Optimization method for power grid system construction based on cascading failure simulation

Abstract

The application provides a power grid system construction optimization method based on cascade failure simulation, which comprises the following steps: firstly, designing an initial network model of a power grid system to be constructed; then, performing cascade failure simulation on the network model containing failure links, including current redistribution and link state updating, to obtain a network model after cascade failure; finally, judging whether the initial network model meets the design requirements according to the network model after cascade failure; the application considers the time delay characteristics of power grid cascade failure in reality and the automatic recovery function of the power grid in the cascade failure process, and realizes the design from the perspective of current through the network model after cascade failure, thus increasing the practical significance; the application can be used in the construction process of the real power grid and can also be used for safety detection of the existing power grid; when the real power grid is constructed, the application can be used to check whether the designed power grid meets the construction requirements.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of cascading failure of power grid systems, and particularly relates to a power grid system construction optimization method based on cascading failure simulation. BACKGROUND

[0002] In daily life, power grids play an increasingly important role, and people's daily life cannot be separated from power grids, and many important activities also cannot be separated from power grid support. Once a power grid fails, it will cause very serious consequences. At the same time, power grid failure provides a challenge for the construction of the power grid. When constructing the power grid, many factors need to be considered, such as economy, environmental protection and the like, and the primary consideration is the safety of the power grid, which is the bottom line. If a major failure occurs in the operation of the power grid, not only will it have a subsequent impact, but also it will waste resources used in the construction of the power grid. Therefore, in the process of constructing the power grid system, ensuring the safety of the power grid is the focus of construction. Whether it is to reduce losses or to provide help for the construction of the power grid system, it is necessary to understand the cascading failure process of the power grid in detail, so as to put forward a practical and effective method.

[0003] The brief process of cascading failure of the power grid is that the failure of part of the nodes / links in the power grid will affect the adjacent nodes / links, the adjacent nodes / links will be affected by the failed nodes / links and thus fail, these newly failed nodes / links will continue the above-mentioned process of affecting the adjacent nodes / links, and thus a cascading effect is produced, eventually leading to the failure from a small range to the entire network.

[0004] Many researchers study the cascading failure process of the power grid from various angles, some of which study the overall process of cascading failure, and some of which study the factors affecting cascading failure. In Physica A: Statistical Mechanics and its Applications, Vol. 402, pp. 169-179, et al. think that in the study of the influence of topology on the robustness of the power grid, the pure topology method is mainly used to evaluate, which cannot capture the essence of power flow. Therefore, they put forward a measure, i.e. effective graph resistance, to link the topology of the power grid with its robustness to cascading failure caused by deliberate attacks, and also consider the basic characteristics of the power grid, such as current distribution according to Kirchhoff's law, and then in order to prove its applicability, the index is applied to the IEEE118 bus power system to improve its robustness to cascading failure. et al. analyze the cascading failure of the power grid from the perspective of topology, and study the cascading failure of the power grid on the basis of pure topology combined with part of the electrical properties, but the combined electrical properties are not deep enough, and lack the characteristics of the power grid in reality, and the value in actual application is relatively weak.

[0005] Jin et al. introduced impedance network as a prototype model to describe cascading failure in IEEE Transactions on Industrial Electronics, vol. 68, no. 1, pp. 632-641. A current re-distribution method was proposed to analyze the cascading failure mechanism of the circuit. The current re-distribution coefficients of two typical fault modes, open circuit and short circuit, were proposed to determine the influence of the failed component on the remaining components. Further, a health confidence value was introduced to evaluate the health state of the impedance network. Finally, the cascading failure behavior was illustrated by examples and the effectiveness of the proposed cascading failure model was verified. Jin et al. combined some characteristics of the power grid and proposed a current re-distribution mechanism, but did not obviously combine some other attributes of the power grid in addition to the electrical properties such as electrical power and current, and did not add some phenomena that may occur when the power grid fails in reality, and has weak practical significance.

[0006] Huang et al. considered the cascading failure and recovery process, including the sequential manner of repairing the failed nodes, in IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 51, no. 1, pp. 400-411. Huang et al. mainly studied the recovery after cascading failure, but before that, a cascading failure model needed to be designed as the research object. Considering that the power grid may be discovered by staff and repaired, or the power grid automatically adjusts and terminates the cascading failure in advance to reduce losses when the cascading failure occurs in reality, an automatic recovery mechanism was simply added to the cascading failure model to design a new failure model and use it. Huang et al. considered the self-recovery in the process of cascading failure of the power grid in the research, but did not focus on in-depth research, only proposed assumptions, and did not consider how the power grid is scheduled to achieve the self-recovery of nodes / links. SUMMARY

[0007] In view of the deficiencies of the prior art, the present application proposes a power grid system construction optimization method based on cascading failure simulation, comprising:

[0008] Step 1: design an initial network model of the power grid system to be constructed;

[0009] Step 2: perform cascading failure simulation on the network model containing failed links, including current re-distribution and link state update, to obtain a network model after cascading failure;

[0010] Step 3: determine whether the initial network model meets the design requirements according to the network model after cascading failure.

[0011] The step 1 comprises:

[0012] Step 1.1: creating an initial network model of the power grid system according to the layout of the power grid;

[0013] Step 1.2: initializing parameters of the initial network model; the parameters include power generation of nodes, current in each link, capacity of each link, recovery probability of each link, tolerance parameter of each link, recovery ratio of self-recovery, and degree of probability decrease of self-recovery; wherein the capacity of the link is represented as:

[0014] cap (v,w) = (1 + a) * I (v,w) (1)

[0015] In the formula, cap (v,w) is the capacity of the link (v, w); I (v,w) is the current of the link (v, w) when the initial power grid flows; a is the tolerance parameter of the link (v, w).

[0016] The step 2 includes:

[0017] Step 2.1: selecting an initial failure link set for failure simulation of the initial network model;

[0018] Step 2.2: re-distributing current for the network model containing the failure link;

[0019] Step 2.3: updating network data according to the current state of the link;

[0020] Step 2.4: repeating steps 2.2-2.3 until the state of the link in the power grid does not change, and obtaining a network model after cascading failure.

[0021] The step 2.1 is specifically described as:

[0022] If it is to reconstruct the power grid system for the area where the power grid system is constructed: selecting the most frequently failed part of the link according to the operation state of the previous power grid system for failure simulation;

[0023] If it is to construct the power grid system for the area where the power grid system is not constructed for the first time: determining the priority of failure of each link according to the amount of power consumption of the node and the number of edges of the node, and selecting the most optimal part of the link for failure simulation.

[0024] The step 2.2 includes:

[0025] Step 2.2.1: determining the current threshold λ according to the design requirements of the power grid system;

[0026] Step 2.2.2: calculating the weight W of the link according to the number of common neighbors of the two end nodes of the link;

[0027] W = len(nearby(v)∩nearby(w)) + 1 (2)

[0028] where v, w are the start point and end point of the link respectively; nearby() represents all the neighboring nodes; len() represents the number;

[0029] Step 2.2.3: Re-distribute the current in the power grid with changed link state, according to the weight W of each link and the current threshold λ of the power grid, spread the current in the failed link to other links in the power grid with the link itself as the starting point; in the process of current re-distribution, if a loop occurs, and when the current on the loop is less than the threshold λ, set the current to 0, and end the current distribution process.

[0030] The step 2.3 is specifically expressed as:

[0031] Step 2.3.1: Update the current link state;

[0032] Step 2.3.2: Determine whether there is an overload link in the network model, if there is, continue to execute step 2.3.3; if not, jump out of the iteration process and end the update of the link state;

[0033] Step 2.3.3: Change the overload link with overload time exceeding the time delay in the network model to a failed link, and delete it from the network model, and execute 2.3.4;

[0034] Step 2.3.4: Select a target link set to be restored according to the current value of the self-recovery probability, and execute step 2.3.5;

[0035] Step 2.3.5: Perform self-recovery operation on each target link.

[0036] The step 2.3.1 is specifically expressed as:

[0037] For the link Q1 changed from overload state to normal state, the update process is: change the state record of the link Q1 from overload to normal, delete the state record of the link Q1 from the state record of the overload link, and delete the initial failure time and time delay attributes on the link Q1;

[0038] For the link Q2 changed from normal state to overload state, the specific update process is as follows:

[0039] Step S1: Calculate the time delay θ of the link (v, w) (v,w) :

[0040]

[0041] In the formula, Deg(v, w) represents the sum of the degrees of the two end nodes v and w of the link (v, w); AveDeg(G) represents the average degree of the power grid G; β is a time delay parameter, β ∈ [0, 1];

[0042] Step S2: modifying the state, changing the state record of the link Q2 from normal to overload, adding the link Q2 on the record of the overload link, and finally adding the initial failure time and the time delay attribute to the link Q2 and recording the corresponding data.

[0043] The step 2.3.5 is specifically expressed as: adjusting the power supply part of the link, dividing the current supplied to the link to other links, and reducing the current on the link according to formula (4);

[0044] result (v,w) =cap (v,w) *δ (4)

[0045] In the formula, result (v,w) represents the current value on the link (v, w) after adjustment; cap (v,w) represents the capacity of the link (v, w); and δ represents the recovery ratio, δ ∈ (0, 1];

[0046] According to formula (5), the self-recovery probability η of each link is updated (new) :

[0047] η (new) =η (old) *η de η de ∈[0,1] (5)

[0048] In the formula, η (new) represents the updated self-recovery probability of the link; η (old) represents the self-recovery probability of the link before updating; and η de represents the degree of reduction of the self-recovery rate.

[0049] The beneficial effects of the present application are:

[0050] The application provides a power grid system construction optimization method based on cascade failure simulation, selects a power grid as a carrier for studying cascade failure, does not consider electrical characteristics such as voltage and electric power, simulates the circulation of current in the power grid and the change of current when the power grid changes, deeply understands the process of cascade failure through current, proposes a current redistribution method, considers the time delay characteristics of the power grid in the process of cascade failure in reality and the automatic recovery function of the power grid in the process of cascade failure, realizes the network model of cascade failure from the perspective of current during the research and design, and increases the practical significance. The application mainly acts on the construction process of the real power grid, and can also be used for safety detection of the existing power grid. When the real power grid is constructed, the application can be used to check whether the designed power grid meets the construction requirements and is safe, if not, the power grid is modified according to the test results of the application and is checked again, the above process is repeated, and finally the power grid design meeting the requirements and being safe is obtained. For the power grid that has been built, the application can simulate the condition of the power grid when a sudden situation is encountered in the operation of the power grid, find out the weak part of the power grid, and protect the weak part. BRIEF DESCRIPTION OF DRAWINGS

[0051] Figure 1 The transformation process of the power grid link in the application;

[0052] Figure 2 The phase flow chart of the power grid system construction optimization method based on cascade failure simulation in the application. DETAILED DESCRIPTION

[0053] The application is further described below in combination with the drawings and specific implementation examples. The purpose of the application is to design based on current and propose a cascade failure model of the power grid combining time delay and automatic recovery, so as to optimize the construction process of the power grid system. The model creates the power grid by combining the complex network theory, takes the commonly used cascade failure model as the research basis, adds the current attribute to the model, displays the characteristics of the power grid, considers part of the characteristics of the real power grid, selects the time delay of the overload of the connection and the automatic recovery of the power grid, simulates and realizes the automatic recovery function of the power grid through current, and adds the concept and design of time delay in the model. It should be noted that the cascade failure in the power grid is divided into node failure and link failure, and the cascade failure model of the power grid proposed in the application only considers link failure. The state change of the link in the power grid is as shown in Figure 1 .

[0054] As shown in Figure 2 , a power grid system construction optimization method based on cascade failure simulation comprises the following steps.

[0055] Step 1: design an initial network model of the power grid system to be constructed; the step 1 comprises the following steps.

[0056] Step 1.1: Create an initial network model of the power grid system according to the power grid layout; extract important information according to design requirements, such as the coverage of the power grid, i.e. which areas need to be built in the power grid; links that must be connected in order to maintain power supply to the entire power grid, dependency relationships between nodes within the region; power consumption in different parts of the power grid region. Design a power grid layout and create an initial network model according to the above information;

[0057] Step 1.2: Initialize the parameters of the initial network model; the parameters include the power generation of the nodes, the current in each link (including the power generation of the power generation node and the current passing through the load node, determined by the power generation of the power station), the capacity of each link (determined according to the demand of the power grid to be created and the current of the initial power grid), the recovery probability of each link (determined from the impact on surrounding buildings after failure, the importance of the building represented by each node is determined according to the degree of dependence of the building on electricity, and different initial recovery probabilities are designed according to different importance levels), the tolerance parameter of each link, the recovery proportion of self-recovery, and the degree of decline of self-recovery probability; wherein the capacity of the link is represented as:

[0058] cap (v,w) = (1 + a) * I (v,w) (1)

[0059] In the formula, cap (v,w) is the capacity of the link (v, w); I (v,w) is the current of the link (v, w) when the initial power grid flows; a is the tolerance parameter of the link (v, w), 0 < a < 1, determined according to the demand of the power grid.

[0060] Step 2: Perform cascading failure simulation on the network model containing failed links, including current redistribution and link state update, to obtain the network model after cascading failure; said step 2 includes:

[0061] Step 2.1: Select an initial failed link set for failure simulation of the initial network model; specifically expressed as:

[0062] If the power grid system is reconstructed in the area where the power grid system is constructed: select the links with the most failure times according to the operation status of the previous power grid system for failure simulation; for example, the failure of the power grid in the past operation, various events that may affect the power grid, etc. are important factors that cause the failure of the power grid links.

[0063] If it is the first time to build the power grid system for the area which has not built the power grid system: according to the amount of power consumption of the node and the number of edges of the node, the priority of each link failure is determined, and the most preferred part of the link is selected for failure simulation; the fewer the number of edges of the node, the greater the probability of failure in the process of cascading failure;

[0064] Step 2.2: current redistribution of the network model containing the failed link; including:

[0065] Step 2.2.1: determine the current threshold λ according to the design requirements of the power grid system; the threshold is designed according to the demand of the construction of the power grid, and is moderate as much as possible under the condition of meeting the demand;

[0066] Step 2.2.2: in the current redistribution stage, the current changes caused by the failed link, the current is redistributed to each link in the power grid, and because of the importance, carrying capacity and other attributes of each link are different, the allocated current is also different, the power grid will allocate the current according to the weight of each link, and the weight W of the link is calculated according to the number of common neighbors of the two end nodes of the link;

[0067] W = len (nearby (v) ∩ nearby (w)) + 1 (2)

[0068] In the formula, v and w are the start point and end point of the link respectively; nearby() represents all the neighbor nodes; len() represents the number;

[0069] Step 2.2.3: current redistribution operation is performed on the current of the power grid with changed link state, according to the weight W of each link and the current threshold λ of the power grid, the current in the failed link is diffused to other links in the power grid with the link itself as the starting point; in the process of current redistribution, if a loop appears, and when the current on the loop is less than the threshold λ, the current flowing through is set to 0, and the current distribution process is ended;

[0070] Step 2.3: update the network data according to the current state of the link; specifically:

[0071] Step 2.3.1: update the current link state; specifically:

[0072] For the link Q1 which changes from overload state to normal state, the update process is: changing the state record of the link Q1 from overload to normal, deleting the state record of the link Q1 from the state record of the overload link, and deleting the initial failure time and time delay attribute of the link Q1;

[0073] For the link Q2 which changes from normal state to overload state, the specific update process is as follows:

[0074] Step S1: Calculate the time delay θ of the link (v, w) (v,w) :

[0075]

[0076] wherein Deg(v, w) represents the sum of the degrees of the two end nodes v and w of the link (v, w); AveDeg(G) represents the average degree of the power grid G; β is a time delay parameter, β ∈ [0, 1];

[0077] Step S2: Modify the state, change the state record of the link Q2 from normal to overload, add the link Q2 on the record of the overload link, and finally add the initial failure time and the time delay attribute to the link Q2 and record the corresponding data;

[0078] Step 2.3.2: Determine whether there is an overload link in the network model, if yes, continue to execute step 2.3.3; if no, jump out of the iteration process and end the update of the link state;

[0079] Step 2.3.3: Change the overload link with the overload time exceeding the time delay in the power grid to a failure link and delete it from the power grid, and execute 2.3.4;

[0080] Step 2.3.4: Select a target link set to be recovered according to the current value of the self-recovery probability, and execute step 2.3.5;

[0081] Step 2.3.5: Perform the self-recovery operation on each target link; the specific expression is: adjust the power supply part of the link, divide the current supplied to the link into other links, thereby reducing the current on the link, at the same time, the current accepted by the terminal of the link will be reduced, thereby causing the output current to be reduced, and the power supply of the corresponding area in the power grid will be reduced according to formula (4) to reduce the current on the link;

[0082] result (v,w) =cap (v,w) *δ (4)

[0083] wherein result (v,w) represents the current value on the adjusted link (v, w); cap (v,w) represents the capacity of the link (v, w); δ represents the recovery ratio, that is, the current on the link is reduced to a certain proportion of the capacity, and the recovery process is stopped, and the value of the recovery ratio δ is determined by the demand of the power grid, δ ∈ (0, 1];

[0084] When the target link is recovered, the self-recovery probability of the target link will be reduced, and the self-recovery probability η (new) :

[0085] η(new) = η (old) * η de η de ∈[0,1] (5)

[0086] In the formula, η (new) represents the updated link self-recovery probability; η (old) represents the self-recovery probability of the link before updating; η de represents the degree of decline of the self-recovery rate, which is obtained from the demand of the constructed power grid;

[0087] Step 2.4: Steps 2.2-2.3 are repeatedly executed until the link state in the power grid does not change, and the network model after cascading failure is obtained.

[0088] Step 3: Determine whether the initial network model meets the design requirements according to the network model after cascading failure. According to the data of the power grid model after cascading failure obtained in step 2, it is determined whether the current power grid meets the requirements analyzed in step 1. If it meets, it is indicated that the initial power grid model proposed in step 1 meets the requirements, which is the final result. If it does not meet, it is indicated that the initial power grid cannot meet the task requirements, and a new power grid model needs to be designed according to the area with large failure impact in this failure simulation.

[0089] The application can help deepen the understanding of the power grid, especially the cascading failure in the power grid, and can increase the related theoretical reserves of cascading failure, which provides help for the subsequent research on cascading failure of the power grid. From the application level, the cascading failure model is used to weigh how to build a more robust power grid when building the power grid, to help quickly build a power grid with strong robustness, so as to save the construction cost, and the model can also be used to detect the performance of the existing power grid, test the resistance of the power grid to cascading failure, and thus come up with corresponding countermeasures or prevent the influence of cascading failure in advance.

Claims

1. A power grid system construction optimization method based on a cascading failure simulation, characterized by, The method comprises the following steps: Step 1: designing an initial network model of a power grid system to be constructed; Step 2: performing cascade failure simulation on the network model containing failed links, including current redistribution and link state updating, to obtain a network model after cascade failure; The step 2 comprises: Step 2.1: selecting an initial failed link set for failure simulation of the initial network model; The step 2.1 is specifically expressed as: If the power grid system is reconstructed in a region where the power grid system has been constructed: selecting the most frequently failed partial links according to the operation condition of the previous power grid system for failure simulation; If the power grid system is initially constructed in a region where the power grid system has not been constructed: determining the priority of each link failure according to the power consumption of the node and the number of edges of the node, and selecting the most optimal partial links for failure simulation; Step 2.2: performing current redistribution on the network model containing failed links; The step 2.2 comprises: Step 2.2.1: determining a current threshold λ according to the design requirements of the power grid system; Step 2.2.2: calculating the weight W of the link according to the number of common neighbors of the two nodes at the ends of the link; W = len (nearby (v) ∩ nearby (w)) + 1 (2) In the formula, v and w are the starting point and the ending point of the link respectively; nearby() represents all the neighbor nodes; len() represents the quantity; Step 2.2.3: performing a current redistribution operation on the current of the power grid with the link state changed, and according to the weight W of each link and the current threshold λ of the power grid, the current in the failed link is diffused to other links in the power grid with the link itself as the starting point; in the current redistribution process, if a loop occurs, and when the current on the loop is less than the threshold λ, the current flowing through is set to 0, and the current distribution process is ended; Step 2.3: updating the network data according to the current state of the link; The step 2.3 is specifically expressed as: Step 2.3.1: updating the current link state; The step 2.3.1 is specifically expressed as: For the link Q1 changed from an overload state to a normal state, the updating process is: changing the state record of the link Q1 from overload to normal, deleting the state record of the link Q1 from the state record of the overload link, and deleting the initial failure time and the time delay attribute of the link Q1; For the link Q2 changed from a normal state to an overload state, the specific updating process is as follows: Step S1 : Calculate the time delay Θ of the link (v, w) (v,w) : In the formula, Deg(v,w) represents the sum of the degrees of the two nodes v and w of the link (v,w); AveDeg(G) represents the average degree of the power grid G; β is a time delay parameter, and β ∈ [0,1]; Step S2: modifying the state, changing the state record of the link Q2 from normal to overload, adding the link Q2 to the record of the overload link, and finally adding the initial failure time and the time delay attribute to the link Q2 and recording the corresponding data; Step 2.3.2: judging whether there is an overload link in the network model, if there is, continuing to execute step 2.3.3; if not, jumping out of the iteration process and ending the updating of the link state; Step 2.3.3: Change the overload link whose overload time exceeds the time delay in the network model into a failed link, and delete it from the network model, and execute step 2.3.4; Step 2.3.4: Select a target link set to be recovered according to the current value of the self-recovery probability, and execute step 2.3.5; Step 2.3.5: Perform self-recovery operation on each target link; The step 2.3.5 is specifically described as follows: Adjust the power supply part of the link, divide the current supplied to the link to other links, and reduce the current on the link according to formula (4); result (v,w) = cap (v,w) * delta (4) where result (v,w) denotes the adjusted current value on the link (v,w); cap (v,w) denotes the capacity of the link (v,w); δ denotes the proportion of recovery, δ ∈ (0,1] The self-recovery probability η of each link is updated according to formula (5) (new) : η (new) = η (old) * η de η de ∈ [0,1] (5) In the formula, η (new) represents the updated link self-recovery probability; η (old) represents the self-recovery probability of the link before updating; η de represents the degree of decline of the self-recovery rate; Step 2.4: Repeat steps 2.2 to 2.3 until the link state in the power grid does not change, and obtain the network model after cascading failure; Step 3: Determine whether the initial network model meets the design requirements according to the network model after cascading failure.

2. The method of claim 1, wherein, The step 1 includes: Step 1.1: Create an initial network model of the power grid system according to the layout of the power grid; Step 1.2: Initialize the parameters of the initial network model; the parameters include the power generation of the node, the current in each link, the capacity of each link, the recovery probability of each link, the tolerance parameter of each link, the recovery ratio of self-recovery, and the degree of self-recovery probability reduction; wherein the capacity of the link is represented as: cap (v,w) = (1 + a) * I (v,w) (1) where cap (v,w) is the capacity of link (v, w); I (v,w) is the initial grid flow through link (v, w); and a is the tolerance parameter of link (v, w).

Images