energy penetration rate. This clarifies the control strategies for different scenarios, ensuring that the strategies are adapted to the characteristics of each zone. Table 2 shows the situation of each zone within the distribution transformer area. Table 3 shows the control situation of each zone within the distribution transformer area.

[0107] Table 2

[0108]

[0109] Table 3

[0110]

[0111] in, The predicted output of renewable energy to the distribution transformer area. This represents the predicted load value for the distribution transformer area.

[0112] Specifically, the excess renewable energy power of a distribution transformer area can be determined based on its predicted renewable energy output and load. The overall renewable energy absorption rate target can then be determined based on this excess renewable energy power and the predicted renewable energy output of the distribution transformer area. Specifically, the overall absorption rate target is determined based on the following formula:

[0113]

[0114] ( hour, )

[0115] in, The target for controlling the curtailment rate of renewable energy;

[0116] in, The excess power of renewable energy in the distribution transformer area refers to the portion of renewable energy output that cannot be absorbed by the grid and needs to be regulated through power curtailment or energy storage. The predicted output of new energy to the distribution transformer area.

[0117] The following security constraints must be met:

[0118] ;

[0119] in, Real-time power (kVA) of the distribution transformer area. The rated capacity of the distribution transformer area;

[0120] Real-time bus voltage (V). .

[0121] S130: For each zone, the excess power that the zone needs to absorb is determined based on the global absorption rate target, the zone weight coefficient, the predicted value of the renewable energy output of the distribution transformer area, and the renewable energy penetration rate of the zone.

[0122] In this embodiment, when regulating each partition, the global absorption rate needs to be broken down to generate a partition strategy, and then each partition is controlled based on the strategy.

[0123] Optionally, determining the excess power to be absorbed by a zone based on the global absorption rate target, the zone weight coefficient, the predicted renewable energy output of the distribution transformer area, and the zone renewable energy penetration rate includes: determining the zone absorption rate target based on the global absorption rate target and the zone weight coefficient; determining the zone renewable energy excess power based on the predicted renewable energy output of the distribution transformer area, the predicted load of the distribution transformer area, and the zone renewable energy penetration rate; and determining the excess power to be absorbed by the zone based on the zone renewable energy excess power and the zone absorption rate target.

[0124] Specifically, the calculation can be performed based on the following formula:

[0125]

[0126] ;

[0127] in, Excess power (kW) that needs to be absorbed by the zone; The target for the absorption rate of wastewater in each zone is (%). The zoning weighting coefficients are: 0.25 for residential zones, 0.3 for commercial zones, and 0.45 for industrial zones, with a total of 1. The excess power of renewable energy in each zone (kW) is allocated by the zone according to the renewable energy penetration rate of the zone.

[0128] Among them, the regional absorption rate target The formula is: Regional excess power of new energy ,in, To address the excess power of new energy sources in the distribution transformer area, The new energy penetration rate is calculated as follows: New energy penetration rate = installed capacity (or output ratio) of new energy in the region / total installed capacity (or total output) of new energy in the distribution transformer area.

[0129] After determining the excess power that each zone needs to absorb, values ​​can be sent to each zone in the form of commands so that each zone can control the corresponding equipment. For example, after receiving the target of "absorbing 47.5kW of excess power", the industrial zone breaks it down into the command of "35kW energy storage charging + 12.5kW wind turbine load increase", with the priority set to "high" and the execution time limit "immediate".

[0130] S140: Based on the excess power to be absorbed by the partition, the new energy regulation weight, and the predicted value of the new energy output of the partition, control the new energy output of the partition, and based on the excess power to be absorbed by the partition, the energy storage regulation weight, the predicted load of the partition, and the predicted value of the new energy output of the partition, control the energy storage charging and discharging power of the partition, and based on the excess power to be absorbed by the partition, control the controllable load of the partition; wherein, the energy storage charging and discharging power of the partition is constrained based on the predicted value of the partition's energy storage SOC.

[0131] In this embodiment, controlling the renewable energy output of a zone based on the excess power to be absorbed by the zone, the renewable energy regulation weight, and the predicted renewable energy output of the zone includes: determining a renewable energy output reduction coefficient based on the excess power to be absorbed by the zone and the renewable energy regulation weight; determining the renewable energy output of the zone based on the renewable energy output reduction coefficient and the predicted renewable energy output of the zone; and controlling the renewable energy output of the corresponding zone based on the renewable energy output of the zone.

[0132] Specifically, the output of new energy sources is controlled by the following algorithm formula:

[0133]

[0134] in, The renewable energy output (kW) of a zone can be the renewable energy output command received by the zone. The reduction coefficient for new energy output (0-0.3 for overcapacity scenarios, 0 for undercapacity scenarios); The weighting for new energy regulation is 0.1 for residential zones, 0.15 for commercial zones, and 0.2 for industrial zones; among which, the response delay is <50ms and the power tracking accuracy is ≤±2% of the command value.

[0135] In this embodiment, optionally, controlling the energy storage charging and discharging power of the partition based on the excess power to be absorbed by the partition, the energy storage regulation weight, the partition load forecast, and the partition renewable energy output forecast includes: determining a first target power based on the excess power to be absorbed by the partition and the energy storage regulation weight, and selecting the larger value between the first target power and the maximum energy storage charging power as the partition's energy storage charging power; determining a target difference between the partition load forecast and the partition renewable energy output forecast, and selecting the larger value between the target difference and zero as the partition renewable energy output gap; determining a second target power based on the partition renewable energy output gap and the energy storage regulation weight, and selecting the larger value between the second target power and the maximum discharge power as the partition energy storage discharge power.

[0136] Specifically, the energy storage charging and discharging power of the zones is controlled based on the following formula system:

[0137] ;

[0138] in, The energy storage charging and discharging power (kW) can be the charging and discharging power command received by the zone, which is used to control the energy storage device. Energy storage regulation weights (0.3 for residential zones, 0.4 for commercial zones, and 0.5 for industrial zones);

[0139] in, This represents the maximum charging power for energy storage. This represents the maximum discharge power of the energy storage.

[0140] in, ;

[0141] in, The regional renewable energy output gap (kW) represents the portion of the regional load demand that renewable energy output cannot meet. If the result is non-positive, the gap is 0. This represents the zone load forecast (kW), which is the expected electricity demand of that zone at a certain moment. The predicted output of new energy in a given zone (kW) is the expected power generation of new energy sources such as wind power and photovoltaics in that zone. During the control process, the response delay is <30ms, the energy storage SOC control accuracy is ±1%, and the charging and discharging current ripple is ≤5%.

[0142] In this embodiment, controlling the controllable load of the partition based on the excess power to be absorbed by the partition includes: determining the power level to be absorbed by the partition based on the excess power to be absorbed by the partition, controlling the charging start and stop time of the charging piles in the partition and controlling the air conditioners in the partition to increase the preset temperature based on the power level.

[0143] Specifically, staggered charging times: adjusting charging start times. With end time For example, residential zoning instructions

[0144] ;

[0145] Air conditioner power adjustment: By modifying the set temperature Adjust the power using the following formula:

[0146] ;

[0147] ;

[0148] in: The air conditioner power adjustment amount (kW) indicates the amount of power that the air conditioner needs to adjust during the control process. This refers to the air conditioner power adjustment coefficient, which is the amount of power change corresponding to each unit adjustment step. "This is the unit for adjusting the air conditioner's speed step."

[0149] The set temperature adjustment amount (°C) for the air conditioner indicates the range of temperature adjustment for the air conditioner. This indicates the absolute value range of the temperature adjustment range for the air conditioner, meaning that the adjustment range of the air conditioner's set temperature is between 2 and 3 adjustment steps (â).

[0150] In this embodiment, the constraint on the predicted SOC value of the zoned energy storage is specifically reflected in the following: if the predicted SOC value of the zoned energy storage is greater than 75%, then the energy storage charging power (e.g., not exceeding 80% of the conventional charging power) is used to prevent overcharging; if the predicted SOC value of the zoned energy storage is less than 75%, then the energy storage charging power is limited to 75% to prevent overcharging. Over-discharge threshold The energy storage discharge power will be limited (e.g., not exceeding 80% of the normal discharge power) to prevent over-discharge.

[0151] In this embodiment, if there is a surplus of new energy in a zone and energy storage charging is required, but the predicted energy storage SOC is greater than 70%, the consumption task will be allocated to controllable loads (such as peak-shifting of charging piles in residential zones and adjustment of air conditioning load) to reduce the pressure on energy storage charging.

[0152] If the predicted regional renewable energy deficit necessitates energy storage discharge, and if the predicted regional energy storage SOC is less than 30%, then the proportion of energy storage discharge should be reduced, prioritizing renewable energy output or grid power purchases to supplement the deficit, thus avoiding insufficient energy storage SOC to sustain regulation. In the multi-objective optimization algorithm, if the predicted regional energy storage SOC is close to the safety boundary (…), or The system dynamically reduces the weight of "energy storage charging and discharging costs" in the overall objective function and increases the weight of "controllable load adjustment" and "photovoltaic output adaptation," so that the optimization strategy prioritizes non-energy storage resources and ensures energy storage security.

[0153] S150: Monitor the operating status data of the distribution transformer area. If the operating status data does not meet the preset safety constraints, adjust the distribution transformer area accordingly.

[0154] In this embodiment, after executing S140, the real-time operating data of each partition can be monitored. If the deviation between the real-time operating data and the target data is less than the preset deviation threshold, global control can be omitted, and the operating status data of the distribution transformer area can be monitored. If the operating status data does not meet the preset safety constraints, the distribution transformer area can be controlled.

[0155] Specifically, for the distribution transformer area, the actual bus voltage can be collected every 50ms. Calculate the voltage deviation:

[0156]

[0157] like This triggered emergency intervention; among them, For voltage deviation, This is the actual bus voltage. This is the rated voltage of the busbar;

[0158] Specifically, the higher the voltage limit ( ):

[0159] ;

[0160] ;

[0161] Among them, for voltage exceeding the upper limit, the energy storage charging power can be increased by 50%, the output of new energy sources can be reduced by 20%, and if it is not restored within 1 second, 20% of the controllable load will be further shut down;

[0162] The lower the voltage limit ( ):

[0163] ;

[0164] ;

[0165] Specifically, for voltages exceeding the lower limit, the energy storage discharge power can be increased by 50%, shutting down 10-20kW of non-core loads. If the loads do not recover within one second, the energy storage discharge power is increased to its maximum value. Provides emergency charging power for energy storage; This is the standard charging power for energy storage. Contributing urgently to new energy; Contribute to the regular development of new energy sources; For emergency discharge power of energy storage; This is the normal discharge power for energy storage; This refers to the power of the out-of-service load.

[0166] In this embodiment, for transformer overload protection, the low-voltage side current of the transformer is sampled every 50ms. Calculate the load factor:

[0167] ;in, Load rate; Overload protection is triggered. This refers to the rated current on the low-voltage side of the distribution transformer.

[0168] Specifically, for mild overload (105% < ≤110%)

[0169] ;

[0170] in, This refers to the shutdown power range of industrial heating equipment. This refers to the reduction ratio of power output from renewable energy sources (photovoltaics). For mild overload, shut down 20-30kW industrial heating equipment (reducing renewable energy output by 10%-15%), and within 5 seconds... Reduced to ≤105%;

[0171] For heavy overload ( ):

[0172] ;

[0173] in, This refers to the range of total load power that needs to be reduced during severe overload. This represents the percentage reduction in total photovoltaic output under severe overload conditions. For severe overload, 50-80kW of controllable industrial loads can be shut down, reducing photovoltaic output to 50%, while energy storage is activated and discharging (30-40kW), reducing output within 1 second. When the value drops to ≤105%, a local audio-visual alarm (volume 100dB, red indicator light constantly on) and a remote 4G SMS notification are triggered.

[0174] The technical solution provided in this application determines the predicted values ​​of renewable energy output, energy storage SOC, and load within a distribution transformer area, and establishes a global absorption target for the distribution transformer area. For each area, the global absorption target is broken down to determine the excess power that the area needs to absorb. Based on the excess power to be consumed by the area, the renewable energy control weight, and the predicted renewable energy output, the renewable energy output of the area is controlled. Furthermore, based on the excess power to be consumed by the area, the energy storage control weight, the predicted load, and the predicted renewable energy output, the energy storage charging and discharging power of the area is controlled. By controlling the controllable load of a zone based on the excess power it needs to consume, the differences between zones are taken into account. The excess power that each zone needs to absorb is determined, and each zone is controlled based on the excess power it needs to absorb and other corresponding indicators. This allows for rapid adaptation to fluctuations in renewable energy and can significantly improve the renewable energy absorption rate. Furthermore, by constraining the energy storage charging and discharging power of a zone through the predicted value of its energy storage SOC, and by monitoring the operating status data of the distribution transformer area, the distribution transformer area can be regulated when preset safety constraints are not met, thereby improving the safety of the distribution transformer area.

[0175] Figure 2 This is a flowchart of a new energy collaborative consumption control method provided in this application embodiment. Based on the above embodiment, this embodiment may further include: monitoring the real-time operating data of each zone; if the deviation between the real-time operating data and the target data is greater than a preset deviation threshold, determining the new energy consumption rate, operating cost, and environmental benefits of the distribution transformer area; normalizing and weighting the new energy consumption rate, operating cost, and environmental benefits to obtain a total objective function; solving the total objective function based on a multi-objective optimization algorithm to obtain an optimal control scheme, and performing global optimization of the distribution transformer area based on the optimal control scheme.

[0176] like Figure 2 As shown, the method provided in this application embodiment includes the following steps:

[0177] S210: Determine the predicted values ​​of renewable energy output, state of charge (SOC) of energy storage, and load for each zone within the distribution transformer area, and determine the predicted values ​​of renewable energy output and load for the distribution transformer area.

[0178] S220: Determine the global absorption rate target based on the predicted output of new energy sources in the distribution transformer area and the predicted load of the distribution transformer area.

[0179] S230: For each zone, the excess power that the zone needs to absorb is determined based on the global absorption rate target, the zone weight coefficient, the predicted value of the renewable energy output of the distribution transformer area, and the renewable energy penetration rate of the zone.

[0180] S240: Based on the excess power to be absorbed by the partition, the new energy regulation weight, and the predicted value of the new energy output of the partition, control the new energy output of the partition, and based on the excess power to be absorbed by the partition, the energy storage regulation weight, the predicted load of the partition, and the predicted value of the new energy output of the partition, control the energy storage charging and discharging power of the partition, and based on the excess power to be absorbed by the partition, control the controllable load of the partition; wherein, the energy storage charging and discharging power of the partition is constrained based on the predicted value of the partition's energy storage SOC.

[0181] S250: Monitor the real-time operating data of each zone. If the deviation between the real-time operating data and the target data is greater than a preset deviation threshold, determine the local consumption rate of new energy, operating cost and environmental benefits of the distribution transformer area based on the real-time operating data.

[0182] In this embodiment, real-time operating data of each partition can be monitored every 100ms. Real-time operating parameters can include real-time output values ​​of new energy sources, real-time SOC values ​​of energy storage, etc., and the deviation is calculated based on the following formula:

[0183]

[0184] If the deviation exceeds a preset deviation threshold, global optimization and control are implemented. A non-dominated sorting genetic algorithm is used to balance the renewable energy absorption rate, economic costs, and environmental benefits to generate the optimal control scheme. Power deviation; Real-time power; The target power is the power command sent to the device.

[0185] In this embodiment, the optimization objective is to maximize the local renewable energy consumption rate of the distribution transformer area. The renewable energy consumption rate of each zone is calculated based on the following formula:

[0186] ;

[0187] ;

[0188] in, The power consumption of new energy sources in each zone (kW);

[0189] in, ;in To determine the local load power of a zone, the smaller value between the actual output of renewable energy in the zone and the load in the zone is taken to ensure that local consumption does not exceed local demand.

[0190] in, The actual output value of new energy sources in the region is obtained directly from the output forecast data or actual monitoring data of new energy sources (wind power and photovoltaic).

[0191] in, This represents the difference between the actual output of renewable energy in a given area and the power consumed in that area (kW).

[0192] In this embodiment, the renewable energy absorption rate of the distribution transformer area It can be a weighted combination of the local consumption rate of new energy in each district and the corresponding zoning weight coefficient, to obtain the comprehensive local consumption rate of new energy.

[0193] In this embodiment, one of the optimization objectives is to minimize the operating cost of the distribution transformer substation, which is calculated based on the following formula:

[0194] ;

[0195] in, For the operating costs of the distribution transformer area; The energy storage charging and discharging cost for the distribution transformer area is calculated based on the following formula:

[0196] ;

[0197] in, The grid power procurement cost for the distribution transformer area is calculated based on the following formula:

[0198] ;

[0199] in, and These are the energy storage charging power and the energy storage discharging power (kW), respectively. and : Charging time and discharging time (h);

[0200] in, Yuan / kWh is the grid electricity price. Yuan / kWh represents the cost of energy storage losses.

[0201] Power (kW) procured for grid power. The duration of power procurement (h).

[0202] In this embodiment, one of the optimization objectives may be to maximize environmental benefits, which are calculated based on the following formula:

[0203] ;

[0204] in, For environmental benefits (kgCO2), that is, the reduction in carbon dioxide emissions;

[0205] The duration of the statistical period is in hours (h).

[0206] kgCO2 / kWh: Emission reduction coefficient per unit of fossil energy.

[0207] S260: Normalize and weight the new energy absorption rate, operating cost and environmental benefits to obtain the overall objective function, solve the overall objective function based on the multi-objective optimization algorithm to obtain the optimal control scheme, and perform global optimization of the distribution transformer area based on the optimal control scheme.

[0208] In this embodiment, the overall objective function is:

[0209] ;

[0210] in, , This is the normalized benchmark value for the grid power cost of the transformer area at full load for 24 hours; The normalized baseline value for emission reduction of new energy sources at full power output for 24 hours; The objective function value (dimensionless) ranges from 0 to 1; the larger the value, the better the control scheme. , ,and These are the weighting coefficients, with values ​​of 0.4, 0.3, and 0.3 respectively.

[0211] In this embodiment, the objective function is solved using a multi-objective algorithm as follows:

[0212] (1) Population initialization: generation There are several feasible solutions, each corresponding to a set of control parameters (energy storage charging power). Energy storage and discharge power Load transfer power The control parameters meet the constraints:

[0213] ;

[0214] (2) Non-dominated sorting: For each solution Calculate the number of times you are dominated (Number of Dominant Solutions) and Dominant Set (The set of solutions dominated by this solution), partition the Pareto front:

[0215] First Frontier: (No other solutions exist);

[0216] Second frontier: Only dominated by the first frontier solution ( );

[0217] This process continues until all solutions have been partitioned.

[0218] (3) Crowding degree calculation: Calculate the crowding degree of the solution within its respective frontier to avoid the algorithm converging to a local optimum:

[0219]

[0220] in, To solve The degree of congestion;

[0221] in, and They are respectively and The kth objective function value;

[0222] in, , They are the first The maximum and minimum values ​​of the objective function;

[0223] Among them, the front boundary solution (To ensure we are not eliminated).

[0224] Selection-crossover-mutation:

[0225] Choice: Roulette wheel selection method, probability of choice. ( (Rankings based on cutting-edge trends)

[0226] (4) Crossover: Single-point crossover, crossover probability ;

[0227] (5) Mutation: random mutation, mutation probability ;

[0228] (6) Iteration Termination: During the iteration After that, the solution with the largest objective function value in the first frontier is output as the optimal control scheme, and the control scheme is used to control each partition.

[0229] S270: Monitor the operating status data of the distribution transformer area. If the operating status data does not meet the preset safety constraints, adjust the distribution transformer area accordingly.

[0230] For details on the other steps, please refer to the above embodiments.

[0231] Based on the above embodiments, the deviation of the renewable energy absorption rate of the distribution transformer area is calculated every hour. If the deviation of the renewable energy absorption rate is greater than a preset threshold, the partition weight coefficient is adjusted. The renewable energy prediction deviation, renewable energy absorption rate, line loss rate, operating cost and environmental benefits of the distribution transformer area are calculated daily. Based on the renewable energy prediction deviation, renewable energy absorption rate, line loss rate, operating cost and environmental benefits of the distribution transformer area, the attention coefficient of the dynamic prediction model and the weight coefficient in the overall objective function are adjusted.

[0232] Specifically, real-time collection of control execution data, calculation of control deviations, and continuous optimization are achieved, as follows:

[0233] (1) Short-term correction (seconds to minutes): when voltage deviation At the same time, adjust the power regulation coefficient of the voltage control module:

[0234] ;

[0235] in, and These are the power regulation coefficients before and after the voltage control module correction.

[0236] (2) Mid-term optimization (hourly): Calculate the deviation of the renewable energy consumption rate of the distribution transformer area every hour. , The actual renewable energy absorption rate refers to the proportion of renewable energy power actually absorbed to the total renewable energy output during the statistical period. The target absorption rate is a preset target for the proportion of new energy consumption. If... Adjust partition weight coefficients ;

[0237] (3) Long-term iteration (daily): Based on the daily operating data (new energy consumption rate, line loss rate, operating cost, environmental benefits), fine-tune the attention coefficient of the dynamic prediction model and the target weight of the multi-objective optimization algorithm to ensure that the system adapts to changes in operating conditions over a long period of time.

[0238] Table 4 records the collaborative process of the technical solutions in the embodiments of this application. For details, please refer to Table 4.

[0239] Table 4

[0240]

[0241] The technical solution provided in this application embodiment can realize closed-loop control of the entire process from data acquisition to optimized absorption in the distribution transformer area, effectively solve the problems of randomness and fluctuation in new energy output, and take into account safety, economy and environmental protection goals, providing reliable technical support for the large-scale access of new energy in the distribution transformer area.

[0242] The solution provided in this application addresses the problems of traditional data acquisition lag and single-dimensionality by frequently collecting renewable energy output data, load data, energy storage data, and meteorological data, combined with outlier removal, data compression, and renewable energy-load coupling feature extraction. This requires protection of the acquisition mechanism, preprocessing algorithm, and data security transmission verification logic. The application proposes a dynamic prediction model, introducing a multi-head attention mechanism, dual-task output of renewable energy output prediction and energy storage SOC prediction, and a multi-task loss function to improve prediction accuracy. Furthermore, this application can divide residential, commercial, and industrial zones according to renewable energy penetration rates and formulate differentiated consumption strategies, achieving millisecond-level equipment response and rapid adaptation to new conditions. Energy fluctuations; through a multi-objective collaborative optimization algorithm, under the constraints of distribution transformer overload and voltage stability, the algorithm balances the renewable energy absorption rate, operating costs, and environmental benefits, constructs a total objective function, solves for the optimal control scheme, and performs global optimization based on the optimal control scheme. Under the premise of meeting preset safety conditions, the distribution transformer area is adjusted, which can significantly improve the local renewable energy absorption rate and comprehensively solve the problems of low absorption efficiency, high safety risks, and economic and environmental imbalance of existing technologies. Through short-term voltage coefficient correction, medium-term model adjustment of zoning weight coefficients, and long-term adjustment of dynamic prediction model and weight coefficients in the total objective function, the efficient absorption and safe operation of renewable energy in the distribution transformer area are finally achieved.

[0243] Figure 3 This is a structural block diagram of a new energy consumption control device provided in an embodiment of this application, such as... Figure 3 As shown, the device includes:

[0244] The prediction module 310 is used to determine the predicted value of renewable energy output, the predicted value of state of charge (SOC) of energy storage, and the predicted value of load of each zone in the distribution transformer area, and to determine the predicted value of renewable energy output and the predicted value of load of the distribution transformer area.

[0245] The first determining module 320 determines the global absorption rate target based on the predicted value of new energy output of the distribution transformer area and the predicted value of load of the distribution transformer area;

[0246] The second determining module 330 is used to determine the excess power that needs to be absorbed by each zone based on the global absorption rate target, the zone weight coefficient, the predicted value of the new energy output of the distribution transformer area, and the new energy penetration rate of the zone.

[0247] The control module 340 is used to control the renewable energy output of the partition based on the excess power to be absorbed by the partition, the renewable energy regulation weight, and the predicted renewable energy output of the partition; and to control the energy storage charging and discharging power of the partition based on the excess power to be absorbed by the partition, the energy storage regulation weight, the predicted load of the partition, and the predicted renewable energy output of the partition; and to control the controllable load of the partition based on the excess power to be absorbed by the partition; wherein the energy storage charging and discharging power of the partition is constrained based on the predicted energy storage SOC value of the partition.

[0248] The control module 350 is used to monitor the operating status data of the distribution transformer area. If the operating status data does not meet the preset safety constraints, the control module 350 controls the distribution transformer area.

[0249] In an alternative embodiment, the apparatus further includes a global optimization module for controlling the controllable load of the partition based on the excess power to be absorbed:

[0250] Monitor the real-time operating data of each zone. If the deviation between the real-time operating data and the target data is greater than a preset deviation threshold, determine the renewable energy consumption rate, operating cost, and environmental benefits of the distribution transformer area.

[0251] The new energy absorption rate, operating costs, and environmental benefits are normalized and weighted to obtain the overall objective function.

[0252] The overall objective function is solved using a multi-objective optimization algorithm to obtain the optimal control scheme, and the distribution transformer area is then globally optimized based on the optimal control scheme.

[0253] In an optional embodiment, determining the predicted values ​​of renewable energy output, state of charge (SOC) of energy storage, and load of each zone within the distribution transformer area includes:

[0254] Collect new energy data, load data, energy storage data, and meteorological data from various zones within the distribution network area, and extract features from the new energy data, load data, and energy storage data to obtain time-series feature data;

[0255] The time-series feature data is input into the dynamic prediction model to obtain the predicted values ​​of renewable energy output, load, and energy storage SOC for each region; wherein, the dynamic prediction model includes a multi-head attention layer and a bidirectional LSTM layer.

[0256] The loss function of the dynamic prediction model is determined by a weighted combination of new energy output loss, energy storage SOC loss, L2 regularization term, smoothing loss, and load loss. Specifically, the new energy output loss is determined based on the actual and predicted values ​​of new energy output in the region; the energy storage SOC loss is determined based on the actual and predicted values ​​of energy storage SOC in the region; the smoothing loss is determined based on the predicted values ​​of new energy output and energy storage SOC in the region; the L2 regularization term is determined by the mapping weight matrix in the dynamic prediction model; and the load loss is determined based on the actual and predicted values ​​of load in the region.

[0257] In an optional embodiment, determining the excess power to be absorbed by a zone based on the global absorption rate target, the zone weight coefficient, the predicted renewable energy output of the distribution transformer area, and the zone's renewable energy penetration rate includes:

[0258] The partition absorption rate target is determined based on the global absorption rate target and the partition weight coefficient;

[0259] The excess power of renewable energy in a given area is determined based on the predicted output of renewable energy in the given area, the predicted load of the given area, and the renewable energy penetration rate of the given area.

[0260] The excess power of new energy in the region is determined based on the region's renewable energy surplus power and the region's absorption rate target.

[0261] In an optional embodiment, controlling the renewable energy output of a region based on the excess power to be absorbed by the region, the renewable energy regulation weight, and the predicted renewable energy output of the region includes:

[0262] The renewable energy output reduction coefficient is determined based on the excess power that needs to be absorbed by the aforementioned zones and the renewable energy regulation weights.

[0263] The renewable energy output of a region is determined based on the renewable energy output reduction coefficient and the predicted renewable energy output of the region, and renewable energy output control is performed on the corresponding region based on the renewable energy output of the region.

[0264] In an optional embodiment, controlling the energy storage charging and discharging power of the zone based on the excess power to be absorbed by the zone, the energy storage regulation weight, the zone load forecast, and the zone renewable energy output forecast includes:

[0265] The first target power is determined based on the excess power to be absorbed by the partition and the energy storage regulation weight, and the larger value between the first target power and the maximum energy storage charging power is selected as the energy storage charging power of the partition.

[0266] Determine the target difference between the predicted load value of the zone and the predicted output value of the zone's new energy sources, and select the larger value between the target difference and zero as the zone's new energy output gap;

[0267] The second target power is determined based on the regional renewable energy output gap and the energy storage regulation weight, and the larger value between the second target power and the maximum discharge power is selected as the regional energy storage discharge power.

[0268] In an optional embodiment, the controllable load of the partition based on the excess power to be absorbed by the partition includes:

[0269] Based on the excess power that the zone needs to absorb, the power level that the zone needs to absorb is determined, and based on the power level, the charging start and stop time of the charging piles in the zone is controlled, and the air conditioner in the zone is controlled to increase the preset temperature.

[0270] In an alternative embodiment, the device further includes an adjustment module for:

[0271] The deviation of the renewable energy absorption rate of the distribution transformer area is calculated every hour. If the deviation of the renewable energy absorption rate is greater than the preset absorption deviation threshold, the zoning weight coefficient is adjusted.

[0272] The daily statistics include the new energy prediction deviation, new energy absorption rate, line loss rate, operating cost, and environmental benefits of the distribution transformer area. Based on the new energy prediction deviation, new energy absorption rate, line loss rate, operating cost, and environmental benefits of the distribution transformer area, the attention coefficient of the dynamic prediction model and the weight coefficient in the overall objective function are adjusted.

[0273] like Figure 3 As shown in the figure, this application provides an electronic device, including a processor 111, a communication interface 112, a memory 113, and a communication bus 114, wherein the processor 111, the communication interface 112, and the memory 113 communicate with each other through the communication bus 114.

[0274] Memory 113 is used to store computer programs;

[0275] In one embodiment of this application, when the processor 111 executes a program stored in the memory 113, it implements the method provided in any of the foregoing method embodiments, including:

[0276] Determine the predicted values ​​of renewable energy output, state of charge (SOC) of energy storage, and load of each zone within the distribution transformer area.

[0277] Based on the predicted values ​​of renewable energy output and load of the respective zones, the predicted values ​​of renewable energy output and load of the distribution transformer area are determined, and the global absorption rate target is determined based on the predicted values ​​of renewable energy output and load of the distribution transformer area.

[0278] For each zone, the excess power that the zone needs to absorb is determined based on the global absorption rate target, the zone weight coefficient, the predicted value of the renewable energy output of the distribution transformer area, and the renewable energy penetration rate of the zone.

[0279] The system controls the renewable energy output of a region based on the excess power to be absorbed by that region, the renewable energy regulation weight, and the predicted renewable energy output of that region. It also controls the energy storage charging and discharging power of that region based on the excess power to be absorbed by that region, the energy storage regulation weight, the predicted load of that region, and the predicted renewable energy output of that region. Furthermore, it controls the controllable load of that region based on the excess power to be absorbed by that region. The energy storage charging and discharging power of that region is constrained by the predicted energy storage SOC of that region.

[0280] Monitor the operating status data of the distribution transformer area. If the operating status data does not meet the preset safety constraints, adjust the distribution transformer area accordingly.

[0281] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the method provided in any of the foregoing method embodiments.

[0282] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown 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 this embodiment according to actual needs.

[0283] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented using software plus a general-purpose hardware platform, or of course, using hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0284] The above embodiments are merely illustrative examples and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations. However, obvious variations or modifications derived therefrom are still within the scope of protection of this application.

Claims

1. A method for controlling the consumption of new energy sources, characterized in that, include: Determine the predicted values ​​of renewable energy output, state of charge (SOC) of energy storage, and load of each zone within the distribution transformer area, and determine the predicted values ​​of renewable energy output and load of the distribution transformer area. The global absorption rate target is determined based on the predicted new energy output and load of the distribution transformer area. For each zone, the excess power that the zone needs to absorb is determined based on the global absorption rate target, the zone weight coefficient, the predicted value of the renewable energy output of the distribution transformer area, and the renewable energy penetration rate of the zone. The renewable energy output of a region is controlled based on the excess power to be absorbed by the region, the renewable energy regulation weight, and the predicted renewable energy output of the region. The energy storage charging and discharging power of a region is also controlled based on the excess power to be absorbed by the region, the energy storage regulation weight, the predicted load of the region, and the predicted renewable energy output of the region. Furthermore, the controllable load of a region is controlled based on the excess power to be absorbed by the region. The energy storage charging and discharging power of a region is constrained by the predicted energy storage SOC of the region. Monitor the operating status data of the distribution transformer area; if the operating status data does not meet the preset safety constraints, adjust the distribution transformer area. The determination of the predicted values ​​of renewable energy output, state of charge (SOC) of energy storage, and load of each zone within the distribution transformer area includes: Collect new energy data, load data, energy storage data, and meteorological data from various zones within the distribution network area, and extract features from the new energy data, load data, and energy storage data to obtain time-series feature data; The time-series feature data is input into the dynamic prediction model to obtain the predicted values ​​of renewable energy output, load, and energy storage SOC for each region; wherein, the dynamic prediction model includes a multi-head attention layer and a bidirectional LSTM layer. The loss function of the dynamic prediction model is determined by a weighted combination of new energy output loss, energy storage SOC loss, L2 regularization term, smoothing loss, and load loss. Specifically, the new energy output loss is determined based on the actual and predicted values ​​of new energy output in the region; the energy storage SOC loss is determined based on the actual and predicted values ​​of energy storage SOC in the region; the smoothing loss is determined based on the predicted values ​​of new energy output and energy storage SOC in the region; the L2 regularization term is determined based on the mapping weight matrix in the dynamic prediction model and the weight matrix of the bidirectional LSTM layer; and the load loss is determined based on the actual and predicted values ​​of the regional load.

2. The method according to claim 1, characterized in that, Following the controllable load of the controllable load zone based on the excess power to be absorbed, the following is also included: Monitor the real-time operating data of each zone. If the deviation between the real-time operating data and the target data is greater than a preset deviation threshold, determine the renewable energy consumption rate, operating cost, and environmental benefits of the distribution transformer area. The new energy absorption rate, operating costs, and environmental benefits are normalized and weighted to obtain the overall objective function. The overall objective function is solved using a multi-objective optimization algorithm to obtain the optimal control scheme, and the distribution transformer area is then globally optimized based on the optimal control scheme.

3. The method according to claim 1, wherein determining the excess power to be absorbed by a zone based on the global absorption rate target, the zone weight coefficient, the predicted value of renewable energy output of the distribution transformer area, and the zone renewable energy penetration rate comprises: The partition absorption rate target is determined based on the global absorption rate target and the partition weight coefficient; The excess power of renewable energy in a given area is determined based on the predicted output of renewable energy in the given area, the predicted load of the given area, and the renewable energy penetration rate of the given area. The excess power of new energy in the region is determined based on the region's renewable energy surplus power and the region's absorption rate target.

4. The method according to claim 1, characterized in that, The control of renewable energy output in a region based on the excess power to be absorbed by the region, the renewable energy regulation weight, and the predicted renewable energy output of the region includes: The renewable energy output reduction coefficient is determined based on the excess power that needs to be absorbed in the aforementioned zones and the renewable energy regulation weights. The renewable energy output of a region is determined based on the renewable energy output reduction coefficient and the predicted renewable energy output of the region, and renewable energy output control is performed on the corresponding region based on the renewable energy output of the region.

5. The method according to claim 1, characterized in that, The method of controlling the energy storage charging and discharging power of a zone based on the excess power to be absorbed by the zone, the energy storage regulation weight, the zone load forecast, and the zone renewable energy output forecast includes: The first target power is determined based on the excess power to be absorbed by the partition and the energy storage regulation weight, and the larger value between the first target power and the maximum energy storage charging power is selected as the energy storage charging power of the partition. Determine the target difference between the predicted load value of the zone and the predicted renewable energy output value of the zone, and select the larger value between the target difference and zero as the renewable energy output gap of the zone. The second target power is determined based on the regional renewable energy output gap and the energy storage regulation weight, and the larger value between the second target power and the maximum discharge power is selected as the regional energy storage discharge power.

6. The method according to claim 1, characterized in that, The controllable load of the controllable load of the partition based on the excess power to be absorbed by the partition includes: Based on the excess power that the zone needs to absorb, the power level that the zone needs to absorb is determined, and based on the power level, the charging start and stop time of the charging piles in the zone is controlled, and the air conditioner in the zone is controlled to increase the preset temperature.

7. The method according to claim 2, characterized in that, Also includes: The deviation of the renewable energy absorption rate of the distribution transformer area is calculated every hour. If the deviation of the renewable energy absorption rate is greater than the preset absorption deviation threshold, the zoning weight coefficient is adjusted. The daily statistics include the new energy prediction deviation, new energy absorption rate, line loss rate, operating cost, and environmental benefits of the distribution transformer area. Based on the new energy prediction deviation, new energy absorption rate, line loss rate, operating cost, and environmental benefits of the distribution transformer area, the attention coefficient of the dynamic prediction model and the weight coefficient in the overall objective function are adjusted.

8. An electronic device, characterized in that, It includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method as described in any one of claims 1-7.

9. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed in a computer, causes the computer to perform the method described in any one of claims 1-7.

Images