Power supply switching method, device, equipment and storage medium of all-electric energy station

Abstract

The present invention belongs to the field of integrated energy technology, and discloses a power supply switching method, device, equipment and storage medium of an all-electric energy station, the method comprising: obtaining the current load of the all-electric energy station at the current time point and the current power generation power of the first power supply device; performing load prediction based on the historical load and current load of the all-electric energy station, and determining the predicted load at a future time point; performing power prediction based on the current power generation power, and determining the predicted power generation power at a future time point; when the predicted power generation power cannot meet the predicted load, determining the power generation power of the second power supply device; when the power generation power of the second power supply device meets the predicted load, switching from the first power supply device to the second power supply device. The present invention can determine in advance whether it is necessary to switch the power supply device, and there is enough time before the switch to further determine whether the standby second power supply device is suitable for power supply. The switching operation is simple, and the stable operation of the all-electric energy station can be ensured.

Description

Technical Field

[0001] The present invention belongs to the field of integrated energy technology, and specifically relates to a power supply switching method, device, equipment and storage medium for an all-electric energy station. Background Art

[0002] An all-electric energy station is a distributed power supply energy station that integrates renewable resources such as solar energy and wind energy. It utilizes a variety of different types of power generation methods such as photovoltaic power generation and wind power generation to achieve multi-energy power supply. The coordinated cooperation of multiple energy sources improves the power supply reliability of the all-electric energy station. In addition, through the refined management of multiple energy sources, efficient energy utilization can be achieved, and low carbon emissions can be achieved.

[0003] Since photovoltaic power generation, wind power generation and other power supply methods have certain limitations, it is necessary to switch the power supply based on the actual situation. For example, when there is sufficient sunlight during the day, switch to photovoltaic power generation mode; when the wind speed is high, switch to wind power generation; or use off-peak electricity at night to achieve peak load shifting.

[0004] Prior art document 1 (CN118674201A) provides a method and device for grid interaction of all-electric energy stations based on the energy characteristics of high-density buildings. By considering the building materials, building structures, and distribution density of high-density buildings in the target area, and combining the historical user living conditions, historical electricity consumption data, and historical weather data of high-density buildings to determine the user electricity consumption model, the electricity consumption data considering temperature control in the future time period is determined; based on the electricity consumption data considering temperature control in the future time period, according to the energy attributes of different energy stations associated with high-density buildings in the target area, the demand response data in the future time period is sent to the corresponding power grid to achieve interaction with the power grid, thereby improving the overall operation efficiency of the power grid. It can be seen that prior art document 1 focuses on the grid interaction of all-electric energy stations. On this basis, with the continuous increase in the penetration rate of new energy, further challenges have been raised to the technical solution of the grid interaction of all-electric energy stations in terms of power switching.

[0005] Prior art document 2 (CN118608187A) discloses a method and system for optimizing and controlling new energy vehicle energy stations in megacities. By constructing an operating model for new energy vehicle energy stations and considering energy management and price setting methods, the optimization and control of new energy vehicle energy stations in megacities are carried out to improve the operating income of new energy vehicle energy stations and the utilization rate of demand-side resources, and support the safe, economical and efficient operation of megacities' power grids. It can be seen that prior art document 2 focuses on the issue of energy supply and demand balance in new energy vehicle energy stations. On this basis, as the charging behavior of new energy vehicles poses a more severe challenge to the energy management of energy stations, the output prediction and timely switching of various energy equipment and the matching of meeting power supply demand need to be solved urgently. Summary of the invention

[0006] The present invention provides a power supply switching method, device, equipment and storage medium of an all-electric energy station to achieve switching between power supply equipment using renewable energy in the all-electric energy station.

[0007] In a first aspect, the present invention provides a power supply switching method for an all-electric energy station, comprising:

[0008] Obtain the historical load of the all-electric energy station, the current load at the current time point, and the current power generation power of the first power supply device; the first power supply device is the device currently supplying power to the all-electric energy station;

[0009] Perform load prediction based on the historical load of the all-electric energy station and the current load to determine the predicted load at a future time point, including: determining the load change cycle of the all-electric energy station, taking the load change cycle as a unit, generating multiple load sequences based on the historical load of the all-electric energy station and the current load, the load sequence including multiple loads sorted by time, setting a weight coefficient for each load sequence, the sum of all the weight coefficients being 1, performing weighted processing on the multiple load sequences based on the weight coefficients to generate an effective load sequence for the load change cycle, inputting the effective load sequence into a pre-trained load prediction model, and determining the predicted load at a future time point based on the prediction result of the load prediction model;

[0010] Perform power prediction based on the current generated power to determine the predicted generated power of the first power supply device at the future time point;

[0011] In the case where the predicted power generation cannot meet the predicted load, determining the power generation power of the second standby power supply device;

[0012] When the power generation power of the second power supply device meets the predicted load, the first power supply device is switched to the second power supply device, and the second power supply device supplies power to the all-electric energy station.

[0013] In some optional implementations, the i-th payload in the effective payload sequence is expressed by the following formula (1):

[0014] (1)

[0015] Where:

[0016] is the i-th payload in the effective payload sequence,

[0017] is the i-th load in the k-th load sequence, , The number of loads included for the load sequence;

[0018] For the The weight coefficient of the load sequence, , expressed as the following formula (2):

[0019] (2)

[0020] Where:

[0021] is the number of payload sequences.

[0022] In some optional implementations, the load prediction model is used to perform iterative dilation and convolution processing on the effective load sequence, and determine the predicted load at a future time point based on the corresponding iterative dilation and convolution processing results.

[0023] In some optional implementations, the load prediction model includes: an embedding mapping layer, an output layer, and a plurality of sequentially arranged iterative dilated convolution structures; the iterative dilated convolution structure includes a plurality of sequentially arranged dilated convolution layers;

[0024] The embedding mapping layer is used to perform embedding mapping processing on the effective load sequence to generate a corresponding load feature vector sequence;

[0025] The iterative dilated convolution structure is used to perform iterative dilated convolution processing on the input feature vector sequence, and output the feature vector sequence after the iterative dilated convolution processing; the feature vector sequence input to the first iterative dilated convolution structure is the load feature vector sequence;

[0026] The output layer is used to perform linear processing on the input feature vector sequence to generate a prediction result corresponding to the effective load sequence; the feature vector sequence input to the output layer is a superposition of the effective load sequence and the feature vector sequence output by the last iterative dilated convolution structure;

[0027] The dilated convolution layer is used to perform dilated convolution processing on the input feature vector sequence and output the feature vector sequence after the dilated convolution processing; and the jth feature vector in the feature vector sequence output by the i-th dilated convolution layer is obtained by performing dilated convolution processing on the jth feature vector, the jth feature vector and the jth feature vector in the feature vector sequence input to the i-th dilated convolution layer. The feature vector, The jth feature vector in the feature vector sequence output by the i-th dilated convolution layer is obtained by convolution of feature vectors. It is expressed as the following formula (3):

[0028] (3)

[0029] Where:

[0030] represents the convolution kernel of the i-th dilated convolutional layer,

[0031] represents the convolution operation,

[0032] , , They represent the feature vector sequence input to the i-th dilated convolutional layer, the j-th feature vector, the The feature vector, feature vectors.

[0033] In some optional implementations, after determining the power generation power of the second standby power supply device, the method further includes:

[0034] Determine a maximum load corresponding to the power generation power of the second power supply device, and use the larger value of the current load and the predicted load as a reference load;

[0035] Determine whether the difference between the maximum load and the reference load is greater than a preset threshold; if the difference between the maximum load and the reference load is greater than the preset threshold, determine that the generated power of the second power supply device meets the predicted load.

[0036] In some optional implementations, the process of setting the preset threshold includes:

[0037] Constructing multiple first fuzzy sets corresponding to the power generation of the second power supply device, and setting corresponding first membership functions; constructing multiple second fuzzy sets corresponding to the load of the all-electric energy station, and setting corresponding second membership functions; constructing multiple third fuzzy sets corresponding to the preset threshold, and setting corresponding third membership functions;

[0038] Determine the corresponding relationship between the first fuzzy set, the second fuzzy set and the corresponding third fuzzy set, and construct a fuzzy control rule;

[0039] According to the first membership function of the first fuzzy set, it is determined that the power generation of the second power supply device belongs to the first membership of the first fuzzy set, and the first fuzzy set corresponding to the non-zero first membership is used as the first valid fuzzy set; according to the second membership function of the second fuzzy set, it is determined that the reference load belongs to the second membership of the second fuzzy set, and the second fuzzy set corresponding to the non-zero second membership is used as the second valid fuzzy set;

[0040] Determine, according to the fuzzy control rule, a third valid fuzzy set among the plurality of third fuzzy sets, which has a corresponding relationship with the first valid fuzzy set and the second valid fuzzy set;

[0041] Determining a third membership degree corresponding to the third valid fuzzy set according to the first membership degree of the first valid fuzzy set and the second membership degree of the second valid fuzzy set; the third membership degree corresponding to the third valid fuzzy set is between the first membership degree of the first valid fuzzy set and the second membership degree of the second valid fuzzy set;

[0042] In the third membership function of the third valid fuzzy set, the region whose membership is less than the third membership is taken as a valid region, and all the valid regions of the third valid fuzzy sets are superimposed to form a superimposed region;

[0043] The threshold corresponding to the center of gravity of the superimposed area is used as the preset threshold.

[0044] In a second aspect, the present invention provides a power supply switching device for an all-electric energy station, comprising:

[0045] An acquisition module, used to acquire the current load of the all-electric energy station at a current time point, and the current power generation power of a first power supply device; the first power supply device is a device currently supplying power to the all-electric energy station;

[0046] A prediction module is used to perform load prediction based on the historical load of the all-electric energy station and the current load to determine the predicted load at a future time point; perform power prediction based on the current power generation to determine the predicted power generation at the future time point;

[0047] A preprocessing module, used for determining the power generation power of the standby second power supply device when the predicted power generation power cannot meet the predicted load;

[0048] A switching module is used to switch from the first power supply device to the second power supply device when the power generation power of the second power supply device meets the predicted load, so that the second power supply device supplies power to the all-electric energy station.

[0049] In a third aspect, the present invention provides a computer device, comprising: a memory and a processor, the memory and the processor are communicatively connected to each other, computer instructions are stored in the memory, and the processor executes the power supply switching method of the all-electric energy station according to the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0050] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the power supply switching method for an all-electric energy station of the above-mentioned first aspect or any corresponding embodiment thereof.

[0051] The embodiment of the present invention predicts the load of the all-electric energy station and the power generation of the first power supply device currently in use, and can determine in advance whether the current first power supply device can continue to supply power, so as to determine in advance whether it is necessary to switch the power supply device; and before switching, there is enough time to further determine whether the standby second power supply device is suitable for power supply, and when the second power supply device can supply power, the switch is made. This method can realize the direct switching of two renewable resource power supply devices without other switching actions, is simple to operate, and can also ensure the stable operation of the all-electric energy station. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the related technologies, the drawings required for use in the specific embodiments or the related technical descriptions will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0053] Figure 1 is a flow chart of a power supply switching method of an all-electric energy station according to an embodiment of the present invention;

[0054] Figure 2 is a schematic diagram of power supply switching of an all-electric energy station according to an embodiment of the present invention;

[0055] Figure 3 is a flow chart of another power supply switching method of an all-electric energy station according to an embodiment of the present invention;

[0056] Figure 4 is a structural schematic diagram of a load prediction model according to an embodiment of the present invention;

[0057] Figure 5 is a schematic diagram of an iterative dilated convolution structure according to an embodiment of the present invention;

[0058] Figure 6 are fuzzy sets and corresponding membership functions set for power generation, load and threshold according to an embodiment of the present invention;

[0059] Figure 7 is a schematic diagram of a superposition area according to an embodiment of the present invention;

[0060] Figure 8is a structural block diagram of a power supply switching device of an all-electric energy station according to an embodiment of the present invention;

[0061] Fig. 9 It is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0062] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0063] All-electric energy stations allow for multiple renewable energy sources to be used for power supply. When one renewable resource is no longer suitable for power supply, you can try to switch to another renewable resource for power supply. However, at this time, since another renewable resource is also not suitable for power supply, the power supply will be unstable. At present, you can first switch from one renewable resource to the mains system for power supply. After that, if it is determined that another renewable resource can supply power, you can switch from the mains system to another renewable resource for power supply. This switching method is complicated and requires multiple switchings. It will also affect the service life of electrical components and increase the failure rate.

[0064] An embodiment of the present invention provides a power supply switching method for an all-electric energy station. By predicting the load and power generation conditions of the all-electric energy station, it is possible to determine in advance whether the current power supply equipment can continue to supply power, and there is sufficient time before switching to determine whether another power supply equipment is suitable for power supply, thereby achieving rapid switching of two renewable resource power supply equipment.

[0065] Figure 1 is a flow chart of a power supply switching method for an all-electric energy station according to an embodiment of the present invention. Figure 1 As shown, the method flow includes the following steps S101 to S105.

[0066] Step S101: Obtain the current load of the all-electric energy station at the current time point and the current power generation power of the first power supply device; the first power supply device is the device currently supplying power to the all-electric energy station.

[0067] In the implementation of the present invention, the all-electric energy station can supply power to a variety of loads during operation, so that the corresponding load amount can be determined; and at the current time point, the load amount of the all-electric energy station, that is, the current load amount, can be determined. For example, if the all-electric energy station is applied to a residential building, the sum of the load amounts of all devices currently using electricity (such as lights, air conditioning lights) in the residential building at the current time point can be determined, and the sum of the load amounts is used as the current load amount.

[0068] Moreover, at the current time point, a certain power supply device can supply power to the all-electric energy station. For the convenience of description, the power supply device is referred to as the "first power supply device". Since this embodiment mainly involves switching of equipment powered by renewable resources, the first power supply device is specifically a power supply device powered by renewable resources, such as photovoltaic power generation equipment, wind power generation equipment, etc.

[0069] The power generation power of the first power supply device at the current time point, ie, the current power generation power, may also be determined.

[0070] Step S102: Perform load forecasting based on the historical load and current load of the all-electric energy station to determine the forecasted load at a future time point.

[0071] Since the load of the all-electric energy station changes dynamically, when some electrical equipment is started or the power supply to some electrical equipment is stopped, the load will change. In the embodiment of the present invention, in order to be able to predict the load of the all-electric energy station, in addition to determining the current load, it is also necessary to obtain the historical load of the all-electric energy station at previous time points; for example, the all-electric energy station can record the load collected each time, store the load and the corresponding time point, so as to form a historical load; when performing subsequent load prediction, the corresponding historical load can be directly read from the storage space.

[0072] Based on the historical load and current load of the all-electric energy station, the load change of the all-electric energy station can be determined, and then based on the load change, the load of the all-electric energy station can be predicted, so that the load at a future time point can be determined, that is, the predicted load.

[0073] For example, a load prediction model for load prediction can be pre-built, and based on the load prediction model, the predicted load of the all-electric energy station at a future time point can be predicted. For example, the future time point can be a time point a certain time (e.g., 1 minute, 5 minutes, etc.) after the current time point.

[0074] It can be understood that in order to predict the load more accurately, historical loads at multiple historical time points are required.

[0075] Step S103: Perform power prediction based on the current generated power to determine the predicted generated power at a future time point.

[0076] In the embodiment of the present invention, similar to load prediction, power prediction can also be performed based on the current power generation. Specifically, the power generation law of the first power supply device can be determined, and the power generation power prediction can be achieved based on the power generation law.

[0077] Among them, similar to load prediction, the historical power generation of the first power supply device at the historical time point can be determined, and the power prediction can be performed by combining the historical power generation and the current power generation, so as to determine the predicted power generation at the future time point. Alternatively, since the power generation of the first power supply device may be affected by natural conditions, for example, photovoltaic power generation equipment is affected by natural conditions such as light intensity and solar angle, power prediction can also be performed based on the change of natural conditions, so as to determine the change of power generation between the current time point and the future time point, and the predicted power generation at the future time point can be determined by combining the current power generation.

[0078] In this embodiment, in order to ensure the comparability of the predicted load and the predicted power generation, the two are the prediction results at the same time point, that is, when the load prediction and the power generation prediction are performed, they correspond to the same future prediction point.

[0079] Step S104: when the predicted power generation cannot meet the predicted load, determine the power generation power of the second standby power supply device.

[0080] In the embodiment of the present invention, if the predicted power generation of the first power supply device at a future time point is large, it can generally meet the predicted load, that is, if the first power supply device is continued to be used for power supply, it can still provide power to the corresponding power-consuming equipment in the future, and there is no need to switch the power supply device. On the contrary, if the predicted power generation of the first power supply device at a future time point is small, it may not be able to meet the predicted load, and the first power supply device is no longer suitable for continuing to supply power, so it is necessary to switch the power supply device at this time.

[0081] For example, the predicted power generation is the power generation of the first power supply device at a future time point, and the predicted power generation corresponds to the load that can be supported, that is, if the first power supply device continues to be used for power supply at a future time point, the first power supply device can supply power to the power consumption device of the load. By comparing the load with the above-mentioned predicted load, it can be relatively simple to determine whether the predicted power generation can meet the predicted load. For example, if the predicted load exceeds the load corresponding to the predicted power generation, it can be determined that the predicted power generation cannot meet the predicted load.

[0082] If the predicted power generation cannot meet the predicted load, you can try to switch to another power supply device. For ease of description, this power supply device is called the second power supply device. It can be understood that at the current time point, the second power supply device is a backup power supply device. And before performing the switching operation, the power generation of the second power supply device is determined first.

[0083] Among them, similar to the first power supply device, the second power supply device is also a power supply device that uses renewable resources for power supply. For example, the first power supply device is a photovoltaic power generation device and the second power supply device is a wind power generation device, or the first power supply device is a wind power generation device and the second power supply device is a photovoltaic power generation device. This embodiment does not limit this.

[0084] Step S105: When the power generation power of the second power supply device meets the predicted load, switch from the first power supply device to the second power supply device, and the second power supply device supplies power to the all-electric energy station.

[0085] In an embodiment of the present invention, if the power generation power of the second power supply device meets the predicted load capacity, it means that if the second power supply device is used for power supply, it can drive these power loads. Therefore, before the future time point arrives, the second power supply device can be used for power supply in advance, that is, the power supply equipment used by the all-electric energy station is switched from the first power supply equipment to the second power supply equipment; after the switch, the second power supply equipment supplies power to the all-electric energy station, and the corresponding first power supply equipment is converted into a backup power supply equipment for the all-electric energy station; for example, the first power supply equipment stops working, or charges the energy storage system of the all-electric energy station, etc., which can be determined based on actual conditions.

[0086] If the power generation power of the second power supply equipment also does not meet the predicted load, and the all-electric energy station has other power supply equipment that uses renewable resources for power supply, such as hydrogen power supply equipment, etc., it is possible to continue to determine whether the power generation power of other power supply equipment meets the predicted load. If so, switch to the other power supply equipment for power supply; if all power supply equipment evenly meets the requirements, the all-electric energy station can switch to the AC power supply mode, and be provided with electricity by stable AC power to ensure that the all-electric energy station can operate stably.

[0087] Figure 2 A schematic diagram of switching power supply for an all-electric energy station is shown. Figure 2 As shown, the all-electric energy station includes photovoltaic power generation equipment and wind power generation equipment, and the all-electric energy station can also be connected to the mains. In addition, the photovoltaic power generation equipment is currently used to power the load of the all-electric energy station, that is, the first power supply equipment is the photovoltaic power generation equipment. Figure 2 The black solid line in the figure represents the power supply circuit of the load.

[0088] If it is determined through prediction that the power supply capacity of the photovoltaic power generation equipment has decreased, it can be determined whether the backup wind power generation equipment can meet the load capacity, that is, the second power supply equipment is the wind power generation equipment; if the power supply capacity of the wind power generation equipment meets the demand, it is determined that it can be switched, that is, switched to the power supply of the wind power generation equipment, and its power supply circuit is referred to Figure 2 On the contrary, if the power supply capacity of the wind power generation equipment is also insufficient, it cannot be switched, but is switched to the mains power supply. Figure 2 Shown in the lower right corner.

[0089] The power supply switching method of the all-electric energy station provided by the embodiment of the present invention can determine in advance whether the current first power supply device can continue to supply power by predicting the load of the all-electric energy station and the power generation power of the first power supply device currently in use, so as to determine in advance whether it is necessary to switch the power supply device; and before switching, there is enough time to further determine whether the standby second power supply device is suitable for power supply, and when the second power supply device can supply power, the switch is performed. This method can realize the direct switching of two renewable resource power supply devices without other switching actions, is simple to operate, and can also ensure the stable operation of the all-electric energy station.

[0090] In this embodiment, a power supply switching method of an all-electric energy station is provided. Figure 3 is a flow chart of a power supply switching method for an all-electric energy station according to an embodiment of the present invention. Figure 3 As shown, the process includes the following steps.

[0091] Step S201: Obtain the current load of the all-electric energy station at the current time point and the current power generation power of the first power supply device; the first power supply device is the device currently supplying power to the all-electric energy station.

[0092] For details, see Figure 1 The relevant description of step S101 of the illustrated embodiment will not be repeated here.

[0093] Step S202: Perform load forecasting based on the historical load and current load of the all-electric energy station to determine the forecasted load at a future time point.

[0094] Specifically, the above step S202 "performs load forecasting based on the historical load and current load of the all-electric energy station to determine the forecast load at a future time point" includes steps S2021 to S2024.

[0095] Step S2021: Determine the load change cycle of the all-electric energy station.

[0096] Since the load of the all-electric energy station is sudden, for example, when the electrical equipment is turned on or off, the load of the all-electric energy station will change suddenly, thus affecting the accuracy of the load prediction. In addition, the load change is generally periodic. In the embodiment of the present invention, the load of the all-electric energy station is predicted in combination with the periodic change law of the load, so as to reduce the problem of poor prediction accuracy caused by sudden load changes.

[0097] Specifically, the load variation law of the all-electric energy station can be determined based on the load of the all-electric energy station over a period of time, thereby determining the corresponding load variation cycle. For example, the load variation of the all-electric energy station is basically similar every day, so the load variation cycle can be specifically 1 day.

[0098] Step S2022: Taking the load change cycle as a unit, generate multiple load sequences according to the historical load and current load of the all-electric energy station; the load sequence includes multiple loads sorted by time.

[0099] In an embodiment of the present invention, taking the current time point as the latest time point, the historical load amounts within multiple load change cycles can be determined; and the corresponding load amounts within each load change cycle can be determined, so that multiple load amount sequences can be constructed based on the load change cycle as a unit, and each load amount sequence includes multiple load amounts sorted by time.

[0100] For example, if the load change cycle is 1 day, taking the current time point as the latest time point, and extracting the load of 7 days in the past, 7 load sequences can be constructed, each of which includes all the loads in one day. It can be understood that the last load of the last load sequence is the current load.

[0101] Step S2023: Combine multiple load sequences to generate an effective load sequence of a load variation period.

[0102] In the embodiment of the present invention, load prediction is performed based on load sequences of multiple load change cycles, which can improve the accuracy of the prediction result, but the required amount of data is large; in order to reduce the complexity of load prediction and improve the training efficiency of the load prediction model, multiple load sequences are merged into a load sequence of a load change cycle, and load prediction is performed based on the merged load sequence. For ease of description, the merged load sequence is called an effective load sequence.

[0103] Optionally, the above step S2023 "combining multiple load sequences to generate an effective load sequence of a load change period" may specifically include steps A1 to A2.

[0104] Step A1: Set a weight coefficient for each load sequence; the sum of all weight coefficients is 1.

[0105] In this embodiment, since the newer the load sequence is, generally, closer to the future load changes, its reference value is higher. Therefore, when setting the weight coefficient for each load sequence, it is set based on the newness of the load sequence. The newer the load sequence, the larger the corresponding weight coefficient.

[0106] Specifically, if there are s load sequences, that is, the number of load sequences is s, for the kth load sequence, let its weight coefficient be In this embodiment, the weight coefficient of the kth load sequence is It is specifically expressed as follows:

[0107] (1)

[0108] Where:

[0109] is the number of payload sequences,

[0110] For the The weight coefficient of the load sequence, .

[0111] Based on this formula, we can see that the larger the k, the newer the load sequence, and its corresponding weight coefficient Also bigger.

[0112] And, all weight coefficients The sum is expressed by the following formula (2):

[0113] (2)

[0114] Where:

[0115] All weight coefficients The sum of all weight coefficients The sum is 1.

[0116] Step A2: performing weighted processing on multiple load sequences according to weight coefficients to generate an effective load sequence of a load variation period.

[0117] In the embodiment of the present invention, if the load sequence includes n loads, and represents the i-th load in the k-th load sequence, i=1,2,…,n, then the i-th load in multiple load sequences can be weighted and the weighted processing result can be used as the i-th load in the effective load sequence , expressed as the following formula (3):

[0118] (3)

[0119] Where:

[0120] is the i-th payload in the effective payload sequence,

[0121] is the weight coefficient of the kth load sequence,

[0122] is the i-th load in the k-th load sequence, ,

[0123] s is the number of payload sequences.

[0124] In this embodiment, by weighting multiple payload sequences, the required effective payload sequence can be determined simply and quickly; and the weight coefficient used in the weighting process is It is related to time, so that the effective load sequence can contain more features of relatively new load sequences, which is beneficial to subsequent load prediction.

[0125] Step S2024, input the effective load sequence into the pre-trained load prediction model, and determine the predicted load at a future time point based on the prediction result of the load prediction model.

[0126] In this embodiment, a load prediction model for load prediction is pre-trained. After the effective load sequence is determined, the effective load sequence is used as input data of the load prediction model, so that the load prediction model can be used to perform load prediction on the effective load sequence, thereby determining the predicted load at a future time point.

[0127] Optionally, the load prediction model can be implemented based on RNN (Recurrent Neural Network), such as LSTM (Long Short-Term Memory) network, but because RNN needs to process in sequence, the processing efficiency is low; in order to improve the prediction efficiency of the load prediction process, this embodiment constructs the load prediction model based on iterative dilated convolutional neural network (IDCNN).

[0128] Specifically, the load prediction model is used to perform iterative dilation convolution processing on the effective load sequence, and determine the predicted load at a future time point according to the corresponding iterative dilation convolution processing result. In this embodiment, since the iterative dilation convolution processing can be implemented in parallel, the prediction efficiency of the load prediction model can be improved; and the iterative dilation convolution processing method has a larger perception field, which can cover most or even all of the load in the effective load sequence, and can ensure the accuracy of the prediction result.

[0129] Optionally, Figure 4 A structural schematic diagram of the load prediction model is shown; Figure 4 As shown, the load prediction model includes: an embedding mapping layer 301, an output layer 303 and a plurality of iterative dilated convolution structures 302 arranged in sequence; the iterative dilated convolution structure 302 includes a plurality of dilated convolution layers 3021 arranged in sequence. Figure 4 In the example, the load prediction model includes two iterative dilated convolution structures 302, and each iterative dilated convolution structure 302 includes three dilated convolution layers 3021 arranged in sequence.

[0130] Among them, the embedding mapping layer 301 is used to perform embedding mapping processing on the effective payload sequence. The payload in the effective payload sequence can be vectorized by Embedding, and the feature vector corresponding to each payload can be determined, so as to generate a corresponding feature vector sequence, that is, a payload feature vector sequence.

[0131] The iterative dilated convolution structure 302 is used to perform iterative dilated convolution processing on the input feature vector sequence and output the feature vector sequence after the iterative dilated convolution processing; the feature vector sequence input to the first iterative dilated convolution structure 302 is a load feature vector sequence.

[0132] The output layer 303 is used to perform linear processing on the input feature vector sequence to generate a prediction result corresponding to the effective load sequence; the feature vector sequence input to the output layer 303 is the superposition of the effective load sequence and the feature vector sequence output by the last iterative dilated convolution structure 302.

[0133] In the embodiment of the present invention, since the length of the effective load sequence is long, it is necessary for the iterative dilated convolution structure 302 to have more layers (for example, the number of dilated convolution layers 3021 is large), which will increase the network depth and easily cause problems such as gradient disappearance or gradient explosion. Therefore, in this embodiment, the residual structure is adopted, such as Figure 4 As shown, a residual connection is also established between the embedding mapping layer 301 and the output layer 303.

[0134] Specifically, Figure 4As shown, after the embedding mapping layer 301 generates a load feature vector sequence, the load feature vector sequence can be input into the first iterative dilated convolution structure 302, that is, the input feature vector of the first iterative dilated convolution structure 302 is the load feature vector sequence. The first iterative dilated convolution structure 302 performs iterative dilated convolution processing on the load feature vector sequence, outputs the processed feature vector sequence, and uses the feature vector sequence as the input feature vector of the second iterative dilated convolution structure 302, and continues to perform iterative dilated convolution processing until the last iterative dilated convolution structure 302 completes the processing.

[0135] Furthermore, the feature vector sequence output by the last iterative dilated convolution structure 302 is superimposed on the load feature vector sequence to form a superimposed sequence, which is used as the input of the output layer 303, and the output layer performs linear processing on the superimposed sequence using an activation function, etc., and finally generates a prediction result corresponding to the effective load sequence; for example, the output layer 303 may include a fully connected layer.

[0136] The prediction result includes the predicted load at a future time point. For example, the prediction result includes the load at multiple time points in the future, and the predicted load corresponding to the future time point can be determined.

[0137] In this embodiment, the iterative dilated convolution structure 302 implements iterative dilated convolution processing on the feature vector sequence using multiple dilated convolution layers 3021. Each dilated convolution layer 3021 is used to perform dilated convolution processing on the input feature vector sequence and output the feature vector sequence after the dilated convolution processing.

[0138] Furthermore, the jth feature vector in the feature vector sequence output by the i-th dilated convolutional layer 3021 is obtained by performing the operation of the i-th dilated convolutional layer 3021 on the feature vector sequence input to the i-th dilated convolutional layer 3021. The feature vector, The feature vector, The jth feature vector in the feature vector sequence output by the i-th dilated convolution layer is obtained by convolution of feature vectors. It is expressed as the following formula (4):

[0139] (4)

[0140] Where:

[0141] represents the convolution kernel of the i-th dilated convolutional layer,

[0142] represents the convolution operation,

[0143] , , They represent the feature vector sequence input to the i-th dilated convolutional layer, the j-th feature vector, the The feature vector, feature vectors.

[0144] In the embodiment of the present invention, for the i-th dilated convolution layer 3021 in the iterative dilated convolution structure 302, As the dilation step size of the i-th dilated convolution layer 3021; ​​and, since the future load is not related to the historical load, the feature vector output by the dilated convolution layer 3021 is only related to the input feature vector in the previous time. Specifically, for the i-th dilated convolution layer 3021, based on the j-th feature vector and the j-th feature vector in its input feature vector sequence, The feature vector, feature vectors, and comprehensively determine the jth feature vector in the output feature vector sequence.

[0145] Figure 5 A schematic diagram of iterative dilated convolution structure 302 is shown, as Figure 5 As shown, the iterative dilated convolution structure 302 includes three dilated convolution layers 3021, namely the first dilated convolution layer, the second dilated convolution layer, and the third dilated convolution layer.

[0146] For the first dilated convolutional layer, i=1, so its dilation step size is , for the last eigenvector, if the eigenvector sequence includes n eigenvectors, then j=n; Figure 5 As shown, if In the feature vector sequence input to the first dilated convolutional layer, the i-th feature vector can be obtained by , the n-1th eigenvector , the n-2th eigenvector Convolution processing is performed to determine the nth eigenvector in the eigenvector sequence output by the first dilated convolution layer. The remaining eigenvectors are similar to this and will not be described here.

[0147] Similarly, for the second dilated convolutional layer, i=2, its dilation step size is For the feature vector sequence input to the second dilated convolutional layer, the nth feature vector in the feature vector sequence output by the second dilated convolutional layer can be determined by performing convolution processing on the nth feature vector, the n-2th feature vector, and the n-4th feature vector.

[0148] For the third dilated convolutional layer, i=3, its dilation step size is For the feature vector sequence input to the third dilated convolutional layer, the nth feature vector in the feature vector sequence output by the third dilated convolutional layer can be determined by performing convolution processing on the nth feature vector, the n-4th feature vector, and the n-8th feature vector therein.

[0149] By sequentially processing multiple dilated convolutional layers, iterative dilated convolution processing of an input feature vector sequence (such as a load feature vector sequence) can be achieved.

[0150] By training the load prediction model, the convolution kernel of each dilated convolution layer can be obtained, and then the dilated convolution processing can be implemented based on the convolution kernel. Specifically, if the convolution kernel of the i-th dilated convolution layer is , and input into the feature vector sequence of the i-th dilated convolutional layer, the j-th feature vector, the The feature vector, The feature vectors are , , , then the jth feature vector in the feature vector sequence output by the i-th dilated convolutional layer is , which is specifically expressed as formula (4).

[0151] Step S203: Perform power prediction based on the current power generation to determine the predicted power generation at a future time point.

[0152] For details, see Figure 1 The relevant description of step S103 of the illustrated embodiment will not be repeated here.

[0153] Step S204: when the predicted power generation cannot meet the predicted load, determine the power generation power of the second standby power supply device.

[0154] For details, see Figure 1 The relevant description of step S104 of the illustrated embodiment will not be repeated here.

[0155] Step S205: When the power generation power of the second power supply device meets the predicted load, switch from the first power supply device to the second power supply device, and the second power supply device supplies power to the all-electric energy station.

[0156] For details, see Figure 1 The relevant description of step S105 of the illustrated embodiment will not be repeated here.

[0157] In some optional embodiments, the method further includes a process of determining whether the power generation of the second power supply device meets the predicted load capacity. Specifically, the process includes the following steps B1 to B2.

[0158] Step B1: Determine the maximum load corresponding to the power generation of the second power supply device, and use the larger value of the current load and the predicted load as a reference load.

[0159] Step B2: Determine whether the difference between the maximum load and the reference load is greater than a preset threshold; if the difference between the maximum load and the reference load is greater than the preset threshold, determine that the power generation power of the second power supply device meets the predicted load.

[0160] In an embodiment of the present invention, for a first power supply device that is in operation, a power generation prediction model for the first power supply device can be constructed to predict the power generation power of the first power supply device at a future time point; however, since the second power supply device is currently in a standby state and is not easy to predict accurately, this embodiment pre-sets a threshold for comparison, i.e., a preset threshold, and determines whether the second power supply device can supply power based on the preset threshold.

[0161] Specifically, after determining the power generation of the second power supply device, the load corresponding to the power generation can be determined, that is, the maximum load that the second power supply device can currently support. Since the maximum load is currently determined, it does not mean that the power-consuming device can also be powered at a future time point. Therefore, in this embodiment, the maximum load is compared with the larger value of the current load and the predicted load. For ease of description, the larger value of the current load and the predicted load is called the reference load. For example, if the predicted load is greater than the current load, the reference load is also the predicted load.

[0162] Among them, if the difference between the maximum load and the reference load is greater than the preset threshold, it can be considered that at the current time point and the future time point, the second power supply device is likely to supply power to the power-consuming device, that is, the power generation power of the second power supply device meets the predicted load. On the contrary, if the difference between the maximum load and the reference load is less than the preset threshold, it can be considered that at the current time point or the future time point, if the second power supply device is used for power supply, the power supply capacity is likely to be insufficient, so the power generation power of the second power supply device cannot meet the predicted load, that is, the second power supply device cannot be switched to supply power.

[0163] Optionally, in order to ensure the accuracy of judging whether the second power supply device is suitable for power supply, the preset threshold should not be fixed, that is, the preset threshold is a dynamic threshold; for example, if the power generation power of the second power supply device is larger, or the reference load of the all-electric energy station is larger, the preset threshold should be larger, so as to ensure the accuracy of the judgment result as much as possible. In this embodiment, the preset threshold is set based on the actual situation by means of fuzzy control. Specifically, the process of setting the preset threshold can include the following steps C1 to C7.

[0164] Step C1: Construct multiple first fuzzy sets corresponding to the power generation power of the second power supply device, and set the corresponding first membership function; construct multiple second fuzzy sets corresponding to the load of the all-electric energy station, and set the corresponding second membership function; construct multiple third fuzzy sets corresponding to the preset threshold, and set the corresponding third membership function.

[0165] Step C2: Determine the corresponding relationship between the first fuzzy set, the second fuzzy set and the corresponding third fuzzy set, and construct the fuzzy control rules.

[0166] Step C3: According to the first membership function of the first fuzzy set, determine that the power generation power of the second power supply equipment belongs to the first membership of the first fuzzy set, and take the first fuzzy set corresponding to the non-zero first membership as the first valid fuzzy set; according to the second membership function of the second fuzzy set, determine that the reference load belongs to the second membership of the second fuzzy set, and take the second fuzzy set corresponding to the non-zero second membership as the second valid fuzzy set.

[0167] Step C4: determining, according to the fuzzy control rule, a third valid fuzzy set among the plurality of third fuzzy sets, which has a corresponding relationship with the first valid fuzzy set and the second valid fuzzy set.

[0168] Step C5: Determine a third membership corresponding to a third valid fuzzy set according to the first membership of the first valid fuzzy set and the second membership of the second valid fuzzy set; the third membership corresponding to the third valid fuzzy set is between the first membership of the first valid fuzzy set and the second membership of the second valid fuzzy set.

[0169] Step C6: In the third membership function of the third valid fuzzy set, the region whose membership is less than the third membership is taken as the valid region, and the valid regions of all the third valid fuzzy sets are superimposed to form a superimposed region.

[0170] Step C7: taking the threshold corresponding to the center of gravity of the superimposed area as the preset threshold.

[0171] In an embodiment of the present invention, since the maximum load of the second power supply device is positively correlated with the power generation power of the second power supply device, a suitable preset threshold can be determined based on the power generation power of the second power supply device and the reference load; wherein, multiple fuzzy sets are set for each parameter, and a corresponding membership function is set for each fuzzy set.

[0172] Specifically, multiple fuzzy sets, i.e., first fuzzy sets, are set for the power generation power of the second power supply equipment, and a corresponding membership function, i.e., first membership function, is set for each first fuzzy set; similarly, multiple second fuzzy sets are set for the load capacity of the all-electric energy station, and corresponding second membership functions are set; multiple third fuzzy sets are set for the preset threshold, and corresponding third membership functions are set.

[0173] Figure 6 The fuzzy sets and corresponding membership functions set for the three parameters of power generation, load and threshold are shown. Figure 6 As shown, the power generation of the second power supply equipment is provided with three first fuzzy sets, namely, low, medium and high, the load of the all-electric energy station is provided with three second fuzzy sets, namely, small, medium and large, and the threshold is provided with five third fuzzy sets, namely, A, B, C, D and E. Moreover, the membership function of each fuzzy set adopts a triangular membership function, and other types of membership functions may also be adopted, which may be determined based on actual needs.

[0174] Furthermore, a fuzzy control rule can be constructed, which specifically includes the correspondence between the first fuzzy set, the second fuzzy set and the corresponding third fuzzy set, that is, the combination of the first fuzzy set and the second fuzzy set corresponds to which third fuzzy set. For example, the first fuzzy set "low" and the second fuzzy set "small" have a corresponding relationship with the third fuzzy set A.

[0175] After determining the power generation P of the second power supply device, the membership of the power generation P and each first fuzzy set, i.e., the first membership, can be determined based on the first membership function of each first fuzzy set; it can be understood that some first memberships may be 0. Figure 6 As shown in the figure, if the power generation P is large, its first membership degree to the first fuzzy set "low" is 0, or in other words, the power generation P does not belong to the first fuzzy set "low", and the first fuzzy sets with the first membership degree of 0 do not need to be considered in subsequent calculations. For the convenience of description, the first fuzzy set corresponding to the non-zero first membership degree is taken as the first valid fuzzy set of the power generation P. Similarly, after determining the reference load L, the second membership degree of the reference load L to each second fuzzy set can also be determined, and the second fuzzy set corresponding to the non-zero second membership degree is taken as the second valid fuzzy set corresponding to the reference load L.

[0176] Moreover, based on the fuzzy control rule, a third fuzzy set corresponding to the first valid fuzzy set and the second valid fuzzy set can be determined, that is, the third valid fuzzy set. It can be understood that since the number of the first valid fuzzy set and the second valid fuzzy set may be multiple, the number of the third valid fuzzy set may also be multiple.

[0177] For a combination of the first valid fuzzy set and the second valid fuzzy set, the third membership corresponding to the corresponding third valid fuzzy set can be determined based on the membership of the two (the first membership and the second membership). The third membership corresponding to the third valid fuzzy set is between the first membership of the first valid fuzzy set and the second membership of the second valid fuzzy set. For example, the third membership corresponding to the third valid fuzzy set is the average of the first membership of the first valid fuzzy set and the second membership of the second valid fuzzy set, or the smaller value, the larger value, etc.; in general, the smaller value of the two can be used as the third membership corresponding to the third valid fuzzy set. Based on the same method, the third membership corresponding to each third valid fuzzy set can be determined.

[0178] The third valid fuzzy set can be used for defuzzification to determine the size of the preset threshold. Specifically, for each third valid fuzzy set, a partial area with a membership less than the third membership can be used as a valid area, and the valid areas of all third valid fuzzy sets are superimposed to form a final superposition area. Afterwards, the center of gravity of the superposition area can be determined, and the threshold corresponding to the center of gravity can be used as the final preset threshold.

[0179] Figure 7 A schematic diagram of the superposition area is shown; Figure 7 As shown in , if the third fuzzy sets B and C are both third valid fuzzy sets, and the third membership of the third valid fuzzy set B is 0.25, and the third membership of the third valid fuzzy set C is 0.5, then the superposition area formed by the superposition of the two valid areas can be specifically as follows Figure 7 In addition, if the threshold corresponding to the centroid of the superimposed area is TH, the preset threshold can be set to TH, that is, the preset threshold corresponding to the generated power P and the reference load L is TH, and then based on the preset threshold TH, it can be determined whether the generated power of the second power supply device meets the predicted load.

[0180] The power supply switching method of the all-electric energy station provided by the embodiment of the present invention combines multiple load sequences of the all-electric energy station to generate an effective load sequence. Based on the effective load sequence, it can cope with the load mutation to a certain extent and improve the accuracy of load prediction. The load prediction model performs iterative dilation convolution processing on the effective load sequence, which can realize parallel processing and improve the prediction efficiency of the load prediction model; the load prediction model adopts a residual structure, which can effectively avoid the gradient problem; each dilation convolution layer performs convolution processing based on the input feature vector with a previous time, which is not affected by future features, and also enables iterative dilation convolution to cover more features corresponding to historical load.

[0181] Whether the second power supply device can supply power is determined by a preset threshold value, and there is no need to predict the power generation power of the second power supply device, and the implementation method is simple; corresponding fuzzy sets and membership functions are set for the power generation power, load and threshold, and the corresponding preset threshold value can be determined based on the actual power generation power and reference load of the second power generation device. This process does not require human participation, can minimize the influence of human experience, and can more accurately determine the appropriate preset threshold value.

[0182] The above describes in detail the power supply switching method process of the all-electric energy station. The method can also be implemented by a corresponding device. The structure and function of the device are described in detail below.

[0183] In this embodiment, a power supply switching device for an all-electric energy station is also provided. The device is used to implement the above embodiments and preferred implementations. The descriptions that have been made are not repeated here. Figure 8 As shown, including:

[0184] An acquisition module 701 is used to acquire the current load of the all-electric energy station at the current time point and the current power generation power of the first power supply device; the first power supply device is the device currently supplying power to the all-electric energy station;

[0185] The prediction module 702 is used to perform load prediction based on the historical load of the all-electric energy station and the current load to determine the predicted load at a future time point; perform power prediction based on the current power generation to determine the predicted power generation at the future time point;

[0186] A preprocessing module 703, configured to determine the power generation power of a second standby power supply device when the predicted power generation power cannot meet the predicted load;

[0187] The switching module 704 is used to switch from the first power supply device to the second power supply device when the power generation power of the second power supply device meets the predicted load, so that the second power supply device supplies power to the all-electric energy station.

[0188] In some optional implementations, the prediction module 702 performs load prediction based on the historical load of the all-electric energy station and the current load to determine the predicted load at a future time point, including:

[0189] Determining a load variation cycle of the all-electric energy station;

[0190] Taking the load change cycle as a unit, generating a plurality of load sequences according to the historical load of the all-electric energy station and the current load; the load sequence includes a plurality of loads sorted by time;

[0191] Combining the multiple load sequences to generate an effective load sequence for the load change period;

[0192] The effective load sequence is input into a pre-trained load prediction model, and the predicted load at a future time point is determined according to the prediction result of the load prediction model.

[0193] In some optional implementations, the prediction module 702 combines the multiple load sequences to generate the effective load sequence of the load change period, including:

[0194] Setting a weight coefficient for each load sequence; the sum of all the weight coefficients is 1;

[0195] Performing weighted processing on the multiple load sequences according to the weight coefficient to generate an effective load sequence of the load change period;

[0196] Furthermore, the i-th payload in the effective payload sequence It is expressed as the following formula (5):

[0197] (5)

[0198] Where:

[0199] is the i-th payload in the effective payload sequence,

[0200] is the weight coefficient of the kth load sequence,

[0201] is the i-th load in the k-th load sequence, ,

[0202] s is the number of payload sequences.

[0203] The weight coefficient of the kth load sequence It is specifically expressed as follows:

[0204] (6)

[0205] Where:

[0206] is the number of payload sequences,

[0207] For the The weight coefficient of the load sequence, .

[0208] In some optional implementations, the load prediction model is used to perform iterative dilation and convolution processing on the effective load sequence, and determine the predicted load at a future time point based on the corresponding iterative dilation and convolution processing results.

[0209] In some optional implementations, the load prediction model includes: an embedding mapping layer, an output layer, and a plurality of sequentially arranged iterative dilated convolution structures; the iterative dilated convolution structure includes a plurality of sequentially arranged dilated convolution layers;

[0210] The embedding mapping layer is used to perform embedding mapping processing on the effective load sequence to generate a corresponding load feature vector sequence;

[0211] The iterative dilated convolution structure is used to perform iterative dilated convolution processing on the input feature vector sequence, and output the feature vector sequence after the iterative dilated convolution processing; the feature vector sequence input to the first iterative dilated convolution structure is the load feature vector sequence;

[0212] The output layer is used to perform linear processing on the input feature vector sequence to generate a prediction result corresponding to the effective load sequence; the feature vector sequence input to the output layer is a superposition of the effective load sequence and the feature vector sequence output by the last iterative dilated convolution structure;

[0213] The dilated convolution layer is used to perform dilated convolution processing on the input feature vector sequence and output the feature vector sequence after the dilated convolution processing; and the jth feature vector in the feature vector sequence output by the i-th dilated convolution layer is obtained by convolving the jth feature vector, the j-2i-1th feature vector, and the j-2ith feature vector in the feature vector sequence input to the i-th dilated convolution layer, and the jth feature vector in the feature vector sequence output by the i-th dilated convolution layer is It is expressed as the following formula (7):

[0214] (7)

[0215] Where:

[0216] represents the convolution kernel of the i-th dilated convolutional layer,

[0217] represents the convolution operation,

[0218] , , They represent the feature vector sequence input to the i-th dilated convolutional layer, the j-th feature vector, the The feature vector, feature vectors.

[0219] In some optional implementations, the device further includes a judgment module, configured to:

[0220] After the preprocessing module 703 determines the power generation power of the standby second power supply device, the maximum load corresponding to the power generation power of the second power supply device is determined, and the larger value of the current load and the predicted load is used as a reference load;

[0221] Determine whether the difference between the maximum load and the reference load is greater than a preset threshold; if the difference between the maximum load and the reference load is greater than the preset threshold, determine that the generated power of the second power supply device meets the predicted load.

[0222] In some optional implementations, the process of setting the preset threshold includes:

[0223] Constructing multiple first fuzzy sets corresponding to the power generation of the second power supply device, and setting corresponding first membership functions; constructing multiple second fuzzy sets corresponding to the load of the all-electric energy station, and setting corresponding second membership functions; constructing multiple third fuzzy sets corresponding to the preset threshold, and setting corresponding third membership functions;

[0224] Determine the corresponding relationship between the first fuzzy set, the second fuzzy set and the corresponding third fuzzy set, and construct a fuzzy control rule;

[0225] According to the first membership function of the first fuzzy set, it is determined that the power generation of the second power supply device belongs to the first membership of the first fuzzy set, and the first fuzzy set corresponding to the non-zero first membership is used as the first valid fuzzy set; according to the second membership function of the second fuzzy set, it is determined that the reference load belongs to the second membership of the second fuzzy set, and the second fuzzy set corresponding to the non-zero second membership is used as the second valid fuzzy set;

[0226] Determine, according to the fuzzy control rule, a third valid fuzzy set among the plurality of third fuzzy sets, which has a corresponding relationship with the first valid fuzzy set and the second valid fuzzy set;

[0227] Determining a third membership degree corresponding to the third valid fuzzy set according to the first membership degree of the first valid fuzzy set and the second membership degree of the second valid fuzzy set; the third membership degree corresponding to the third valid fuzzy set is between the first membership degree of the first valid fuzzy set and the second membership degree of the second valid fuzzy set;

[0228] In the third membership function of the third valid fuzzy set, the region whose membership is less than the third membership is taken as a valid region, and all the valid regions of the third valid fuzzy sets are superimposed to form a superimposed region;

[0229] The threshold corresponding to the center of gravity of the superimposed area is used as the preset threshold.

[0230] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0231] An embodiment of the present invention also provides a computer device. Fig. 9 is a schematic diagram of the structure of a computer device provided by an optional embodiment of the present invention, such as Fig. 9 As shown, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process instructions executed in the computer device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Fig. 9 A processor 10 is taken as an example.

[0232] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.

[0233] The memory 20 stores instructions executable by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0234] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0235] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.

[0236] The computer device further comprises a communication interface 30 for the computer device to communicate with other devices or a communication network.

[0237] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0238] A part of the present invention may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present invention through the operation of the computer. Those skilled in the art should understand that the existence of the computer program instruction in a computer-readable medium includes, but is not limited to, a source file, an executable file, an installation package file, etc., and accordingly, the way in which the computer program instruction is executed by the computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium may be any available computer-readable storage medium or communication medium accessible to the computer.

[0239] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A power supply switching method for an all-electric energy station, characterized in that: The method comprises the following steps: Obtain the historical load of the all-electric energy station, the current load at the current time point, and the current power generation power of the first power supply device; the first power supply device is the device currently supplying power to the all-electric energy station; The method comprises: performing load prediction according to the historical load of the all-electric energy station and the current load to determine the predicted load at a future time point, including: determining the load change cycle of the all-electric energy station, taking the load change cycle as a unit, generating a plurality of load sequences according to the historical load of the all-electric energy station and the current load, wherein the load sequence comprises a plurality of loads sorted in time, setting a weight coefficient for each of the load sequences, the sum of all the weight coefficients being 1, performing weighted processing on the plurality of load sequences according to the weight coefficients, generating an effective load sequence of the load change cycle, inputting the effective load sequence into a pre-trained load prediction model, and determining the predicted load at a future time point according to the prediction result of the load prediction model; the i-th load in the effective load sequence is represented by the following formula (1): Where: is the i-th payload in the effective payload sequence, is the i-th load in the k-th load sequence, , The number of loads included for the load sequence; For the The weight coefficient of the load sequence, , expressed as the following formula (2): Where: is the number of payload sequences; Perform power prediction based on the current generated power to determine the predicted generated power of the first power supply device at the future time point; In the case where the predicted power generation cannot meet the predicted load, determining the power generation power of the second standby power supply device; When the power generation power of the second power supply device meets the predicted load, the first power supply device is switched to the second power supply device, and the second power supply device supplies power to the all-electric energy station.

2. The power supply switching method of the all-electric energy station according to claim 1, characterized in that: The load prediction model is used to perform iterative dilation and convolution processing on the effective load sequence, and determine the predicted load at a future time point according to the corresponding iterative dilation and convolution processing result.

3. The power supply switching method of the all-electric energy station according to claim 2, characterized in that: The load prediction model comprises: an embedding mapping layer, an output layer and a plurality of sequentially arranged iterative dilated convolution structures; the iterative dilated convolution structure comprises a plurality of sequentially arranged dilated convolution layers; The embedding mapping layer is used to perform embedding mapping processing on the effective load sequence to generate a corresponding load feature vector sequence; The iterative dilated convolution structure is used to perform iterative dilated convolution processing on the input feature vector sequence, and output the feature vector sequence after the iterative dilated convolution processing; the feature vector sequence input to the first iterative dilated convolution structure is the load feature vector sequence; The output layer is used to perform linear processing on the input feature vector sequence to generate a prediction result corresponding to the effective load sequence; the feature vector sequence input to the output layer is a superposition of the effective load sequence and the feature vector sequence output by the last iterative dilated convolution structure; The dilated convolution layer is used to perform dilated convolution processing on the input feature vector sequence and output the feature vector sequence after the dilated convolution processing; and the jth feature vector in the feature vector sequence output by the i-th dilated convolution layer is obtained by performing dilated convolution processing on the jth feature vector, the jth feature vector and the jth feature vector in the feature vector sequence input to the i-th dilated convolution layer. The feature vector, The jth feature vector in the feature vector sequence output by the i-th dilated convolution layer is obtained by convolution of feature vectors. It is expressed as the following formula (3): Where: represents the convolution kernel of the i-th dilated convolutional layer, represents the convolution operation, , , They represent the feature vector sequence input to the i-th dilated convolutional layer, the j-th feature vector, the The feature vector, feature vectors.

4. The power supply switching method of the all-electric energy station according to claim 1, characterized in that: After determining the power generation power of the second standby power supply device, the method further includes: Determine a maximum load corresponding to the power generation power of the second power supply device, and use the larger value of the current load and the predicted load as a reference load; Determine whether the difference between the maximum load and the reference load is greater than a preset threshold; if the difference between the maximum load and the reference load is greater than the preset threshold, determine that the generated power of the second power supply device meets the predicted load.

5. The power supply switching method of the all-electric energy station according to claim 4, characterized in that: The process of setting the preset threshold includes: Constructing multiple first fuzzy sets corresponding to the power generation of the second power supply device, and setting corresponding first membership functions; constructing multiple second fuzzy sets corresponding to the load of the all-electric energy station, and setting corresponding second membership functions; constructing multiple third fuzzy sets corresponding to the preset threshold, and setting corresponding third membership functions; Determine the corresponding relationship between the first fuzzy set, the second fuzzy set and the corresponding third fuzzy set, and construct a fuzzy control rule; According to the first membership function of the first fuzzy set, it is determined that the power generation of the second power supply device belongs to the first membership of the first fuzzy set, and the first fuzzy set corresponding to the non-zero first membership is used as the first valid fuzzy set; according to the second membership function of the second fuzzy set, it is determined that the reference load belongs to the second membership of the second fuzzy set, and the second fuzzy set corresponding to the non-zero second membership is used as the second valid fuzzy set; Determine, according to the fuzzy control rule, a third valid fuzzy set among the plurality of third fuzzy sets, which has a corresponding relationship with the first valid fuzzy set and the second valid fuzzy set; Determining a third membership degree corresponding to the third valid fuzzy set according to the first membership degree of the first valid fuzzy set and the second membership degree of the second valid fuzzy set; the third membership degree corresponding to the third valid fuzzy set is between the first membership degree of the first valid fuzzy set and the second membership degree of the second valid fuzzy set; In the third membership function of the third valid fuzzy set, the region whose membership is less than the third membership is taken as a valid region, and all the valid regions of the third valid fuzzy sets are superimposed to form a superimposed region; The threshold corresponding to the center of gravity of the superimposed area is used as the preset threshold.

6. A power supply switching device for an all-electric energy station, which runs the power supply switching method for an all-electric energy station as claimed in any one of claims 1 to 5, characterized in that: The power supply switching device comprises: An acquisition module, used to acquire the current load of the all-electric energy station at a current time point, and the current power generation power of a first power supply device; the first power supply device is a device currently supplying power to the all-electric energy station; A prediction module is used to perform load prediction based on the historical load of the all-electric energy station and the current load to determine the predicted load at a future time point; perform power prediction based on the current power generation to determine the predicted power generation at the future time point; A preprocessing module, used for determining the power generation power of the standby second power supply device when the predicted power generation power cannot meet the predicted load; A switching module is used to switch from the first power supply device to the second power supply device when the power generation power of the second power supply device meets the predicted load, so that the second power supply device supplies power to the all-electric energy station.

7. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the power supply switching method of the all-electric energy station described in any one of claims 1 to 5 by executing the computer instructions.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the power supply switching method of the all-electric energy station described in any one of claims 1 to 5.

Images