CN114698119B - 5G communication/cloud edge computing resource collaborative allocation method for distribution network distributed protection system - Google Patents

Abstract

The present invention discloses a method for collaborative allocation of 5G communication/cloud-edge computing resources of a distribution network distributed protection system, the steps of which include: 1. Establishing three modes of distributed protection devices for judging whether they are in an abnormal state; 2. Establishing a cloud-edge computing delay model and a 5G communication model of a distribution network distributed protection monitoring and intelligent evaluation system; 3. Establishing an optimization objective function with minimal traffic loss; 4. Solving the optimal solution of the optimization objective function, and using the optimal solution to obtain the optimal solution for collaborative allocation of 5G communication/cloud-edge computing resources of the distribution network system, thereby minimizing 5G communication traffic loss. The present invention solves the problem of 5G communication traffic loss in the distribution network distributed system with time constraints and intelligent evaluation computing power constraints by optimizing the collaborative allocation of 5G communication/cloud-edge computing resources of the distribution network distributed system.

Description

5G communication/cloud-edge computing resource collaborative allocation method for distributed protection system of distribution network

Technical Field

The present invention relates to a method for collaborative allocation of state data 5G communication/cloud-edge computing resources, and belongs to the field of wireless communications.

Background Art

In recent years, the rapid development of 5G communication technology has provided a low-latency, highly reliable information channel for distribution network protection services. With the development of information technology and the increasing demand for diversified power grids, improving the flexibility and elasticity of power grid operation has become an urgent need for the power system. Among them, 5G D2D technology allows information exchange between end users by allowing spectrum reuse, which can not only greatly improve spectrum utilization, alleviate the load pressure of 5G base stations, and improve the flexibility of power grid operation, but also improve the total number of D2D users and the communication quality of the distribution network.

With the in-depth application of power Internet of Things technology, further mining the value of a large amount of data generated on the device side has become a development trend. This trend has dramatically increased the requirements for data storage, processing and transmission on the device side, and it is necessary to deploy lightweight computing services on the device side to process information. Lightweight computing on the device side integrates network, storage, computing and other resources, and can provide data services on the device side of the distribution network system to improve system operation efficiency. However, lightweight computing is only applicable to scenarios such as real-time, short-cycle data analysis and local decision-making, and is insufficient for the overall analysis of the distribution network. The main station computing platform is suitable for big data analysis of non-real-time, long-cycle data. Therefore, collaborative computing has shown its advantages.

The current research on collaborative computing is mainly focused on edge computing, mobile edge computing and multi-access edge computing. In the power sector, there are also studies that introduce cloud-edge collaboration into applications such as energy management, demand response, load forecasting, and intelligent operation and maintenance. Although the needs of power application scenarios are taken into account, they all focus on the theory of collaborative computing itself, and do not consider the massive application of distribution network equipment. Therefore, the combination of collaborative computing with the distribution network system computing resources, communication resource distribution, and equipment configuration, so collaborative optimization has become a difficult problem in the research.

Zeng Zhi and others from the School of Electrical Engineering and Automation at Jiangxi University of Science and Technology analyzed the factors that affect path impedance, constructed a road section impedance function, and proposed a path allocation method based on an improved Logit model. Based on the 5G communication platform and data transmission system, the data information of each path is collected, and the calculation results are sent to the vehicle through the allocation method. Although the article proposes a path allocation method, this allocation method does not utilize the computing resources on the vehicle side and the resource reuse between vehicles, and the flexibility of the vehicle network operation is low.

Li Ruihua from Henan Institute of Science and Technology installed a certain number of sensors with perception functions in a certain area according to the actual situation, and completed the design of the software and hardware circuits for online detection of transmission lines. Through online detection of transmission lines, the operating status of transmission lines can be grasped in real time on the edge side, and the operating information of transmission lines can be displayed on the upper computer interface, so as to grasp the information of transmission lines in time, reduce the workload of staff and improve efficiency. Although the article uses the edge side to process circuit information, it does not use cloud-edge collaboration and cannot solve the disadvantage of large cloud latency.

Bai Yuyang and others from the School of Electrical Engineering and Automation of Wuhan University introduced the development background and key technologies of edge computing, explained the functions and characteristics of cloud-edge collaboration and edge-edge collaboration, and proposed using cloud-edge collaboration, edge-edge collaboration, edge intelligence and other technologies to solve the problems faced by power systems, such as high real-time performance, short data cycles, and complex tasks. Although the article takes into account the needs of power application scenarios, it focuses on the theory of collaborative computing and does not consider the combination of collaborative computing with the distribution of computing resources and communication resources in the distribution network system and the status of equipment configuration.

Summary of the invention

In view of the above-mentioned shortcomings in the prior art, the present invention provides a 5G communication/cloud-edge computing resource collaborative allocation method for a distribution network distributed protection system, in order to minimize the communication traffic loss while meeting the constraints of data transmission delay, thereby improving the resource utilization of the distribution network system and ensuring the 5G communication quality of the distribution network and the cloud-edge collaborative computing capabilities.

In order to achieve the above-mentioned object of the invention, the present invention adopts the following technical scheme:

The present invention discloses a 5G communication/cloud-edge computing resource collaborative allocation method for a distribution network distributed protection system. The distribution network distributed protection monitoring and intelligent evaluation system comprises N distributed protection devices, a 5G base station, and a cloud server; each distributed protection device comprises a sensor, a 5G communication module, and an edge-side embedded intelligent evaluation module, and monitors its own status data through the sensor; the cloud server comprises a cloud-side intelligent evaluation software module, and is characterized in that it comprises the following steps:

Step 1: Establish three modes of the distributed protection device to determine whether it is in an abnormal state;

Mode 1: Use the embedded intelligent evaluation module on the edge to evaluate whether the device itself is abnormal.

Mode 2: The 5G communication module transmits its own status data to the edge-side embedded intelligent evaluation module of other adjacent distributed protection devices through 5G D2D communication to evaluate whether it has any abnormality.

Mode 3: The 5G communication module transmits its own status data to the cloud-based intelligent evaluation software module to evaluate whether it has any abnormalities.

Step 2: Establish the 5G communication model and cloud-edge computing delay model of the distribution network distributed protection monitoring and intelligent evaluation system;

Step 2.1: Establish the 5G communication model according to formula (2-1) to formula (2-4):

Mode 1:

Mode 2:

Mode 3:

In formula (2-1) to formula (2-4), _Rn is the amount of status data of the nth distributed protection device evaluating its own equipment, 1≤n≤N, The amount of status data to evaluate whether the distributed protection device itself has abnormal status through the first mode; The amount of state data transmitted to the embedded intelligent evaluation module of other N-1 adjacent distributed protection devices through the second mode to evaluate whether the distributed protection device itself has abnormal status data; The amount of status data transmitted to the cloud intelligent evaluation software module through the third mode to evaluate whether its own distributed protection device has abnormal status;

Step 2.2: Use equations (2-5) to (2-8) to construct the cloud-edge computing delay model:

T＝max(T ₁ , T ₂ , T ₃ ) (2-5)

In formula (2-5) to formula (2-8), T is the total delay time, T ₁ , T ₂ , and T ₃ are the delay times generated by the amount of state data for evaluating whether the distributed protection device itself is abnormal through the first mode, the second mode, and the third mode, respectively; is 1024; V is the rate of status data transmitted by 5G D2D communication to evaluate whether its own distributed protection device has abnormality through the second mode; B is the channel bandwidth; v ₁ , v ₂ , and v ₃ are the calculation rates of status data of evaluating whether its own distributed protection device has abnormality through the first mode, the second mode, and the third mode, respectively, and v ₃ is much larger than v ₁ and v ₂ ;

Step 3: Construct the traffic loss objective function Y of the cloud-edge computing resources and 5G communication resources of the distribution network distributed protection monitoring and intelligent evaluation system. According to the limiting factors of computing resources and communication resources, establish the constraint conditions of the intelligent evaluation computing capacity, and according to the real-time requirements of the intelligent evaluation, establish the delay constraint conditions;

Step 3.1, use equation (3-1)-equation (3-3) to construct the objective function Y of state data transmission flow loss:

In formula (3-1) and formula (3-2), y(s) is the flow loss caused by the transmission of state data through the sth mode; is the amount of status data for evaluating whether the distributed protection device of the sth mode is abnormal; θ(s) is the traffic loss generated by the sth mode for evaluating whether the distributed protection device of the sth mode is abnormal per kilobyte of status data transmitted through the 5G base station; and θ(1) = 0, θ(2) = 0;

Step 3.2: Use equations (3-4) to (3-5) to construct delay constraints and intelligent evaluation computing capacity constraints:

T≤T _min (3-4)

In formula (3-4), T _min is the minimum delay time;

Step 4: Use the jointly improved Hungarian optimization algorithm to solve the optimal solution for the coordinated allocation of 5G communication resources and cloud-edge computing resources;

Step 4.1, construct an N×N matrix R, where the element in the i-th row and the j-th column of the matrix R is recorded as _Rij , where _Rij represents the amount of state data sent by the i-th sensor to the N-j+1-th edge-side embedded intelligent evaluation module in the N edge-side embedded intelligent evaluation modules through the first mode or the second mode to evaluate whether its distributed protection device has abnormal status; i = 1, 2, ..., N, j = 1, 2, ..., N;

Step 4.2: Use formula (4-1) to obtain the matrix R' so that each row of the matrix R' contains zero elements:

In formula (4-1), Represents the minimum element of the i-th row in the matrix R;

Step 4.3, use formula (4-2) to obtain the matrix R' so that each row of the matrix R' contains zero elements:

In formula (4-2), represents the minimum element in the jth column of the matrix R';

Step 4.4: Use formula (4-3) to get the minimum element Then, by exchanging the elements in the _ith row and the jth row in the matrix R” _ij , all the smallest elements in the matrix R” _ij are arranged on the diagonal:

In formula (4-2), Represents the smallest element in the matrix R';

Step 4.5, when the diagonal elements of the matrix R” are all zero, that is, R” _ij = 0, i = j, and the elements outside the diagonal of the matrix R” reach the maximum value, the maximum amount of status data is obtained in the first mode and the second mode, when N sensors evaluate their own distributed protection devices in N edge-side embedded intelligent evaluation modules to see whether abnormalities occur. While minimizing the 5G traffic loss of status data transmission, the optimal solution for the collaborative allocation of 5G communication/cloud-edge computing resources is obtained.

Compared with the prior art, the present invention has the following beneficial effects:

1. The present invention designs three modes for the distributed protection device in the distribution network system to judge whether it is in an abnormal state. The three modes are run simultaneously to process the state data used to evaluate whether the distributed protection device itself is abnormal at a higher speed and efficiency, thereby effectively reducing the 5G traffic loss of state data transmission under the limitation of delay and computing power, improving the resource utilization rate in the distribution network system and the working efficiency of the edge-side embedded intelligent evaluation module, alleviating the pressure on the 5G base station, and improving the 5G communication quality/cloud-edge collaborative computing capability of the distribution network system.

2. The present invention establishes a 5G communication model and a cloud-edge computing delay model corresponding to the three modes in the distribution network system. Through the 5G communication model and the cloud-edge computing delay model, the edge-side embedded intelligent evaluation module and the cloud-side intelligent evaluation software module can more accurately calculate and evaluate the computing power and delay time, and at the same time provide more accurate constraints for evaluating whether the transmission flow loss of abnormal state data volume of its own distributed protection device occurs, thereby improving the efficiency and accuracy of state data volume transmission and calculation, as well as the working efficiency and accuracy of the edge-side embedded intelligent evaluation module and the cloud-side intelligent evaluation software module.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a diagram of a 5G communication/cloud-edge computing resource collaborative allocation model of the present invention;

FIG. 2 is a flow chart of the steps of the method of the present invention.

DETAILED DESCRIPTION

In this embodiment, as shown in FIG1 , a distribution network distributed protection system includes N distributed protection devices, a 5G base station, and a cloud server; each distributed protection device includes a sensor, a 5G communication module, and an edge-side embedded intelligent evaluation module; the cloud server includes a cloud intelligent evaluation software module, and has:

Distributed protection device: In the distributed protection device, the sensor generates status data by monitoring the status data of the distribution network protection device and the voltage and current data at both ends. The status data includes but is not limited to: the voltage/current data of the protection device, the temperature/humidity data of the protection device, the current data of the protection device, etc. The status data is transmitted to the cloud/embedded edge side according to the required intelligent evaluation and computing capabilities through the 5G communication hardware module. The embedded edge sides can share the status data of the distributed protection device through 5GD2D (Device to Device) communication. The intelligent evaluation module in the embedded edge side downloads, calculates and processes the status data, and transmits the processing results to the protection device module to realize the protection monitoring and intelligent evaluation of the distribution network distributed protection device;

In this embodiment, as shown in FIG2 , a 5G communication/cloud-edge computing resource collaborative allocation method for a distribution network distributed protection system includes:

Step 1: Set three modes of the distributed protection device to determine whether it is in an abnormal state;

Mode 1: Use the embedded intelligent evaluation module on the edge to evaluate whether the device itself is abnormal.

Mode 2: The 5G communication module transmits its own status data to the edge-side embedded intelligent evaluation module of other adjacent distributed protection devices through 5G D2D communication to evaluate whether it has any abnormality.

Mode 3: The 5G communication module transmits its own status data to the cloud-based intelligent evaluation software module to evaluate whether it has any abnormalities.

Among them, the edge-side embedded intelligent evaluation module and the cloud-based intelligent evaluation software module have different intelligent evaluation computing capabilities and calculation accuracy; in the second and third modes, the traffic costs of 5G communication data transmission to the cloud and D2D communication are also different;

Step 2: Establish the 5G communication model and cloud-edge computing delay model of the distribution network distributed protection system;

Step 2.1: Establish a 5G communication model based on equations (2-1) to (2-4):

Mode 1:

Mode 2:

The third mode:

In formula (2-1) to formula (2-4), _Rn is the amount of status data of the nth distributed protection device evaluating its own equipment, 1≤n≤N, The amount of status data to evaluate whether the distributed protection device itself has abnormal status through the first mode; The amount of state data transmitted to the embedded intelligent evaluation module of other N-1 adjacent distributed protection devices through the second mode to evaluate whether the distributed protection device itself has abnormal status data; The amount of status data transmitted to the cloud intelligent evaluation software module through the third mode to evaluate whether its own distributed protection device has abnormalities; the unit of the status data amount is kilobytes;

Step 2.2: Use equations (2-5) to (2-8) to build a cloud-edge computing delay model:

T＝max(T ₁ , T ₂ , T ₃ ) (2-5)

In formula (2-5) to formula (2-8), T is the total delay time, _T1 , _T2 , and _T3 are the delay time caused by the amount of status data for evaluating whether the distributed protection device itself has abnormality through the first mode, the second mode, and the third mode, respectively, in seconds; θ represents the conversion between kilobytes and megabytes, and the value is 1024; V is the rate of status data for evaluating whether the distributed protection device itself has abnormality through 5G D2D communication transmission through mode 2, in kilobytes/second; B is the channel bandwidth, in megabytes/second; _v1 , _v2 , and _v3 are the calculation rates of status data for evaluating whether the distributed protection device itself has abnormality through the first mode, the second mode, and the third mode, respectively, and _v3 is much larger than _v1 and _v2 , in kilobytes/second;

Step 3: Use steps 3.3 to 3.2 to construct the objective function of minimizing the traffic loss of state data transmission:

Step 3.1, use equation (3-1)-equation (3-3) to construct the objective function Y with the minimum state data transmission flow loss:

In formula (3-1) and formula (3-2), y(s) is the flow loss caused by the transmission of state data through the sth model, in yuan; is the amount of status data for evaluating whether the distributed protection device of the sth mode is abnormal; θ(s) is the traffic fee generated by the sth mode for transmitting each kilobyte of status data for evaluating whether the distributed protection device of the sth mode is abnormal through the 5G base station, and the unit is yuan/kilobyte; since the first and second modes do not transmit status data through the 5G base station, no traffic fee generated by transmitting status data through the 5G base station is generated, that is, θ(1) = 0, θ(2) = 0, by maximizing the amount of status data for evaluating whether the distributed protection device of the sth mode is abnormal and Minimize 5G traffic loss Y;

Step 3.2: Use formula (3-4)-formula (3-5) to construct delay constraints and intelligent evaluation computing capacity constraints:

T≤T _min (3-4)

In formula (3-4), T _min is a constant, indicating the minimum delay time.

Step 4: Use the jointly improved Hungarian optimization algorithm to solve the optimal solution for the coordinated allocation of 5G communication resources and cloud-edge computing resources;

Step 4.1, construct an N×N matrix R, where the element in the i-th row and the j-th column of the matrix R is denoted as _Rij , where _Rij represents the amount of state data sent by the i-th sensor to the N-j+1-th edge-side embedded intelligent evaluation module through the first mode or the second mode to evaluate whether its distributed protection device has abnormal conditions; i = 1, 2, ..., N, j = 1, 2, ..., N;

Step 4.2: Use formula (4-1) to make each row of matrix R contain zero elements, and obtain matrix R':

In formula (4-1), Represents the minimum element of the i-th row in the matrix R;

Step 4.3: Use formula (4-2) to make each row of matrix R contain zero elements, and obtain matrix R'':

In formula (4-2), represents the minimum element in the jth column of the matrix R';

Step 4.4: Use formula (4-3) to compare the j-th column element in the matrix R' with each row element to obtain the minimum element. Then, by exchanging the elements of the _ith row and the jth row in the matrix R” _ij , all the smallest elements in the matrix R” _ij can be replaced. Arranged on the diagonal:

In formula (4-2), Represents the smallest element in the matrix R';

Step 4.5, when all diagonal elements of the matrix R” are zero, that is, R” _ij = 0, i = j, the elements R” _ij outside the diagonal of the matrix R” take the maximum value, that is, when i≠j, so as to obtain the maximum state data volume R” _ij for N sensors to evaluate their own distributed protection devices in N edge-side embedded intelligent evaluation modules in the first mode and the second mode, and obtain the minimum state data volume for each sensor to evaluate its own distributed protection device in the third mode through the cloud-based intelligent evaluation software module through (2-4). While maximizing the communication/computing resources of the edge-side embedded intelligent evaluation module and minimizing the 5G traffic loss of state data transmission, the optimal solution for the coordinated allocation of 5G communication/cloud-edge computing resources is obtained.

Claims (1)

1. A method for collaboratively allocating 5G communication and cloud-edge computing resources of a distribution network distributed protection system, wherein the distribution network distributed protection monitoring and intelligent evaluation system comprises N distributed protection devices, 5G base stations, and cloud servers; each distributed protection device comprises a sensor, a 5G communication module, and an edge-side embedded intelligent evaluation module, and monitors its own status data through the sensor; the cloud server comprises a cloud-side intelligent evaluation software module, and is characterized in that it comprises the following steps: Step 1: Establish three modes of the distributed protection device to determine whether it is in an abnormal state; Mode 1: Use the embedded intelligent evaluation module on the edge to evaluate whether the device itself is abnormal. Mode 2: The 5G communication module transmits its own status data to the edge-side embedded intelligent evaluation module of other adjacent distributed protection devices through 5G D2D communication to evaluate whether it has any abnormality. Mode 3: The 5G communication module transmits its own status data to the cloud-based intelligent evaluation software module to evaluate whether it has any abnormalities. Step 2: Establish the 5G communication model and cloud-edge computing delay model of the distribution network distributed protection monitoring and intelligent evaluation system; Step 2.1: Establish the 5G communication model according to formula (2-1) to formula (2-4): Mode 1: Mode 2: The third mode: In formula (2-1) to formula (2-4), _Rn is the amount of status data of the nth distributed protection device evaluating its own equipment, 1≤n≤N, The amount of status data to evaluate whether the distributed protection device itself has abnormal status through the first mode; The amount of state data transmitted to the embedded intelligent evaluation module of other N-1 adjacent distributed protection devices through the second mode to evaluate whether the distributed protection device itself has abnormal status data; The amount of status data transmitted to the cloud intelligent evaluation software module through the third mode to evaluate whether its own distributed protection device has abnormal status; Step 2.2: Use equations (2-5) to (2-8) to construct the cloud-edge computing delay model: T＝max(T ₁ , T ₂ , T ₃ ) (2-5) In formula (2-5) to formula (2-8), T is the total delay time, T ₁ , T ₂ , and T ₃ are the delay times generated by the amount of state data for evaluating whether the distributed protection device itself is abnormal through the first mode, the second mode, and the third mode, respectively; is 1024; V is the rate of status data transmitted by 5G D2D communication to evaluate whether its own distributed protection device has abnormality through the second mode; B is the channel bandwidth; v ₁ , v ₂ , and v ₃ are the calculation rates of status data of evaluating whether its own distributed protection device has abnormality through the first mode, the second mode, and the third mode, respectively, and v ₃ is greater than v ₁ and v ₂ ; Step 3: Construct the traffic loss objective function Y of the cloud-edge computing resources and 5G communication resources of the distribution network distributed protection monitoring and intelligent evaluation system. According to the limiting factors of computing resources and communication resources, establish the constraint conditions of the intelligent evaluation computing capacity, and according to the real-time requirements of the intelligent evaluation, establish the delay constraint conditions; Step 3.1, use equation (3-1)-equation (3-3) to construct the objective function Y of state data transmission flow loss: In formula (3-1) and formula (3-2), y(s) is the flow loss caused by the transmission of state data through the sth mode; is the amount of status data for evaluating whether the distributed protection device of the sth mode is abnormal; θ(s) is the traffic loss generated by the sth mode for evaluating whether the distributed protection device of the sth mode is abnormal per kilobyte of status data transmitted through the 5G base station; and θ(1) = 0, θ(2) = 0; Step 3.2: Use equations (3-4) to (3-5) to construct delay constraints and intelligent evaluation computing capacity constraints: T≤T _min (3-4) In formula (3-4), T _min is the minimum delay time; Step 4: Use the jointly improved Hungarian optimization algorithm to solve the optimal solution for the coordinated allocation of 5G communication resources and cloud-edge computing resources; Step 4.1, construct an N×N matrix R, where the element in the i-th row and the j-th column of the matrix R is recorded as _Rij , where _Rij represents the amount of state data sent by the i-th sensor to the N-j+1-th edge-side embedded intelligent evaluation module in the N edge-side embedded intelligent evaluation modules through the first mode or the second mode to evaluate whether its distributed protection device has abnormal status; i = 1, 2, ..., N, j = 1, 2, ..., N; Step 4.2: Use formula (4-1) to obtain the matrix R' so that each row of the matrix R' contains zero elements: In formula (4-1), Represents the minimum element of the i-th row in the matrix R; Step 4.3, use formula (4-2) to obtain the matrix R' so that each row of the matrix R' contains zero elements: In formula (4-2), represents the minimum element in the jth column of the matrix R'; Step 4.4: Use formula (4-3) to get the minimum element Then, by exchanging the elements in the _ith row and the jth row in the matrix R” _ij , all the smallest elements in the matrix R” _ij are arranged on the diagonal: In formula (4-3), Represents the smallest element in the matrix R'; Step 4.5, when the diagonal elements of the matrix R” are all zero, that is, R” _ij = 0, i = j, and the elements outside the diagonal of the matrix R” reach the maximum value, the maximum amount of status data is obtained in the first mode and the second mode, when N sensors evaluate their own distributed protection devices in N edge-side embedded intelligent evaluation modules to see whether abnormalities occur. While minimizing the 5G traffic loss of status data transmission, the optimal solution for the coordinated allocation of 5G communication and cloud-edge computing resources is obtained.