Resource optimization scheduling method and system based on multi-port flexible interconnection

Abstract

The application discloses a resource optimization scheduling method and system based on multi-port flexible interconnection, relates to the technical field of power system engineering, and comprises the following steps: collecting first power distribution data of a first object; encoding the first power distribution data to determine the first object code; obtaining second structure data according to the first object code and constructing a second object; and inputting the first power distribution data into the second object to obtain a prediction result and perform scheduling. The method can be used for predicting the harmonics generated in the power distribution network when the power distribution network is flexibly expanded, and the power resource scheduling scheme of the power distribution network can be adjusted according to the predicted harmonics and the architecture of the power distribution network. In this way, the power distribution network can not only adapt to the access of various new energy, but also realize efficient and flexible power regulation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power system engineering, in particular to a resource optimization scheduling method and system based on multi-port flexible interconnection. BACKGROUND

[0002] When the environment of the distribution network encounters extreme environmental impact in a short time, the new energy equipment that supplies power to the distribution network may be aged or damaged due to changes in environmental factors. For example, the main equipment constituting the distribution network, such as overhead lines, power cables, switches, ring network cabinets, distribution transformers, etc., are often affected by wind, sun, rain, etc. in outdoor environment for a long time, and are prone to aging and have high failure rate.

[0003] Different types of new energy will have multiple impacts on the distribution network when they are connected to the distribution network. Specifically, it may affect power quality, for example, wind power and photovoltaic power generation, which are greatly affected by weather and environment, have the characteristics of intermittency and volatility. This instability will cause voltage fluctuations and affect voltage qualification rate. When new energy is connected to the distribution network, part of the output power on the original network line is supplied by distributed power supply, and the voltage at each node on the branch point line will be enhanced, which may cause the voltage level to be too high. In addition, the power elements and rectifier-inverter equipment provided on the new energy generation equipment will generate direct current and harmonics. Once the harmonics enter the power grid, the voltage of the distribution network will be deformed, which will seriously damage the quality of electric energy in actual application and shorten its service life. At the same time, harmonics will also affect the accuracy of measuring instruments and measuring equipment, and even cause the power system automatic device to operate incorrectly, affecting the normal operation of the power system. In addition, it may also affect the network structure. After the new energy is connected to the distribution network, the existing one-way power supply radial network will be transformed into a network interconnected between multi-electric weak rings and users, which fundamentally changes the distribution form of the power grid. This change makes the power grid structure more complex, and the load direction and problem are more difficult to predict, and the network damage is directly related to the specific location of the load and power supply and the actual reserves of the power supply. SUMMARY

[0004] In view of the above problems, the present application is proposed.

[0005] Therefore, the technical problem solved by the present application is that the existing technology may not be able to effectively execute scheduling instructions such as peak regulation, frequency regulation and load reduction in the actual power resource scheduling process, which may pose a threat to the safe and stable operation of the power grid.

[0006] To solve the above technical problems, the present application provides the following technical scheme: a resource optimization scheduling method based on multi-port flexible interconnection, comprising:

[0007] collecting first distribution data of a first object;

[0008] encoding the first power distribution data to determine a first object code;

[0009] obtaining second structure data according to the first object code, and constructing a second object;

[0010] inputting the first power distribution data into the second object to obtain a prediction result and performing scheduling.

[0011] As a preferred scheme of the resource optimization scheduling method based on multi-port flexible interconnection, the first object first power distribution data is collected at a plurality of sampling time points at a preset sampling frequency.

[0012] As a preferred scheme of the resource optimization scheduling method based on multi-port flexible interconnection, the first power distribution data is encoded, including encoding the first power distribution data at each time point to obtain a first power distribution data attribute code.

[0013] As a preferred scheme of the resource optimization scheduling method based on multi-port flexible interconnection, the first object code is determined, including determining a device code of the first object for the first power distribution data attribute code at each time point.

[0014] As a preferred scheme of the resource optimization scheduling method based on multi-port flexible interconnection, the second structure data is obtained according to the first object code, including obtaining an inter-device attribute between devices of the first object to obtain an inter-device edge weight, and constructing graph data at each time point.

[0015] As a preferred scheme of the resource optimization scheduling method based on multi-port flexible interconnection, the second object is constructed, including training a prediction model according to the graph data at each time point, the prediction model being used to predict operation data of the power distribution network to obtain the second object.

[0016] Preferably, the prediction model is a long short-term memory model, and the model includes a series of reused neurons.

[0017] As a preferred scheme of the resource optimization scheduling method based on multi-port flexible interconnection, the first power distribution data is input into the second object to obtain a prediction result and perform scheduling, including, when formulating a resource scheduling strategy, first collecting the first power distribution data in real time and encoding to obtain real-time encoding, inputting the real-time encoding into the prediction model of the second object trained to a convergence state to obtain predicted operation data, and formulating a scheduling strategy according to the predicted first power distribution data.

[0018] As a preferred scheme of the resource optimization scheduling system based on multi-port flexible interconnection provided by the application, wherein: comprising a data acquisition module, a data encoding module, a data prediction module, a scheduling decision module;

[0019] The data acquisition module is used for collecting power distribution data of the first object; using a distributed sensor and a smart meter connected, the first power distribution condition data is obtained at a preset sampling frequency;

[0020] The data encoding module is used for encoding the collected first power distribution data, generating attribute encoding and determining object encoding, comprising a data processing unit and an object encoding unit;

[0021] The data prediction module is used for constructing second structure data based on the equipment encoding of the first object, training a prediction model, and generating a second object; comprising a graph data construction unit for generating graph data according to the attributes between the devices of the first object and the edge weights between the devices; a prediction model training unit adopts a long short-term memory model trained by graph data;

[0022] The scheduling decision module is used for generating a prediction result based on real-time power distribution data and formulating a resource optimization scheduling strategy.

[0023] A computer device comprising a memory and a processor, the memory stores a computer program, and the processor executes the computer program to realize the steps of the resource optimization scheduling method based on multi-port flexible interconnection.

[0024] A computer readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to realize the steps of the resource optimization scheduling method based on multi-port flexible interconnection.

[0025] The application has the following beneficial effects: the resource optimization scheduling method based on multi-port flexible interconnection provided by the application quantitatively characterizes the power resource configuration of the power distribution network by collecting environmental data of the environment where the power distribution network is located, human data of the area where the power distribution network is located, and operation data in the operation process of the power distribution network. And according to the environmental data, the human data and the operation data, a harmonic prediction model is trained, which is used to predict the harmonics generated in the power distribution network when the power distribution network is flexibly expanded, and the power resource scheduling scheme of the power distribution network is adjusted according to the predicted harmonics and the architecture of the power distribution network. In this way, the power distribution network can not only adapt to the access of various new energy, but also realize efficient and flexible power regulation. And before the topology construction rule and design method are proposed, simulation is carried out to verify the theory and perform simulation test, so as to ensure the effectiveness and reliability in actual application. In addition, the strategy, balance regulation and multi-end coordination control technology need to be finely designed to ensure the stability and efficiency of the system under complex and variable operating conditions. BRIEF DESCRIPTION OF DRAWINGS

[0026] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort on the basis of these drawings.

[0027] Figure 1 A flowchart of a resource optimization scheduling method based on multi-port flexible interconnection provided for the first embodiment of the present application.

[0028] Figure 2 A self-encoding model schematic diagram of a resource optimization scheduling method based on multi-port flexible interconnection provided for the second embodiment of the present application.

[0029] Figure 3 A structure schematic diagram of a neuron of a resource optimization scheduling method based on multi-port flexible interconnection provided for the second embodiment of the present application.

[0030] Figure 4 A prediction model structure diagram of a resource optimization scheduling method based on multi-port flexible interconnection provided for the second embodiment of the present application. DETAILED DESCRIPTION

[0031] In order to make the above objectives, features and advantages of the present application more apparent and understandable, the specific embodiments of the present application will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without any creative effort should fall within the scope of protection of the present application.

[0032] Embodiment 1, refer to Figures 1-2 For an embodiment of the present application, a resource optimization scheduling method based on multi-port flexible interconnection is provided, comprising:

[0033] S1: collecting first power distribution data of a first object.

[0034] Further, the collecting of the first power distribution data of the first object comprises presetting a sampling frequency and collecting at multiple sampling time points.

[0035] In the embodiments of the present application, the first object can be a new energy power supply device (for example, a wind turbine, a photovoltaic generator), a line of a power distribution network, an overhead line, a power cable, a switch, a ring network cabinet, a distribution transformer, etc.

[0036] Further, environmental factors can affect the stability of the power distribution network operation. Specifically, changes in environmental factors can cause changes in the output power of new energy power supply equipment (e.g., wind turbines, photovoltaic generators). For example, when the environment in which the power distribution network is located encounters wind, rain, and lightning, the output power of the wind turbine can increase, and the output power of the photovoltaic generator can decrease. And the line of the power distribution network is easily affected by natural factors such as wind, rain, and lightning. For example, lightning can cause power distribution lines to catch fire, power poles to collapse, and other problems. Strong winds can also negatively affect the normal operation of power distribution lines, and rainy and snowy weather can also pose a threat to power supply stability.

[0037] In the embodiments of the present application, collecting the first power distribution condition data of the first object can be collecting environmental data, human data, and power distribution network operation data at multiple sampling times according to a preset sampling frequency.

[0038] The environmental data includes, but is not limited to, temperature, humidity, wind power, and the number of thunderstorms. The human data at least includes population density. The power distribution network operation data includes the device type of the plurality of new energy power supply equipment in the power distribution network, the output power of the new energy power supply equipment, the output current of the new energy power supply equipment, the output voltage of the new energy power supply equipment, the supply area of the new energy power supply equipment, the supplied population of the new energy power supply equipment, and the power supply data sequence of the new energy power supply equipment.

[0039] In an optional embodiment, collecting power distribution data can also be a data collection enhancement method that uses unmanned aerial vehicle inspection technology to collect environmental and equipment data in the power distribution network area in real time that cannot be covered by fixed sensing equipment; and uses mobile sensing terminals (such as intelligent vehicles or handheld devices) to temporarily collect power distribution data and environmental conditions in a specific area. Through this multi-source, multi-level collection method, comprehensive and accurate data support can be provided for the system, and high-quality data foundation can be provided for subsequent coding, modeling, prediction, and scheduling links. This data collection method is particularly suitable for complex power distribution network systems with multi-port flexible interconnection.

[0040] Further, with economic development, the population density in some areas is increasing year by year, and the electricity load is also increasing year by year. If the construction of the power distribution network lags behind the demand for electricity in economic development, problems such as overloading of the power distribution line and insufficient capacity of the distribution transformer may occur, and in severe cases, low voltage may even occur. If the network structure of the power distribution network is unreasonable, such as excessively long lines without necessary branches, it can cause excessive loss of the power distribution line, reduce the output voltage at the end of the line, and affect the power supply voltage and power supply quality. If the diameter of the first end of the line is too small, it can also affect the output of the electric energy, and in severe cases, it can cause the power distribution line to fuse. In the actual power resource scheduling process, there may be a situation that the scheduling instructions for peak regulation, frequency regulation, and load reduction are not effectively executed, which can threaten the safe and stable operation of the power grid.

[0041] It should be noted that the environmental data, the cultural data and the power distribution network operation data can be collected at multiple sampling time points according to a preset sampling frequency.

[0042] S2: encode the first power distribution data to determine the first object code.

[0043] Further, the encoding of the first power distribution data includes encoding the first power distribution data at each time point to obtain a first power distribution data attribute code. The determination of the first object code includes determining the device code of the first object for the first power distribution data attribute code at each time point.

[0044] In the embodiment of the application, the encoding of the first power distribution data can be encoding the environmental data, the cultural data and the power distribution network operation data at each time point respectively to obtain environmental encoding, cultural encoding and power distribution network attribute encoding.

[0045] Specifically, the environmental data, the cultural data and the power distribution network operation data at each time point can be encoded by using a pre-trained auto-encoding model. The schematic diagram of the auto-encoding model is shown as follows. Figure 2

[0046] In an optional embodiment, the encoding of the power distribution data can also be a statistical feature analysis of the original data at each time point by a statistical model-based aggregation encoding method, and the original data is encoded into a set of statistical features. The specific steps are as follows: calculating the statistical properties (such as mean, maximum, variance) of the environmental data; performing cluster analysis on the cultural data to generate category encoding; extracting distribution characteristics (such as frequency distribution, distribution peak, etc.) of the power distribution network operation data, and generating attribute encoding based on these characteristics.

[0047] Embodiment 2, refer to Figure 1 , Figure 3 , Figure 4 In an embodiment of the application, a resource optimization scheduling method based on multi-port flexible interconnection is provided, comprising:

[0048] S3: obtaining second structure data according to the first object code, and constructing a second object.

[0049] Further, the second structure data obtained according to the first object code includes obtaining the inter-device edge weight according to the inter-device attribute of the first object, and constructing the graph data at each time point.

[0050] ​In the embodiment of the present application, the second structure data obtained according to the first object coding can be that, for each time, the device coding of the plurality of new energy power supply devices in the power distribution network is determined according to the environment coding, the cultural coding and the attribute coding, the edge weight between the devices is determined according to the line attribute between the devices, and the graph data of each time is constructed according to the device coding and the edge weight.

[0051] Further, for each new energy power supply device, the attribute information and the power supply data sequence are spliced in a preset order to obtain the graph data corresponding to each new energy power supply device. For example, the data corresponding to each node in the graph data includes [substation type, output power, output current, output voltage, supplied area, supplied population, power supply data sequence], and the edge weight between different nodes is used to represent the association relationship between different new energy devices, including [line length, line rated power, line age].

[0052] In the embodiment of the present application, the constructing the second object can be that, a prediction model is trained according to the graph data of each time, the prediction model is used to predict the operation data of the power distribution network, and the second object is obtained;

[0053] Preferably, the prediction model is a long short-term memory model, which includes a series of reused neurons, and the structure of the neuron is as follows Figure 3 The model structure is as shown in the following figure Figure 4 Wherein:

[0054] Figure 3 In the formula, x t is the input graph data at time t;

[0055] h t represents the prediction result, i.e. the label, which is obtained jointly by C t and o t , and the formula is as shown in the following figure;

[0056] f t is the forgetting data, and the value range is 0 to 1. The way to obtain f t includes: based on the sigmoid function, x t and h t-1 (the prediction result at time t-1) are calculated to obtain a value between 0 and 1, and the forgetting data f t is closer to 0, which means that more predicted labels predicted by the first sample data at the previous time need to be forgotten, and the forgetting data is closer to 1, which means that less predicted labels predicted by the first sample data at the previous time need to be forgotten. Specifically, W f and b f are the weight and bias corresponding to the forgetting data f t .

[0057] it is input data (the name is input data), obtain i t The way includes: based on sigmoid function, x t And h t-1 (t-1 moment prediction result) are calculated to obtain a value between 0 and 1; W i And b i It is the weight and bias corresponding to the input data.

[0058] Is the initial state value of the neuron at t moment; W c And b c It is the weight and bias corresponding to the initial state value.

[0059] O t Is the hidden output data at t moment. The method for obtaining o t It includes: based on sigmoid function, x t And h t-1 (t-1 moment prediction result) are calculated to obtain a value between 0 and 1; W o And b o It is the weight and bias corresponding to the hidden output data.

[0060] Ct represents the state value of the neuron at t moment.

[0061] In Figure 4 Figure 3 Figure 4 , although multiple neurons are shown, in fact, there is only one neuron in real existence, which is recycled. For example, when training the temperature prediction model, the basic data at t moment and the basic data at t+1 moment can be collected, the initial state value and the initial prediction result value are randomly set in advance, the neuron corresponding to t moment is input, the prediction result at t moment and the output state data are obtained; At t+1 moment, the input of the neuron becomes the prediction result at t moment, the state data output at t moment and the basic data at t+1 moment, and the cycle continues until the training is completed. According to the cycle life, discharge depth and temperature in the prediction result at t moment and the basic data at t moment, the loss value is determined; according to the cycle life, discharge depth and temperature in the prediction result at t+1 moment and the basic data at t+1 moment, the loss value at t+1 moment is determined

[0062] In an optional embodiment, the construction of the second object can also be the implementation steps of the pre-training model based on self-supervised learning as follows:

[0063] Step 1, for large-scale unlabeled power grid map data, design self-supervised learning tasks, such as node prediction task (predict node features) or edge prediction task (predict the relationship between nodes), learn high-quality feature representation of nodes and edges in the graph through these tasks.

[0064] Step 2, using a graph neural network (such as GCN, GraphSAGE) to encode the graph data, extract the structural features and attribute features of the graph, and convert them into low-dimensional embedding vectors to construct the feature embedding representation.

[0065] Step 3, input these pre-trained feature embeddings as initialization parameters into the LSTM model, so that the LSTM can obtain higher initial prediction ability from the structural information of the graph data.

[0066] Step 4, further optimize the pre-training initialized LSTM model, use the labeled data of the power distribution network for supervised training to improve the model performance, complete the construction of the second object, which is used for high-precision prediction of the operation data of the power distribution network, thereby significantly improving the prediction accuracy and generalization ability.

[0067] It should be noted that the prediction model is preferably a long short-term memory model, and the model includes a series of reused neurons.

[0068] S4: inputting the first power distribution data into the second object to obtain a prediction result and performing scheduling.

[0069] In the embodiments of the present application, the step of inputting the first power distribution data into the second object to obtain a prediction result and performing scheduling can be, when formulating a resource scheduling strategy, first collecting the first power distribution data in real time and encoding to obtain real-time encoding, inputting the real-time encoding into the prediction model of the second object trained to a convergent state, obtaining predicted operation data, and formulating a scheduling strategy according to the predicted first power distribution data.

[0070] The predicted operation data is used to represent the harmonics, output current, and output voltage generated by different new energy power supply equipment when running, which are predicted from the currently collected environmental data, cultural data, and power distribution network operation data.

[0071] Furthermore, a scheduling strategy is formulated according to the predicted operation data.

[0072] Specifically, a harmonic elimination strategy is formulated according to the predicted harmonic data. The output power of different new energy power supply equipment is adjusted according to the predicted output current and output voltage, so as to ensure that the adjustment of the output current and output voltage is as small as possible, and the adjusted output current and output voltage meet the load demand of the power distribution network.

[0073] In an optional embodiment, the input prediction model obtains the prediction result and scheduling can also be that the first power distribution data collected in real time is inputted into the prediction model of the second object trained to the convergent state after coding, and the prediction result is corrected in combination with the historical operation data to generate more accurate predicted operation data; the operation parameters of the new energy power supply equipment are dynamically adjusted according to the predicted operation data and the corrected result, including the working mode switching, output power distribution and operation priority setting of the equipment.

[0074] Specifically, the harmonic management strategy is implemented according to the predicted harmonic data, for example, the operation of the active filter device is started or optimized to reduce the harmonic content in the power distribution network; the power distribution strategy between devices is further optimized according to the predicted output current and output voltage, and the load of part of the devices is selectively reduced and the high-efficiency, low-fluctuation devices are preferentially enabled, so as to realize the overall operation efficiency improvement of the power supply equipment, while ensuring that the adjustment range is as small as possible and meets the real-time load demand of the power distribution network.

[0075] It should be noted that when formulating the resource scheduling strategy, the environmental data, cultural data and power distribution network operation data are first collected in real time and coded to obtain real-time coding, which is inputted into the prediction model trained to the convergent state to obtain the predicted operation data.

[0076] Embodiment 3: The following is an embodiment of the present application, which provides a resource optimization scheduling method based on multi-port flexible interconnection. In order to verify the beneficial effects of the present application, economic benefit calculation and simulation experiments are used for scientific demonstration.

[0077] First, weather environment (wind speed, visibility, precipitation, light intensity, light duration) - influence - new energy power - influence - harmonic size human environment (power station radiation range, population density, power consumption duration, peak power consumption) - influence - power distribution network architecture - influence - harmonic propagation

[0078] Demand research: research the current high-density community power distribution status and new energy access situation, and clarify the specific demand of flexible expansion and optimization configuration. Flexible interconnection topology construction rule and design method: through simulation and case analysis, explore the flexible interconnection topology construction rule under the access of new energy, and form the design guide.

[0079] Modulation strategy and control technology research: research modulation strategy, balance adjustment and multi-end coordinated control technology to ensure the stable operation of equipment under the access of multiple new energies. Implementation of multi-end coordinated control technology: in the complex system of multiple new energy access and multiple port interaction, how to realize the efficient and stable multi-end coordinated control technology to ensure the overall performance and safety of the system is one of the difficulties of the research.

[0080] Effectiveness of field test and feedback optimization: The field test environment is complex, and data collection and analysis are difficult. How to effectively use field data feedback to optimize the system and ensure the practical application value of research results is another difficulty that needs to be overcome.

[0081] During the experiment, environmental data (wind speed, visibility, precipitation, light intensity), human data (electricity usage time, peak electricity power), and power distribution network operation data (new energy equipment output power, voltage, current) were collected according to the set sampling frequency.

[0082] Test scenario selection: A high-density community with a total area of 2 square kilometers and a population of about 15,000 people was selected. Distributed photovoltaic systems and energy storage devices were installed in the community, with a total power of 10 MW and 2 MW, respectively, connected to multiple ports of the power distribution network. Real-time monitoring equipment such as wind speed sensors, light intensity meters, and current sensors were equipped with a sampling frequency of 10 minutes per time. Environmental factors (wind speed, light intensity, etc.) and human factors (population density, peak electricity usage, etc.) were set as input variables, and 7-day multi-port power distribution data were collected.

[0083] Real-time input of newly collected power distribution data into the prediction model to obtain the harmonic, current, and voltage prediction values of the new energy equipment operation; dynamic adjustment of equipment output power and operation mode based on prediction data to ensure output stability and satisfaction of power supply demand; adjustment of specific equipment harmonic characteristics using a tuner based on predicted harmonic data to reduce harmonic propagation.

[0084] Table 1 Experimental data table

[0085]

[0086] Data analysis and result comparison: Through the above experimental data, it can be found that the optimization strategy based on the method of the present invention effectively improves the operation stability and efficiency of the power distribution system:

[0087] Significant harmonic suppression effect: After optimization (power distribution ports E and F), the total harmonic distortion (THD) is reduced to 2.5% and 2.8%, respectively, which is about 40% lower than the unoptimized state (average THD is 4.37%). This indicates that the prediction model has high accuracy and practicality in harmonic characteristic analysis and modulation strategy development.

[0088] After the optimization strategy is implemented, the adjustment range of the equipment output power of each power distribution port is significantly reduced. For example, the output power of the power distribution port E is stabilized at 520 kW, and the prediction power error is reduced to ±2.5%. This balanced adjustment greatly improves the operating life of the equipment. In the experiment, the output power of the new energy equipment is dynamically adjusted to meet the peak power demand of the power distribution network while reducing the risk of overload. For example, after optimization, the output power of the power distribution port F is stabilized from 400 kW to 450 kW, and the output current and voltage fluctuations are significantly reduced.

[0089] In one embodiment of the present application, a resource optimization scheduling system based on multi-port flexible interconnection is provided, including a data acquisition module, a data encoding module, a data prediction module, and a scheduling decision module.

[0090] The data acquisition module is used to acquire power distribution data of a first object. Distributed sensors and smart meters are used to obtain first power distribution data at a preset sampling frequency.

[0091] The data encoding module is used to encode the collected first power distribution data to generate attribute encoding and determine object encoding, including a data processing unit and an object encoding unit.

[0092] The data prediction module is used to construct second structure data based on the equipment encoding of the first object, train a prediction model, and generate a second object. The graph data construction unit generates graph data based on the attributes and edge weights between the devices of the first object. The prediction model training unit trains a long short-term memory model using the graph data.

[0093] The scheduling decision module is used to generate prediction results based on real-time power distribution data and develop resource optimization scheduling strategies.

[0094] If the functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the present application or the parts that essentially contribute to the prior art or parts of the technical solutions can be embodied in the form of a software product, which is stored in a storage medium and includes instructions to make a computer device (which can be a personal computer, server, or network device) execute all or part of the steps of the embodiments of the present application. The aforementioned storage medium includes U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disk or optical disk, and various storage medium that can store program codes.

[0095] The logic and / or steps represented in the flow diagrams or otherwise described herein, for example, can be considered as a sequence of executable instructions, and can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. In the context of this specification, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. The computer-readable medium can be, for example but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, device, or propagation medium.

[0096] More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electrical connection (electronic) having one or more wires, a portable computer diskette (magnetic), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can also be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, for example via an optical scanner, then compiled, interpreted, or otherwise processed, and stored in a computer memory in a form that can be later executed by a computer. In some embodiments, the machine-readable medium can be a computer- readable storage medium.

[0097] It should be understood that aspects of the application can be implemented in hardware, software, firmware or combinations thereof. In the above embodiments, various steps or methods can be implemented in software or firmware that is stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, and in another embodiment, any of the following can be used: discrete logic circuitry having logic gates for implementing logic functions upon data signals, application specific integrated circuits having appropriate combinational logic gates, programmable gate arrays (PGA), field programmable gate arrays (FPGA), and so forth. It should be noted that the foregoing embodiments are merely examples of implementations of the technical solutions of the present application and are not limiting. Although the present application has been described in detail with reference to the preferred embodiments, it should be understood that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present application, and all such modifications or replacements should be encompassed in the scope of the claims of the present application.

[0098] It should be noted that the above examples are only used to illustrate the technical solutions of the present application but not limit the present application. Although the present application is described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or equivalently replaced, without departing from the spirit and scope of the technical solutions of the present application, which should be covered in the scope of the claims of the present application.

Claims

1. A resource optimization scheduling method based on multi-port flexible interconnection, characterized in that, include: Collect the first power distribution data of the first object; Encode the first power distribution data to determine the first object code; The second structure data is obtained based on the encoding of the first object, and the second object is constructed. The first power distribution data is input into the second object to obtain the prediction result and then dispatched. The step of obtaining the second structure data based on the encoding of the first object includes obtaining the inter-device edge weights based on the inter-device attributes of the first object and constructing graph data at each time step. Constructing a second object, implemented using a pre-trained model based on self-supervised learning, includes: For large-scale unlabeled distribution network map data, self-supervised learning tasks are designed, including node prediction tasks or edge prediction tasks. Learn high-quality feature representations of nodes and edges in a graph through node prediction or edge prediction tasks; Graph neural networks are used to encode graph data, extract structural and attribute features of the graph, and transform the structural and attribute features into low-dimensional embedding vectors to construct feature embedding representations. The pre-trained feature embeddings are used as initialization parameters input into the LSTM model, enabling the LSTM to obtain initial prediction capabilities from the structural information of the graph data. The pre-trained and initialized LSTM model is further optimized by using labeled data from the power distribution network for supervised training to improve the performance of the LSTM model and complete the construction of the second object.

2. The resource optimization scheduling method based on multi-port flexible interconnection as described in claim 1, characterized in that: The first power distribution status data of the first object to be collected includes a preset sampling frequency and collection at multiple sampling times.

3. The resource optimization scheduling method based on multi-port flexible interconnection as described in claim 2, characterized in that: The encoding of the first power distribution data includes encoding the first power distribution data at each time moment to obtain the first power distribution data attribute code.

4. The resource optimization scheduling method based on multi-port flexible interconnection as described in claim 3, characterized in that: The determination of the first object code includes determining the device code of the first object for the first power distribution data attribute code at each time moment.

5. The resource optimization scheduling method based on multi-port flexible interconnection as described in claim 4, characterized in that: The construction of the second object includes training a prediction model based on the graph data at each time point. The prediction model is used to predict the operating data of the distribution network to obtain the second object. The prediction model is a long short-term memory model, which includes cascaded, reusable neurons.

6. The resource optimization scheduling method based on multi-port flexible interconnection as described in claim 5, characterized in that: The step of inputting the first power distribution data into the second object to obtain the prediction result and perform scheduling includes, when formulating a resource scheduling strategy, firstly collecting the first power distribution data in real time and encoding it to obtain real-time encoding, inputting the real-time encoding into the prediction model of the second object trained to the convergence state to obtain prediction operation data, and formulating a scheduling strategy based on the predicted first power distribution data.

7. A system employing the resource optimization scheduling method based on multi-port flexible interconnection as described in any one of claims 1 to 6, characterized in that: It includes a data acquisition module, a data encoding module, a data prediction module, and a scheduling decision module; The data acquisition module is used to collect power distribution data of the first object; it uses a connection between distributed sensors and smart meters to acquire the first power distribution status data at a preset sampling frequency. The data encoding module is used to encode the collected first power distribution data, generate attribute codes, and determine object codes, including a data processing unit and an object encoding unit; The data prediction module is used to construct second structure data based on the device code of the first object, train the prediction model, and generate the second object; it includes a graph data construction unit that generates graph data based on the device attributes and device edge weights of the first object; and a prediction model training unit that trains a long short-term memory model using the graph data. The scheduling decision module is used to generate prediction results and formulate resource optimization scheduling strategies based on real-time power distribution data.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the resource optimization scheduling method based on multi-port flexible interconnection as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the resource optimization scheduling method based on multi-port flexible interconnection as described in any one of claims 1 to 6.

Images