A supercomputing charging pile system

Abstract

The supercomputing charging pile system provided by the embodiment of the present invention includes: a power distribution main station device, a fusion terminal device and a plurality of charging piles; each charging pile reports work information and computing power information to the power distribution main station through the fusion terminal device; the power distribution main station device obtains the flow calculation task issued by the supercomputing terminal, and at the same time, according to the work information and computing power information, decomposes the flow calculation task into a plurality of subtasks, and determines the work information corresponding to each subtask; wherein the subtask corresponds to the charging pile one by one; each subtask and the work information corresponding to each subtask are issued to each charging pile; the charging pile obtains the processing result according to each subtask and the work information corresponding to each subtask, and reports the processing result to the supercomputing terminal through the fusion terminal device and the power distribution main station device in sequence. The present invention can effectively alleviate the computing pressure of the supercomputing center by utilizing the distributed computing power of the charging piles and handing over a part of the flow tasks of the supercomputing to the distributed computing power.

Description

Technical Field

[0001] The present invention belongs to the technical field of network computing power, and in particular relates to a supercomputing charging pile system. Background Art

[0002] With the rapid growth of electric vehicles and the large-scale access of charging piles to the power grid, the complexity of power flow calculation has increased dramatically. Existing technologies mainly rely on centralized computing resources for power flow calculation, usually processing related tasks on dedicated servers in large data centers or power grid dispatch centers.

[0003] With the rapid increase in the number of electric vehicles and the large-scale access of charging piles to the power grid, the complexity of power grid power flow calculation has increased significantly. Now it relies more on centralized computing resources such as large data centers or dedicated servers in power grid dispatching centers to handle power flow calculation tasks. The current centralized cluster deployment architecture of supercomputing centers leads to a large body and a sharp increase in energy consumption in the pursuit of computing power, which greatly limits its application scenarios. After the large-scale application of V2G (Vehicle-to-Grid), on the one hand, a large number of electric vehicles transmit electricity to the power grid through V2G technology, which will make the power flow state of the power grid more complex and changeable. The centralized computing mode of traditional supercomputing centers will face greater challenges, and the computing pressure will be further increased. In addition, its high energy consumption and limited application scenarios will be more prominent when dealing with the complex power grid power flow calculation brought by V2G. Summary of the invention

[0004] In view of this, the present invention provides a supercomputing charging pile system, which aims to solve the problem in the prior art that the power flow state of the power grid has become more complex and changeable, and the centralized computing mode of the traditional supercomputing center will face greater challenges.

[0005] A first aspect of the present invention provides a supercomputing charging pile system, comprising: a power distribution master station device, a fusion terminal device and a plurality of charging piles; the power distribution master station device is connected to the supercomputing terminal; the power distribution master station device is connected to at least one fusion terminal device; each fusion terminal device is connected to a plurality of charging piles;

[0006] Each charging pile is used to report working information and computing power information to the power distribution master station through the fusion terminal device;

[0007] The power distribution master station device is used to obtain the power flow calculation task issued by the supercomputing terminal, and at the same time, according to the work information and computing power information, decompose the power flow calculation task into multiple subtasks, and determine the work information corresponding to each subtask; wherein the subtask corresponds to the charging pile one by one; each subtask and the work information corresponding to each subtask are issued to each charging pile;

[0008] The charging pile is also used to obtain processing results based on each subtask and the work information corresponding to each subtask, and report the processing results to the supercomputing terminal through the fusion terminal device and the distribution master station device in turn.

[0009] In a possible implementation, the power distribution master station device is used to: divide the power flow calculation task into a plurality of minimum calculation tasks;

[0010] Input each minimum computing task, computing power information, and work information into the neural network to obtain the allocation probability of each minimum computing task to each charging pile, and form an allocation probability vector of each minimum computing task; wherein the allocation probability is determined according to the computing amount of each minimum computing task, the remaining computing power of each charging pile, and the first association relationship between each minimum computing task and the work information;

[0011] According to the allocation probability vectors of each minimum computing task, a decomposition scheme is determined to decompose the power flow computing task into multiple subtasks.

[0012] In a possible implementation, the power distribution master station device is used to:

[0013] According to the type of electrical quantity, node type, node location and time step, the power flow calculation task is divided into multiple minimum calculation tasks.

[0014] In a possible implementation, the power distribution master station device is used to:

[0015] Select one allocation probability from each allocation probability vector;

[0016] Calculating the allocation evaluation value according to the second association relationship between each selected allocation probability and each minimum computing task;

[0017] Select the plan with the highest allocation evaluation value and determine the decomposition plan.

[0018] In a possible implementation, the power distribution master station device is used to:

[0019] If, according to the selected allocation probability, each charging pile is assigned at most one subtask, then the selected allocation probabilities are added together to obtain the allocation evaluation value;

[0020] If there are at least two subtasks assigned to a charging pile according to the selected allocation probability, the selected allocation probabilities corresponding to the subtasks on the charging pile are multiplied and then added to other selected allocation probabilities to obtain an allocation evaluation value.

[0021] In a possible implementation, the power distribution master station device is used to:

[0022] According to the electrical quantity type, node type, node position and time step corresponding to each subtask, the work information corresponding to each subtask is determined.

[0023] In a possible implementation, the charging pile is also used for:

[0024] Subtract the reserved computing power from the rated computing power of the charging pile to get the remaining computing power of the charging pile.

[0025] In a possible implementation, the charging pile is also used for:

[0026] In response to a computing power supply request of the electric vehicle, obtaining first computing power information from an onboard terminal of the electric vehicle;

[0027] Subtract the reserved computing power from the rated computing power of the charging pile to get the available computing power of the charging pile;

[0028] The remaining computing power of the charging pile is determined according to the available computing power and the first computing power information.

[0029] In a possible implementation, the charging pile is also used for:

[0030] Obtain computing power usage information of each vehicle terminal;

[0031] Calculate the amount of electricity exchange based on computing power usage information;

[0032] Report the power exchange amount to the power distribution master station device.

[0033] In a possible implementation, the power distribution master station device is used to:

[0034] Obtain the first computing power information;

[0035] Calculate the uncertainty index of vehicle charging;

[0036] The reserved computing power of the power distribution master station device is determined according to the first computing power information and the uncertainty index.

[0037] The fourth aspect of the present invention provides a computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the supercomputing charging pile system of the first aspect above are implemented.

[0038] The supercomputing charging pile system provided by the embodiment of the present invention includes: a power distribution main station device, a fusion terminal device and a plurality of charging piles; the power distribution main station device is connected to the supercomputing terminal; the power distribution main station device is connected to at least one fusion terminal device; each fusion terminal device is connected to a plurality of charging piles; each charging pile is used to report work information and computing power information to the power distribution main station through the fusion terminal device; the power distribution main station device is used to obtain the flow calculation task issued by the supercomputing terminal, and at the same time, according to the work information and computing power information, the flow calculation task is decomposed into a plurality of subtasks, and the work information corresponding to each subtask is determined; wherein the subtask corresponds to the charging pile one by one; each subtask and the work information corresponding to each subtask are issued to each charging pile; the charging pile is also used to obtain the processing result according to each subtask and the work information corresponding to each subtask, and the processing result is reported to the supercomputing terminal through the fusion terminal device and the power distribution main station device in sequence. The present invention can effectively alleviate the computing pressure of the supercomputing center by utilizing the distributed computing power of the charging pile and handing over a part of the flow task of the supercomputing to the distributed computing power. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0040] Figure 1 It is a structural schematic diagram of a supercomputing charging pile system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0041] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present invention. However, it should be clear to those skilled in the art that the present invention may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present invention.

[0042] Figure 1 This is a flow chart of the implementation of the supercomputing charging pile system provided by the embodiment of the present invention. Figure 1 As shown, the supercomputing charging pile system includes: a power distribution master station device 11, a fusion terminal device 12 and a plurality of charging piles 13; the power distribution master station device 11 is connected to the supercomputing terminal; the power distribution master station device 11 is connected to at least one fusion terminal device 12; each fusion terminal device 12 is connected to a plurality of charging piles 13;

[0043] Each charging pile 13 is used to report working information and computing power information to the power distribution master station device 11 through the fusion terminal device 12;

[0044] The power distribution master station device 11 is used to obtain the power flow calculation task issued by the supercomputing terminal, and at the same time, according to the work information and computing power information, decompose the power flow calculation task into multiple subtasks, and determine the work information corresponding to each subtask; wherein the subtask corresponds to the charging pile 13 one by one; each subtask and the work information corresponding to each subtask are issued to each charging pile 13;

[0045] The charging pile 13 is also used to obtain processing results according to each subtask and the work information corresponding to each subtask, and report the processing results to the supercomputing terminal through the fusion terminal device 12 and the power distribution master station device 11 in sequence.

[0046] In the embodiment of the present invention, the power distribution master station device: This device bears important information exchange responsibilities in the entire charging pile system. It is mainly responsible for issuing strategies with the fusion terminal device, that is, conveying relevant charging strategies, power supply strategies, etc. to the fusion terminal device. At the same time, it will also report its own status information to the fusion terminal device so that the entire system can fully grasp its operation status.

[0047] The fusion terminal device plays a key role in connecting the upper and lower parts of the system. On the one hand, it maintains close contact with the power distribution master station device through real-time communication, receives strategies issued by the power distribution master station device, and feeds back its own status information to the power distribution master station device. On the other hand, it will also use the same real-time communication method to perform load distribution operations with itself (there may be duplication in the original text here, and it is speculated that it should be related to the charging pile system and other parts), reasonably distribute the power load according to the actual situation, and ensure the smooth operation of the entire charging system. In addition, it will also report its own status information during the load distribution process so that other parts of the system can understand its working status.

[0048] The charging pile system is a key part of providing charging services directly for electric vehicles, including V2G charging piles.

[0049] The V2G charging pile has a unique internal circuit structure, which consists of a bidirectional AC / DC converter and a bidirectional DC / DC converter.

[0050] Bidirectional AC / DC Converter:

[0051] The bidirectional AC / DC converter plays the role of the front-stage circuit in the V2G charging pile. Its main function is to realize the bidirectional conversion between alternating current (AC) and direct current (DC), and can convert the input AC into DC for subsequent circuit processing; at the same time, under certain circumstances, it can also convert DC into AC to realize the reverse transmission of electric energy, which provides an important circuit foundation for the orderly power supply function of electric vehicles.

[0052] The bidirectional DC / DC converter is the post-circuit of the V2G charging pile. It is mainly responsible for further processing and adjusting the DC power converted by the bidirectional AC / DC converter. It also has the ability of bidirectional conversion and can optimize the voltage, current and other parameters of the DC power according to actual needs to better meet the requirements of electric vehicle charging and power supply to the grid (in V2G mode).

[0053] In an embodiment of the present invention, each charging pile has the ability to report working information and computing power information to the fusion terminal device. The working information covers various real-time status data of the charging pile during the actual operation process, such as whether the vehicle is being charged, the specific power of the charging, the charging time, the battery power of the connected vehicle, etc. This information can intuitively reflect the current impact of the charging pile on the power grid flow. The computing power information mainly involves the computing power-related parameters of the charging pile's own hardware, such as processor performance, memory size, etc. By reporting this information, the distribution master station device can understand how much computing power each charging pile can undertake. For example, a charging pile with a high-performance processor and sufficient memory can relatively undertake more complex and computationally intensive power flow calculation subtasks.

[0054] In an embodiment of the present invention, the charging pile will receive each subtask and corresponding work information issued from the power distribution master station device via the fusion terminal device, and then carry out calculation work based on these contents. For example, a charging pile receives a subtask to calculate the impact of its own charging on the current of the branch where it is located, as well as its own charging power, basic parameters of the branch where it is located, and other work information related to it. It will use its own computing resources to perform calculations according to the corresponding power flow calculation method to obtain the processing results. After that, the charging pile will report this processing result to the supercomputing terminal through the fusion terminal device and the power distribution master station device in turn, so that the supercomputing terminal can summarize the calculation results of each charging pile, and finally complete the analysis of the entire power grid power flow calculation task, so as to achieve accurate grasp and effective management of the power grid power flow situation involving a large number of charging piles.

[0055] In an embodiment of the present invention, the supercomputing terminal generates a power flow calculation task according to the overall operation requirements of the power grid and sends it to the distribution master station device. Based on the work information and computing power information collected from each charging pile, the distribution master station device decomposes the power flow calculation task into multiple subtasks, and determines the work information corresponding to each subtask to ensure that each subtask can be adapted to the corresponding charging pile, and then sends the subtask and the corresponding work information to each charging pile in turn. After receiving the task, each charging pile performs a specific calculation operation, and after obtaining its own processing results, it reports the results layer by layer in the order of the fusion terminal device and the distribution master station device, and finally feeds back to the supercomputing terminal. By integrating these calculation results from different charging piles, the supercomputing terminal can complete the power flow calculation analysis of the power grid area where the entire supercomputing charging pile system is located, thereby providing strong data support for the reasonable scheduling and stable operation of the power grid.

[0056] In some embodiments, the power distribution master station device is used to: divide the power flow calculation task into multiple minimum computing tasks; input each minimum computing task, computing power information, and work information into a neural network to obtain the allocation probability of each minimum computing task to each charging pile, forming an allocation probability vector for each minimum computing task; wherein the allocation probability is determined based on the computing amount of each minimum computing task, the remaining computing power of each charging pile, and the first association relationship between each minimum computing task and the work information; based on the allocation probability vector of each minimum computing task, determine a decomposition plan to decompose the power flow calculation task into multiple subtasks.

[0057] In some embodiments, the power distribution master station device is used to divide the power flow calculation task into multiple minimum calculation tasks according to the type of electrical quantity, node type, node location and time step.

[0058] In an embodiment of the present invention, for the node voltage calculation part in the power flow calculation task, it is divided into units of each node in the power grid. For example, in a power grid with multiple nodes (including charging pile access nodes), the node voltage equation is constructed according to Kirchhoff's current law. The solution for each node voltage can be regarded as a minimum calculation task, and the voltage of each node is calculated separately through matrix operations. The same is true for branch current calculation. According to Ohm's law, the current is calculated for each branch in the power grid. The current calculation of each branch is regarded as a minimum calculation task, and these tasks are independent of each other and can be processed in parallel. In terms of power calculation, the power calculation for each node or branch can be divided into a minimum calculation task, and the corresponding voltage and current values ​​need to be obtained first during the calculation. These tasks are relatively independent of each other, which is convenient for subsequent distribution.

[0059] Taking the power supply range of a substation as an example, all node voltage calculation, branch current calculation, and power calculation tasks in the low-voltage side power supply area of ​​a substation are divided into a group of minimum calculation tasks. For example, for multiple 10kV distribution line areas powered by a 110kV substation, all contents related to power flow calculation in the area, including bus voltage calculation on the low-voltage side of the substation, current and power calculation of the 10kV distribution line, and voltage, current, and power calculation of the charging pile access point in the area, are grouped as a minimum calculation task group.

[0060] When divided by functional areas, such as commercial areas, residential areas, industrial areas, etc., the power flow calculation tasks in each functional area are divided into a minimum calculation task. Taking the commercial area as an example, the power flow calculation of the power grid composed of all charging piles and surrounding power-consuming equipment in the commercial area, including the load calculation of the distribution transformer in the commercial area, the calculation of the impact of the charging power of each store charging pile on the local power grid, and the power flow interaction calculation between other power-consuming equipment such as public lighting in the commercial area and the charging piles, is divided into a minimum calculation task.

[0061] If it is a dynamic power flow calculation, a fixed time step is set, such as every 15 minutes as a time step. Then the power flow calculation corresponding to each time step, including the voltage change calculation of all nodes within this time step, the dynamic adjustment calculation of branch currents, and the dynamic change calculation of power, can be regarded as a minimum calculation task. There is a time series relationship between these minimum calculation tasks, and the calculation result of the previous time step can be used as the initial condition for the calculation of the next time step.

[0062] After the minimum computing tasks are divided, each minimum computing task is used as a subtask corresponding to a certain charging pile, or more than two minimum computing tasks are used as subtasks corresponding to a certain charging pile and are allocated to the charging pile for calculation.

[0063] In the embodiment of the present invention, the characteristics of each minimum computing task are sorted out, including the size of the computing amount (such as the scale of matrix operations in the node voltage computing task, the amount of voltage and impedance data obtained in the branch current computing task, etc.). At the same time, the first association relationship between the minimum computing tasks is clarified, for example, the task of calculating the branch current is closely related to the task of calculating the node voltages at both ends of the branch, and this association relationship is used as part of the input information.

[0064] In an embodiment of the present invention, the computing power information of each charging pile is collected, including the computing speed of the processor (the number of floating-point operations that can be performed per second), the memory size and the remaining available memory, the read and write speed of the storage device, etc. This information can accurately reflect the scale and complexity of the computing tasks that each charging pile can undertake. Summarize the working information of each charging pile, such as the current charging status (charging, idle waiting, charging paused, etc.), charging power, charging time, and the remaining battery power of the connected vehicle. These working information not only reflects the actual impact of the charging pile on the current power grid, but also indirectly reflects its remaining computing resources and the correlation with each part of the current calculation task.

[0065] In an embodiment of the present invention, a suitable neural network architecture is constructed, such as a multi-layer perceptron or a neural network with an attention mechanism. The number of input layer nodes is determined according to the sum of the feature dimensions of the minimum computing task, computing power information, and work information. The intermediate hidden layer performs nonlinear transformation through a suitable activation function to learn the implicit patterns and rules between different input factors. The output layer corresponds to each charging pile and outputs the allocation probability of each minimum computing task being assigned to the charging pile. When calculating the allocation probability, the neural network will comprehensively consider multiple factors, as follows:

[0066] Computational factors: For the minimum computing tasks with large computational load, they tend to be assigned to charging piles with strong computing power (such as fast processor speed and sufficient memory). For example, for a node voltage calculation task that requires a large number of matrix operations, the neural network will assign a higher probability to charging piles with strong computing power based on the processor performance and memory of the charging pile.

[0067] Remaining computing power factor: Real-time evaluation of the remaining computing power of each charging station to avoid task overload. If a charging station is currently performing other tasks and has limited remaining computing power, then the probability of it being assigned to a new minimum computing task will be reduced accordingly. For example, a charging station with less remaining memory space will have a probability of being assigned to a minimum computing task that requires a large amount of memory to store data that is close to zero.

[0068] Task association factor: Consider the association between minimum computing tasks. For closely related minimum computing tasks, try to assign them to the same or adjacent charging piles to facilitate data interaction and collaborative computing during the calculation process. For example, for a set of minimum computing tasks that calculate multiple node voltages and related branch currents in a certain area, the neural network will give priority to assigning them to charging piles that are close to the area and have computing power that meets the requirements.

[0069] Generate allocation probability vector: After the operation of the neural network, for each minimum computing task, a corresponding allocation probability value for each charging pile will be obtained, and these probability values ​​constitute the allocation probability vector of the minimum computing task. For example, for a minimum computing task, the probability of being allocated to charging pile A is 0.3, the probability of being allocated to charging pile B is 0.5, and the probability of being allocated to charging pile C is 0.2, then its allocation probability vector is [0.3, 0.5, 0.2].

[0070] Determine the decomposition scheme according to the allocation probability vector, and decompose the power flow calculation task into multiple subtasks. For each minimum calculation task, determine the assigned charging pile according to the principle of the maximum probability value in its allocation probability vector. For example, for the minimum calculation task mentioned above, the probability value of charging pile B in its allocation probability vector is 0.5, which is the largest, so this minimum calculation task is assigned to charging pile B. At the same time, on the basis of maximizing the probability allocation, the overall task load of each charging pile is comprehensively considered to avoid the phenomenon that some charging piles are overloaded and some charging piles are idle. The ratio of the total calculation amount of the minimum calculation tasks assigned to each charging pile to its total remaining computing power can be calculated, and the ratio of each charging pile can be made as close as possible to achieve relatively balanced utilization of resources. At the same time, for those closely related minimum calculation tasks, it is necessary to ensure that the assigned charging piles are convenient for data interaction and collaborative calculation when allocating them. Integrate multiple minimum calculation tasks assigned to the same charging pile, and these integrated task sets constitute a subtask. For example, charging pile A is assigned three minimum computing tasks, namely, partial node voltage calculation in a certain area, relevant branch current calculation and reactive power calculation in the area. Then these three minimum computing tasks are integrated together to form a subtask assigned to charging pile A.

[0071] In some embodiments, the power distribution master station device is used to: select one allocation probability from each allocation probability vector; calculate the allocation evaluation value based on the second correlation relationship between each selected allocation probability and each minimum computing task; select the plan with the highest allocation evaluation value and determine the decomposition plan.

[0072] In the embodiment of the present invention, one allocation probability is selected from each allocation probability vector.

[0073] For each minimum computing task, there is a corresponding allocation probability vector, which represents the probability of this minimum computing task being allocated to each charging pile. For example, assuming there are three charging piles A, B, and C, for the minimum computing task M1, its allocation probability vector is [P(M1,A)=0.3, P(M1,B)=0.5, P(M1,C)=0.2], and for the minimum computing task M2, its allocation probability vector is [P(M2,A)=0.4, P(M2,B)=0.3, P(M2,C)=0.3], etc. At this time, randomly select an allocation probability from these vectors, such as selecting P(M1,B)=0.5 from the vector of M1, and selecting P(M2,A)=0.4 from the vector of M2, etc., so that a set of preliminary allocation scheme examples are obtained, but this is only one of many possible combinations.

[0074] In an embodiment of the present invention, the second association relationship refers to the association between each minimum computing task and the working information of the charging pile. For example, if a minimum computing task is to calculate the impact of charging of a group of charging piles in a certain area on the surrounding voltage stability, and there are several charging piles working in the area, then this minimum computing task is closely related to the working information of these working charging piles. For another example, calculating the power loss on a distribution line is directly related to the charging power, charging status and other working information of the charging piles connected to the line. By combing and quantifying these associations, a correlation matrix or correlation function representing the degree of association can be obtained.

[0075] By repeating the above process of selecting the allocation probability from the allocation probability vector and calculating the allocation evaluation value multiple times, multiple different allocation schemes and their corresponding allocation evaluation values ​​can be obtained. For example, the evaluation value of the scheme obtained by the first selection is 0.9, the evaluation value of the scheme obtained by the second selection is 0.7, the evaluation value of the scheme obtained by the third selection is 1.1, and so on.

[0076] Compare these allocation evaluation values ​​and select the highest one. For example, if the third choice has the highest evaluation value of 1.1, then the allocation plan corresponding to the third choice is determined as the final decomposition plan. This decomposition plan specifies which charging pile each minimum computing task is specifically assigned to, thereby decomposing the entire power flow calculation task into multiple subtasks, each of which is executed by the corresponding charging pile, so as to achieve the purpose of using distributed charging pile computing resources to complete complex power flow calculation tasks, and fully consider the correlation between the minimum computing task and the charging pile work information in the allocation process, thereby improving the rationality and effectiveness of task allocation.

[0077] In some embodiments, the power distribution master station device is used to: if according to the selected allocation probability, each charging pile is assigned at most one subtask, then the selected allocation probabilities are added to obtain an allocation evaluation value; if according to the selected allocation probability, there are at least two subtasks assigned to the charging pile, then the selected allocation probabilities corresponding to the subtasks on the charging pile are multiplied, and then added with other selected allocation probabilities to obtain an allocation evaluation value.

[0078] In an embodiment of the present invention, first, the power distribution master station device needs to check the task allocation scheme formed according to the currently selected allocation probability to confirm whether the condition that the number of subtasks assigned to each charging pile is at most one is met. In other words, check each charging pile one by one to see if there is a charging pile that is assigned more than one subtask. If there is no such situation, then enter the allocation evaluation value calculation process corresponding to this situation. Under the premise of meeting the above conditions, the power distribution master station device will add the allocation probabilities selected by each minimum computing task to obtain the allocation evaluation value. For example, assuming that there are three minimum computing tasks M1, M2, and M3, the corresponding selected allocation probabilities are P(M1, A) = 0.3 (indicating the probability that the minimum computing task M1 is assigned to charging pile A), P(M2, B) = 0.4, and P(M3, C) = 0.2. At this time, according to the rules, the three selected allocation probabilities are added together, that is, the allocation evaluation value V = P(M1, A) + P(M2, B) + P(M3, C) = 0.3 + 0.4 + 0.2 = 0.9. Through this calculation method, a quantitative evaluation value corresponding to the allocation scheme is obtained, which can be used to compare with other possible allocation schemes in the future to determine which scheme is better.

[0079] Similarly, the distribution master station device first checks the task allocation scheme determined by the currently selected allocation probability, this time to see whether there is a certain (or some) charging pile that is assigned at least two subtasks. If such a charging pile is found to exist, the allocation evaluation value is calculated according to the rules of this situation. For those charging piles that are assigned at least two subtasks, the distribution master station device will multiply the selected allocation probabilities corresponding to each subtask assigned to the charging pile. For example, assuming that two subtasks are assigned to the charging pile D, which come from the minimum computing tasks M4 and M5, respectively, and the corresponding selected allocation probabilities are P(M4, D) = 0.5 and P(M5, D) = 0.6, then first multiply these two probabilities to obtain a comprehensive probability value of the charging pile D, that is, P(M4, D) × P(M5, D) = 0.5 × 0.6 = 0.3. After completing the probability multiplication of all charging piles with at least two subtasks, the results of these multiplications are added together with the selected allocation probabilities corresponding to other charging piles (i.e., charging piles that are only assigned to one subtask at most), and finally the allocation evaluation value of the entire allocation scheme is obtained. For example, in addition to charging pile D, charging pile E is also assigned a subtask, and the corresponding selected allocation probability is P(M6,E)=0.2, then the allocation evaluation value of the allocation scheme is V=P(M4,D)×P(M5,D)+P(M6,E)=0.3+0.2=0.5. That is, the more subtasks are assigned to a charging pile, the smaller its allocation probability is, thereby avoiding allocating too many computing tasks to a charging pile.

[0080] In some embodiments, the power distribution master station device is used to determine the work information corresponding to each subtask according to the electrical quantity type, node type, node location and time step corresponding to each subtask.

[0081] 1. Determine working information based on electrical quantity type

[0082] Voltage related subtasks:

[0083] If the subtask is to calculate the voltage of a specific node or area in the power grid, the corresponding work information will include the voltage-related charging pile work data. For example, the charging power of the charging pile being charged, because the charging power will affect the voltage of the node. High-power charging may cause the node voltage to drop, so the real-time charging power information of the charging pile is crucial for the voltage calculation subtask.

[0084] In addition, the number and status of charging piles connected near the node or area (whether they are idle, charging paused, etc.) will also affect the voltage distribution. For example, when multiple charging piles are connected at the same time and in different working states, the equivalent impedance of the power grid will change, thereby affecting the voltage amplitude. Therefore, the access status information of these charging piles needs to be used as the working information of the voltage subtask.

[0085] Current related subtasks:

[0086] For the subtask of calculating branch current, the work information mainly involves the current data of the charging pile connected to the branch. If the charging pile has a built-in current monitoring function, its real-time current value will be directly used for branch current calculation. For example, in a distribution line with multiple charging piles connected, the output current of each charging pile is the key data for calculating the total current of the branch.

[0087] At the same time, the charging state conversion of the charging pile (such as switching from low-power charging to high-power charging) will cause a sudden change in current. This state change information should also be included in the scope of working information because it will affect the dynamic changes of branch current in different time periods.

[0088] Power related subtasks:

[0089] The subtask of calculating active power and reactive power requires obtaining the power factor information of the charging pile. The power factor may be different for different types of charging piles or at different charging stages. For example, some early models of charging piles have a low power factor and consume more reactive power, which will affect the power balance of the power grid. Therefore, the power factor of the charging pile and its changes over time are important working information for the power subtask.

[0090] The relationship curve between the charging power and time of the charging pile is also key work information. By analyzing this curve, we can understand the power demand of the charging pile at different time points, so as to accurately calculate the power distribution of the entire power grid at different times. For example, for a charging pile cluster in a commercial area, the power demand during daytime business hours and non-business hours at night is quite different. This time characteristic is very important for the calculation of the power subtask.

[0091] 2. Determine work information based on node type

[0092] Ordinary nodes (non-substation nodes):

[0093] For charging piles connected to ordinary nodes, their working information includes their own charging data and the load characteristic information of the node. For example, in addition to the charging power, current and other data of the charging pile, the load size and change rules of other conventional electrical equipment on the node (such as street lights, small commercial stores, etc.) also need to be considered. Because these conventional loads and the charging pile load jointly affect the changes in the electrical quantity of the node.

[0094] The historical voltage and current data of the node are also important working information. By analyzing the historical data, the changing trend of the node electrical quantity under the current working state of the charging pile can be predicted. For example, if the historical data shows that the voltage of the node fluctuates greatly during a specific time period, and it coincides with the high-power charging period of the charging pile at that time, then when performing the relevant subtask calculations, it is necessary to focus on the impact of the current working conditions of the charging pile during that period on the node voltage.

[0095] Substation node (or important hub node):

[0096] The working information of the charging piles connected near the substation node needs to consider the operating parameters of the substation in addition to its own charging data. For example, the capacity of the main transformer of the substation, the current load rate and other information. Because these parameters determine the substation's carrying capacity and regulation ability for the surrounding charging piles. If the substation load rate is high, the charging power changes of the surrounding charging piles may have a greater impact on the output voltage and current of the substation, so this information is essential for the calculation of related subtasks.

[0097] The voltage regulation strategy of the substation and the operating status of the reactive compensation equipment should also be used as work information. For example, when the substation performs voltage regulation operations, it will affect the voltage input of the charging piles connected near it, which in turn affects the charging behavior of the charging piles and the electrical contribution to the grid. Therefore, understanding these operating strategies and equipment status of the substation can more accurately calculate the subtasks related to the substation node.

[0098] 3. Determine work information based on node location

[0099] Central location node in the region:

[0100] If the charging pile is connected to the central node in the area, its working information should comprehensively consider the distribution of electrical quantity in the entire area. For example, the overall charging power distribution of all charging piles in the area, the direction and size of current in each branch, etc. Because the change of electrical quantity at the central node will have a great impact on the operation of the power grid in the entire area, and it will also be affected by the comprehensive influence of other nodes and charging piles in the area.

[0101] The topological structure information of the power transmission lines in the region is also important working information. Understanding the connection mode, length, impedance and other parameters of the line can better analyze the power transmission characteristics from the central node to other nodes, as well as the change law of electrical quantity under different working conditions of charging piles. For example, when the charging power of a charging pile at a certain edge position in the region changes, the line topological structure information can be used to infer the path and degree of influence on the voltage and current of the central node.

[0102] Regional edge location nodes:

[0103] For charging piles connected to nodes at the edge of the region, their working information focuses on the interaction with the electrical quantity of the adjacent region. For example, the impact of the output voltage and current of the substation in the adjacent region on the edge nodes of this region, and the feedback impact of the working status of the charging piles on the edge nodes of this region on the power grid of the adjacent region. For example, at the boundary between two commercial areas, the high-power charging of the charging piles on one side may affect the voltage stability of the other side through the edge nodes, so the relevant electrical quantity information of both sides needs to be used as working information.

[0104] The reliability of power supply of edge nodes is also critical. For example, whether the node has a backup power supply, the redundancy of power supply, etc. Because edge nodes are relatively more susceptible to external factors (such as natural disasters, upper grid failures, etc.), the reliability of their power supply will affect the continuous operation capability of the charging pile, and thus affect the calculation of related subtasks. For example, if the reliability of power supply of the edge node is low, then when calculating the subtasks related to the node, it is necessary to consider the impact of the charging pile stopping working due to power outages on the electrical quantity of the grid.

[0105] 4. Determine work information based on time step

[0106] Static power flow calculation (single time point):

[0107] For the static power flow calculation subtask (i.e., power flow calculation only for a specific time point), the working information is mainly the real-time working status data of the charging pile at that time point. For example, the charging power, current, voltage and other data of each charging pile at a certain moment. These data reflect the contribution and impact of the charging pile on the electrical quantity of the power grid at that moment.

[0108] The total load and distribution of the power grid at that time (other loads except charging piles) are also working information. For example, at noon, in addition to the charging load of the charging piles, the power load of air conditioners, lighting and other equipment in the commercial area is at its peak. When performing the static power flow calculation subtask at that time, it is necessary to comprehensively consider the interaction between these loads and the charging pile load.

[0109] Dynamic power flow calculation (multiple time steps):

[0110] In the dynamic power flow calculation subtask, it is necessary to obtain the working state change data of the charging pile in multiple time steps. For example, the charging power change curve and current change curve of the charging pile in each time step (such as one time step every 15 minutes). These data can reflect the dynamic changes of the load of the charging pile at different times, which is very important for analyzing the dynamic changes of the power flow of the power grid over a period of time.

[0111] At the same time, it is also necessary to consider the load forecast information of the power grid in different time steps (including charging piles and other conventional loads). For example, based on historical data and real-time monitoring data, the load growth or reduction trend of the entire power grid in the next time step can be predicted, so that the direction and magnitude of the change of electrical quantity can be estimated in advance when performing the dynamic power flow calculation subtask, improving the accuracy and foresight of the calculation.

[0112] In some embodiments, the charging pile is also used to: subtract the reserved computing power from the rated computing power of the charging pile to obtain the remaining computing power of the charging pile.

[0113] In an embodiment of the present invention, in order to ensure the stability of the charging pile when performing the charging function and other basic operation tasks, a part of the computing power needs to be reserved. Because in the actual operation process, the charging pile must first ensure the safe and stable charging of the electric vehicle, which involves real-time monitoring of the battery status, adjusting the charging current and voltage, and other operations, all of which require certain computing resources to support. If all the computing power is used for other additional computing tasks (such as participating in power grid flow calculation, etc.), the charging function may be stuck, abnormal or even faulty, affecting the user's normal charging experience. For example, when the charging pile is processing complex flow calculation tasks at the same time and there is no reserved computing power, the charging current may not be adjusted in time, resulting in problems such as battery charging too fast or too slow. The size of the remaining computing power directly determines the scale and complexity of the additional computing tasks that the charging pile can undertake. In the supercomputing charging pile system, when the power distribution master station device decomposes the flow calculation task into multiple subtasks and assigns them to each charging pile, it is necessary to refer to the remaining computing power of each charging pile to ensure that the assigned subtasks do not exceed the computing power range of the charging pile, and avoid the failure of task execution or the impact on the normal charging function of the charging pile. For example, a charging pile with a higher remaining computing power can be assigned relatively complex and computationally intensive power flow calculation subtasks, such as tasks involving the joint calculation of multiple node voltages and branch currents; while a charging pile with a lower remaining computing power can only undertake some simple and computationally intensive subtasks, such as voltage monitoring calculations of a single node, thereby achieving rational utilization of computing power resources within the entire system, improving the efficiency and reliability of distributed computing, and facilitating the smooth implementation of power grid power flow calculations and other tasks.

[0114] In some embodiments, the charging pile is also used to: obtain first computing power information from the on-board terminal of the electric vehicle in response to a computing power supply request of the electric vehicle; subtract the reserved computing power from the rated computing power of the charging pile to obtain the available computing power of the charging pile; and determine the remaining computing power of the charging pile based on the available computing power and the first computing power information.

[0115] In the embodiments of the present invention, under the general trend of rapid development of the current automobile industry, electrification, intelligence and networking have become significant features. In order to achieve highly automated driving functions, smart electric vehicles need to accurately perceive complex driving environments, reasonably plan driving routes, and make reliable decisions. In order to meet the above-mentioned complex and high-real-time computing requirements, today's smart electric vehicles are generally equipped with advanced and powerful intelligent computing chips. Such high computing power enables the vehicle to quickly process various complex information during driving, ensuring the smooth operation of automatic driving or assisted driving functions.

[0116] However, when a smart car is not driving, such as when it is parked for charging or parked for a long time in a parking lot, the computing power originally used for driving-related functions is idle. Because at these moments, the vehicle no longer needs to process tasks such as driving environment perception, path planning and decision-making in real time, and those high-end NPUs and GPUs with powerful computing capabilities basically have no computing tasks to perform.

[0117] Therefore, with the permission of the car owner, the computing power of the vehicle, that is, the first computing power information, can be obtained to supplement the distributed computing power of the charging pile. Specifically, the remaining computing power of the charging pile needs to be updated. After completing the subtask allocation, the charging pile will hand over part of the calculation to the on-board terminal.

[0118] In some embodiments, the charging pile is also used to: obtain computing power usage information of each vehicle-mounted terminal; calculate the power exchange amount based on the computing power usage information; and report the power exchange amount to the power distribution master station device.

[0119] In an embodiment of the present invention, a portion of the computing power of the vehicle can be borrowed to provide a portion of electricity exchange services for the car owner, that is, the corresponding amount of electricity is exchanged according to the usage of the computing power, so that the car owner can actively participate in the computing power supply. That is, after the electric car is connected to the power grid, the computing power of the network increases with the access volume, but at the same time, the computing power of the electric car is also added to the network. The computing power of the supercomputing network elastically expands and contracts with the pressure of the power information network, effectively reducing the computing pressure of the supercomputing center.

[0120] In some embodiments, the power distribution master station device is used to: obtain first computing power information; calculate the uncertainty index of vehicle charging; and determine the reserved computing power of the power distribution master station device based on the first computing power information and the uncertainty index.

[0121] In the embodiment of the present invention, electric vehicles may enter / exit at any time, and there is a certain uncertainty. Therefore, the power distribution master station device needs to reserve computing power. When the electric vehicle fails to withdraw computing power as planned, the computing task of the vehicle terminal is handed over to the power distribution master station device to complete to avoid affecting the power network. Among them, the uncertainty index can be a preset value, or it can be based on the historical electric vehicle entry / exit information at any time, combined with the prediction value of the long short-term memory network, which is not limited here. The more computing power provided by the vehicle terminal and the higher the uncertainty index, the greater the computing power that the power distribution master station device needs to reserve.

[0122] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0123] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0124] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. A supercomputing charging pile system, characterized in that: include: Power distribution master station device, integrated terminal device and multiple charging piles; The power distribution master station device is connected to the supercomputing terminal; the power distribution master station device is connected to at least one fusion terminal device; Each fusion terminal device is connected to a plurality of charging piles; Each charging pile is used to report working information and computing power information to the power distribution master station device through the fusion terminal device; The power distribution master station device is used to obtain the power flow calculation task issued by the supercomputing terminal, and at the same time, according to the work information and the computing power information, decompose the power flow calculation task into multiple subtasks, and determine the work information corresponding to each subtask; wherein the subtask corresponds to the charging pile one by one; each subtask and the work information corresponding to each subtask are issued to each charging pile; The charging pile is also used to obtain processing results according to each subtask and the work information corresponding to each subtask, and report the processing results to the supercomputing terminal through the fusion terminal device and the power distribution master station device in sequence.

2. The supercomputing charging pile system according to claim 1, characterized in that: The power distribution master station device is used for: Dividing the power flow calculation task into a plurality of minimum calculation tasks; Input each minimum computing task, the computing power information, and the working information into a neural network to obtain the allocation probability of each minimum computing task to each charging pile, and form an allocation probability vector of each minimum computing task; wherein the allocation probability is determined according to the computing amount of each minimum computing task, the remaining computing power of each charging pile, and the first association relationship between each minimum computing task and the working information; According to the allocation probability vectors of each minimum computing task, a decomposition scheme is determined to decompose the power flow computing task into a plurality of subtasks.

3. The supercomputing charging pile system according to claim 2, characterized in that: The power distribution master station device is used for: The power flow calculation task is divided into a plurality of minimum calculation tasks according to the type of electrical quantity, the type of node, the location of the node and the time step.

4. The supercomputing charging pile system according to claim 2, characterized in that: The power distribution master station device is used for: Select one allocation probability from each allocation probability vector; Calculating the allocation evaluation value according to the second association relationship between each selected allocation probability and each minimum computing task; Select the plan with the highest allocation evaluation value and determine the decomposition plan.

5. The supercomputing charging pile system according to claim 4, characterized in that: The power distribution master station device is used for: If, according to the selected allocation probabilities, each charging pile is allocated with at most one subtask, then the selected allocation probabilities are added together to obtain the allocation evaluation value; If at least two subtasks are assigned to a charging pile according to the selected allocation probability, the selected allocation probabilities corresponding to the subtasks on the charging pile are multiplied and then added to other selected allocation probabilities to obtain the allocation evaluation value.

6. The supercomputing charging pile system according to claim 1, characterized in that: The power distribution master station device is used for: According to the electrical quantity type, node type, node position and time step corresponding to each subtask, the work information corresponding to each subtask is determined.

7. The supercomputing charging pile system according to claim 1, characterized in that: The charging pile is also used for: Subtract the reserved computing power from the rated computing power of the charging pile to get the remaining computing power of the charging pile.

8. The supercomputing charging pile system according to claim 1, characterized in that: The charging pile is also used for: In response to a computing power supply request of the electric vehicle, obtaining first computing power information from an onboard terminal of the electric vehicle; Subtract the reserved computing power from the rated computing power of the charging pile to get the available computing power of the charging pile; The remaining computing power of the charging pile is determined according to the available computing power and the first computing power information.

9. The supercomputing charging pile system according to claim 8, characterized in that: The charging pile is also used for: Obtain computing power usage information of each vehicle terminal; Calculate the amount of electricity exchange based on the computing power usage information; The power exchange amount is reported to the power distribution master station device.

10. The supercomputing charging pile system according to claim 8, characterized in that: The power distribution master station device is used for: Obtaining the first computing power information; Calculate the uncertainty index of vehicle charging; The reserved computing power of the power distribution master station device is determined according to the first computing power information and the uncertainty index.

Images